Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
METHOD FOR IDENTIFYING AND EVALUATING COMMON CAUSE FAILURES OF SYSTEM COMPONENTS
Document Type and Number:
WIPO Patent Application WO/2019/201715
Kind Code:
A1
Abstract:
A method and system for identifying and evaluating common cause failures (CCF) of system components (C), wherein at least one analytical artifact (AA) and machine readable sys- tem related spatial and/or topological data and/or machine readable system related lifecycle data are processed to ana- lyze automatically a susceptibility of system components (C) to common cause failure based on common cause failure influ- encing factors (CCFIF).

Inventors:
KAUKEWITSCH CHRISTOF (DE)
HEILMANN REINER (DE)
ZELLER MARC (DE)
Application Number:
EP2019/059143
Publication Date:
October 24, 2019
Filing Date:
April 10, 2019
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
SIEMENS AG (DE)
International Classes:
G06Q10/00
Foreign References:
CN107038520A2017-08-11
Download PDF:
Claims:
Patent Claims

1. A method for identifying and evaluating common cause fail ures (CCF) of system components (C) ,

wherein at least one analytical artifact (AA) and machine readable system related spatial and/or topological data and/or machine readable system related lifecycle data are processed to analyze automatically a susceptibility of system components (C) to common cause failure based on common cause failure influencing factors (CCFIF) .

2. The method according to claim 1 wherein the analytical ar tifact (AA) comprises

a machine readable safety, reliability or availability relat ed analytical artifact (AA) generated on the basis of a ma chine readable functional description (mrFD) of the technical system (SYS) of interest and at least one system evaluation criterion (SysEC) .

3. The method according to claim 1 or 2, wherein system com ponents of the technical system of interest (SYS) have asso ciated machine readable functional descriptions (mrFD) in cluding port definitions and component failure mode descrip tions processed to generate automatically the analytical ar tifact (AA) in response to at least one applied system evalu ation criterion (SysEC) .

4. The method according to any of the preceding claims 1 to 3, wherein the analytical artifact (AA) comprises

a fault tree,

a Markov chain,

a combination of fault tree(s) and Markov chain (s),

an FMEA,

an FMECA or

an FMEDA.

5. The method according to any of the preceding claims 1 to

4, wherein the system related machine readable lifecycle data (mrLD) comprises data regarding system design, system test ing, system history, system component history, training data, data regarding planned or implemented operation procedures and/or maintenance procedures concerning the system of inter est .

6. The method according to any of the preceding claims 1 to

5, wherein a quantitative common cause failure result (CCFR) is calculated on the basis of a machine readable common cause failure model (CCFMOD) provided for common cause failure in fluencing factors (CCFIFs) .

7. The method according to claim 6 wherein the common cause failure model (CCFMOD) comprises the IEC 61508 common cause failure model used to calculate a beta factor as a quantita tive common cause failure result (CCFR) .

8. The method according to claim 6 or 7 wherein a counter measure is automatically triggered in response to the calcu lated common cause failure result (CCFR) .

9. The method according to any of the preceding claims 1 to

8, wherein the machine readable spatial or topological data (mrSD, mrTD) comprises data regarding an arrangement and/or placement of system components (C) within the system of in terest (SYS), in particular position coordinates and/or dis tances between system components (C) .

10. The method according to any of the preceding claims 1 to

9, wherein the common cause failure influencing factors

(CCFIFs) comprise

spatial proximity,

design principles,

technologies ,

manufacturers , lifecycle facts,

level of redundancy,

diversity within implemented redundancies,

complexities .

11. The method according to any of the preceding claims 2 to 10 wherein the system evaluation criterion (SysEC) comprises a reliability criterion,

an availability criterion, and/or

a safety criterion.

12. The method according to claim 4, wherein the fault tree and/or Markov chain are provided by

transforming at least one system evaluation crite rion (SysEC) into one or more corresponding relevant state patterns at ports forming a system boundary of the system of interest (SYS) and by

generating the fault tree and/or Markov chain on the basis of the relevant state patterns and on the basis of the failure modes (FM) of the components of the system of in terest (SYS) .

13. The method according to claim 4, wherein the analytical artifact is provided by

transforming at least one system evaluation crite rion (SysEC) into one or more corresponding relevant state patterns at ports at the system boundary and/or at ports in side of the system of interest (SYS) and by

generating the analytical artifact (AA) on the ba sis of the relevant state patterns and on the basis of the component failure modes descriptions of the components (C) of the system of interest (SYS) .

14. The method according to claim 13 comprising

transforming the system evaluation criterion

(SysEC) into at least one system state that can be represent ed by a state pattern, applying the at least one state pattern to input ports and output ports of the system boundary of the system of interest (SYS) ,

deriving relevant system failure events by automat ically taking into account the failure propagation mechanisms based on the machine readable functional descriptions (mrFD) of the system' s components (C) including their failure mode descriptions, and

assembling the derived failure events to generate the fault tree and/or Markov chain used as an analytical ar tifact (AA) .

15. The method according to any of the preceding claims 2 to

14, wherein the machine readable functional description

(mrFD) of a system component (C) comprises

port definitions of input and output ports of the compo nent,

component failure modes,

an internal state or states of the component,

a failure rate,

a maintenance activity,

an inspection interval,

a mean down time, and/or

a mean time to repair.

16. The method according to any of the preceding claims 1 to

15, wherein the components (C) of the system of interest (SYS) comprise

hardware components,

software components to be executed by hardware com ponents such as CPUs

and

embedded components.

17. A system for identification and evaluation of common cause failures (CCF) of a technical system of interest (SYS) comprised of system components (C) ,

said identification and evaluation system comprising a database (DB) which stores a digital twin of the system of interest (SYS) including machine readable system related spatial and/or topological data (mrSD, mrTD) and/or machine readable system related lifecycle data (mrLD) and

a processing unit adapted to process at least one analytical artifact (AA) and the machine readable system re lated spatial and/or topological data and/or machine readable system related lifecycle data to analyze automatically a sus ceptibility of system components (C) of said system of inter est (SYS) to common cause failures based on common cause failure influencing factors (CCFIFs) .

18. The system according to claim 17 wherein the processing unit comprises at least one processor adapted to calculate a quantitative common cause failure result (CCFR) on the basis of a machine readable common failure cause model (CCFMOD) provided for the common cause failure influencing factors (CCFIFs) .

19. The system according to claim 18 wherein the processing unit comprises a user interface (UI) adapted to output the calculated common cause failure result (CCFR) and/or a con trol interface adapted to output a control signal in response to the calculated common cause failure result (CCFR) .

20. The system according to any of the preceding claims 17 to 19, wherein the analytical artifact (AA) is generated auto matically by a processing unit on the basis of machine reada ble functional descriptions (mrFD) of system components (C) stored in the database (DB) in response to at least one ap plied system evaluation criterion (SysEC) .

Description:
Description

Method for identifying and evaluating common cause failures of system components

The invention relates to a method and system for identifying and evaluating common cause failures of system components of an investigated technical system of interest such as an in dustrial facility.

A technical system can comprise a plurality of different com ponents, in particular hardware components connected to each other via wired or wireless links. The system can comprise several sub systems which in turn can include different kind of hardware and/or software components to be executed by hardware components. A technical system has to fulfil differ ent kinds of criteria. A technical system has to meet prede termined safety, reliability, availability or maintainability criteria to meet a technical standard or to fulfil obliga tions from a contract. The increasing complexity of technical systems makes it more difficult to develop, analyse, monitor and control them. A technical system can potentially be harm ful to humans or other facilities. Different kinds of safety analysis techniques can be used to assess a potential risk of an industrial system. For instance, failure modes and effects analysis FMEA can be used for failure analysis of an investi gated system. FMEA involves analysis of components, assem blies and sub systems of an investigated system to identify failure modes as well as their causes and effects and to de fine activities such as mitigation measures etc. In a conven tional failure mode and effect analysis for each component the failure modes and their resulting effects on the rest of the system can be recorded in a specific FMEA worksheet.

Conventionally, analytical artifacts, AA, such as an FMEA ta ble documenting the outcome of an FMEA are generated manually by domain experts. Consequently, significant efforts, costs and time for experts are involved for developing and analys ing complex technical systems. Safety, reliability or availability related analytical arti facts, AA, are results when a complex technical systems has been analyzed with respect to the fulfilment of corresponding safety, reliability or availability requirements that may for instance stem from contractual obligations. The safety, reli ability or availability related analytical artifacts, AA, for technical systems such as failure modes and effect analysis, fault tree or Markov chains are conventionally generated manually for instance by reliability, availability, maintain ability or safety (RAMS) experts or related teams. Signifi cant efforts, costs and time must be spent especially in case of complicated or challenging applications.

Common cause failures should be considered and evaluated in safety, reliability and availability analyses. The corre sponding analytical artifacts shall reflect the effects of common cause failures.

Common cause failures are failures resulting from common causes. Common causes can comprise a wide variety of causes relating to technical, lifecycle related and/or spatial or topology aspects between different components C of a complex technical system of interest. Factors which influence common cause failures can comprise design principles, technologies, component manufacturers, component lifecycle resources, com ponent redundancies, component diversities and even component complexities. For instance, in a factory where pressure in a fluid tank is monitored by two different pressure sensor com ponents there can be different kinds of common cause failure influencing factors. For instance, if the two pressure sen sors attached to the fluid tank are implemented by the same sensor technology or use the same measuring principle it is more likely that both sensors fail in a specific failure sce nario. The same is true if both pressure sensors attached to the fluid tank are manufactured by the same component manu facturer. For instance, if the manufacturer has low standards for specific aspects of quality control it is likely that two pressure sensors manufactured by this manufacturer may both fail during specific operational conditions. Even if the sen sors are manufactured by a manufacturer with high quality standards the fact that they originate from the same design and manufacturing processes does increase the probability that both pressure sensors may fail simultaneously e.g. under certain conditions or in a specific state of operation in the technical system. Another possible influencing factor for a common cause failure is that both sensors implement similar measuring techniques. This increases also the probability of a common cause failure. Another possible influencing factor for a common cause failure is the applied lifecycle procedure such as maintenance. For instance, if two sensor components attached to a fluid tank are maintained by the same mainte nance team this does increase the probability that a common cause failure might occur as well e.g. because a faulty re pair or calibration procedure might have been applied to both sensors. This may for instance be loose fixation screws for the sensors leading to susceptibility with respect to oszil- lations, i.e. both sensors could be positioned improperly or fall off simultaneously.

During planning and/or operation of an industrial facility or technical system of interest it is important to identify and evaluate common cause failures of system components in the investigated system. If sources of common cause failures are not identified this can lead to common failures of system components or sub-systems or even to a failure of the entire investigated technical system of interest.

Accordingly, it is an object of the present invention to pro vide a method and system for identifying and evaluating com mon cause failures of system components of a system of inter est .

This object is achieved according to a first aspect of the present invention by a method comprising the features of claim 1. The method provides according to the first aspect of the pre sent invention a method for identifying and evaluating common cause failures, CCF, of system components, C, of a system,

SYS, wherein at least one analytical artifact, AA, a machine readable system related spatial and/or topological data and/or machine readable system related lifecycle data are processed to analyze automatically a susceptibility of system components to common cause failure based on common cause failure influencing factors, CCFIFs.

In a possible embodiment of the method according to the first aspect of the present invention the analytical artifact, AA, comprises a machine readable safety, reliability or avail ability related analytical artifact generated on the basis of a machine readable functional description, mrFD, of the tech nical system of interest, SYS, and at least one system evaluation criterion, Sys-EC.

In a further possible embodiment of the method according to the first aspect of the present invention system components, C, of the technical system of interest, SYS, have associated machine readable functional description including port defi nitions and component failure mode descriptions processed to generate automatically the analytical artifact, AA, in re sponse to at least one applied system evaluation criterion, SYS-EC.

In a possible embodiment of the method according to the first aspect of the present invention the analytical artifact com prises a fault tree, a Markov chain or a combination of fault trees and Markov chains, an FMEA (Failure Modes and Effects Analysis) , an FMECA (Failure Modes, Effects and Criticality Analysis) and/or an FMEDA (Failure Modes, Effects and Diag nostic Analysis (FMEDA) ) .

In a further possible embodiment of the method according to the first aspect of the present invention the system related machine readable lifecycle data comprises data regarding sys tem design, system testing, system history, system component history, training data, data regarding planned or implemented operation procedures and/or maintenance procedures concerning the system of interest.

In a still further possible embodiment of the method accord ing to the first aspect of the present invention a quantita tive common cause failure result is calculated on the basis of a machine readable common cause failure model provided for common cause failure influencing factors, CCFIFs.

In a still further possible embodiment of the method accord ing to the first aspect of the present invention the common cause failure model comprises the IEC 61508 common cause failure model used to calculate a beta factor as a quantita tive common cause failure result.

In a still further possible embodiment of the method accord ing to the first aspect of the present invention a counter measure is automatically triggered in response to the calcu lated common cause failure result.

In a still further possible embodiment of the method accord ing to the first aspect of the present invention the machine readable spatial or topological data comprises data regarding an arrangement and/or placement of system components within the system of interest, in particular position coordinates and/or distances between system components.

In a still further possible embodiment of the method accord ing to the first aspect of the present invention the common cause failure influencing factors related to

spatial proximity,

design principles,

technologies ,

manufacturers ,

lifecycle facts, level of redundancies,

diversities within implemented redundancies, and

complexities .

In a further possible embodiment of the method according to the first aspect of the present invention the system evalua tion criterion comprises

a reliability criterion,

an availability criterion, and/or

a safety criterion.

In a still further possible embodiment of the method accord ing to the first aspect of the present invention the fault tree and/or Markov chain are provided by

transforming at least one system evaluation criterion into one or more corresponding relevant state patterns at ports forming a system boundary of the system of interest, SYS, and by

generating the fault tree and/or Markov chain on the basis of the relevant state patterns and on the basis of the failure modes of the components of the system of interest.

In a still further possible embodiment of the method accord ing to the first aspect of the present invention the analyti cal artifact, AA, is provided by

transforming at least one system evaluation criterion, SYS- EC, into one or more corresponding relevant state patterns at ports of the system boundary and/or at ports inside of the system of interest and by

generating the analytical artifact, AA, on the basis of the relevant state patterns and on the basis of the component failure modes descriptions of the components of the system of interest .

In a still further possible embodiment of the method accord ing to the first aspect of the present invention the method further comprises transforming the system evaluation criterion, SYS-EC, into at least one system state that can be represented by a state pattern,

applying at least one state pattern to input ports and output ports of the system boundary of the system of interest, deriving relevant system failure events by automatically us ing the failure propagation mechanisms based on the machine readable functional descriptions, mrFD, of the system compo nents including their failure mode descriptions, and

assembling the derived failure events to generate the fault tree and/or Markov chain used as an analytical artifact, AA.

In a still further possible embodiment of the method accord ing to the first aspect of the present invention the machine readable functional description, mrFD, of a system component comprises

port definitions of input and output ports of the system com ponent,

component failure modes,

an internal state or states of the component,

a failure rate,

a maintenance activity,

an inspection interval,

a mean down time, and/or

a mean time to repair.

In a still further possible embodiment of the method accord ing to the first aspect of the present invention the system components of the system of interest comprise

hardware components, software components to be executed by hardware components such as CPUs, and embedded components.

The invention further provides according to a second aspect a system for identification and evaluation of common cause failures, CCFs, of a technical system of interest, SYS, com prising the features of claim 17. The invention provides according to the second aspect a sys tem for identification and evaluation of common cause fail ures of a technical system of interest comprising system com ponents, said identification and evaluation system comprising a database which stores a digital twin of the system of in terest including machine readable system related spatial and/or topological data and/or machine readable system re lated lifecycle data and

a processing unit adapted to process at least one analytical artifact, AA, and the machine readable system related spatial and/or topological data and/or machine readable system re lated lifecycle data to analyze automatically a susceptibil ity of system components of said system of interest to common cause failure based on common cause influencing factors, CCFIFs .

In a possible embodiment of the system according to the sec ond aspect of the present invention the processing unit com prises at least one processor adapted to calculate a qualita tive or quantitative common cause failure result, CCFR, on the basis of a machine readable common cause failure model, CCFMOD, provided for common cause failure influencing fac tors, CCFIFs.

In a further possible embodiment of the system according to the second aspect of the present invention the processing unit comprises a user interface, UI, adapted to output the calculated common cause failure result and/or a control in terface adapted to output a control signal in response to the calculated common cause failure result, CCFR.

In a still further possible embodiment of the system accord ing to the second aspect of the present invention an analyti cal artifact, AA, is generated automatically by a processing unit on the basis of machine readable functional descrip tions, mrFD, of system components stored in the database in response to at least one applied system evaluation criterion, SYS-EC. In the following possible embodiments of the different as pects of the present invention are described in more detail with reference to the enclosed figures.

Fig. 1 shows an exemplary technical system of interest which can be defined and investigated by the method and system according to the present invention;

Fig. 2 shows a further exemplary technical system which can be defined and investigated by a method and system according to the present invention;

Fig. 3 shows a flowchart of a possible exemplary embodi ment of a method for providing an analytical arti fact which can be used by a method for identifying and evaluating common cause failures of system com ponents according to an aspect of the present in vention;

Fig. 4 to 18 show tables for illustrating the operation of he method and system according to the present in vention for a specific exemplary technical system according to Fig. 2 to be investigated.

Fig. 19 shows a block diagram for illustrating a possible exemplary embodiment of a system for identification and evaluation of common cause failures, CCFs, of a technical system of interest according to an aspect of the present invention;

Fig. 20 shows a block diagram of an exemplary system of in terest to illustrate the operation of the method and system according to the present invention;

Fig. 21 shows a schematic diagram for illustrating machine readable topological data and/or machine readable spatial data which can be used by the method and system according to the present invention to calcu late a susceptibility of the system components to common cause failure influencing factors;

Fig. 22 to 25 are tables illustrating the operation of the method and system according to the present inven tion for a specific exemplary technical system to be investigated.

As can be seen in the diagram illustrated in Fig. 1 an inves tigated technical system SYS can comprise several components C. The components C can comprise sub components such as switches, sensors or actuators, software components to be executed by hardware components such as CPUs and embedded components comprising both hardware and embedded software to run the respective hardware component. Each component C can comprise input ports and output ports for connecting the re spective component with other components of the defined and investigated system. In the illustrated exemplary system of Fig. 1 the system comprises three components Cl, C2, C3. The system SYS comprises a system boundary SYS-B which comprises the interface ports of the system to other systems. In the illustrated example of Fig. 1 the first component Cl com prises an input port to receive a signal and two output ports which are connected to other components C2, C3 of the inves tigated system SYS. The second component C2 comprises a sin gle input port connected to the first output port of the first component Cl. The second component C2 comprises a sin gle output port which forms an external port or interface at the system boundary SYS-B of the system. The third component C3 also comprises a single input port and a single output port. The single input port of the component C3 is connected internally to the second output port of the first component Cl as shown in Fig. 1. The output port of the third component C3 forms an external port or interface at the system boundary SYS-B of the investigated system SYS. As can be seen in Fig.

1 the system boundary SYS-B comprises in the illustrated em bodiment three ports, i.e. the input port of the first compo- nent Cl, the output port of the second component C2 and the output port of the third component C3. The components C within the system SYS can comprise different kinds of compo nents C including hardware components, software components to be executed by hardware components such as CPUs and embedded components. The hardware components can comprise all analogue or digital components. The components C, C2, C3 each have an associated machine readable functional description including the port definitions as well as component failure mode de scriptions which are processed to generate automatically an analytical artifact used for development and/or analysis of the investigated technical system of interest SYS in response to at least one applied system evaluation criterion. These analytical artifacts comprise in a possible embodiment a fault tree, a Markov chain, a combination of fault tree(s) and Markov chain (s), an FMEA table, or an FMECA table, an FMEDA table. These analytical artifacts are generated auto matically on the basis of a full functional description of the system of interest SYS including a machine readable de scription of the failure modes of its components or sub sys tems. Further, the analytical artifacts can be generated for different kinds of evaluation criteria such as for instance safety, reliability, maintainability and/or availability cri teria .

The different components C such as components Cl, C2, C3 of the system SYS illustrated in Fig. 1 can be supplied by dif ferent suppliers which offer their technical components or sub systems with a standardized generic functional descrip tion which is machine readable. It is also possible to use a semiformal functional description with OMG SysML or AADL or EAST-ADL. The functional description does include input and output port definitions and additional information data about failure modes, their causes and corresponding failure proper ties, in particular failure rates, preventive and corrective maintenance activities and test-related data. In a possible embodiment each component C comprises an associated func tional description. This functional description of the compo- nent C can comprise the port definitions of the input and output ports of the respective component C, all component failure modes of the component, at least one internal state of the respective component, a failure rate, maintenance ac tivities, an inspection interval and/or a mean down time (MDT) and/or a mean time to repair (MTTR) of the component.

The different vendors or suppliers providing components C for the complex technical system SYS can in future use a stan dardized common generic functional description of the respec tive component C which can be stored in a memory. In a possi ble embodiment the machine readable functional description of the respective component C within the system SYS of interest can be stored in a local memory of the component C. The local memory can be integrated within the component. In a possible embodiment the functional description can be read from the memory by a reader. In a still further possible embodiment the functional description for a component can be output via ports of the system SYS in response to a specific request ap plied to the respective component or even applied to the sys tem SYS. In a specific implementation the system boundary SYS-B may comprise a specific pin or port used to read out a functional description from different components C of the in vestigated system SYS. In an alternative implementation the functional description can also be read out from a local mem ory output by the component using internal connections and an output port of a component of the system SYS such as the out put port of the second or third components C2, C3 illustrated in the example of Fig. 1. For instance, the functional de scription of components Cl, C2 can be read from the output port of component C2 and the functional description of compo nents Cl, C3 can be read from the output port of component C3.

The functional descriptions extracted from the system SYS of interest can be supplied to a processor or processing unit of an investigating system which can be used for analysing, monitoring and/or even controlling the technical system SYS of interest or a larger system that may comprise the techni cal system SYS using at least one automatically generated analytical artifact. This analytical artifact can be gener ated from the associated machine readable functional descrip tions of the different components C within the investigated system SYS. In an alternative embodiment machine readable functional description of a component C within the system SYS of interest can also be stored in a cloud, in a remote data storage or database. Each component C of the investigated system SYS comprises an associated machine readable func tional description. The different machine readable functional descriptions of all components C forming part of the investi gated system SYS can be evaluated or processed by a process ing unit to generate automatically at least one analytical artifact for at least one applied system evaluation crite rion. This analytical artifact can be used for developing and/or analysing and/or controlling the investigated techni cal system SYS or a larger system that may comprise the sys tem SYS. In the illustrated example of Fig. 1 the overall system description also includes the connection from the first output port of the first component Cl to the input port of component C2 as well as the connection from the second output port of component Cl to the input port of component C3. It is also possible to embed the overall functional de scription into the corresponding system environment and take into account the relevant operational conditions of the sys tem. In a preferred embodiment a composition of components or sub systems to the overall investigated system SYS can be conducted automatically. The generic functional description of the component or sub system can reflect standardized im plementation concepts or architectural patterns, e.g. with respect to sensor circuit, signalling or communication proto cols which facilitates the system development. The relevant component functionality of the system component can be se lected and enriched using architectural patterns if needed in the course of the system definition and system development.

It is possible that application classes for programmable or configurable components are predefined for selection from a repository .

In a possible embodiment system-specific information data can be added. This may for instance relate to a fact that an un contained fire leads to a very long mean down time MDT of the investigated system since the whole system has to be rebuilt. Critical combinations of the events can also be specified with corresponding mean down times MDT in order to assist a complex safety, reliability or availability analysis. A more comprehensive description of the component or sub system how ever may already include these kinds of dependencies in the functional description of the respective components.

The automatically generated analytical artifact can be used for a safety analysis, a reliability analysis, an availabil ity analysis, a maintainability analysis or further evalua tions of the system. Accordingly, the investigated system can be investigated to check whether the system fulfils certain system evaluation criteria. These system evaluation criteria can comprise reliability criteria, availability criteria, maintainability criteria, safety criteria or further evalua tion criteria. In a possible embodiment the system evaluation criterion can be derived automatically from a technical safety standard or contractual obligations stored in a data base. Further, the system evaluation criteria can be derived in a possible embodiment automatically from a machine read able contract. For example, a safety criterion may be based on a national or international standard and a reliability criterion, a maintainability criterion or an availability criterion can be based on specific contractual obligations that can be transformed and translated into at least one state pattern. The corresponding state patterns have to be applied to relevant input ports and/or output ports of the investigated system. With these assignments it is possible to derive relevant failure events by following the method de scribed below. The derived relevant failure events are then assembled to the required safety, reliability or availability related analytical artifact, as for instance a fault tree model. The imminent failure propagation has to be respected in order to identify the relevant failure events and to as semble the corresponding fault tree. The procedure can be conducted automatically by resorting to the overall func tional system or solution description or to a subset of it depending on the evaluation criterion. The generated analyti cal artifact can comprise in a possible embodiment a fault tree or a Markov chain. A fault tree is generated automati cally on the basis of the machine readable functional de scriptions of all components C forming part of the investi gated system SYS. These functional descriptions include port definitions and component failure modes of the different com ponents. A failure mode relates to a specific manner or way of which a failure occurs. A failure mode can describe a failure state of the respective item (or function) under con sideration. A failure mode is the result of a failure mecha nism. The components' or sub systems' failure modes can con tribute directly or according to a more sophisticated logic to failure events on system level. This may involve logic combinations with other components C or sub systems of the investigated system SYS according to a specifically imple mented evaluation logic.

In a possible embodiment the functional description of each component C includes data about failure modes, FM, as well as about preventive and/or corrective maintenance activities and even test-related data on component or sub system level. This information can be aggregated in order to establish mainte nance related artifacts such as required resources, required tooling, training requirements or corresponding plans. Pre ventive and/or corrective maintenance activities can specify tools, skills, training requirements or repair time etc. In order to determine a relevant down time it is required to re spect specific circumstances of the operational concept such as additional logistical overhead etc. Data with respect to the failure modes can also indicate if and how those failure modes can be detected. This data can be used for generation of test cases.

With respect to a safety analysis of the investigated system SYS relevant safety functions can explicitly be defined with their functional dependencies in course of the system devel opment .

The standardized nature of most safety criteria can be ex ploited to prepare an ontology to be used for the system de sign in order to match the corresponding terms or at least in order to facilitate associated mapping processes between the functional description of the system and the corresponding state pattern. In case that the safety, reliability, avail ability or maintainability related analytical artifacts do not meet predetermined quantitative target values it may be required to further detail the functional description or even to modify the system architecture or the operational concept and to repeat the analysis thereafter until the required tar get values are reached.

Fig. 2 shows a diagram for illustrating a possible exemplary embodiment of a system SYS of interest to be investigated.

The illustrated system of Fig. 2 comprises a system to detect fire and to disconnect a target system from a high voltage power supply grid. The target system can comprise one or sev eral power consumption entities and can be disconnected from the high voltage power supply grid by an actuator component C4 of the investigated technical system SYS. The investigated system SYS comprises in the illustrated embodiment four dif ferent components Cl, C2, C3, C4. The investigated system consists of two infrared sensors Cl, C2, a CPU component C3 and an actuator component C4. The actuator component C4 is capable of opening and closing a connection to an external high voltage power supply grid HVG. This is a measure estab lished in order to contain or to support the extinction of fire . A potential source of infrared radiation IR which may be caused by fire shows statistical behaviour. In the illus trated example the input ports of the infrared sensor compo nents Cl, C2 are facing the potential IR source (fire) . In the illustrated figure of Fig. 2 the arrows indicate if the respective port forms an input port or an output port.

Both sensor components Cl, C2 are dedicated to sense the in frared radiation source and to transmit a message "fire" or "no fire" to the connected CPU controller C3. Each sensor Cl, C2 can have implemented the functionality to detect internal sensor failures and to transmit the signal "internal failure fire sensor" in these cases. For instance, both components Cl, C2 receive a supply voltage from external power supply sources SUP-C1, SUP-C2 as illustrated in Fig. 2. Further, the CPU component C3 also receives an external power supply source SUP-C3 outside the system boundary SYS-B of the inves tigated system SYS.

A logic can be implemented in the CPU component C3 as fol lows. A "HIGH" output level toward the actuator component C4 is only activated in case that both fire detector components Cl, C2 transmit a "no fire" signal to the CPU component C3.

In other cases the output level at the output port of the CPU component C3 is set to "LOW". Both sensor components Cl, C2 as well as the CPU component C3 require dedicated power sup plies to function as intended. Lacking power supply leads to "no signal" at the output port of the fire detector compo nents Cl, C2 or to a "LOW" signal at the CPU output port of component C3 respectively.

The detection system SYS as illustrated in Fig. 2 has to meet different kinds of predefined evaluation criteria. For exam ple, a possible safety criterion might be that the high volt age grid HVG has to be switched off by the actuator C4 if fire is detected. The system failure criterion can be trans formed in a possible embodiment into one or more correspond ing relevant state patterns at the ports forming the system boundary SYS-B of the investigated system SYS. The states at the different ports of the system boundary SYS-B comprise failure criterion fulfilling states and failure criterion not fulfilling states. In a possible embodiment the state pattern can be binary, i.e. high/low.

The potential IR source representing the fire can be repre sented by a state "1" (high) applied to the input ports of the sensor components Cl, C2. If the investigated system SYS operates correctly, the output port of the actuator component C4 has a state "0" (low) . Accordingly, a binary state pattern at these three ports "1-1-0" indicates a correct operation of the system with respect to the system function described above. The safety criterion in this case is not fulfilled in the former case. A state pattern such as "1-1-1" indicating that the investigating system does not switch off the high voltage grid HVG although both sensor components Cl, C2 face an IR source indicates that the safety failure criterion is fulfilled and that the safety function "switch off the high voltage grid HVG by the actuator C4 if fire is detected" of the system SYS has failed. The states of all other ports of system SYS are arbitrary with respect to the evaluation cri terion chosen above i.e. all states of these ports have to be considered. For reasons of simplicity these ports are not de picted here in the corresponding pattern representation. In order to express that all states of a certain port matter one can for instance choose a representation "X" for the corre sponding port in order to express that e.g. in case of binary patterns both states "1" and "0" apply.

Different kind of state patterns can be generated for differ ent system evaluation criteria including safety criteria, re liability criteria, availability criteria and maintainability criteria. In a possible embodiment at least one system evaluation criterion is automatically transformed into one or more corresponding relevant state patterns at the ports form ing the system boundary SYS-B and/or at internal system ports of the investigated system SYS. In a further step a fault tree and/or Markov chain can be generated on the basis of the relevant state patterns and on the basis of the component failure modes of the components C forming part of the inves tigated system SYS. In the illustrated example of Fig. 2 a fault tree can be generated on the basis of relevant state patterns derived from at least one system failure criterion and on the basis of the component failure modes of the compo nents Cl, C2, C3, C4 specified in the associated machine readable functional descriptions of the respective components Cl to C4.

As illustrated in the flowchart of Fig. 3 the automatic gen eration of an analytical artifact, AA, such as a fault tree FT can be performed in different main steps.

In a first step SI the respective system evaluation criterion is transformed into at least one state definition which may for instance be presented in the form of a binary state pat tern. The corresponding evaluation or failure criterion for system SYS can demand that the system is or remains connected to the high voltage power supply grid in case of a fire. In a specific embodiment the system evaluation criterion can be a text based system evaluation criterion which is transformed automatically by a linguistic transformation program into a corresponding relevant state pattern such as a binary or multi-level state pattern.

In a further step S2 the generated state pattern and hence the evaluation criterion is applied to input ports and output ports at the system boundary SYS-B of the system SYS or to internal ports of the system of interest. This may comprise that both fire detector components Cl and C2 are exposed to infrared radiation (IR) due to fire and that the actuator component C4 is or remains in the state "closed" or "con nected to the high voltage power supply grid (HVG)".

In a further step S3 all combinations of events that satisfy the input and output ports state pattern and hence the evaluation criterion are derived by automatically taking into account the failure propagation mechanisms and assembled to generate the fault tree which forms an analytical artifact that can be used for development and/or analysis of the in vestigated technical system. The combination of events that satisfy the evaluation criterion is based on the component failure modes and the functional dependencies of the system SYS .

The generated analytical artifact, AA, can be used by the method according to the present invention to identify and evaluate common cause failures, CCFs, of system components of an investigated system. The analytical artifact, AA, can also be provided by an external source in a machine readable for mat .

The following tables illustrate examples for functional de scriptions of the components C of the investigated system as illustrated in Fig. 2. To demonstrate the process of both the generic version as it may be delivered by a component sup plier or vendor as well as the instantiated version including required modifications, i.e. the implemented version of the component within the system SYS are depicted in the tables if "instantiation" is indicated. It is the lower of the pre sented tables with the exception of the actuator table where for reasons of required space the tables are split on two separate pages.

The failure rates and mean down times MDTs within the tables are fictitious and are only used for demonstration purposes. Further, every function and for every failure mode corres ponding preventive and/or corrective maintenance activities are indicated. By knowing the relevant operational concept it is assumed that the corresponding mean down times MTDs can be calculated from the data of the functional descriptions in cluding port definitions and component failure mode descrip tions , e.g. by adding an appropriate overhead for travel, access or setup. The mean down times MDT for the instantiated components C are a function of the generic values and can be implemented in the operational concept.

Table 1 illustrated in Fig. 4 shows a functional description of the power supply component for the fire sensors Cl, C2. Table 1 shows the functional description of the generic com ponents and the functional description of the instantiated and modified power supply component. The functional descrip tion of the instantiated and modified component emerges from adapting the generic functional description with respect to the specific requirements of the system in interest.

Table 2 illustrated in Fig. 5 shows the functional descrip tion for the power supply for the CPU component C3 in the ge neric and instantiated and modified form.

Table 3 illustrated in Fig. 6A, 6B, 6C shows a functional de scription for the fire sensor components Cl, C2 of the inves tigated system both in generic and instantiated form.

Table 4 illustrated in Fig. 7A to 7F shows a functional de scription of the CPU component C3 in generic and instantiated and modified form. The instantiated and modified functional description accounts for the specific requirements of the system to be built. This entails the relevant logic to be im plemented on the CPU.

Further, table 5 in Fig. 8A, 8B illustrates the functional description of the actuator component C4 in generic form.

Further, table 6 of Fig. 9A, 9B illustrates the functional description of the actuator component C4 in instantiated form.

Failure modes connected to the rows with the indication "no" with respect to "fault detection" can be omitted or filtered since for the corresponding operation request the dormant or sleeping fault is irrelevant. In a further possible embodiment one may choose to omit the corresponding lines from the functional description.

For demonstrating the operation of a method according to the present invention two different exemplary evaluation criteria may be applied to the investigated system SYS as illustrated in Fig . 2.

For example, the following system safety criterion 1 can be applied: "In case of fire the system shall be disconnected from the high voltage power supply grid with a tolerable haz ard rate of THR_Fire_l . "

The definition above implies a present fire. This fire en tails an infrared radiation at the system boundary SYS-B of the investigated system SYS at the input ports of the fire sensor components Cl, C2 shown in Fig. 2. Therefore it is possible to exclude the corresponding states at the input ports of the fire sensor components that refer to "IR not present" .

The system output side is represented by the output port of the actuator component C4. Only in case of a closed connec tion between the output port of the actuator component C4 to wards the high voltage power supply grid HVG a violation of the specified safety criterion is possible. In a possible im plementation by means of a filter function one gets the po tential failure states or events as indicated in following table 7 by excluding all states referring to an open output connection and by excluding sleeping failure modes that are irrelevant to the corresponding switching request and hence cannot be detected in course of this operation.

Table 7 shown in Fig. 10A, 10B illustrates the relevant states for an exemplary system failure criterion.

One can identify two failure modes of the actuator from table 7 that contribute to the event according to a potential vio- lation of the exemplary safety criterion which can be re ferred to as a safety function failure. With respect to fail ure propagation it is additionally possible that functions that constitute output to the actuator's input also contrib ute to system failures according to the safety criterion 1. However as can be seen from table 7 above this is only possi ble in case a "HIGH" level is applied to the input port of actuator component C4. The input port of the actuator compo nent C4 is elected and corresponds to the output port of the CPU component C3 as illustrated in Fig. 2. Consequently the CPU's table can be filtered accordingly.

Table 8 shown in Fig. 11A, 11B illustrates the relevant states for the CPU component C3 for a potential violation of the above exemplary system safety criterion 1.

On the basis of the functional description of table 8 it is possible to identify four additional failure modes of the CPU component C3 that can contribute to the safety failure crite rion. Moreover, functions delivering input to the CPU compo nent C3 can also contribute to failures on a system level by means of failure propagation. This however is only possible as long as the input ports of the CPU component C3 are con nected to the fire sensing components Cl, C2 both receive the signal "no fire" and the CPU power supply works.

By taking into account the fact that a fire and hence infra red radiation must be present in order to be relevant for the above given exemplary safety criterion 1 gets table 9 for the fire sensor component Cl .

Table 9 shown in Fig. 12A, 12B illustrates relevant states of the fire sensor component Cl for the above given exemplary system safety criterion 1. The same applies to fire sensor component C2.

From the above analysis it can be derived that only one fail ure mode as depicted in table 9 contributes to the safety criterion on sensor level. This failure mode is only relevant in case both sensor components erroneously transmit the "no fire" signal to the CPU component C3 due to "missed detec tion". The evaluation logic ensures that only if both sensor components transmit the "no fire" signal to the CPU component C3 a contribution to the analytical artifact according to a potential violation of safety criterion 1 is possible.

The analytical artifact representing a fault tree for a safe ty function can now be written as below. The OR -operator represents an OR gate of the elements listed in between its brackets. This -operator can also be applied to just one element. The corresponding table and its lines for the relevant failure mode are indicated in parenthesis.

Fault Tree

(System "Fire Detection And Disconnection From The Grid", Ap plication of "Safety Criterion 1")

TOPEVENT ( Safety Criterion 1)=

(Fire sensor S1 : OR [24 ] AND Fire sensor S2 : OR [24 ] ) OR 25, 27, 28] OR

Actuator : OR [ 32 , 37]

The numbers in the brackets illustrate rows within the above tables corresponding to failure modes of the respective com ponents .

A fault tree is generated automatically for the TOPEVENT cor responding to a potential violation of the system safety cri terion 1.

A quantitative fault tree evaluation that may be executed by means of commercial fault tree calculation software will yield a result that has to be compared with the safety crite rion 1 in order to finally determine whether the safety cri terion has been achieved or failed. The same applies to the fault trees below. The compilation represents a conservative approximation. This means that the effects of failure modes are entirely propa gated to the system output even though other failure events on this way could diminish their influence. The method dis closed in this patent application can be used to calculate the results precisely by taking into account those failure modes that prevent the investigated system to show a malfunc tion according to the definition of a certain criterion. It is possible to calculate the effects of combined failure modes where one failure mode prohibits another failure mode to propagate. Consequently, a more general description is the following generated fault tree:

Fault Tree

(System "Fire Detection And Disconnection From The Grid", Application Of "Safety Criterion 1", detailed)

TOPEVENT ( Safety Criterion 1)=

(Fire sensor S1 : OR [24 ] ) AND (Fire sensor S2 : OR [24 ] )

(CPU: OR [24, 25, 27, 28] ANDNOT (Actuator : OR [ 31 , 36 ]) ) OR

Actuator : OR [ 32 , 37 ]

Please note that the events "Fire sensor S1:OR[20]", "Fire sensor S2:OR[20]" as well as "CPU:OR[18]" correspond to events that are outside of the relevant system boundary SYS- B. In many cases these external events can be considered as being ideal, i.e. without any failure mode. Nevertheless, the method according to the present invention allows to respect the influence in case that corresponding probabilistic data is given or can be estimated.

By means of the method according to the present invention the relevant failure modes can be selected and composed according to a relevant state pattern. With respect to computing the corresponding results one has to take into account the nature of the failure modes, i.e. dormant or sleeping failures have to be calculated by taking into account their test or opera tional interval and hence their so-called time at risk (TAR) . For illustrating the operation of the method according to the present invention a second exemplary evaluation criterion with respect to reliability may be applied to the system.

The corresponding reliability criterion 2 is given as fol lows :

"The mean number of unintended disconnections from the high voltage power supply grid per year caused by the system shall not exceed NUD (number of unintended disconnec

tions) _PerCalendarYear_l "

Note that this definition of the criterion only focuses on disconnections that are caused without a fire being present since a disconnection due to fire is intended. Moreover being unable to reconnect to the high voltage power supply grid HVG (e.g. due to a dormant fault) is excluded by definition of the reliability criterion.

Table 10 of Fig. 13A, 13B shows the relevant states of the actuator component C4 of the system SYS for a potential vio lation of the reliability criterion stated above.

Hence the actuator component C4 contributes to the relevant events only with its failure mode "opens uncommandedly" . All other failure modes can be excluded with respect to this re liability criterion.

Moreover it can be derived from the list of dataset above that only a "LOW" signal level that is equivalent to "no sig nal" at the input port of the actuator C4 needs to be regard ed with respect to a potential failure propagation. Since the input port of the actuator C4 corresponds to the output port of the CPU component C3 the CPU list can be filtered accord ingly. Table 11 shown in Fig. 14A, 14B, 14C below illustrates the relevant states of the CPU component C3 for the reliability criterion stated above.

Failure modes of the CPU component C3 that have to be re flected in this reliability related analytical artifact are the ones in lines 20 and 26. Lines 19, 22 and 23 in table 11 represent potentially propagated failure modes of functions connected to the input port of the CPU component.

According to Fig. 2 the power supply of the CPU component C3 is located outside the relevant system boundary SYS-B. Conse quently faults caused by the CPU power supply component (or function) SUP-C3 do not relate to the reliability criterion 2 and shall not be attributed to the investigated system SYS as indicated in line 19 of table 11 and hence can be left out for generating the corresponding analytical artifact.

Contrary to this the events indicated in lines 22, 23 of ta ble 11 represent output states of the fire sensor components Cl, C2. Those are located inside the relevant system boundary and hence belong to the investigated system SYS as illustrat ed in Fig . 2.

It can be derived that the relevant input signal on the input ports respectively must be NOT "no fire" which means that all other signal apart from the "no fire" signal must be regard ed. The power supplies SUP-C2, SUP-C2 for the sensor compo nents Cl, C2 are both located outside the relevant system boundary SYS-B. Hence their contributions can be ignored with respect to the reliability related analytical artifact. The remaining failure events caused by sensor component Cl are depicted in Table 12.

Table 12 shown in Fig. 15A, 15B illustrates the relevant states of fire sensor component Cl for a potential violation of the above stated reliability criterion 2. A situation with an infrared signal IR being present at input port of sensor component Cl must be excluded since this cor responds to a normal intended sensor operation and does not constitute an unintended disconnection from the high voltage power supply grid HVG. The same holds for fire sensor compo nent C2.

Please note that all remaining failure modes of the list of both sensor components do contribute to the analytical arti fact since the implemented CPU logic outputs a "LOW" level at its output port in case that any of the transmitted fire sen sor signals deviates from the "no fire" signal.

The analytical artifact formed by a fault tree can now be written as below. The corresponding table and its lines for the relevant failure modes are indicated in parenthesis.

Fault Tree

(System "Fire Detection And Disconnection From The Grid", Ap plication Of "Reliability Criterion 2")

TOPEVENT (Reliability Criterion 2)=

[21 , 25 , 26, 27 ] OR Fire sensor

] OR

Actuator : OR [36]

A quantitative fault tree evaluation by means of e.g. a com mercial fault tree calculation software will enable to deter mine if the reliability criterion 2 has been met or failed.

The method according to the present invention can also be used to support the generation of FMEA artifacts or its de rivatives such as Failure Modes and Effects and Criticality Analysis (FMECA) or a Failure Modes, Effects and Diagnostic Analysis (FMEDA) . On the basis of generated or manually elaborated FMEA arti facts it is also possible to generate, elaborate or detail relevant functional descriptions. For FMEA analysis and cor responding derivatives single faults or failures can be in vestigated while the remaining system is supposed to function perfectly .

The aspect of "occurrence" can be addressed by means of a failure rate or the probability of a certain failure mode. Specific environmental conditions may be involved such as for instance temperature, load, switching rate etc.

The aspect of "detectability" can be addressed by correspond ing control functionality of the system, by operations dis closing certain failure states, by self testing or by means of preventive maintenance measures including for instance in spections or testing procedures in order to discover dormant faults within the system.

Further, the aspect of "severity" can be addressed by means of functional dependencies. However its significance can be defined application-specific. A certain classification can be prepared in advance and then applied to the different output results .

It is possible that a supplier of a component may add a de scription of the failure causes for the failure modes of the components or sub systems delivered to the system manufactur er. This allows a more focused procedure during system devel opment, e.g. by means of adding dedicated sensor circuits or enforcing quality measures to detect failures earlier and more reliably.

For the elaborated example a severity classification is used. It may be assumed that an unwanted disconnection from the high voltage power supply grid HVG be rated with a number of e.g. "5" while a fire event that does not lead to a discon nection from the high voltage power grid is rated with a num- ber "9" due to its potential devastating nature and its long reinstatement time. Moreover, it can be supposed that a de tected fire can be fought quickly and hence be kept well con tained. All other failures may be rated with a value "2".

The FMEA severity classification described above corresponds to certain system states or state patterns. Those state pat terns can be used to classify the severity on system failures in an automated way.

The failure modes of fire sensor component Cl are displayed in table 13 shown in Fig. 16A, 16B. Table 13 illustrates the failure modes of fire sensor component Cl for FMEA Analysis.

The logic implemented in the CPU component C3 can evaluate the fire sensor component output signals and can initiate a disconnection from the high voltage power supply grid HVG if the signal of fire sensor component Cl deviates from "no fire". Hence all events illustrated in table 13 with the ex ception of line 24 can be rated with a number "5" directly because they lead to a disconnection from the power supply grid HVG. In case of line 24 of table 13 an infrared source (such as a fire) must be present. With the assumption of only a single failure being present in the entire system SYS it can be concluded that fire sensor component C2 works, detects the infrared source and transmits the fire signal to the CPU component C3. The CPU component C3 in turn can command a dis connection from the high voltage power grid HVG which results in a value "5" rating again. Consequently, all single faults of the fire sensor component Cl lead to a number "5" rating. The same applies to fire sensor component C2.

Failure modes of the CPU component C3 of the system SYS are displayed in table 14. Table 14 shown in Fig. 17A, 17B illus trates the failure modes of the CPU component C3 for FMEA analysis .

The failure modes indicated in lines 19 and 25 of table 14 lead to a number "5" rating. The events indicated in all oth- er lines of the table lead to a number "9" rating since an infrared source must be present. This is the case because the fire sensors must transmit the "fire" signal due to the sin gle failure assumption as indicated before and hence also a fire must be present.

Failure modes of the actuator component C4 are indicated in table 15. Table 15 shown in Fig. 18A, 18B illustrates the failure modes of the actuator component C4 of the system for FMEA analysis.

The dormant or sleeping faults according to the lines 28, 29, 30, 33, 34, 35 of table 15 cannot be detected. Here a normal operation is given despite a failure being present. According to the severity classification described above those failure modes receive a number "2" rating.

The failure mode according to line 32 of table 15 is rated with value "9" because the disconnection from the high volt age power grid HVG is not conducted in case of a present fire ("LOW" level at input represents a fire in case of single faults analysis) .

The same holds for failure modes according to line 37 of ta ble 15 since a connection to the power supply grid HVG is reestablished with a fire being present i.e. the actuator component C4 receives a "LOW" signal which stands for a pre sent fire under the assumption of single system faults.

The failure mode with respect to line 31 of table 15 repre sents a rating of "2" since the actuator component C4 simply remains open.

The failure mode FM according to line 36 of table 15 has to be rated with a value "5" event since it is an uncommanded disconnection from the high voltage power grid HVG. For rea sons of simplicity failures of the power supplies for the fire sensor components Cl, C2 or of the power supply for the CPU component C3 have not been investigated since they form elements outside the relevant system boundary SYS-B.

It is possible to enrich the functional descriptions of the components C with performance the prosattributes such as for instance corresponding time durations. For example maximum durations can be defined. For the exemplary system as illus trated in Fig. 2 the following maximum durations might be given :

Fire sensor components Cl and C2 : detection of infrared source and transmitting a signal: 150 ms

CPU component C3 : processing input signals and commanding output signal: 50 ms

Actuator component C4 : disconnection from the high voltage power supply grid: 100 ms

A further modified safety criterion can be formulated as fol lows :

"In case of fire the system shall be disconnected from the high voltage power supply grid with a tolerable hazard rate of THF_Fire_l within one second" .

The maximum duration of the coupled functions amounts to 300 ms (150 + 50 + 100 ms) which is shorter than one second. As a consequence, the above safety criterion can be met by the system of interest. If the relevant safety criterion indicat ed a reaction time below 300 ms it would be required to per form amendments on component level or to alter the architec ture of the investigated system in order to achieve a shorter overall reaction time which fits to the requirement of the modified safety criterion. In case a remaining time budget as indicated above for safety criterion is provided it is possi ble to optimize the investigated system. One solution can be that the CPU component C3 evaluates the received fire signals for a longer period of time before commanding a disconnection signal to the actuator component C4. This will reduce the number of false alarms by resorting to mean values.

In a possible embodiment the generated analytical artifact, AA, can comprise a Markov chain. This analytical artifact,

AA, can be required in case that the sequence of events is relevant and a more detailed modeling is required to achieve certain target values. Accordingly, it can be possible to im plement a Markov chain logic for the modeling of a safety, reliability or availability related artifact. In this embodi ment additional knowledge on system level can be introduced. This can be for instance the fact that the mean down time MDT of the investigated system SYS in case of an undetected fire is significantly longer than those of all other failure states. Corresponding critical sequences can be specified.

The Markov chain approach is far-reaching since every fault tree can be described by a Markov chain of first order.

The method for providing an analytical artifact, AA, can be formed by a corresponding program or tool to generate one or several analytical artifacts. With the method according to the present invention it is possible to generate one or more analytical artifacts faster. Moreover the generated analyti cal artifacts AA are more precise and have a higher level of consistency. The analytical artifacts AA can be generated with less effort. The analytical artifacts can be generated by means of any kind of evaluation criterion e.g. a safety, a reliability, an availability or a maintainability criterion. Failure modes can prevent other failure modes from causing failures on system level and can also be recognized in order to receive more precise results of the respective analytical artifacts .

After system definition and selection of the relevant system port states according to the applied evaluation criterion manual steps are no longer required to derive or to generate automatically the analytical artifacts. The method and system can also be applied to dynamic systems of interest. For cer tain applications the process of selection of the relevant system states may be supported or partly or even totally be conducted automatically.

The complex investigated systems SYS can comprise any kind of technical systems such as production plants, factories, fa cilities, power distribution systems, vehicles etc. Conse quently there is a huge variety of different use cases for the method according to the present invention. In a possible embodiment the generated analytical artifact AA can be stored for further processing in a memory or database. The generated analytical artifact can be stored in a repository for further use. Moreover the generated analytical artifact AA can also be processed to monitor and/or control automatically compo nents of the system SYS of interest, the entire system or a system that comprises the system depending on an evaluation result of the processed analytical artifact. Accordingly, the invention further provides a system for analyzing, monitoring and/or controlling any kind of technical system of interest SYS which can comprise a plurality of different components. The different components C of the investigated system SYS have ports which are connected to each other via wired or wireless links and comprise associated machine readable func tional descriptions which may be stored in a local or remote memory. The associated machine readable functional descrip tions mrFD comprise port definitions and failure modes (FM) descriptions which are processed by a processor to generate automatically at least one analytical artifact AA. The gener ated analytical artifact AA can in turn be processed to ana lyze, monitor and/or control the system of interest. The ana lytical artifact AA can also be read from a database or re pository .

In a possible embodiment the analytical artifact AA can be generated in the development phase of an investigated system SYS consisting of components C or subsystems connected to each other according to a stored system model of the devel- oped system. This data model of the investigated system can be stored in a database of the development system. Further, a prototype of the investigated system can be analyzed using analytical artifacts AA. Further, a technical system can also be monitored or even controlled by processing analytical ar tifacts AA. When exchanging components C within a system SYS the corresponding analytical artifacts AA generated for the respective system can be adapted or reconfigured automatical ly for the new system. Calculation and generation of the ana lytical artifacts can be performed in a possible embodiment during operation of the investigated system even in real time. The manufactures of the different components C within the system SYS can provide the associated functional descrip tions of the respective component C in a possible embodiment online via internet to the analytical system used for analy zing the respective system of interest. The machine readable functional description of the respective component can be in tegrated within local memory of the manufactured component C and can be read out by a processing unit of the analytical system analyzing the investigated system including the re spective components.

Fig. 19 shows an exemplary embodiment of a system for identi fication and evaluation of common cause failures CCF of a technical system of interest SYS comprised of subsystems and system components C. The identification and evaluation system according to the present invention comprises in the illus trated embodiment a database DB which stores a digital twin of the system of interest. The digital twin of the system of interest does include machine readable system related spatial (mrSD) and/or topological data (mrTD) and/or machine readable system related lifecycle data (mrLD) . The machine readable spatial or topological data can comprise in a possible embod iment data regarding an arrangement and/or placement of sys tem components C of the system of interest. For example, ma chine readable spatial or topological data can comprise posi tion coordinates and/or distances between systems components of the investigated system. The spatial or topological data can comprise data about system or plant site decomposition, in particular location or segments of systems, subsystems and/or system components. The investigated system can be any kind of technical or industrial system consisting of sub systems and hardware and/or software components. The system of interest can comprise hardware components, software compo nents to be executed by hardware components such as CPUs as well as embedded components. The database DB further stores machine readable system related lifecycle data mrLD. Machine readable lifecycle data can comprise for instance data re garding system design, system testing, system history, system component history, training data, data regarding planned or implemented operation procedures and/or maintenance proce dures concerning the respective system of interest.

The digital twin of the system of interest can also comprise in a possible embodiment machine readable functional data mrFD as illustrated in the block diagram of Fig. 19. On the basis of the machine readable functional data mrFD forming a functional description of the technical system of interest, a first processing unit PU1 can generate a machine readable safety, reliability or availability related analytical arti fact AA in response to at least one system evaluation crite rion SYS-EC as shown in Fig. 19. The generated analytical ar tifact AA can comprise in a possible embodiment a fault tree, Markov chain and/or a combination of fault trees and Markov chains. The generated analytical artifact AA can be stored temporarily in a memory for further processing as illustrated in Fig. 19. The machine readable functional description mrFD can comprise in a possible embodiment port definitions and component failure mode descriptions processed by the first processing unit PU1 to generate automatically the analytical artifact AA in response to the at least one applied system evaluation criterion. The system evaluation criterion SYS-EC can comprise for instance a reliability criterion, an availa bility criterion and/or a safety criterion. In a possible em bodiment the first processing unit PU1 generates automatical ly a fault tree FT and/or a Markov chain. In a possible em- bodiment the first processing unit PU1 transforms at least one system evaluation criterion into one or more correspond ing relevant state patterns at ports forming a system bounda ry of the system of interest SYS and then generates the fault tree and/or Markov chain on the basis of the relevant state patterns and on the basis of the component failure modes of the components of the investigated system of interest. In a possible embodiment the system evaluation criterion SYS-EC is transformed into at least one system state that can be repre sented by a state pattern. Then at least one state pattern is applied to input ports and output ports of the system bounda ry of the system of interest. Further, relevant system fail ure events are derived by automatically taking into account or using failure propagation mechanism based on the machine readable functional description mrFD of the system components including their failure mode descriptions. Finally, the de rived failure events are assembled by the first processing unit PU1 to generate the fault tree and/or Markov chain used as an analytical artifact AA which is stored temporarily in a memory for further processing. The machine readable function al description mrFD of the system components used by the first processing unit PU1 to generate automatically the ana lytical artifact AA can comprise for instance port defini tions of input and output ports of the component, failure modes, internal states of the components, failure rates, maintenance activities, inspection intervals, mean down times and/or mean times to repair.

As illustrated in the block diagram of Fig. 19 the second processing unit PU2 of the identifying and evaluation system has access to the generated analytical artifact AA and has further access to the machine readable spatial data, the ma chine readable topological data and the machine readable lifecycle data of the investigated system stored in the data base DB . The processing unit PU2 comprises at least one pro cessor adapted to process the analytical artifact AA and the machine readable system related spatial and/or topological data and/or machine readable system related lifecycle data to analyze automatically a susceptibility of system components of the investigated system SYS to common cause failure influ encing factors CCFIF. Common cause failure influencing fac tors can comprise design principles, technologies, manufac turers, lifecycle resources or facts, level of redundancy, diversities within implemented redundancies as well as com plexities. The processing unit PU2 can calculate in a possi ble embodiment a quantitative common cause failure result CCFR on the basis of a machine readable common cause failure model CCFMOD provided for the common cause failure influenc ing factors CCFIFs. The machine readable common cause failure model CCFMOD can also be stored in a memory or database as illustrated in Fig. 19. In a possible embodiment the common cause failure model can comprise the IEC 61508 common cause failure model which can be used by the processing unit PU2 to calculate a beta factor as a quantitative common cause fail ure result CCFR. The processing unit PU2 and the processing unit PU1 can be integrated in one processing unit.

In a possible embodiment one or more countermeasures can be automatically triggered in response to the calculated common cause failure result CCFR output by the processing unit PU2. The processing unit PU2 can comprise in a possible embodiment a user interface UI used to output the calculated common cause failure result CCFR and/or the analytical artifact in cluding the common cause failure result CCFR to a user. The processing unit PU2 can also comprise a data interface adapted to output a control signal CRTL in response to the calculated common cause failure result CCFR, e.g. by using the analytical artifact including the common cause failure result CCFR. This control signal can be applied to a local or remote control unit to trigger countermeasures in response to the calculated common cause failure result CCFR or in re sponse to the analytical artifact including the common cause failure result CCFR. Different countermeasures can be ranked or listed according to their effectiveness. The countermeas ure can be a technical measure such as switching from one sub-system to another sub-system or segregating certain com- ponents, e.g. using different cable ducts, or an organiza tional measure such as additional training or deployment of different maintenance teams to certain areas of an industrial plant. Common cause failure results CCFRs can be calculated in a possible embodiment during a planning stage of the in vestigated system. In a further embodiment the common cause failure result CCFR can be calculated during operation of the investigated system. In this embodiment the generated control signal can change an operation mode of the investigated in dustrial system in response to the calculated quantitative common cause failure result CCFR. For instance, if the proba bility for common cause failures becomes too high or if the calculation of the analytical artifact (e.g. a fault tree) including common cause failures yields a too high failure probability a corresponding countermeasure can be automati cally triggered by the system illustrated in Fig. 19. The system enables common cause failure evaluations with respect to any kind of safety, reliability or availability related analytical artifacts generated for an investigated technical system and automated procedures. With the method and system according to the present invention it is possible to generate more comprehensive, more precise, more consistent and more reliable results by automated processing of data and analyti cal artifacts. The calculation of the quantitative common cause failure result CCFR can be performed fast to trigger fitting countermeasures with regard to common cause failures CCFs within the investigated system.

The processing unit PU2 can be connected to a user interface UI adapted to list and rank suggested countermeasures in re sponse to the calculated common cause failure result. The processing unit is adapted to evaluate different countermeas ures and to define a countermeasure or sets of countermeas ures that will be satisfactory to meet at least one prede fined system evaluation criterion if this criterion can be met by a reduction of common cause failure effects. The user interface can be used to evaluate different counter measures and to implement a countermeasure or sets of coun termeasures in the design, in the lifecycle and/or in the op erational model that will be satisfactory to meet at least one predefined system evaluation criterion if this criterion can be met by a reduction of common cause failure effects.

The processing unit can be adapted to evaluate different countermeasures by taking into account additional criteria available in the machine readable data such as costs, efforts etc. to derive an optimal countermeasure or optimal sets of countermeasures in the design, in the lifecycle and/or in the operational model that will be satisfactory to meet at least one predefined system evaluation criterion if this criterion can be met by a reduction of common cause failure effects.

The system suppliers or components suppliers can offer their technical components or sub-systems with a standardized de scription, i.e. the so called digital twin. This standardized description can comprise a semiformal functional description such as SysML including functional and/or relevant spatial information. The description can include input and output port definitions as well as additional information about failure modes, their causes and corresponding failure proper ties such as for instance failure rates or preventive and corrective maintenance activities. These generic descriptions are stored in the database DB and are available to the first processing unit PU1 and the second processing unit PU2.

In the following the method and system according to the pre sent invention are described in more detail with respect to the exemplary investigated system illustrated in Fig. 20. The investigated system SYS can provide an overall functionality to disconnect itself from an external high voltage HV power grid in case of fire as illustrated in Fig. 20. In the illus trated example the investigated system comprises one infrared sensor SI as a first system component and a second sensor S2 as a second system component. In the illustrated example the second sensor S2 can be implemented in a first alternative similar to the first infrared sensor SI and in a second al ternative 2 by using a different measuring principle or tech nology, e.g. as a smoke sensor or smoke detector. The illus trated investigated system SYS of Fig. 20 further comprises a central processing unit CPU and an actuator which is supposed to be capable of opening and closing a connection to the ex ternal high voltage power grid. This is a measure established in order to contain or extinct a fire. Other measures may al so be useful but they do not form part of the example illus trated in Fig. 20. Note that Fig. 20 illustrates a simplified model since a more complex activation circuit may be required in order to control a circuit breaker. In addition a combina tion of a circuit breaker and a disconnector can be imple mented usually in order to be sure of avoiding unintended re connections of the power supply grid. Moreover, reconnecting of the grid normally requires an additional allowance signal i.e. by a main control system. This does avoid a reconnection during fire extinction or cleaning-up activities. Neverthe less, for demonstration purposes the simplified system of Fig. 20 is used. The outer box as illustrated in Fig. 20 does identify a system boundary of the investigated system SYS.

The direction of the arrows does indicate whether a port forms an input port or an output port. All sensors SI, S2 il lustrated in Fig. 20 are dedicated to detect a fire and to transmit a message "fire" or "no fire" to the connected con troller (CPU) . In a possible embodiment the logic implemented in the controller is as follows. A "HIGH" output level to wards the actuator is only activated in case that both fire detectors do transmit the "no fire" signal to the CPU. In other cases the output level shall be set to "LOW". In the example shown in Fig. 20 it is furthermore assumed that both sensors SI, S2 as well as the CPU of the investigated system require dedicated power supplies to function as intended. Lacking power supply shall lead to "no signal" at the output of the fire detectors or to a "LOW" signal at the CPU output, respectively . To perform the method for identifying and evaluating common cause failures system components of an investigated system as illustrated in Fig. 20 the second processing unit PU2 has ac cess to an analytical artifact AA representing the investi gated system derived from machine readable functional de scription mrFD of the investigated system SYS. The analytical artifact AA derived from the machine readable functional data mrFD of the investigated system SYS can comprise a fault tree .

Formula 1

(Fire sensor Sl:OR[24]) AND (Fire sensor S2:OR[24]) OR

CPU: OR [24, 25, 27, 28] OR

Actuator : OR [ 32 , 37 ] "

In the above formula 1 there is only one "AND" gate and no "M-out-of-N" gate with M greater than 1. As a consequence fire sensors SI and S2 of the investigated system form candi dates for a common cause evaluation which will be carried out subsequently .

Topologically the HGV power station according to Fig. 21 can be separated into a peripheral section and a central control room. The distance within the cable ducts may reach for in stance 500 m in typical plants. This information or data can form part of the topological and/or spatial machine readable data of the digital twin stored in the database DB of the evaluation system. Examples for other technical domains such as semiconductors might contain wiring distances of only a few micrometers.

Fig. 21 illustrates the topological, spatial data for the first alternative by using a top down view as well as X and Y coordinates with every increment representing 100 m.

In the illustrated example (alternative 1) it can be assumed that both fire sensors SI, S2 are detectors for infrared ra diation and that both cables are put in the same cable duct 1 and that subsequently potential disturbing influences on the cables 1 and 2 may lower the reliability of fire detection.

As can be seen from Fig. 21 sensor SI and sensor S2 are physi-cally separated with a distance of around 100 m.

In order to demonstrate the effects of different common cause failure influencing factors, CCFIFs, on the fire detection system an evaluation of the cable connections between sensors SI and sensors S2 and the CPU as illustrated in Fig. 21 is included .

The machine readable data can contain technical information about the components of the investigated system. For sake of simplicity only a total failure rate is indicated in the ta ble illustrated in Fig. 22. Fig. 22 does depict a simplified example. Fig. 22 shows reliability data as well as design or logistic information data. For each component of the investi gated system failure data, design principle, supplier or man ufacturer as well as other machine readable data is stored in the table shown in Fig. 22. For instance, the first sensor SI has a total failure of 1270 fit and is formed by an infrared sensor. Sensor SI is manufactured by SensorSupplier_l , and has the spatial location within system X2Y1 as shown in Fig. 21.

For the second sensor S2 two alternatives are included in the table of Fig. 22. In the first alternative the second sensor SI is an infrared sensor whereas in the second alternative the second sensor S2 is a smoke sensor. Further, both sensors S1,S2 according to the second alternative are manufactured by two different sensor suppliers. Actually the second sensor S2 is placed at the same location within the system for both al ternatives as also indicated in the table of Fig. 22. The first cable 1 is extending through cable duct 1. The second cable 2 is extending in the first alternative through the same cable duct 1. In the second alternative the cable 2 is extending through another cable duct 2 separate from cable duct 1 where cable 1 is located.

Development, manufacturing, installation and commissioning, operation, maintenance and so on form lifecycle phases of each component in the investigated system. Fig. 23 shows an example for lifecycle data LD which can be used by the method and system according to the present invention. Each lifecycle step or phase is conducted by a certain amount of resources in terms of persons in the required roles, methods to be ap plied, tools etc. Lifecycle steps determine the quality and hence dependability of system components or functions. If system components comprise similar lifecycle data LD for par ticular steps these components may suffer the same weaknesses that may potentially lead to common cause failures CCFs . For the example given here it is assumed that both fire sensors SI, S2 need calibration after their installation. When an er ror occurs during calibration the effect of redundancy on re liability might be reduced and hence the probability of a common cause failure might be increased.

In a possible embodiment for calculating a susceptibility of system components to common cause failure influencing factors a common cause failure model can be used by the second pro cessing unit PU2 as illustrated also in Fig. 19. In a possi ble implementation the common cause failure model CCFMOD can be a complex common cause failure model such as the IEC 61508 common cause failure model used to calculate a beta factor as a quantitative common failure cause result CCFR.

Fig. 24 shows a machine readable common cause failure model of the investigated system illustrated in Fig. 21 for the first alternative i.e. where the second sensor S2 is formed by an infrared sensor and hence both sensors SI, S2 are iden tical. In the exemplary system according to alternative 1 the common cause failure model CCFMOD processed by the processing unit PU2 yields for the applied common cause failure influ encing factor CCFIFs a common cause failure result CCFR i.e. a factor Gamma = 9 %. This means that it can be assumed that

9 % of the corresponding faults may lead to a system failure due to common cause failures despite the implemented redun dant element or sensor. This can also be seen from Formula 2 (Alternative 1) as below:

Formula 2 - Alternative 1

((Fire sensor Sl:OR[24] OR Cable 1) AND ((Fire sensor

S2 : OR [24]) OR Cable 2) OR

((Fire sensor S2:OR[24] OR Cable 1) AND Gamma= 9%) OR

CPU: OR [24, 25, 27, 28] OR

Actuator : OR [ 32 , 37]"

It is of note that for simplification purposes a conservative estimation has been made in the given example, i.e. Gamma = 9 % is applied to the entire fire detection sensor despite the fact that the sensors are physically separated with only the cables being located in the same cable duct.

Based on a knowledge database of past industry best practice and the stored digital twin, suggestions can be generated for a system designer or analyst to mitigate or eliminate common cause failure influencing factors. Alternatively, an imple mentation of alternatives can be suggested or the alterna tives can be even implemented automatically. The most suita ble alternative depending on the safety, reliability or availability target and potentially further criteria (such as cost, availability of parts or resources, time constraints etc.) can be evaluated for a final implementation of the sys tem. In case of the exemplary system illustrated in Fig. 21 the method and system according to the present invention can suggest the responsible user the following countermeasures:

• put cable 2 in a second cable duct (e.g. in cable duct 2 together with the cable to component OUT) ,

• check whether a second sensor based on a different meas uring principle can be used and/or

• dispatch a different technician to calibrate sensor S2. As a consequence of the suggestions generated automatically by the evaluation system according to the present invention the following changes or countermeasures can be assumed to be implemented in the investigated technical system:

• cable 2 is put into another cable duct (cable duct 2) and is hence separated from cable 1 and/or

• for sensor S2 a smoke detector is chosen which imple

ments a measuring principle different from infrared ra diation. In addition this smoke detector can be produced by a different manufacturer. In the given example the failure rate of the smoke sensor is supposed to be 2000 fit which is slightly higher than the failure rate of an infrared sensor.

• it may be assumed in the exemplary system that a second technician may not be available for sensor calibration. In this case the corresponding suggestion generated by the evaluation system according to the present invention may not be implemented.

Fig. 25 shows a table for a machine readable common cause failure model mrCCFMOD for the second alternative where the second sensor S2 is implemented as a smoke sensor. As a re sult of the mitigation or elimination of common cause factors the common cause failure result CCFR or factor amounts to Gamma = 3 % .

A corresponding fault tree FT is displayed below.

Formula 3 - Alternative 2

((Fire sensor Sl:OR[24] OR Cable 1) AND ((Fire sensor

S2 : OR [24] OR Cable 2) OR

((Fire sensor S2:OR[24] OR Cable 1) AND Gamma=3%) OR

CPU: OR [24, 25, 27, 28] OR

Actuator : OR [ 32 , 37]"

Please note that fire sensor S2 in Formula 3 above refers to a smoke detector and no longer to an infrared sensor. The method and system for identifying and evaluating common cause failures CCF can be applied to any investigated tech nical system consisting of sub-systems and/or components. The technical or industrial system can be for instance a produc tion facility or a power plant. The investigated system may also comprise mobile systems such as complex vehicles or any other complex devices. With the method and system according to the present invention susceptibility of system components to common cause failures based on common cause influencing factors CCFIFs can be calculated automatically. Further, countermeasures to reduce the susceptibility of system compo nents to common cause failures can be triggered according to a calculated quantitative common cause failure result CCFR such as a specific common cause failure result factor, in particular the beta factor of the IEC 61508 common cause failure model. The countermeasures can comprise technical countermeasures and/or organizational countermeasures. For instance, a requirement for a system to be built may specify that certain components forming a functional redundancy shall be maintained by different maintenance teams or persons. Fur ther, components can be bought from different manufacturers. The method and system according to the present invention can be employed in a planning or design phase but also in an op eration phase of the investigated system. The investigated system can be represented by a digital twin stored in a local or remote database DB as shown in Fig. 19.

The system according to the present invention can output an indication or description of the found susceptibility of one or more system components of the system to specific common cause failure influencing factors. The indication generated by the system can comprise structured data and/or results in text form such as: "at fluid tank both pressure sensors A, B building a redundancy in the system setup have been manufac tured by the same manufacturer M and work according to same measurement principle, and consequently these sensors A, B have a certain susceptibility on common cause failures of the investigated system. The impact is for instance beta = 5 %." In this simple example the probability that the two mentioned sensors A, B would fail at the same time can be 5 % multi plied by the probability that a single sensor fails. The method and system according to the present invention used for identifying and evaluating common cause failures for system components increases the safety, reliability and/or availa bility of investigated technical systems significantly. Fur ther, the susceptibility of system components of the analyzed system to common cause failure influencing factors CCFIFs can be calculated automatically and fast. This allows to trigger countermeasures even during run time of the investigated sys tem. The method according to the present invention can be im plemented by a program executed on one or more processing units as also illustrated in context with Fig. 19. The method can form part of a control software used for controlling a technical system. Further, the method can also form part of a design or planning tool used for designing or planning a technical system which may include organizational require ments .