Continuous flow engine data stream and battery-based power storage facility data stream processing method and data man agement device

The present method refers to a method to process data streams of multiple continuous flow engines and/or data streams of a battery-based power storage facility utilizing a data manage ment device to store and forward the data in a highly effi cient way. Furthermore, the present invention refers to a da ta management device being adapted to realize such method. Additionally, the present invention refers to a system con taining multiple data management devices to establish a data network. Furthermore, the present invention refers to a com puter program product to cause an inventive data management to execute an inventive method.

While the generation, storage and distribution of data is a common matter today several fields till suffer from grave drawbacks to make use of all data being generated and stored. Continuous flow engines create a huge amount of data like sensor data during monitoring said continuous flow engine. Modern continuous flow engines provide an even increased amount of detailed data to be reviewed and analyzed. While such data is highly beneficially to early identify problems and note chances for optimizations it is a big challenge to make such data available as the amount rises. This even in creases in case the utilization of the data of multiple sites is intended. Herein, existing solutions lack a suitable pro cessing system to make such data available.

Similar problems are observed for battery-based power storage facilities consisting of a plurality of connected batteries utilized to flexibly store electric power. Corresponding fa cilities suffer from multiple batteries being required to be coupled to provide the demanded voltage and the like. Howev er, minor defect in a single battery results in the whole coupled branch of batteries being required to be shut down. Monitoring such battery-based power storage facility, howev er, is a major challenge to be solved as the flood to data to be monitored requires a highly sophisticated and extensive server system rendering the benefit obtained by such facility quite low. Also, such facility requires an improved possibil ity to handle the data available and make good use of it to avoid significant losses while keeping the monitoring effort at bay.

This and further problems are solved by the products and methods as disclosed hereafter and in the claims. Further beneficial embodiments are disclosed in the dependent claims and the further description and figures. These benefits can be used to adapt the corresponding solution to specific needs or to solve additional problems.

According to one aspect the present invention refers to a method of processing data streams of at least two, more pre ferred at least four, even more preferred at least seven, continuous flow engines and/or data streams of a battery- based power storage facility containing at least 10 battery arrangements utilizing a data management device containing a processing unit and a data storage, wherein the at least 10 battery arrangements are working independently from each oth er to provide the required power, wherein

- the at least two continuous flow engines and/or the at least 10 battery arrangements provide data streams, prefera bly sensor data streams, to be analyzed,

- the data management device retrieves the data streams of the at least two continuous flow engines and/or the at least 10 battery arrangements and stores data of the data streams as historical data in a data storage,

- the data management device upon request forwards historical data, wherein the historical data is forwarded in a binary format. It was noted that the application in the context of continu ous flow engines is typically especially emphasized. Accord ing to a further aspect the present invention refers to a method of processing data streams of at least two, more pre ferred at least four, even more preferred at least seven, continuous flow engines utilizing a data management device containing a processing unit and a data storage, wherein

- the at least two continuous flow engines provide data streams, preferably sensor data streams, to be analyzed,

- the data management device retrieves the data streams of the at least two continuous flow engines and stores data of the data streams as historical data in a data storage,

- the data management device upon request forwards historical data, wherein the historical data is forwarded in a binary format.

It was noted that utilizing the binary format for such data is surprisingly beneficial. The data acquired for such con tinuous flow engines is typically characterized by data sets being beneficially processed and transported this way. Here in, not only a significant reduction of the data amount could be achieved. In fact, switching to such data format enables to significantly the overall processing of data as well as retrieval. Although, this is not to be understood to limit the present invention it is assumed that the average data created for such continuous flow engine is beneficially pro cessed in such way resulting in some significant benefit when applying the inventive method and especially the specific em bodiments as specified hereafter to a continuous flow engine system.

The phrase "data stream provided by a continuous flow engine" naturally does not only refer to data streams generated by the continuous flow engine itself. The person skilled in the art understands such phrase to refer to data streams related to such continuous flow engine. Herein, the auxiliaries re quired to utilize such continuous flow engine are also providing data streams like the temperature of the fuel, the burning and boiling processes in a boiler providing steam for a steam turbine, the temperature of the air in the air inlet of a gas turbine, the oil pressure of hydraulic components controlling various elements required for the utilization of such continuous flow engine, the composition of a fluid to be compressed by a compressor, and the like.

According to a further aspect the present invention refers to a data management device containing a processing unit and at least one data storage, wherein the data management device is adapted to realize an inventive method, wherein the data man agement device contains a processing unit and a data storage, wherein the data storage contains an API being adapted to output data in a binary format.

According to a further aspect the present invention refers to a system containing at least two, more preferred at least three, even more preferred at least five, inventive data man agement devices, wherein the at least two data management de vices are adapted to exchange historical data.

According to a further aspect the present invention refers to a computer program product, tangibly embodied in a machine- readable storage medium, including instructions operable to cause an inventive data management device to execute an in ventive method.

To simplify understanding of the present invention it is re ferred to the detailed description hereafter and the figures attached as well as their description. Herein, the figures are to be understood being not limiting the scope of the pre sent invention, but disclosing preferred embodiments explain ing the invention further.

Fig. 1 shows a scheme of an inventive method showing the data processing within an industrial power plant and forwarding said data to be remotely analyzed. Preferably, the embodiments hereafter contain, unless speci fied otherwise, at least one processor and/or data storage unit to implement the inventive method.

Unless specified otherwise terms like "calculate", "process", "determine", "generate", "configure", "reconstruct" and com parable terms refer to actions and/or processes and/or steps modifying data and/or creating data and/or converting data, wherein the data are presented as physical variable or are available as such.

The term "data storage" or comparable terms as used herein, for example, refer to a temporary data storage like RAM (Ran dom Access Memory) or long-term data storage like hard drives or data storage units like CDs, DVDs, USB sticks and the like. Such data storage can additionally include or be con nected to a processing unit to allow a processing of the data stored on the data storage.

Continuous flow engines are characterized by a continuous stream flowing through said engine. For example, such contin uous flow engine utilizes a stream of fluid flowing through the engine to rotate a rotor that in turn converts the kinet ic energy into electricity. Such fluid stream can be generat ed by means of burning a fuel using burner in a gas turbine or boiling a liquid like water in a steam generator. Examples of corresponding continuous flow engines are gas turbines and steam turbines. Alternatively, such continuous flow engine utilizes energy to move a fluid in said continuous flow en gine like in a compressor. Herein, such compressor is often associated to a fridge being an application of the daily life. However, compressors are also utilized in industrial applications like in the chemical industry to continuously compress a fluid like in a cracking process to allow a fur ther processing. It was noted that such continuous flow engines benefit sig nificantly from the inventive method. It is assumed that the specific data created in such engines is especially suitable to be processed herewith. Supposedly, the amount of data col lected for such engines, the data typically collected, the behavior of these engines and correlated created data, as well as many other factors result in the surprisingly high benefit obtained when applying the inventive method in this context.

A battery-based power storage facility contains a plurality of battery arrangement consisting of connected batteries to provide the required output. Such battery-based power storage facilities can be utilized to compensate for a lack of power resulting, for example, from fluctuations during the power generation utilized renewable energy. Herein, a significant number of, for example, lithium batteries need to be coupled to provide the required output. While such system is flexible and a very good addition to supporting the renewable energies with high-speed adaptable power generation units like gas turbines the battery-based power storage facility suffers from the required processing of data to monitor such facility and its batteries.

According to one aspect the present invention refers to a method as described above.

According to further embodiments it is preferred that values contained in the data streams are at least partially stored as changes over time. Herein, the absolute value is no stored continuously, but is typically stored as change in relation to a foregoing state. For example, how the value changed in comparison to the foregoing value noting down the change and the time said change occurred. For example, it can be pre ferred that least 90%, more preferred at least 95%, even more preferred at least 99%, of the correspondingly stored data is stored as value change, based on the total amount of data sets of said value data as contained in the historical data base.

According to further embodiments it is preferred that the da ta streams are buffered in RAM data storage of the data man agement device. Despite the relatively high price it is sur prisingly beneficial to utilize RAM storage for buffering the data streams. It was noted that the increased swiftness to easily adapt and speed up the data processing easily out weighs the increase price and effort in this context.

According to further embodiments it is preferred that the at least one data stream of each continuous flow engine, prefer ably all data streams of each continuous flow engine, and/or the data streams of the battery-based power storage facility more preferred for each tag, are buffered in reserved RAM da ta storage areas of the data management device. For example, this can be realized by extending a Microsoft.Net class li brary. Typically, it is preferred that such buffering is ap plied to continuous flow engine-based systems.

According to further embodiments it is preferred that a cir cular buffer is used for buffering the data streams. Typical ly, it is preferred that at least 50%, more preferred at least 75%, even more preferred 90%, of the data streams are buffered utilizing a circular buffer. Typically, it is pre ferred that all data streams are buffered by such circular buffer. For the sake of completeness, it is pointed out that the aforementioned percentage of the data streams is based on the number of data streams and not the amount of data of said data streams.

According to further embodiments it is preferred that the circular buffer is a self-maintaining circular buffer. Such self-maintaining circular buffer covers all value changes of a tag for a specified period of time. Typically, it is pre ferred to prevent an unlimited growth of such circular buffer by defining a size and/or time range. This secures that it is always possible to swiftly and concurrently access the near time history of each tag.

According to further embodiments it is preferred that it is preferred that the buffered data is stored as value changes over time. It was noted that apparently the highly continuous utilization of the continuous flow engines and the behavior of the batteries in a battery-based power storage facility allow to significantly reduce the required data storage this way. This allows to easily realize the quick data storage utilizing a low amount of RAM data storage enabling to pro vide a significant amount of system resources, for example, to simultaneously process that data or respond to, for exam ple, a request received from a remote analytical device.

According to further embodiments it is preferred that the da ta management device utilizes a http/2 based application pro gramming interface (API) to request the historical data. For example, such API can be gRPC. For many applications it is typically preferred that the processed data is encoded uti lizing protocol buffer encoding. It was noted that such API in combination with the specified data sources provides high performance. Furthermore, such API enables to allow an easy access by means of a different computer simply by opening a port as required. Additionally, it was noted that such API is able to run using almost any operating system.

According to further embodiments it is preferred that the da ta management device utilizes an API application programming interface (API), wherein the API can be utilized to create plugins for retrieving data from the industrial plant. Here in, the plugins can be created by internal users being locat ed, for example, at the site of the continuous flow engine and data management device or by external users being located remotely. The term "plugin" as used herein preferably refers to a class library that implements a specified interface and implements code to retrieve data from a specific data source or a specific protocol. Such plugin preferably connects cer tain data sources and the data management device to transfer data to a specified location preferably including a pro cessing of the data before it is transferred from the data management device to a specified location. Especially, the possibility of the inventive method to also allow such crea tion of plugins by an external user is highly beneficial.

According to further embodiments it is preferred that the plugin utilizes at least one communication protocol stored on the data management device. Herein, commonly available proto cols can be utilized. For example, such protocols can be se lected from MODBUS, IEC 61850, IEC 60870, S7 protocol (RFC 1006), or OSIsoft PI AF SDK.

According to further embodiments it is preferred that certif icates are exchanged between a client and the data management device before the data management device forwards the histor ical data to the client. Such client can be, for example a centralized analytical device retrieving historical data from a plurality of data management devices or a user interface utilized by a user to retrieve historical data locally or preferably remotely.

According to further embodiments it is preferred that re questing and forwarding the historical data utilizes an SSL authentication and/or a TLS authentication between a client requesting the historical data and the data management de vice. It was noted that corresponding authentication system allow to very beneficially realize the data exchange between a client and the data management device.

According to further embodiments it is preferred that the historical data is at least partially cached by the data man agement device. Typically, it is preferred that predefined priority data is cached. For example, such priority data can be selected based on commonly requested and/or most recent data being typically be very often requested within a short time frame to, for example, take action in case an alert re quires an immediate response. It was noted that the inventive method allows to reduce the workload, for example, of availa ble RAM data storage available. Such RAM data storage or com parable data storage can, thus, be beneficially utilized to cache an amount of historical data enabling to further im prove the benefit obtained. For example, under normal condi tions of typical applications it becomes possible to cache the historical data to be written to smoothen the writing process of the historical data as well as allow a more time- consuming data processing resulting in a delayed storing from time to time. Furthermore, corresponding historical most likely to be requested can be retrieved in advance to further increase the historical data retrieval. Surprisingly, such means can be utilized for the applications as described here in despite the typically very limited resources based on the processing power set free by means of the inventive method.

According to further embodiments it is preferred that the historical data being stored on the data storage contains historical sensor data, wherein the historical sensor data is at least partially protected against tampering. Typically, it is preferred that all historical sensor data is protected against tampering. For example, such protection can be achieved by restricting the access of the historical sensor data, for example, by encrypting the data. Additionally or alternative, such historical sensor data can be protected against tampering by enabling it identify changes utilizing, for example, hash keys and the like. It was noted that uti lizing such protection is especially beneficial as changes of the sensor data made for various reasons easily result in significant damages in such industrial plant.

According to further embodiments it is preferred that the historical data is stored in a binary format on the data storage. Directly storing the data in binary format is typi cally beneficial despite the significant additional effort to convert the data when storing the data. Especially, the sig- nificantly reduced size of the data is very beneficial as it was noted that despite the low price of corresponding data storages the speed to retrieve such data on demand is signif icantly increased this way.

According to further embodiments it is preferred that the method contains that the data streams of the at least two continuous flow engines and/or the at least 10 battery ar rangements are at least partially not in a binary format, wherein the method contains the step of converting the data streams into binary format when storing the of the data streams as historical data on a data storage or when forward ing the historical data upon request. It was noted that such conversion despite the overall increased effort is more than compensated by the benefits obtained by the inventive method forwarding the historical data in a binary format.

According to further embodiments it is preferred that the da ta streams of the at least two continuous flow engines and/or the at least 10 battery arrangements are at least partially, preferably completely, stored in a time series data format.

It was noted that such data format is a very efficient way for typical applications to store such data of continuous flow engines or battery arrangements to be processed accord ing to the inventive method.

According to further embodiments it is preferred that the historical data is stored in at least two data formats, pref erably wherein one format is a time series data format, more preferred timescale time series data format. Such time series data base format is typically preferred to be combined with storing further data in binary data format. It was noted that storing the historical data in a time series data format is typically beneficial and enables to more efficiently organize and retrieve such data on short notice. Furthermore, it was noted that PostgreSQL time series data base format are typi cally very beneficial in this context. One example of such format is timescale. It was noted that such time series data base format provides many benefits when utilized for the in ventive method. For example, it allows to keep the speed of storing new data independent from the size of the available data base. This is especially beneficial for the application cases as specified herein as not only the amount of data gen erated per day is tremendous, but this generation of data al so keeps on for a very prolonged period of time only briefly interrupted by maintenance or upgrade actions. Therefore, utilizing such format is very beneficial for such applica tions. Furthermore, it was noted that the data of such appli cations can be very efficiently retrieved from such data base format. Therefore, it was noted that taking into account the intended purpose such data base format surprisingly provided a very good overall performance.

According to further embodiments it is preferred that, the historical data is stored on the data storage in at least two different data archives, wherein at least one of the at least two different data archives contain raw data, and wherein at least one of the at least two data archives contains data generated by applications. It was noted that such split of the data into different archives is surprisingly beneficial. The benefit like the increased accessibility and writing speed as observed is attributed to the data created in the context of continuous flow engines and that the amount of da ta as well as data values in this context fits very well to provide such beneficial split. However, it has to be under stood that this assumption is not meant to limit the inven tion, as the corresponding reason seems to be based on a num ber of reasons.

According to further embodiments it is preferred that the da ta management device automatically retrieves at least one raw data stream and at least one application data stream, wherein such raw data stream is a continuous data stream preferably generated by a sensor, and wherein such application data stream is data generated by an application. Such application data stream can be continuous or discontinuous, however, is characterized in that it does not represent a measure charac teristic, but an output of an application being tasked to op erate in connection to the continuous flow engine. For exam ple, such application data can be controls parameters, alerts and/or evaluations of the state of the continuous flow en gine.

According to further embodiments it is preferred that the da ta streams contain high frequency data. Such high frequency data is characterized by a high number of values in a given time period. Typically, it is preferred that such high fre quency data provide at least 10 values per second, more pre ferred at least 30 values per second, even more preferred at least 100 value per second. For example, such high frequency data is vibrational data of a gas turbine or the like. It was noted that especially processing such data according to the method of the present invention is very beneficial and allows to, for example, achieve significant reduction of the access speed of historical data and the like. Especially, such em bodiments can be beneficially combined with further embodi ments herein like storing the data as value changes over time. For high frequency data of continuous flow engines a tremendous reduction of the file size can be achieved even further increasing the access speed of historical data.

According to further embodiments it is preferred that the da ta streams contain at least one raw data stream, wherein the at least one raw data stream is processed to provide an adapted frequency. For example, the frequency of a raw data stream is reduced to reduce the processing power to further process the raw data stream. It was noted that sensor data in the specified applications is not necessarily required to provide the originally possible resolution as at least under typical working conditions the sensors provide significant additional data not required to correctly evaluate the situa tion. Typically, it is preferred that such processing is sit uation dependent and allows to increase the frequency accord- ing to the specific situation and, for example, increase the frequency of stored data in alert situations.

According to further embodiments it is preferred that the da ta streams are grouped as at least two groups, wherein at least one of the at least two groups is adapted with regard to the update frequency. Such update frequency characterizes the frequency of new values being stored as historical data. It was noted that many data streams can easily be grouped this way and configured to provide a lower update frequency under specific conditions to adapt such processing with lit tle effort without impairing the insight or reliability of such data. The values of such reduced update frequency modi fied data stream can be, for example, simply be based on val ues being selected from the measured values with an increased time interval. Such method is typically preferred as it re quired little to no processing power and provides reliable data. Alternatively, such reduction can be achieved by providing a mean value for the values within the selected time range. This provide the further benefit to balance out deviations and can be very beneficial for specific applica tions like data streams providing a significant fluctuation.

According to further embodiments it is preferred that the da ta management device sends instructions to at least a part of the sources of the data streams, wherein the instructions contain a request to adapt the update frequency of the data streams. Surprisingly, such adaption can be easily automized allowing to significantly reduce the data streams and simpli fy their processing.

According to further embodiments it is preferred that the da ta management device creates a statistical dataset. Herein such statistical dataset contains at least one, more pre ferred at least two of the statistical values selected from the group consisting of a number of data sets of at least one data stream being summarized to provide the adapted update frequency, a number of values providing a specified quality, and a time dataset indicating the last change of a value of a data stream. Typically, it is especially preferred to create a time dataset indicating the last change of a value of a da ta stream.

According to further embodiments it is preferred that histor ical data is midterm historical data. The term "midterm" his torical data preferably refers to historical data of at most the past 4 months, more preferred at most the past 3 months. It was noted that the inventive system is especially suitable to process such data in corresponding applications.

According to further embodiments it is preferred that the da ta management device communicates with a client, wherein the communication includes requesting and forwarding the historical data, wherein the communication is at least partially, preferably completely, secured against unallowed access. For example, such communication can be encrypted to ensure that in case of a remotely located client servers in between are unable to acquire the historical data. Such encryption can be based on a hardware encryption or software encryption. For many up grade cases it is typically preferred to include a hardware encryption in the data management device to be directly in stalled. However, even in such case it is typically preferred to keep the option to switch to a software encryption in case the hardware encryption is corrupted or otherwise impaired.

According to further embodiments it is preferred that the da ta management device forwards the historical data to a fur ther data management device receiving data streams from dif ferent continuous flow engines. Typically, it is even pre ferred that distanced data management devices communicate with each other. Herein, some remote system can be provided exchanging data at remote locations to improve, for example, predictions utilizing correspondingly acquired data. According to further embodiments it is preferred that the at least two continuous flow engines contain at least two, more preferred at least four, even more preferred at least seven, gas turbines. It was noted that the data like especially sen sor data acquired for such gas turbines is especially benefi cially processed and stored utilizing the inventive method.

According to further embodiments it is preferred that the continuous flow engine is a compressor, a gas turbine, a steam turbine, or a combined cycle power device, more pre ferred a gas turbine, a steam turbine, or a combined cycle power device, even more preferred a gas turbine. As already stated above applying the inventive method for such applica tions is typically especially beneficial.

According to a further aspect the present invention refers to a data management device containing a processing unit and at least one data storage, wherein the data management device is adapted to realize an inventive method, wherein the data management device contains a processing unit and a data storage, wherein the data storage contains an API being adapted to output data in a binary format.

According to further embodiments it is preferred that the da ta management device is adapted to execute plugins being cre ated utilizing an API stored on the data management device.

It was noted that such data management device is highly bene ficial to even further improve the possibilities and benefits provided by the inventive data management device.

According to further embodiments it is preferred that the da ta management device contains at least one protocol to be utilized by a plugin. Herein, commonly available protocols can be utilized. For example, such protocols can be selected from MODBUS, IEC 61850, IEC 60870, S7 protocol (RFC 1006), or OSIsoft PI AF SDK. According to a further aspect the present invention refers to a system containing at least two, more preferred at least three, even more preferred at least five, inventive data man agement devices, wherein the at least two data management de vices are adapted to exchange historical data.

According to further embodiments it is preferred that the da ta management devices are adapted to simultaneously forward historical data to a data analysis device.

According to a further aspect the present invention refers to a computer program product, tangibly embodied in a machine- readable storage medium, including instructions operable to cause an inventive data management device to execute an in ventive method.

The following detailed description of the figure uses the figure to discuss illustrative embodiments, which are not to be construed as restrictive, along with the features and fur ther advantages thereof.

Figure 1 shows a scheme of an inventive method showing the data processing within an industrial power plant and forward ing said data to be remotely analyzed. Herein, the gas tur bines 2a, 2b, 2c, 2a', 2b', 2c' as examples of the continuous flow engines are located in two industrial power plants 3,

3'. The gas turbines 2a, 2b, 2c, 2a', 2b', 2c' generate a plurality of data streams that are forwarded to the corre sponding data management device 1, 1' located at the corre sponding industrial plant 3, 3'. The corresponding data man agement device 1, 1' processes the incoming data streams and stores the data as historical data in a data base 7, 1'. Herein, the processed data is encoded utilizing protocol buffer encoding.

The data streams are buffered by the data management devices 1, 1' using RAM data storage allowing to quickly store and retrieve the data to easily adapt the processing of said data as required. Herein, one data management device 1 utilizes a circular buffer for 80% of the data streams. The remaining data streams are buffered using a linear buffer. The second data management device 1' utilizes circular buffers for all data streams. While the performance of the second data man agement device 1' is further improved the performance of the first data management device 1 is still satisfying based on the lower amount of data provided and the higher RAM data storage available for said data management device 1. It is planned to further upgrade the industrial power plant 3 and the interrelated components including to change all buffers to circular buffers for the future.

The circular buffers of both data management devices 1, 1' are self-maintaining circular buffers. Herein, the buffers as limited to buffer a defined time interval, wherein the inter val is further adapted to be eventually reduced in the case the unused RAM data storage falls below 5% of the available RAM data storage.

The data management device 1, 1' are communicating with each other as well as forward the data in binary format to the an alytical device 4. Herein, the data management devices 1, 1' enable to simultaneously request the corresponding data from both industrial plants 3, 3' as required. The API used by the data management device 3, 3' is http/2 based enabling a sur prisingly beneficial requesting or providing of the histori cal data. The data is forwarded directly from data management device 1 and data management device 1' without requiring to first forward the required data to a single location. In this context, the data management devices 1, 1' and the analytical device exchange certificates before transmitting data to en sure that the data safety is secured.

The present invention was only described in further detail for explanatory purposes. However, the invention is not to be understood being limited to these embodiments as they repre sent embodiments providing benefits to solve specific prob- lems or fulfilling specific needs. The scope of the protec tion should be understood to be only limited by the claims attached.