Technical field

[01] The present invention relates to method of monitoring a process of circulation of solid material in a circulating fluidized bed reactor according to the preamble of claim 1.

[02] The present invention relates to control system for monitoring a process of circulation of solid material in a circulating fluidized bed reactor according to the preamble of the second independent claim.

Background art

[03] In circulating fluidized bed reactor fine solid material is utilized in the process by fluidizing the solid material to an extent that considerable portion of the material is entrained from the reaction chamber to the at least one solid material separator, from the solid material separator a portion of the solid material separated from gas may be led to a fluidized bed heat exchanger and from the fluidized bed heat exchanger circulated back to the combustion chamber. Such a CFB reactor is well applicable to produce power by combustion of fuel in the CFB reactor for generation of steam, in which case the CFB reactor is usually referred to as a CFB boiler. Likewise, it is also known to utilize CFB reactor for producing product gas, such as gaseous fuel resulting in reactions taken place in the CFB reactor. In case of producing gaseous fuel from solid fuel the CFB reactor is usually referred to as a CFB gasifier.

[04] Challenging fuels may cause particle agglomeration in the fluidized bed material, which can further lead to more severe sintering of the solid material and eventually to blockage in the solids return system and shutdown of the reactor, if corrective procedures are not started in time. For example with continuously fluctuating fuel quality, sintering may be impossible for operators to recognize by following the conventional basic operative routine. [05] US 8292977 discloses a system for controlling a circulatory amount of particles in a circulating fluidized bed furnace wherein particles are circulated between a fluidized bed combustion furnace for heating of the particles and a fluidized bed gasification furnace for gasification of raw material through heating of the raw material by the heated hot particles. Control is based on measurement of pressure in the fluidized bed gasification furnace and controlling exhaust rate from the fluidized bed gasification furnace.

[06] Document JP4254004B2 discloses a controlling fluidization rate based on estimating external circulation solids in circulating fluidized bed boiler. The estimation is based on measuring temperature and pressure in the reactor and an external fluidized bed superheater heat exchanger as well as outlet steam temperature and amount of de-superheating water.

[07] JP4443481 B2 relates to a system for diagnosing a flow medium clogging comprising a plurality of differential pressure gauges that measure the differential pressure at a predetermined location in the flow medium circulation path, a plurality of thermometers that measure the temperature at a predetermined location in the flow medium circulation path, provide a determination result display means for displaying the fact that the flow medium is clogged and the location where the flow medium is clogged when the determination means determines that the flow medium is clogged.

[08] Even if circulating fluidized bed (CFB) reactor has it benefit over other combustion technologies, in particular agglomeration problem relating to the solid material in CFB technology is of concern since it can lead to unscheduled shutdowns of the plant.

[09] An object of the invention is to provide method of and control system of monitoring a process of circulation of solid material in a circulating fluidized bed reactor by means of which unintentional shutdowns can be avoided or at least minimized. Disclosure of the Invention

[010] Objects of the invention can be met substantially as is disclosed in the independent claims and in the other claims describing more details of different embodiments of the invention.

[011] According to the invention method of monitoring a process of circulation of solid material in a circulating fluidized bed reactor, which reactor comprising a reaction chamber, at least one solid material separator, a return path between said at least one solid material separator and the reaction chamber and in which method, the process of circulation of solid material comprises arranging solid material to be entrained by gas flow in the reaction chamber, and to entrain further from the reaction chamber to the at least one solid material separator and passing solid material from the solid material separator via the return path to the reaction chamber. The method comprises at least the following step: a. selecting process variable of the process of circulating of solid material in the return path, and selecting performance indicators of the process of circulation of solid material amongst the selected process variables for each performance indicator of the process of circulation of solid material, b. creating a multivariate model for each performance indicator, using history data of the process variables and the performance indicators of the process of circulation of solid material c. determining a modelled value of the performance indicators, by applying current measured values of the process variables to the multivariate model d. comparing the modelled value of each performance indicator to a respective measured value of each performance indicator and inspect presence of an anomaly between the modelled value and the measured value.

[012] When the modelled value of a performance indicator is determined by making use of on-line values process variables and the modelled value of the performance indicator is compared to on-line values of respective performance indicator, the method provides an effect by means of which possible problems, such as bed quality related and/or risk of sintering of bed material, and/or risk of potential bed material blockages in circulation of solid material can be effectively foreseen so that remedial actions can be taken early enough to maintain the process operational. [013] According to a preferable aspect of the invention, the method comprises combustion of fuel in the circulating fluidized bed reactor, that is in a circulating fluidized bed boiler. Thus, according to a preferable aspect of the invention, the reaction chamber is a combustion chamber.

[014] According to another aspect of the invention, the method comprises producing gaseous fuel by transforming combustible substance into gaseous fuel in the circulating fluidized bed reactor, that is in a circulating fluidized bed gasifier.

[015] According to an aspect of the invention, the method may include, in combination with any other step or steps of the method, indicating to the operator of possible remedial actions to be taken, which controls the circumstances to a direction decreasing the tendency of agglomeration of solid material, including at least one of the following: changing the fuel mixture by reducing a share of fuels prone to form agglomerates introducing additives such as clay (kaolin, for example), or increasing the amount of such additives, raising an agglomeration temperature of the bed and/or reducing load of the reactor and/or first reducing load and then increasing the load, which also decreases agglomeration tendency increasing the feed of make-up material, such as sand, to the reaction chamber or fluidized bed heat exchanger, discharging bottom ash from the reaction chamber or increasing the discharging rate of bottom ash, which removes agglomerates from the reaction chamber solid material removal from return path which also makes it possible to remove agglomerated bed material.

[016] According to an aspect of the invention, the comparison between the modelled value of the performance indicators to a respective measured value of each performance indicator in step d. above may be based on at least one of the following: difference, absolute value of the difference, or ratio. [017] According to an aspect of the invention the process of circulation of solid material comprises passing solid material from the separator directly to the reaction chamber via a loop seal, wherein at least i. Pressure difference of a loop seal in the return path ii. Temperature in the loop seal in the return path in the circulation of solid material are selected as the performance indicators of the process of circulation of solid material.

[018] This aspect relates to an embodiment of the invention where the CFB reactor comprises a loop seal in the return path and the selected performance indicators provides efficient manner of monitoring the process of circulation of solid material.

[019] According to an aspect of the invention the process of circulation of solid material comprises passing solid material from the separator directly to the reaction chamber via the return path, wherein at least i. Pressure difference in the return path ii. Temperature in the return path in the circulation of solid material are selected as the performance indicators of the process of circulation of solid material.

[020] This aspect relates to an embodiment of the invention where the return path of the CFB reactor is not provided with a loop seal or the solid material is directed to the reaction chamber from a position upstream the loop seal.

[021] According to an aspect of the invention the process of circulation of solid material comprises passing solid material from the separator via a fluidized bed heat exchanger to the reaction chamber, wherein at least i. Pressure difference of a loop seal in the return path ii. Temperature in the loop seal in the return path in the circulation of solid material iii. Pressure difference of a fluidized bed heat exchanger, iv. Temperature of solid material downstream a heat exchange unit in the fluidized bed heat exchanger, are selected as the performance indicators of the process of circulating of solid material. [022] This aspect relates to an embodiment of the invention where the selected performance indicators cover a CFB reactor comprising a loop seal and a fluidized bed heat exchanger in the return path. The selected performance indicators provide efficient manner of monitoring the process of circulation of solid material at some critical positions of the process.

[023] According to an aspect of the invention, the temperature of solid material downstream the fluidized bed heat exchange unit is measured at the bottom portion of the fluidized bed heat exchanger. According to another embodiment of the invention, the temperature of the solids in the fluidized bed heat exchanger is measured at the bottom portion of the fluidized bed above fluidization nozzles. Generally, the term “downstream the heat exchange unit” may be understood as downstream or in below heat exchange tubes of the unit, which extend into the fluidized bed heat exchanger. In other words, the temperature of solid material downstream the fluidized bed heat exchange unit may be measured below the heat exchange unit in the fluidized bed heat exchanger. Further in another terms, the temperature of solid material downstream the fluidized bed heat exchange unit may be measured below the heat exchanger tubes in the fluidized bed heat exchanger and above the fluidization nozzles. The fluidized bed heat exchange unit may be chosen according to the need to be, for example, evaporator, superheater or reheater, to name a few.

[024] According to an aspect of the invention the process of circulation of solid material comprises passing solid material from the separator directly to the reaction chamber via a loop seal, wherein at least i. Pressure difference of a loop seal in the return path, and ii. Temperature in the loop seal in the return path in the circulation of solid material are selected as the performance indicators of the process of circulation of solid material, and wherein i. process variables of the performance indicator Pressure difference of the loop seal in the return path comprise: Aggregate reaction gas flow rate fed into the reactor, product gas temperature upstream the loop seal, bed temperature in the reaction chamber, ii. process variables of the performance indicator Temperature in the loop seal in the circulation of solid material comprise: Aggregate reaction gas flow rate fed into the reactor, temperature of the product gas upstream the loop seal, bed temperature in the reaction chamber.

[025] According to an aspect of the invention, in case of circulating fluidized bed boiler, the product gas can be called as flue gas containing products of combustion reactions.

[026] According to an aspect, in case of circulating fluidized bed gasifier, when gasifying carbonaceous fuels, such as biofuels or waste derived fuels, air and/or oxygen as well as steam may be supplied to the reaction chamber so as to generate product gas in which the main components comprise carbon monoxide CO, hydrogen H2, and hydrocarbons C _x H _y . The product gas of the circulating fluidized bed gasifier may be called as syngas.

[027] According to an aspect of the invention the process of circulation of solid material comprises passing solid material from the separator directly to the reaction chamber via a loop seal, wherein at least i. Pressure difference of a loop seal in the return path, and ii. Temperature in the loop seal in the return path in the circulation of solid material are selected as the performance indicators of the process of circulation of solid material, and wherein i. process variables of the performance indicator Pressure difference of the loop seal in the return path comprise: Aggregate combustion gas flow rate fed into the reactor, flue gas temperature upstream the loop seal, bed temperature in the reaction chamber, ii. process variables of the performance indicator Temperature in the loop seal in the circulation of solid material comprise: Aggregate combustion gas flow rate fed into the reactor, temperature of the flue gas upstream the loop seal, bed temperature in the reaction chamber.

[028] According to an aspect of the invention, the reaction gas is air. According to an aspect of the invention, the reaction gas is air, or a mixture of air and recirculated flue gas. According to an aspect of the invention, the reaction gas is pure oxygen. According to an aspect of the invention, the reaction gas is a mixture of oxygen and recirculated product gas. According to an aspect of the invention, in case of circulating fluidized bed boiler, the reaction gas can be called as combustion gas.

[029] According to an aspect of the invention the method comprises gasification of fuel in the CFB reactor and the process of circulation of solid material comprises passing solid material from the separator directly to the reaction chamber via a loop seal, wherein at least i. Pressure difference of a loop seal in the return path, and ii. Temperature in the loop seal in the return path in the circulation of solid material are selected as the performance indicators of the process of circulation of solid material, and wherein i. process variables of the performance indicator Pressure difference of the loop seal in the return path comprise: Aggregate gas flow rate fed into the reactor, product gas temperature upstream the loop seal, bed temperature in the reaction chamber, ii. process variables of the performance indicator Temperature in the loop seal in the circulation of solid material comprise: Aggregate gas flow rate fed into the reactor, temperature of the product gas upstream the loop seal, bed temperature in the reaction chamber. According to an embodiment the circulating fluidized bed reactor is a circulating fluidized bed gasifier and wherein fluidization gas comprising at least one of the following: inert gas, steam, oxygen or mixtures thereof.

[030] According to an aspect of the invention the process of circulation of solid material comprises passing solid material from the separator via a fluidized bed heat exchanger to the reaction chamber, wherein at least i. Pressure difference of a loop seal in the return path ii. Temperature in the loop seal in the return path in the circulation of solid material iii. Pressure difference of a fluidized bed heat exchanger, iv. Temperature of solid material downstream the fluidized bed heat exchange unit are selected as the performance indicators of the process in the step of the process of circulation of solid material, and wherein i. process variables of the performance indicator Pressure difference of the loop seal in the return path comprise: Aggregate reaction gas flow rate fed into the reactor, temperature of product gases upstream the loop seal, bed temperature in the reaction chamber, ii. process variables of the performance indicator Temperature in the loop seal in the circulation of solid material comprise: Aggregate reaction gas flow rate fed into the reactor, temperature of product gas upstream the loop seal, bed temperature in the reaction chamber iii. process variables of the performance indicator Pressure difference of fluidized bed heat exchanger comprise: Aggregate reaction gas flow rate fed into the reactor, temperature in the loop seal in the return path in the circulation of solid material, pressure difference of loop seal, gas flow rate to the fluidized bed heat exchanger, bed temperature in the reaction chamber, iv. process variables of the performance indicator Temperature of fluidized bed heat exchanger comprise: Aggregate reaction gas flow rate fed into the reactor, temperature in the loop seal, pressure difference of the loop seal, gas flow rate to the fluidized bed heat exchanger, bed temperature in the reaction chamber.

[031] According to an aspect of the invention the aggregate reaction gas flow rate is total flow rate of gas flows into the reaction chamber of the CFB reactor.

[032] According to an embodiment where the method comprises combustion of fuel in a presence of air the aggregate reaction gas flow rate is total flow rate of air flows into the reaction chamber of the CFB reactor.

[033] According to an embodiment where the method comprises combustion of fuel in a presence of air the aggregate air flow rate is flow rate comprising primary air flows fed into the reaction chamber. [034] According to an embodiment where the method comprises combustion of fuel in a presence of air the aggregate air flow rate is flow rate comprising primary air and secondary air flows fed into the reaction chamber.

[035] According to an embodiment where the method comprises combustion of fuel in a presence of air the aggregate air flow rate is flow rate comprising primary air, secondary air and tertiary air flows fed into the reaction chamber.

[036] According to an embodiment where the method comprises combustion of fuel in a presence of air the aggregate air flow rate is flow rate comprising primary air and secondary air flows fed into the reaction chamber and air fed into the fluidized bed heat exchanger.

[037] According to an embodiment where the method comprises combustion of fuel in a presence of air the aggregate air flow rate is total flow rate of air flows into the reaction chamber and into the fluidized bed heat exchanger.

[038] According to an embodiment where the method comprises combustion of fuel in a presence of air the aggregate air flow rate is flow rate comprising primary air and secondary air flows fed into the reaction chamber and the fluidized bed heat exchanger and into the loop seal.

[039] According to an embodiment where the method comprises combustion of fuel in a presence of air the aggregate air flow rate is total flow rate of air flows into the reaction chamber, the fluidized bed heat exchanger and into the loop seal.

[040] According to an aspect of the invention the process of circulation of solid material comprises passing solid material from the separator via a fluidized bed heat exchanger to the reaction chamber, wherein at least i. Pressure difference of a loop seal in the return path ii. Temperature in the loop seal in the return path in the circulation of solid material iii. Pressure difference of a fluidized bed heat exchanger, iv. Temperature of solid material downstream the fluidized bed heat exchange unit are selected as the performance indicators of the process in the step of the process of circulation of solid material, and wherein i. process variables of the performance indicator Pressure difference of the loop seal in the return path comprise: Aggregate combustion gas flow rate fed into the reactor, temperature of product gases upstream the loop seal, bed temperature in the reaction chamber, ii. process variables of the performance indicator Temperature in the loop seal in the circulation of solid material comprise: Aggregate combustion gas flow rate fed into the reactor, temperature of product gas upstream the loop seal, bed temperature in the reaction chamber iii. process variables of the performance indicator Pressure difference of fluidized bed heat exchanger comprise: Aggregate combustion gas flow rate fed into the reactor, temperature in the loop seal in the return path in the circulation of solid material, pressure difference of loop seal, combustion gas flow rate to the fluidized bed heat exchanger, bed temperature in the reaction chamber, iv. process variables of the performance indicator Temperature of fluidized bed heat exchanger comprise: Aggregate combustion gas flow rate fed into the reactor, temperature in the loop seal, pressure difference of the loop seal, combustion gas flow rate to the fluidized bed heat exchanger, bed temperature in the reaction chamber.

[041] According to an aspect of the invention the aggregate combustion gas flow rate is total flow rate of combustion gas flows into the reaction chamber of the CFB reactor.

[042] According to an aspect of the invention, the aggregate combustion gas flow rate is flow rate comprising primary combustion gas flows fed into the reaction chamber.

[043] According to a preferable aspect of the invention, the aggregate combustion gas flow rate is flow rate comprising primary combustion gas and secondary combustion gas flows fed into the reaction chamber.

[044] According to another preferable aspect of the invention, the aggregate combustion gas flow rate is flow rate comprising primary combustion gas, secondary combustion gas and tertiary combustion gas flows fed into the reaction chamber.

[045] According to an aspect of the invention, the combustion gas is air and recirculation product gas. According to an aspect of the invention, the combustion gas is oxygen and recirculation product gas. According to an aspect of the invention, the combustion gas is primary air and recirculation product gas. According to an aspect of the invention, the combustion gas is oxygen and recirculation product gas.

[046] According to an aspect of the invention the bed temperature is an average bed temperature in the reaction chamber which is calculated from at least two measurement points in the reaction chamber at least one of which is at a grid level of the chamber.

[047] According to an aspect of the invention creating the multivariate model comprises

- measuring values of predetermined process variables, storing the measured values with a time stamp thus forming the history data of process variables,

- measuring values of performance indicator, storing the measured values with a time stamp thus forming the history data of performance indicators,

- selection of valid history data using predetermined data filters.

[048] According to an aspect of the invention creating the multivariate model comprises

- measuring values of predetermined process variables, storing the measured values with a time stamp thus forming the history data of process variables,

- measuring values of performance indicator, storing the measured values with a time stamp thus forming the history data of performance indicators,

- selecting valid history data using predetermined data filters, and

- updating the multivariate model.

[049] According to an aspect of the invention the data filter is configured to approve data which is older than a pre-set quarantine time. Advantageously, by this way the model will not be taught with potential abnormal operation values, for example, due to start of agglomeration of solid material in the loop seal. In other words, problem-related data is not used in model training.

[050] According to an aspect of the invention the data filter is configured to approve data which is older than two weeks.

[051] According to an aspect of the invention the data filter is configured to approve data which is not older than two months.

[052] According to an aspect of the invention the data filter is configured filter out from history data any data from shut down situations and/or from any abnormal operation based on predefined limits for input variables or an external information of abnormal operation.

[053] According to an aspect of the invention the method comprises at least the following step: a. selecting process variable of the process of circulating of solid material in the return path, and selecting performance indicators of the process of circulation of solid material amongst the selected process variables for each performance indicator of the process of circulation of solid material, b. creating a multivariate model for each performance indicator, using history data of the process variables and the performance indicators of the process of circulation of solid material, which multivariate model is a multivariate linear regression having a first number (N) of measured observations of each process variable and a second number (P) of different process variables of the process of circulation of solid material as follows yi = b ₀ + b Xi'! + b ₂ x _{i 2} +. . . b _P x _{i P} + £t where t = 1,2,...N and method comprising reading history data of yi, Xj,i ,Xj.2, ■ ■ ■ ,Xj, _p where y = performance indicator and Xj,i,Xj,2,... ,Xj, _p are process variables, and solving the constant bo and factors bi,b2,... bp and performing the fitting by minimizing the sum of the squares of the vertical deviations from each data point to the line that fits best for the history data. c. determining a modelled value of the performance indicators, by applying current measured values of the process variables to the multivariate model d. comparing the modelled value of each performance indicator to a respective measured value of each performance indicator and inspect presence of an anomaly between the modelled value and the measured value.

[054] According to an aspect of the invention the method comprises at least the following step: a. selecting process variable of the process of circulating of solid material in the return path, and selecting performance indicators of the process of circulation of solid material amongst the selected process variables for each performance indicator of the process of circulation of solid material, b. creating a multivariate model for each performance indicator, using history data of the process variables and the performance indicators of the process of circulation of solid material, which multivariate model is a multivariate linear regression having a first number (N) of measured observations of each process variable and a second number (P) of different process variables of the process of circulation of solid material as follows yi = b ₀ + b Xi'! + b ₂ x _{i 2} +. . . b _P x _iP + £t where t = 1,2,...N and method comprising reading history data of yi, Xj,i ,Xj.2, ■ ■ ■ ,Xj, _p where y = performance indicator and Xj,i ,Xj,2, ... ,Xj, _p are process variables, and solving the constant bo and factors bi,b2,... bp and performing the fitting by minimizing the sum of the squares of the vertical deviations from each data point to the line that fits best for the history data, wherein creating the multivariate model comprises measuring values of predetermined process variables, storing the measured values with a time stamp thus forming the history data of process variables, measuring values of performance indicator, storing the measured values with a time stamp thus forming the history data of performance indicators, selecting valid history data using predetermined data filters, and updating the multivariate model, c. determining a modelled value of the performance indicators, by applying current measured values of the process variables to the multivariate model d. comparing the modelled value of each performance indicator to a respective measured value of each performance indicator and inspect presence of an anomaly between the modelled value and the measured value.

[055] According to an aspect of the invention the first number of measured observations N is at least 10 times the second number P of different process variables.

[056] According to an aspect of the invention multivariate mode is updated after a period of time triggered by lapse of a constant predetermined time interval, or by a trigger input.

[057] According to an aspect of the invention a risk index for each performance indicator is calculated using the information of presence of anomaly.

[058] According to an aspect of the invention a risk index for each performance indicator is calculated using anomaly between the modelled value and the measured value.

[059] According to an aspect of the invention the reactor comprises at least a first return path a first solid material separator and the reaction chamber and a second return path between a second solid material separator and the reaction chamber, wherein the method concerning the process of circulation of solid material in the first return path, and the method concerning the process of circulation of solid material in the second return path are practised separately to the return paths.

[060] A control system for monitoring a process of circulation of solid material in a circulating fluidized bed reactor between a reaction chamber and at least one solid material separator, and via a return path, comprising a loop seal, to the reaction chamber, and further a control system for controlling the process of circulation of solid material, wherein the control system comprises a performance modelling unit, comprising access to a source history data of performance indicators of the process of circulation of solid material in the return path and process variables for each performance indicator a multivariate model for each performance indicator executable instructions which, when executed in the data processing unit, o update the multivariate model for each performance indicator, using history data of predetermined process variables and the performance indicators of the process of circulation of solid material, resulting in a calibrated multivariate model, a performance diagnostic module, comprising

Inputs for receiving measurement data of process variables and performance indicators of the process of circulation of solid material, executable instructions which, when executed in the data processing unit, o determine a modelled value of the performance indicators, by applying current measured values of the process variables to the calibrated multivariate model o compare the modelled value of each performance indicator to a measured respective value of performance indicator and inspecting a presence of anomaly between the modelled value and the measured value.

[061] According to an aspect of the invention the control system comprising measurement sensors for at least following process variables: pressure sensors for measuring pressure drop in the loop seal; product gas temperature sensor and means for determining aggregate air flow rate to the reactor and bed temperature in the reaction chamber of the reactor.

[062] According to an aspect of the invention circulation of solid material in a circulating fluidized bed reactor between a reaction chamber and at least one solid material separator, and via a fluidized bed heat exchanger in the return path, to the reaction chamber, the control system comprising measurement sensors for at least following process variables: pressure sensors for measuring pressure drop in the loop seal; product gas temperature sensor upstream the loop seal; temperature sensor in the loop seal; pressure sensors for measuring pressure drop in the fluidized bed heat exchanger; temperature sensors for measuring temperature of solid material downstream the fluidized bed heat exchange unit in the fluidized bed heat exchanger; and means for determining aggregate air flow rate to the reactor and bed temperature in the reaction chamber of the reactor. [063] This provides an effect by means of which possible problems in circulation of solid material can be effectively foreseen. Detected anomalies act as precursors for problems in the process of circulation of solid material, such as tendency to sinter.

[064] Availability of CFB reactor is improved and operational cost are reduced due to avoiding of unnecessary shutdowns.

[065] Thus, by monitoring the measured values and the modelled values, a beginning solid material sintering can be detected and measures to heal the process, or at least to avoid the sintering becoming worse, can be taken in good time. In CFB boiler applications, this may help to avoid combustion boiler system shutdowns because of solid material sintering, and also avoid costly reparations. Advantageously, detected anomalies in solid material circulation give information on bed quality, preferably, information whether sintering is taking place in the solid material. Or in other words, it will be possible to receive information about solids circulation related problem that may have a tendency of leading to a shutdown if no remedial action is taken. Therefore, the availability of the reactor may be improved and/or operational costs may be reduced. The method is preferably carried out automatically either in local reactor control system or remotely, preferably in a process intelligence system.

[066] The multivariate model may be an artificial intelligence tool. According to an embodiment of the invention, the multivariate model may be a neural network.

[067] Preferably, the calibration of the model is not performed (i.e. the calibration is omitted) for a predefined time upon detecting an anomaly. In addition to, or alternatively, boiler/reactor shut down situations, abnormal operation and/or abnormal conditions are preferably filtered out or omitted from calibration data. This approach may help to avoid a possible circulation quality problem to contaminate the calibration. This approach can be fine-tuned such that the calibration is not performed for a predefined time upon detecting a local temperature anomaly that fulfils a given threshold. Then only severe enough conditions producing a sufficiently large anomaly signal can be chosen to be led to the skipping of calibration for a predefined time period. [068] According to an aspect of the invention in the method estimating the risk index for a performance indicator is assessed as follows.

- current data of performance indicators (KPI) of circulation of solid material is measured,

- based on the current operation data of the reactor, at least one of the following is computed: i) an average of the performance indicators; ii) standard deviation of measured performance indicators; iii) a difference between maximum measured performance indicator value and minimum measured performance indicator; iv) difference between average performance indicator and measured performance indicators;

- using the computation results from i), ii), iii) and/or iv), preparing a risk index for the performance indicator KPI.

[069] The computation results from i), ii), iii) and/or iv) are compared with corresponding predefined limits so as to get risk indexes for average, standard deviation, difference between maximum and minimum KPI value, and/or difference between average KPI value and measured KPI values.

[070] In computation of deviation of a KPk; k = 1 , ..., K from average KPI the average includes all KPI measurements except the measurement of KPIk.

[071] Preferably, in the method also v) modelled values of KPk; k = 1 , ..., K are computed, and residuals between the measured values of the performance indicators and the modelled values of the performance indicators are computed. The results from the step v) are advantageously also used in the preparing of the risk index, preferably such that residuals are compared with corresponding predefined limit so as to get sintering risk index for KPI residuals.

[072] According to an aspect of the invention in the method estimating the risk index for a performance indicator is assessed such that current data of performance indicators (KPI) of circulation of solid material is measured and v) modelled values of KPk; k = 1 , ..., K are computed, and residuals between the measured values of the performance indicators and the modelled values of the performance indicators are computed. The results from the step v) are advantageously also used in the preparing of the risk index, preferably such that residuals are compared with corresponding predefined limit so as to get sintering risk index for KPI residuals.

[073] The final risk index may then be the maximum of above risk indexes, for example. In this manner, the predictive accuracy of bed sintering index can be still improved.

[074] Advantageously, according to an aspect of the invention, it is possible to locate the most critical location in which the greatest risk for blockage or sintering occurs. This would be particularly advantageous in case of several separator and return path assemblies.

[075] In this connection the term bed temperature refers to a number that is a typical representation of a bed temperature, such as an average bed temperature. Bed temperature can be calculated in different ways, such as arithmetic mean, truncated mean or mid-range, just to mention a few. Calculation may include a desired amount of data in one or several locations in the CFB reactor.

[076] The exemplary embodiments of the invention presented in this patent application are not to be interpreted to pose limitations to the applicability of the appended claims. The verb "to comprise" is used in this patent application as an open limitation that does not exclude the existence of also unrecited features. The features recited in dependent claims are mutually freely combinable unless otherwise explicitly stated. The novel features which are considered as characteristic of the invention are set forth in particular in the appended claims.

Brief Description of Drawings

[077] In the following, the invention will be described with reference to the accompanying exemplary, schematic drawings, in which

Figure 1 illustrates a circulating fluidized bed reactor according to an embodiment of the invention,

Figure 2 illustrates a control system according to an embodiment of the invention, Figure 3 illustrates a circulating fluidized bed reactor according to another embodiment of the invention,

Figure 4 illustrates results obtained with method according to the invention, and Figure 5 illustrates a circulating fluidized bed reactor according to another embodiment of the invention.

Detailed Description of Drawings

[078] In the following the description of figures generally relates to examples of method of air combustion of fuel in a CFB reactor. Even if some minor structural changes may be needed, the CFB reactor and its embodiments described in the figures are as well applicable for producing syngas by practising gasification process in the reactor. Correspondingly the described CFB reactor and its embodiments may be utilized for practising so called oxy-combustion process, meaning combustion with oxygen enriched gas, which may contain air and/or recycled product gas. Figure 1 depicts schematically a circulating fluidized bed reactor 10 being specifically a circulating fluidized bed boiler 10, which is configured to produce superheated steam in a manner known per se. The circulating fluidized bed boiler will be referred to as CFB boiler for sake of conciseness. The CFB boiler 10 comprises a combustion chamber 12, at least one solid material separator 14 and a solid material return channel 16. Generally, the route of separated solid material from the separator back to the combustion chamber is called a return path 15. The combustion chamber 12 comprises tube walls so-called finned tube walls wherein fins are welded between tubes. The wall tubes are connected to a water-steam circuit of boiler system (not shown). The solid material separator 14 is preferably cooled, comprising also tube walls similarly to the combustion chamber 12. The combustion chamber 12 comprises a wind box 18, which is configured to feed fluidization gas, typically air, through nozzles of a grid 20 at the bottom of the combustion chamber 12. The air introduced through the grid acts as fluidization gas and is the primary combustion air. Secondary air may be fed into the combustion chamber 12 at higher level via one or more air inlets 21. As discussed earlier, the fluidizing gas and the combustion gases are usually air, but they can also comprise circulated product gas and/or oxygen or a mixture thereof. It should be understood that generally the term product gas may be understood that the gas exiting the separator 14 is a product gas of the reactions in the reaction chamber. For example, in case the method comprises gasification of fuel material, the product gas is combustible gas - so called syngas, and in case the method comprises combustion of fuel in the reaction chamber and producing steam with released heat, the product gas may be referred to as flue gas. In some practical applications, such as so-called oxygen combustion, part of the flue gas can be recycled back to the reaction chamber as fluidization gas, and it may therefore be referred to as recycling gas.

[079] The wind box 18, and also other air inlets, are in connection with a source of air 24. This is an example of an air operated CFB boiler. There is at least one inlet 22 for fuel in connection with the combustion chamber 12. Operation of the CFB boiler involves a process of circulation of solid material, which in this connection may also be referred to as bed material, as well. Bed material may comprise sand, limestone, and/or clay, that in particular may comprise kaolin, and also unburned fuel. Due to the bed material inside the boiler, CFB boilers have high heat transfer coefficients and substantially uniform temperature distribution and have a considerably low stable combustion temperature. Combustion of fuel in the circulating fluidized bed result in heating, evaporating the water in the water-steam circuit and superheating the steam, which can be used in a manner known as such, for example, production of electric power in a steam turbine generator. The steam cycle is not described here in more detailed manner.

[080] It is characteristic to the CFB boiler that during its operation a process of circulation of solid material is maintained via a route formed by the combustion chamber, a separator and solids return path. Combustion of fuel in the CFB results in high-efficiency combustion of various solid fuels with low emissions, even when burning fuels with completely different calorific values at the same time. Due to fluidization, there is an internal movement of solid material inside the combustion chamber 12 generally upwards at the mid-section of the chamber and downwards flow of solids near the walls, as is depicted by the arrows in the figure 1. Product gases and solid material is flown from the combustion chamber 12 to the solid material separator 14 inlet of which is connected to the combustion chamber by a connector duct 26. The solid material separator 14, which is preferably a cooled cyclone separator, comprises a first outlet 28 for product gas, and second outlet 30 for separated solid material 32. The first outlet 28 of the separator is, in practice, connected to a back pass 40. The back pass comprises a number of heat exchanges which may include an air preheater 46, an economizer 44, and superheaters and reheaters 42. The actual amount of different heat transfer surfaces in each of these components, for example, may be selected for each CFB boiler differently according to actual needs.

[081] The solid material which is transported by the product gases to the separator 14 is separated from the gas as is depicted by the arrow in the figure. The second outlet 30 which may also be referred to as a particle outlet, is connected to a lower part of the combustion chamber 12 by the return path 15 for returning separated solid material back to the combustion chamber 12. The return path 15 is provided with a so called loop seal 32 which prevents back-flow from the combustion chamber 12 to the particle outlet and makes it possible to feed separated solid material controllably back to the combustion chamber 12. The loop seal may also be referred to as a gas seal. Operation of the loop seal is controlled by fluidization air, which can be controllably supplied through an air inlet 23. Thus the circulation of solid material comprises the flow of solid material from the combustion chamber 12 to the solid material separator 14 and from the solid material separator 14 via the return path 15 and the loop seal 32 back to the combustion chamber 12. The process of circulation of solid material is maintained and controlled during the CFB boiler is operating. The general direction of movement of solid material may be used for referring to positions in the CFB boiler, which direction becomes clear in the description above.

[082] In CFB boilers, the solid material circulation may be divided in two categories: internal circulating material flow that means solids material circulating inside the combustion chamber (12) which are depicted schematically with upsidedown U-shaped arrows in Figure 1 and as discussed above, and external circulating material flows that means solid material circulating outside the combustion chamber i.e. in the particle separator and loop seal and return path 15 (in Figure 1). The external circulating material flows may also involve processing of solid material in a fluidized bed heat exchanger, as is shown in the figure 3. Advantageously, according to an aspect of the invention, the solid material in the return path may in other terms be referred as solid material in the external circulation of the CFB boiler.

[083] In order to monitor the process of circulation of solid material the CFB boiler 10 comprises a plurality of sensors for obtaining online data of performance indicators and process variables. The online data becomes history data when it is stored after the time moment of measurement. There are at least following sensors - and point of measurements - arranged to the CFB boiler for practising the method according to the invention: a first pressure sensor 101 arranged upstream to the loop seal 32 but downstream the solid material separator i.e. between the loop seal and the solid material separator 14; a second pressure sensor 102 arranged downstream to the loop seal 32, between the loop seal 32 and the combustion chamber 12;

The purpose of the first and the second pressure sensor is to determine pressure difference provided by the loop seal 32. a first temperature sensor 100 arranged in the loop seal 32; a second temperature sensor 103 arranged upstream to the loop seal 32; it is shown in the figure 1 that the second temperature sensor 103 is arranged at the first outlet 28 of the solid separator 14, which also represent the temperature of the solid material upstream to the loop seal 32 in the sense of in the figure 1 a first air flow rate sensor 105 for measuring the primary air flow rate through the grid 20 of the CFB boiler; a second air flow rate sensors 106 for measuring the secondary air flow rate;

The purpose of the first and the second air flow rate sensors is to determine an aggregate air flow rate fed into the CFB boiler. a number of third temperature sensors 104 in connection with the combustion chamber 12 arranged to determine an bed temperature in the combustion chamber 12.

[084] Additionally there may be an air flow rate sensor 109 for measuring the air flow rate to the loop seal, which may be optionally included in the aggregate air flow rate, as is depicted in the figure 1 . The temperature sensors downstream the second outlet 30 of the separator 14 provide measurement values which represent temperature of the solid material.

[085] The CFB boiler further comprises a control system 48 for handing numerical operations relating to control of the CFB boiler and particularly to monitor the process of circulation of solid material in the CFB boiler. It should be understood that the figure 1 is an exemplary description of process of circulation of solid material via a certain type of return path and description of process of circulation of solid material via one return path. Should the CFB boiler comprises more than one return path, which would often be the case in practise, the invention is applicable to each one of the return paths separately. Advantageously, the CFB boiler may be provided with the solid material separators 14 and corresponding return paths 15 on opposite sides of the combustion chamber 12 (not shown in figures).

[086] In figure 2 there is a general illustration of the method and the control system 48 in the control system by means of which the process of circulation of solid material in a CFB reactor 10 can be monitored such that an indication of a condition relating to the circulation of solid material that could lead to shutting down the reactor is observed early enough to take corrective actions and the shut down the reactor may be avoided.

[087] The control system 48 participates in practising a method of monitoring a process of circulation of solid material in the circulating fluidized bed boiler 10. Control system comprises one or more computers and executable instructions i.e. computer programs which, when executed in the control system 48 performs the method in the circulating fluidized bed boiler 10. The method comprises a. selecting performance indicators of the process of circulation of solid material and process variables for each performance indicator of the process of circulation of solid material, b. calibrating a multivariate model for each performance indicator, using history data of the process variables and the performance indicators of the process of circulation of solid material c. determining a modelled value of the performance indicators, by applying current measured values of the process variables to the multivariate model d. comparing the modelled value of each performance indicator to a respective measured value of each performance indicator and inspect presence of an anomaly between the modelled value and the measured value.

[088] The control system comprises a performance modelling unit 400. The modelling unit 400 comprises executable instructions which, when executed in the control system 48 calibrate the multivariate model for each performance indicator, using history data of predetermined process variables and the performance indicators of the process of circulation of solid material, resulting in a calibrated multivariate model.

[089] The performance modelling unit 400 has, or is provided with an access, such as a data transfer communication with a source of history data 401 of a) performance indicators of the process of circulation of solid material obtained from the CFB boiler, and b) process variables for each performance indicator. The history data is stored in a data media, which is used as the source of history data 401 , is obtained by measuring the values of predetermined process variables 101 , 102, 103,..109 (see figure 1) over a period of time, and storing the measured values with a time stamp thus forming the history data of process variables. Acquiring the history data involves respectively measuring the values of performance indicator, storing the measured values with a time stamp thus forming the history data of performance indicators. The control system 48 comprises a data filter unit 406 which is configured to filter out invalid process data and thus the history data comprises data which has been subjected to filtering process of using predetermined data filters. The history data is therefore data which describes normal operation conditions of the CFB boiler 10.

[090] Advantageously the filtering process 406 may comprise the following conditions or rules. Firstly, a quarantine time for measured data is set so that only the data which is older than a pre-set quarantine time is approved. The quarantine time depends on the case. In some practical applications the quarantine time may be even as short as three to seven days. However preferably the quarantine time is seven to fourteen days, even more preferably at least two weeks. Additionally, it is preferred to filter out data which may be obsolete due to being too old and therefore the predetermined data filter is configured to approve data, which is not older than a predetermined time, advantageously not older than two months. Also, the filter unit is configured to filter out from history data any data from shut down situations and/or data originating from any abnormal operation condition, e.g. based on predefined limits for input variables or an external information setting the data to be unusable or data of abnormal operation.

[091] This way the model is based on history data which represents normal operation conditions. The history data relating to the CFB boiler shown for example in the figure 1 comprises data of process variables and the performance indicators:

- pressure values (sensor 101) upstream the loop seal 32 but downstream the solid material separator i.e. between the loop seal and the solid material separator 14;

- pressure values (sensor 102) downstream to the loop seal 32, between the loop seal 32 and the combustion chamber 12;

- or alternatively, pressure difference over the loop seal 32 (combined sensors 101 ,102)

- temperature (sensor 103) upstream the loop seal 32

- aggregate air flow rate (sensors 105,106) fed into the CFB boiler

- aggregate bed temperature (sensors 104) in the combustion chamber 12.

[092] The performance indicators represent factors which describe the state of the process of circulation of solid material. In case of the embodiment shown in the figure 1 the performance indicators are advantageously i. pressure difference of a loop seal in the return path and ii. temperature in the loop seal in the return path in the circulation of solid material

[093] The data in the source of history data 401 is used as an input for the modelling unit 400, which is configured to prepare and/or calibrate a multivariate model assigned separately for each performance indicator, in this case for two performance indicators. The performance modelling unit 400 thus provides the multivariate model for each performance indicator. The calibration may be repeated at predefined intervals, or periodically. This helps to keep the model actual, reflecting the possible changes caused by normal use of the CFB boiler, but also to changes in fuel quality, the environmental conditions (temperature, ambient humidity, ambient pressure changes) which may lead to operation parameters changing over time. The calibration may be prevented upon detecting an anomaly in the process. In this manner it may be ensured that a problem in bed material circulation that is just developing will not contaminate the calibration and the model.

[094] The model can be constructed by the modelling unit 400 using a multivariate linear regression. In principle, past input values of the model (i.e., history data of measurements) are used for estimating the coefficients of the model. The model is then used to estimate the prevailing situation by making use of current online data and the estimated coefficients.

[095] For example, in linear regression the response variable is expected to be a linear combination of process variables. Multiple linear regression can be used to model the relationship between multiple process variables and performance indicator by fitting a linear equation to history data.

[096] In case of the embodiment shown in the figure 1 a multivariate model for loop seal temperature with N observations is defined as follows:

Yi = b ₀ + biX; i + b ₂ x _{i 2} + b ₃ Xi ₃ + £; , where y denotes the value of a performance indicator as the loop seal temperature, Xj,i is the /th value of temperature upstream the loop seal 32 (sensor 103) Xi, 2 is the /th value of aggregate air flow rate (sensors 105,106) fed into the CFB boiler

Xj,3 is the /th value of bed temperature (sensors 104) in the combustion chamber 12 bo is a constant, bi ... bs are the unknown, KPI-specific coefficients to be estimated, and

£i comprises experimental errors of the model.

[097] In case of the embodiment shown in the figure 1 a multivariate model for loop seal pressure difference with N observations is defined as follows:

Yi = b ₀ + biX^ + b ₂ x; , ₂ + b ₃ Xi ₃ + £; , where y denotes the value of a performance indicator,

Xj,i is the /th value of temperature upstream the loop seal 32 (sensor 103) Xj,2 is the /th value of aggregate air flow rate (sensors 105,106) fed into the CFB boiler

Xj,3 is the /th value of bed temperature (sensors 104) in the combustion chamber 12 bo is a constant, bi ... bs are the unknown, KPI-specific coefficients to be estimated, and

£i comprises experimental errors of the model.

The fitting is performed by minimizing the sum of the squares of the vertical deviations from each data point to the line that fits best for the observed data, that is the optimal coefficient values by minimizing the sum of squared errors.

[098] The modelling unit 400 provides required coefficients of the model which are based on viable history data, to be used for modelling the performance indicators by applying online data of process variable to the model. During the control system 48 and the CFB boiler 10 are in operation the history data comprising data of process variables and the performance indicators is continuously read and stored to the source of history data 401 . The modelling unit 400 is configured to update or calibrate the model i.e. the coefficients of the model in order to learn the model the latest conditions of normal operation of the process of circulation of solid material.

[099] There is also a performance diagnostic module 404 provided in the control system 48. The performance diagnostic module is configured to receive current online data by a source of current data 402 from the CFB boiler of the performance indicators, and the process variables and a newly calibrated model of the performance indicators from the modelling unit 400. The performance diagnostic module 404 comprise instructions to determine a modelled value of the performance indicators by applying current measured values of the process variables to the calibrated multivariate model. Additionally the performance diagnostic module 404 is configured to compare the modelled value of each performance indicator to a measured respective value of performance indicator and inspecting a presence of an anomaly between the modelled value and the measured value. Based on the outcome of the comparison a predetermined measure or measures can be taken is generated as a diagnostic output 408. [0100] The presence of an anomaly and need for remedial actions, can be realized by estimating a risk index of each KPI. The performance diagnostic module 404 may comprises instructions to practise a method estimating the risk index for a performance indicator, which performs following acts:

- current data of performance indicators (KPI) of circulation of solid material is measured,

- based on the current data of the boiler, at least one of the following i) an average of the performance indicators is computed; ii) standard deviation of measured performance indicators is computed; iii) a difference between maximum measured performance indicator value and minimum measured performance indicator is computed; iv) difference between average performance indicator KPI and measured performance indicators is computed ;

- using the computation results from i), ii), iii) and/or iv), preparing a risk index for the performance indicator KPI. The computation results from i), ii), iii) and/or iv) are compared with corresponding predefined limits so as to get risk indexes for average, standard deviation, difference between maximum and minimum KPI, and difference between average KPI and measured KPIs. In computation of deviation of a KPIk from average KPI the average includes all KPI measurements except the measurement of KPIk.

[0101] Preferably, in the method also, or alternatively v) modelled values of KPk; k = 1 , ..., K are computed, and residuals between the measured values of the performance indicators and the modelled values of the performance indicators are computed. The results from the step v) are advantageously also used in the preparing of the risk index, preferably such that residuals are compared with corresponding predefined limit so as to get sintering risk index for KPI residuals.

[0102] The final risk index may then be the maximum of above risk indexes, for example. In this manner, the predictive accuracy of bed sintering index can be still improved.

[0103] The present inventors have observed that in this manner, the resulting risk index provides an indication of a condition in the process of circulation of solid material in a circulating fluidized bed boiler, that could lead to shutting down the boiler unless treated, early enough to take corrective actions such that the need to shut down the boiler may be avoided.

[0104] Optionally the control system comprises a storage of history coefficients of the model 410, where each calibrated model is stored. The performance diagnostic module 404 may comprise a model evaluation function, which check the newly created model and in case newly created model is found to be imperfect, a model from the storage of history coefficients of the model 410 is used until an intact fresh model can be provided.

[0105] Figure 3 depicts schematically a circulating fluidized bed boiler 10 which has at one of its return paths 15 (only one shown for clarity reasons) provided with a fluidized bed heat exchanger 50. It should be understood that the return path 15 shown in the figure 3 may be conceived to be in the same CFB boiler disclosed in the figure 1 , which means that the CFB boiler may be provided with several return paths 15, where more than one of the return paths 15 is preferable provided with a fluidized bed heat exchanger 50. Figure 3 refers also to a practical application where the CFB boiler comprises several return paths 15, all of which are provided with a fluidized bed heat exchanger 50. The method according to the invention is the practised in connection with all of the return paths 15 separately.

[0106] The fluidized bed heat exchanger 50 is arranged to the return path 15 downstream the loop seal 32 in the return channel 16. Solid material flows through the loop seal 32 into the fluidized bed heat exchanger 50 where a bubbling bed of solid material is formed by introducing fluidization air into the fluidized bed heat exchanger 50 through a grid 52 at the bottom thereof. The fluidized bed heat exchanger 50 is provided with a lifting chamber 54 with respective inlet 54 of transport air. The lifting chamber transfers the solid material from the fluidized bed heat exchanger 50 back to the combustion chamber 12 via a return duct 55.

[0107] The fluidized bed heat exchanger 50 is provided with one or more heat exchange units 58, which are preferably connected to, for example the steam cycle. The heat exchange units may be, evaporators, steam superheaters and/or steam reheaters. The heat exchange units comprise heat transfer surface, such as one or more tube bundles inside the bubbling bed of solid material formed the fluidized bed heat exchanger 50.

[0108] In the CFB boiler, the solid material which is transported by the product gases to the separator 14 is separated from the gas as is depicted by the arrow in the figure. The second outlet 30 which may also be referred to as a particle outlet of the separator 14, is connected to a lower part of the combustion chamber 12 by the return path 15 for returning separated solid material back to the combustion chamber 12. The return path 15 is provided with a so called loop seal 32 which prevents back-flow from the combustion chamber 12 to the particle outlet and makes it possible to feed separated solid material controllably forward in the return path 15. Operation of the loop seal is controlled by fluidization air, which can be controllably supplied through an air inlet 23. Thus the circulation of solid material comprises the flow of solid material from the combustion chamber 12 to the solid material separator 14 and from the solid material separator 14 via the return channel 16 to the fluidized bed heat exchanger 50, and from the fluidized bed heat exchanger 50 back to the combustion chamber 12. While the fluidized bed heat exchanger 50 is operated heat is transferred from the solid material to the steam flowing in the heat exchange unit 58 this cooling the solid material prior to its introduction back to the combustion chamber 12.

[0109] In order to monitor the process of circulation of solid material the CFB boiler 10 according to the embodiment of figure 3 comprises a plurality of sensors for obtaining online data of performance indicators and process variables. The online data becomes history data when it is stored after the time moment of measurement. There are at least following sensors - and point of measurements - arranged to the CFB boiler for practising the method according to the invention: a first pressure sensor 101 arranged upstream to the loop seal 32 but downstream the solid material separator i.e. between the loop seal and the solid material separator 14; a second pressure sensor 102 arranged downstream to the loop seal 32, between the loop seal 32 and the fluidized bed heat exchanger 50;

The purpose of the first and the second pressure sensor is to determine pressure difference provided by the loop seal 32. a first temperature sensor 100 arranged in the loop seal 32; a second temperature sensor 103 arranged upstream to the loop seal 32; it is shown in the figure 2 that the second temperature sensor 103 is arranged at the first outlet 28 of the solid separator 14, which also represent the temperature of the solid material upstream to the loop seal 32 in the sense of in the figure 2 a first air flow rate sensor 105 for measuring the primary air flow rate through the grid 20 of the CFB boiler; a second air flow rate sensors 106 for measuring the secondary air flow rate;

The purpose of the first and the second air flow rate sensors is to determine an aggregate air flow rate fed into the CFB boiler. A number of third temperature sensors 104 in connection with the combustion chamber 12 arranged to determine an bed temperature in the combustion chamber 12; a third air flow rate sensor 108 for measuring the air flow rate fed to the fluidized bed heat exchanger 50; a third pressure sensor 110 arranged upstream to the heat exchange unit 58 in the fluidized bed heat exchanger 50; a fourth pressure sensor 112 arranged downstream to the heat exchange unit 58 in the fluidized bed heat exchanger 50 a third temperature sensor 114 arranged downstream to the heat exchange unit 58 in the fluidized bed heat exchanger 50

[0110] The control system 48 described in the figure 2 is applicable to the CFB boiler 10 described in the figure 3 with the necessary modifications relating to data of performance indicators and process variables. The actual, accurate location of sensors can be determined case by case. For example, the third temperature sensor may in some cases be arranged between the lifting chamber 54 and combustion chamber 12, because temperature in the particular location represents the temperature of the solid material downstream the heat exchange unit 58. Similarly, the temperature sensor 103 at the outlet 28 of the separator 14 may be positioned differently, as long as it represents the temperature of the solid material upstream position to the loop seal 32. [0111] When applied to the CFB boiler according to figure 3 the control system 48 participates in practising a method of monitoring a process of circulation of solid material in the circulating fluidized bed boiler 10. Control system comprises one or more computers and executable instructions i.e. computer programs which, when executed in the control system 48 performs the method in the circulating fluidized bed boiler 10. The method comprises a. selecting performance indicators of the process of circulation of solid material and process variables for each performance indicator of the process of circulation of solid material, b. calibrating a multivariate model for each performance indicator, using history data of the process variables and the performance indicators of the process of circulation of solid material c. determining a modelled value of the performance indicators, by applying current measured values of the process variables to the multivariate model d. comparing the modelled value of each performance indicator to a respective measured value of each performance indicator and inspect presence of an anomaly between the modelled value and the measured value.

[0112] The control system comprises a performance modelling unit 400. The modelling unit 400 comprises executable instructions which, when executed in the control system 48 calibrate the multivariate model for each performance indicator, using history data of predetermined process variables and the performance indicators of the process of circulation of solid material, resulting in a calibrated multivariate model.

[0113] The performance modelling unit 400 has, or is provided with an access, such as a data transfer communication with a source of history data 401 of a) performance indicators of the process of circulation of solid material obtained from the CFB boiler, and b) process variables for each performance indicator. The history data is stored in a data media, which is used as the source of history data 401 , is obtained by measuring the values of predetermined process variables 101 ,102,103,..114 (see figure 3) over a period of time, and storing the measured values with a time stamp thus forming the history data of process variables. Acquiring the history data involves respectively measuring the values of performance indicator, storing the measured values with a time stamp thus forming the history data of performance indicators. The control system 48 comprises a data filter unit 406 which is configured to filter out invalid process data and thus the history data comprises data which has been subjected to filtering process of using predetermined data filters. The history data is therefore data which describes normal operation conditions of the CFB boiler 10. The filter unit is described more detailed in connection with the description of the figure 2.

[0114] The history data relating to the CFB boiler shown in the figure 3 comprises data of process variables and the performance indicators:

- pressure values (sensor 101) upstream the loop seal 32 but downstream the solid material separator i.e. between the loop seal and the solid material separator 14;

- pressure values (sensor 102) downstream to the loop seal 32, between the loop seal 32 and the fluidized bed heat exchanger 50;

- or alternatively, pressure difference over the loop seal 32 (combined sensors 101 ,102);

- temperature (sensor 100) in the loop seal 32;

- temperature (sensor 103) upstream to the loop seal 32;

- pressure values (sensor 110) upstream the heat exchange unit 58 in the fluidized bed heat exchanger 50;

- pressure values (sensor 110) downstream to the heat exchange unit 58 in the fluidized bed heat exchanger 50;

- temperature (sensor 114) downstream to the heat exchange unit 58 in the fluidized bed heat exchanger 50;

- aggregate air flow rate (sensor 108) fed into the fluidized bed heat exchanger 50;

- aggregate air flow rate (sensors 105,106) fed into the CFB boiler;

- bed temperature (sensors 104) in the combustion chamber 12

[0115] The performance indicators represent factors which describe the state of the process of circulation of solid material and the fluidized bed heat exchanger. In case of the embodiment shown in the figure 3 the performance indicators are advantageously i. pressure difference of a loop seal in the return path and ii. temperature in the loop seal in the circulation of solid material iii. pressure difference of the fluidized bed heat exchanger, iv. temperature of solid material downstream the fluidized bed heat exchange unit

[0116] The data in the source of history data 401 is used as an input for the modelling unit 400, which is configured to prepare and/or calibrate a multivariate model assigned separately for each performance indicator, in this case for two performance indicators. The performance modelling unit 400 thus provides the multivariate model for each performance indicator. The calibration may be repeated at predefined intervals, or periodically. This helps to keep the model actual, reflecting the possible changes caused by normal use of the CFB boiler, but also to changes in fuel quality, the environmental conditions (temperature, ambient humidity, ambient pressure changes) which may lead to operation parameters changing over time. The calibration may be prevented upon detecting an anomaly in the process. In this manner it may be ensured that a problem in bed material circulation that is just developing will not contaminate the calibration and the model.

[0117] The model can be constructed by the modelling unit 400 using a multivariate linear regression. In principle, past input values of the model (i.e., history data of measurements) are used for estimating the coefficients of the model. The model is then used to estimate the prevailing situation by making use of current online data and the estimated coefficients.

[0118] For example, in linear regression the response variable is expected to be a linear combination of process variables. Multiple linear regression can be used to model the relationship between multiple process variables and performance indicator by fitting a linear equation to history data.

[0119] In case of the embodiment shown in the figure 3 a multivariate model for the temperature of solid material downstream fluidized bed heat exchange unit (FBHX in the following) with N observations is defined as follows: yi = b ₀ + biX; i + b ₂ xj ₂ + b ₃ Xi ₃ + b ₄ x _ii4 + b ₅ x _ii5 + £; , where y denotes the value of a performance indicator,

Xj,i is the /th value of pressure difference over the loop seal 32 (sensors 110,112), Xi, 2 is the /th value of temperature in the loop seal 32 (sensor 100) Xj,3 is the /th value of aggregate air flow rate (sensors 105,106) fed into the CFB boiler

Xj,4 is the /th value of bed temperature (sensors 104) in the combustion chamber 12

Xj,5 is the /th value of air flow rate (sensor 108) fed to the chamber of the fluidized bed heat exchanger 50 bo is a constant, bi ... b5 are the unknown, KPI-specific coefficients to be estimated, and

E comprises experimental errors of the model.

[0120] In case of the embodiment shown in the figure 3 a multivariate model for the pressure difference over the fluidized bed heat exchanger with N observations is defined as follows: yi = b ₀ + biXj.i + b ₂ xj ₂ + b ₃ Xi ₃ + b ₄ x _ii4 + b ₅ x _ii5 + £; , where y denotes the value of a performance indicator,

Xj,i is the /th value of pressure difference over the loop seal 32 (sensors 110,112), , 2 is the /th value of temperature in the loop seal 32 (sensor 100)

Xj,3 is the /th value of aggregate air flow rate (sensors 105,106) fed into the CFB boiler

Xj,4 is the /th value of bed temperature (sensors 104) in the combustion chamber 12

Xj,5 is the /th value of air flow rate (sensor 108) fed to the chamber of the fluidized bed heat exchanger 50 bo is a constant, bi ... b5 are the unknown, KPI-specific coefficients to be estimated, and

£ comprises experimental errors of the model.

The fitting is performed by minimizing the sum of the squares of the vertical deviations from each data point to the line that fits best for the observed data, that is the optimal coefficient values by minimizing the sum of squared errors.

[0121] The modelling unit 400 provides required coefficients of the model which are based on viable history data, to be used for modelling the performance indicators by applying online data of process variable to the model. During the control system 48 and the CFB boiler 10 are in operation the history data comprising data of process variables and the performance indicators is continuously read and stored to the source of history data 401 . The modelling unit 400 is configured to update or calibrate the model i.e. the coefficients of the model in order to learn the model the latest conditions of normal operation of the process of circulation of solid material.

[0122] There is also a performance diagnostic module 404 provided in the control system 48 which is applicable to the CFB boiler provided with one or more fluidized bed heat exchangers also. The description or the performance diagnostic module in connection with figure 2 is therefore applicable to the embodiment of the figure 3 as well.

[0123] Figure 4 the describes results obtained with method of monitoring a process of circulation of solid material in a circulating fluidized bed boiler according to figure 3. Figure 4 discloses online measurement results provided by the third temperature sensor 114. The third temperature sensor is located in the fluidized bed heat exchanger below the heat exchange unit 58, near the grid of the chamber, and thus the curve M114 shows the temperature of the performance indicator iv. -temperature of solid material downstream the fluidized bed heat exchange unit. The other curve P114 shown in the figure 4 depicts modelled values of the performance indicator when current measured values of the process variables are applied to the multivariate model of the KPI . Horizontal axis shows time where the zero-moment is the actual time of shut down of the boiler. As can be seen in the chart, the modelled value of the model shows a deviation from the measured value several hours before the process will be too disturbed to recover, and no remedy action would prevent the shut down. As it becomes clear from the example the model will indicate the coming problem more than 30 hours before the shut down it will be irreversible. Unfavourable conditions can be observed early enough within a time window (hatched area) that is suitably long and sufficiently much in advance before occurrence of the actual problem.

[0124] Figure 5 depicts schematically a circulating fluidized bed boiler 10 which has at one of its return paths 15 (only one shown for clarity reasons) provided with a fluidized bed heat exchanger 50 and with a by-pass path 56 which connects the solid material return path 15 at a location between the loop seal 32 and the fluidized bed heat exchanger 50 to the combustion chamber 12. Thus, the by-pass path 56 is arranged for controllably passing 0-100% of the solid material flow in the return path 15 directly to the combustion chamber, while passing the possible remining portion to the fluidized bed heat exchanger 50. This way process of circulation of solid material comprises two modes: the first one (the bypass mode) in which method is applied to the circulation of solid material directly form the loop seal 32 to the combustion chamber 12 of the CFB boiler (when there is a flow of solid material through the by-pass path), and the second one in which the method is applied to the circulation of solid material form the loop seal 32 to the combustion chamber 12 of the CFB boiler via the fluidized bed heat exchanger 50 (when there is a flow of solid material through the fluidized bed heat exchanger 50). Hence, the embodiment shown in the figure 5 can be understood to be a combination of the embodiments shown in the figures 1 and 3 in a single solid material return path 15 in terms of applying the method according to the invention.

[0125] It is to also to be understood that the return path 15 shown in the figure 5 may be conceived to be in the same CFB boiler disclosed in the figure 1 or figure 3, which means that the CFB boiler may be provided with several return paths 15 with different setup, where more than one of the return paths 15 is preferable provided with a fluidized bed heat exchanger 50. Figure 5 refers also to a practical application where the CFB boiler comprises several return paths 15, all or some of which are provided with a fluidized bed heat exchanger 50 with a bypass path 55. The method according to the invention is the practised in connection with all of the return paths 15 separately.

[0126] In addition, or alternatively to having the by-pass path 56, in the figure 5 there is disclosed a solid material discharge path 56’ connected to the solid material return path 15 at a location between the loop seal 32 and the fluidized bed heat exchanger 50. Location of the discharge path 56’, or its point of extraction of material, may be other than shown here if so desired. By means of the solid material discharge path 56’ it possible to remove 0-100% of the solid material flow from the process of circulation of solid material. The portion of removed material may be later returned back to the reactor 10, as such, or after a desired processing performed to the solid material. [0127] The control system 48 described in the figure 2 is applicable to the CFB boiler 10 described in the figure 5 with the necessary modifications relating to data of performance indicators and process variables.

[0128] An exemplary embodiment of calculation of residual-based KPIs and the risk index for a case wherein a circulating fluidized bed boiler 10 has at one of its return paths 15 provided with a fluidized bed heat exchanger 50. The following steps are taken:

• Creating the KPI models based on pressure difference and temperature in loop seal and pressure difference and temperature in the fluidized bed heat exchange chamber

• Comparing the modelled values of the KPI’s to measured values at the current point of time (t), for example, KPI for a modelled loop seal temperature can be computed as follows

KPIloop seal temp, modelled (0 — Yt ^— t>0 T b^X^ j + b 2 ^x t,2 T bgXf g + £[ where bo is a KPI-specific constant (as solved earlier), and bo... bs are known (as solved earlier) KPI-specific coefficients, and x _t ,i is the fth value of temperature upstream the loop seal 32 (sensor 103) x _t ,2 is the fth value of aggregate air flow rate (sensors 105,106) fed into the CFB boiler x _t ,3 is the fth value of bed temperature (sensors 104) in the combustion chamber 12.

• Comparison being made of computing deviations between the model outputs and the measured values so as to obtain residuals (KPIk,res(t) where k=1 to K (K = number of KPIs) (e.g. modelled temperature - measured temperature in the loop seal that is KPIi ₀₀ pseaitermp,res(t) = KPI loop seal temp, modelled^)- KP oop seal temp, meas (t)) [0129] The residual limits for each KPI type are shown schematically in below table: wherein A, B, C and D depict a predefined limit values. • Calculating the risk index for each KPIs as follows: r _k = 100 X ( I KPI _k ,res(t) - (lup, _k + llo, _k )/2 I ) / ( (lup, _k - l|o, _k )/2 ) where |.| is an absolute value, k= 1 to K (K=number of KPIs)

Calculating the overall risk index Rl as follows:

Rl = max(r _k ), where r _k = single risks and where k= 1 to K (K=number of KPIs).

[0130] As an example, let us assume that residual for temperature in loop seal is KPIioop seal temp, res(t) ⁼ KPIioop seal temp, modelled(t) - KPIioop seal temp, meas (t)=B _up . Then, using above formula, we get for the loop seal temperature risk index: hoop seal temp ^— 100 X ( | Bup ^— (Bup + B|o)/2 | ) / ( (Bup - B|o)/2 ). In case B _up = B and B| _O = -B, then ri _oop seal temp =100 and thus calculation of the overall risk index with above formula results in Rl = max(r _k ) = 100. [0131] The case without the fluidized bed heat exchange chamber goes similarly than above example but omitting the values (KPIs) related to the fluidized bed heat exchange chamber.

[0132] According to an aspect of the invention, the overall risk index may be calculated using at least one of the following formulas: as maximum Rl = max(rk), average Rl=mean(rk), weighted average RI=Wmean(rk) or median Rl=me- dian(r _k ).

[0133] According to a preferable aspect of the invention, risk index for each KPIs is limited to have a maximum value of 100 and lowest value of 0 i.e. rk = [0,... ,100]. So, if absolute value of KPIk is greater than absolute value of lower limit (lio.k) or upper limit (l _up ,k), then rk=100. Generally, if KPIk does not belong to the interval [lio.k, lup.k], then rk=100. It is also possible to have a condition written as if 100 x ( | KPIk(t) - (lup.k + lio,k)/2 | ) / ( (l _up ,k - ho,k)/2 ) > 100, then r _k = 100 and else r _k = 100 x ( | KPI _k (t) - (l _up ,k + ho,k)/2 | ) / ( (l _up ,k - ho,k)/2 ).

[0134] In the above example, table indicate that absolute limit values may be equal, however, it is possible that upper limits and lower limits may be defined differently so that the absolute values of upper and lower limits differ for the corresponding KPI. It should be noted, however, that li ₀ ,k< lup.k-

[0135] The above-described example is made for clarifying purposes only and not meant to limit the scope of claimed invention. Furthermore, instead of residuals, other mathematical comparison is possible, for example, computing a ratio between corresponding values.

[0136] While the invention has been described herein by way of examples in connection with what are, at present, considered to be the most preferred embodiments, it is obvious to the skilled person that, along with the technical progress, the basic idea of the invention can be implemented in many ways. The details mentioned in connection with any embodiment above may be used in connection with another embodiment when such combination is technically feasible. Part list a circulating fluidized bed boiler 10 a combustion chamber 12 a solid material separator 14 a solid material return path 15 a solid material return channel 16 a wind box 18 a grid 20 an air inlet 21 a loop seal fluidization air inlet 23 a source of air 24 a fuel inlet 22 a duct 26 a first outlet 28 a second outlet 30 a loop seal 32 a back pass 40 superheaters and optional reheaters 42 an economizer 44 an air preheater 46 a control system 48 a fluidized bed heat exchanger 50 a grid of the fluidized bed heat exchanger 52 a lifting chamber 54 a return duct 55 a by-pass path 56 a heat exchange unit 58 a first temperature sensor 100 a first pressure sensor 101 a second pressure sensor 102 a second temperature sensor 103 a third temperature sensors 104 a first air flow rate sensor 105 a second air flow rate sensors 106 a third air flow rate sensor 108 a fourth air flow rate sensor 109 a third pressure sensor 110 a fourth pressure sensor 112 a third temperature sensor 114 a performance modelling unit 400 source of history data 401 source of current data 402 a performance diagnostic module 404 a data filter unit 406 a diagnostic output 408 a storage of history coefficients of the model 410