INFORMATION PROCESSING DEVICE, INFORMATION PROCESSING METHOD AND PROGRAM

Abstract

Information processing device (40) comprising: a measured value acquisition unit (401) configured to receive measured values ​​of state variables measured by a plurality of measuring instruments (20); an omission detection unit (402) configured to detect, as a target state quantity, a state quantity whose value is a temporal or spatial omission among state quantities to be processed on the basis of the measured values ​​of a target device (10); an estimation unit (404) configured to estimate each of a plurality of estimated values ​​regarding the target state quantity in which the omission was detected using a plurality of models for explaining the target device (10) based on the measured values; a probability specification unit (406) configured to specify each of a plurality of probabilities of estimated values ​​based on a probability distribution of state variable values ​​with respect to the target state variable; and a management value specifying unit (407) configured to specify a management value to be used for management of the target device (10) based on the plurality of estimated values ​​and the plurality of probabilities.

Description

[Technical field]

[0001] The invention relates to an information processing apparatus, an information processing method and a program. [State of the art]

[0002] In a plant such as a power plant, it is provided that a monitoring device collects state quantity values ​​of a target device constituting the plant, such as temperature and pressure at the time of operation of the target device, and that the collected state quantities are used for maintenance, monitoring and the like of the device.

[0003] It has been proposed that the monitoring device processes the collected state variables in such a way that an operator of the device can easily use them and perform maintenance, monitoring, and the like of the device. For example, Patent Document 1 proposes that when a computer detects omitted parts of the collected state variables of an industrial plant, the computer performs a complementary process and calculates the values ​​of the omitted state variables.

[0004] In addition, it is known that models such as a physical model and a statistical model are used when values ​​of omitted state variables are to be adjusted by a complementary domain process.

[0005] Patent Document 2 discloses a device for real-time monitoring of a substance in a human or animal, the device comprising a sensor that provides a time series of substance level measurements, the measurements indicative of an inferred level of the substance in a part of the human or animal, and a processor that applies an interacting multiple-model strategy to a system model to provide a combined estimate of the inferred substance level from the substance level measurements. The substance may be glucose. The device may also be configured to control the substance using the interacting multiple-model strategy to a system model to provide a combined estimate of a dose to be administered.

[0006] Patent Document 3 discloses that a data sampling part samples N pieces of input / output data on a process. This sampled data is fed to a parameter decision part, and the parameters of multiple process models are obtained through a nonlinear optimization process. Then, a posterior probability arithmetic part obtains a probability density function based on the scattered and observed values ​​obtained from the parameter estimate determined by the part. The probability density function is then operated by Bayes' theorem to obtain the posterior probability of the process model. Thus, optimal process control is achieved by identifying a process model that has the maximum posterior probability. Literature listPatent documents [Patent Document 1] US 2016 / 0004794 A1 [Patent Document 2] US 8,977,504 B2 [Patent Document 3] JP S64-15808 A [Summary of the invention][Technical problem]

[0007] Patent Document 1 discloses that an information processing apparatus performs a complementary process for supplementing a value of an omitted state quantity, but does not describe a specific method for setting a value of an omitted state quantity.

[0008] In addition, although a value assumed by a state variable of a device follows a fixed probability distribution, a value of an omitted state variable is not set based on an estimated value of a state variable estimated using each of the models in view of such a probability distribution.

[0009] The invention has been made in view of the problem described above, and one of its objects is to appropriately set state quantity values ​​used for the management of a target device based on an estimated value of a state quantity calculated using a plurality of models, in view of a probability distribution of the state quantity values. [Solution to the problem]

[0010] The present invention provides an information processing apparatus according to independent claim 1, an information processing method according to independent claim 10, and a program according to independent claim 11. Advantageous modifications can be found in the dependent claims. [Advantageous effects of the invention]

[0011] The information processing apparatus according to at least one aspect of the above-described aspects can appropriately set a value of a state quantity to be used for management of a target device based on an estimated value of a target state quantity calculated using each model in view of a probability distribution of state quantity values. [short description of the drawings] Fig. 1 is a schematic block diagram showing a management system according to a first embodiment. Fig. 2 is a schematic block diagram showing a configuration of a management apparatus according to the first embodiment. Fig. 3 is a flowchart showing operations of the management device according to the first embodiment. Fig. 4 is a diagram showing a specific example of a method for specifying a management value according to the first embodiment. Fig. 5 is a schematic block diagram showing a configuration of a management device according to a second embodiment. Fig. 6 is a flowchart showing operations of a management device according to a second embodiment. Fig. 7 is a diagram showing a specific example of a method for specifying a management value according to the second embodiment. Fig. 8 is a flowchart showing operations of a management device according to a third embodiment. Fig. 9 is a diagram showing a specific example of a method for specifying a management value according to the third embodiment. Fig. 10 is a flowchart showing operations of a management device according to a fourth embodiment. Fig. 11 is a flowchart showing a specific example of a method for specifying a management value according to the fourth embodiment. Fig. 12 is a schematic block diagram showing a configuration of a management device according to a fifth embodiment. Fig. 13 is a flowchart showing operations of the management device according to the fifth embodiment Fig. 14 is a schematic block diagram showing a configuration of a computer according to at least one embodiment. [Description of the Embodiment]First Embodiment

[0012] In the following, embodiments will be described in detail with reference to the accompanying drawings.

[0013] Fig. 1 is a schematic block diagram showing a configuration of a management system according to a first embodiment.

[0014] A management system 1 includes a target device 10, a plurality of measuring instruments 20, a communication device 30, and a management device 40.

[0015] The target device 10 is a device to be managed by the management device 40. Examples of the target device 10 include a gas turbine, a steam turbine, a boiler, a coal gasification furnace, and the like. Additionally, the target device may be an environmental facility, a chemical plant, or a transportation system such as an aircraft.

[0016] The measuring instrument 20 is provided in the target device 10 and measures a state variable of the target device 10.

[0017] The communication device 30 transmits a measured value of the state quantity measured by the measuring instrument to the management device 40 via a network N.

[0018] The management device 40 manages the target device 10 based on the measurement value received from the communication device. The management device 40 is an example of an information processing device. Configuration of the management device

[0019] Fig. 2 is a schematic block diagram showing a configuration of the management apparatus according to the first embodiment.

[0020] The management device 40 includes a measurement value acquisition unit 401, an outlet acquisition unit 402, a storage unit 403, an estimation unit 404, a probability distribution storage unit 405, a probability specification unit 406, a management value specification unit 407, and a management unit 408.

[0021] The measured value acquisition unit 401 receives measured state variable values ​​measured by a plurality of measuring instruments 20 from the communication direction 30.

[0022] The omission detection unit 402 detects a state variable whose value has been omitted from the state variables to be processed based on the plurality of measured values ​​acquired by the measured value acquisition unit 401. Here, the omission of a value refers to a temporal or spatial omission. For example, in a case where the management unit 408 manages a state variable for each time Δt, the omission detection unit 402 detects an omission of a measured value at time T+Δt when one measured value is acquired at time T and one measured value is acquired at time T+2Δt.Furthermore, in a case where the management unit 408 manages a state quantity for each distance Δd, omission of measurement values ​​at a position (0, Δd), a position (Δd, 0), a position (Δd, Δd), a position (Δd, 2Δd), and a position (2Δd, Δd) is detected when measurement values ​​at a position (0, 0), a position (2Δd, 0), a position (0, 2Δd), and a position (2Δd, 2Δd) are acquired.

[0023] The model storage unit 403 stores a plurality of models for explaining the movement of the target device 10. A statistical model and a physical model can be used as models. In addition, a control model and a knowledge model can be used. The statistical model is a model that statistically reproduces the movement of the target device 10 based on a value of a state variable during the past operation of the target device 10. The statistical model is updated based on stored state variables from the past operation. The physical model is a model that reproduces the movement of the target device 10 using a numerical expression (e.g., a thermodynamic equation) that follows the laws of nature, based on design information of the target device 10.

[0024] The estimation unit 404 estimates a value of a state variable for each model stored in the model storage unit 403 based on a measured value measured by the measured value acquisition unit 401. Hereinafter, a state variable to be estimated by the estimation unit 404 is referred to as a target state variable. That is, the estimation unit 404 calculates values ​​of a plurality of target state variables using different models.

[0025] The probability distribution storage unit 405 stores a probability distribution table in which a value of a target state variable and a probability of assuming that value are related to each other. The probability distribution table is obtained in advance using design information of the target device 10 or statistics of past state variables. Meanwhile, the probability distribution storage unit 405 may store a probability distribution function instead of a probability distribution table.

[0026] The probability specifying unit 406 specifies a probability of accepting an estimated value of a target state quantity obtained by the estimating unit 404 for each estimated value based on a probability distribution stored in the probability distribution storing unit 405.

[0027] The management value specifying unit 407 selects one of a plurality of estimated values ​​estimated by the estimating unit based on a probability specified by the probability specifying unit 406, and sets the selected value to be a value (management value) to be used for the management of the target device 10.

[0028] The management unit 408 manages the target device 10 based on the measured value acquired by the measured value acquisition unit 401 and a value specified by the management value specification unit 407. Examples of managing the target device 10 include monitoring whether or not a state quantity of the target device 10 deviates from a permissible operating range, monitoring whether or not an output related to an evaluation item of the target device 10 satisfies a target, outputting a control signal of the target device 10, and the like. Examples of the movement item include the amount of NOx emission, electricity sales revenue, gas temperature, and the like. Operation of the management device

[0029] Fig. 3 is a flowchart showing operations of the management device according to the first embodiment.

[0030] When the management device 40 starts managing the target device 10, the measurement value acquisition unit 401 acquires a measurement value of a state quantity measured by the measuring instrument 20 from the communication device 30 (step S1). Next, the omission detection unit 402 detects an omission in the measurement values ​​acquired by the measurement value acquisition unit 401 (step S2). The estimation unit 404 applies the measurement value acquired by the measurement value acquisition unit 401 to each of a plurality of models to obtain each of the omitted state quantities (target state quantities) for which the omission was detected (step S3).

[0031] Next, the probability specifying unit 406 specifies an occurrence probability for each estimated value based on a probability distribution stored in the probability distribution storage unit 405 (step S4). In addition, the management value specifying unit 407 specifies the highest probability among the probabilities specified by the probability specifying unit and selects an estimated value with respect to the probability to specify a value of a state quantity at which the omission was detected (step S5). In addition, the management unit 403 manages the target device 10 based on the measured value acquired by the measured value acquiring unit 401 and a value specified by the management value specifying unit 407 (step S6).In a case where the target device is a gas turbine, the target device is managed based on the specified management value, for example, by changing a gas turbine output command value, changing the setting of an opening of an IGV, or a fuel flow rate. Specific example of the operation

[0032] Here, a method for specifying a management value according to the first embodiment will be described using a specific example.

[0033] Fig. 4 is a flowchart showing a specific example of a method for specifying a management value according to the first embodiment.

[0034] A case in which a probability distribution of values ​​of the target state variables is a distribution that Fig. 4, and in which the estimation unit 404 outputs an estimated value e1 based on a first model and an estimated value e2 based on a second model, will be described. A curve G1 shown in Fig. 4 is a curve in which a vertical axis represents a probability density and a horizontal axis represents a target state variable. The probability specification unit 406 obtains an occurrence probability of the estimated value e1 based on a probability distribution of a target state variable. In the Fig. 4 is a probability density for the occurrence probability of the estimated value e1=0.2. In addition, the probability specification unit 406 obtains an occurrence probability of the estimated value e2 based on a probability distribution of a target state variable. In the example shown in Fig. 4, a probability density of the occurrence probability of the estimated value is e2=0.3. In addition, the management value specifying unit 407 specifies an estimated value with a higher occurrence probability among the occurrence probabilities of the specified estimated values ​​to be the management value. In the example shown in Fig. 4, since the appearance probability of the estimated value e2 is higher than the appearance probability of the estimated value e1, the management value specifying unit 407 determines a management value to be the estimated value e2. Operations and effects

[0035] In this way, according to the first embodiment, the management device 40 specifies a value to be used for managing the target device 10 from a plurality of estimated values ​​based on a probability distribution of the values ​​of the target state variables. That is, according to the first embodiment, the management device 40 can appropriately set a value of a target state variable to be used for managing the target device 10 based on an estimated value of a target state variable calculated using models, given a probability distribution of the values ​​of the target state variables. Second embodiment

[0036] The management device 40 according to the first embodiment sets a value of a target state quantity to be used for managing the target device 10 based on a probability distribution of values ​​of the target state quantities. On the other hand, the management device 40 according to a second embodiment sets a value of a target state quantity to be used for managing a target device 10 based on a probability distribution of values ​​of evaluation items for managing the target device 10. Examples of the evaluation items include the amount of NOx emission, electric sales revenue, gas temperature, and the like. Configuration of the management device

[0037] Fig. 5 is a block diagram showing a configuration of the management device according to another embodiment.

[0038] A management device 40 according to the second embodiment further includes, in addition to components of the first embodiment, an evaluation value calculation unit 409. The evaluation value calculation unit 409 calculates a value of an evaluation item of the target device 10 by using each of a plurality of estimated values ​​estimated by an estimation unit and a measured value measured by a measured value acquisition unit 401 for each of the estimated values.

[0039] A probability distribution storage unit 405 according to the second embodiment stores a probability distribution table in which a value of an evaluation item and a probability of accepting the value are related to each other.

[0040] A probability specifying unit 406 according to the second embodiment specifies a probability of accepting an evaluation value calculated by the evaluation value calculating unit 409 for each evaluation value based on a probability distribution stored in the probability distribution storing unit 405. Operations of the management device

[0041] Fig. 6 is a flowchart showing operations of the management device according to the second embodiment.

[0042] When the management device 40 starts managing the target device 10, the measurement value acquisition unit 401 acquires a measured value of a state variable from a measuring instrument of a communication device 30 (step S101). Next, the omission detection unit detects the omission of a measured value acquired by the measurement value acquisition unit 401 (step S102). The estimation unit 404 applies the measured value acquired by the measurement value acquisition unit 401 to each of a plurality of models to obtain all the estimated values ​​of all the state variables for which the omission was detected (step S103).

[0043] Next, the evaluation value calculation unit 409 calculates values ​​of a plurality of evaluation items based on each of a plurality of estimated values ​​estimated by the estimation unit 404 (step S104). The values ​​of the evaluation items can be obtained using a function that has values ​​of a plurality of state variables as explanatory variables. The evaluation value calculation unit 409 calculates an evaluation value by substituting a measured value or an estimated value as the explanatory variable of the function.

[0044] Next, the probability specifying unit 406 specifies an occurrence probability for each evaluation value based on a probability assignment stored in the probability storage unit 405 (step S105). In addition, the management value specifying unit 407 specifies the highest probability among the probabilities specified by the probability specifying unit 406 and selects an estimated value to be used for calculating an evaluation value related to the probability to specify a value of a state quantity for which the omission was detected (step S106). In addition, the management unit 408 manages the target device 10 based on a measurement value acquired by the measurement value acquiring unit and a value specified by the management value specifying unit 407 (step S107). Specific example of the operation

[0045] Here, a method for specifying a management value according to the second embodiment will be described using a specific example.

[0046] Fig. Fig. 7 is a flowchart showing a specific example of specifying a management value according to the second embodiment. A case where a probability distribution of values ​​to be evaluated has a Fig. 7, and in which the estimation unit 404 outputs an estimated value e1 based on a first model and an estimated value e2 based on a second model, is described here. A curve G2, which in Fig. 6 is a graph in which a vertical axis represents a probability density and a horizontal axis represents a value of a probability distribution. The evaluation value calculation unit 409 calculates an evaluation value f(e1), which is a value of an evaluation item, based on the estimated value e1. In addition, the evaluation value calculation unit 409 calculates an evaluation value f(e2), which is a value of an evaluation item, based on the estimated value e2. Next, the probability specification unit 406 obtains an occurrence probability of the evaluation value f(e1) based on a probability distribution of values ​​of the evaluation items. In the Fig. 7, the probability density of the occurrence probability for the evaluation value f(e1) is 0.3. In addition, the probability specification unit 406 obtains an occurrence probability of the evaluation value f(e2) based on a probability distribution of values ​​of the evaluation items. In the example shown in Fig. In the example shown in Figure 7, the probability density of the occurrence probability of the evaluation value is f(e2)=0.35. In addition, the management value specification unit specifies an estimated value to be used for calculating a higher occurrence probability among the occurrence probabilities of the specified evaluation values ​​to be the management value. Since in the Fig. 7, the appearance probability of the evaluation value f(e2) is higher than the appearance probability of the evaluation value f(e1), the management value specifying unit 407 sets the estimated value e2 used for calculating the evaluation value f(e2) to be a management value. Operations and effects

[0047] In this way, according to the second embodiment, the management device 40 specifies a value to be used for managing the target device 10 from among a plurality of estimated values ​​based on a probability distribution of the values ​​to be evaluated and calculated based on a target state quantity. That is, according to the second embodiment, the management device 40 can appropriately set the value of a target state quantity to be used for managing the target device based on an estimated value of a target state quantity calculated using each of the models, considering a probability distribution of the values ​​to be evaluated. Third embodiment

[0048] The management device 40 according to the second embodiment sets a value of a target state quantity to be used for managing the target device based on a probability distribution of values ​​of evaluation items for managing the target device 10. Here, depending on an evaluation item, a value of the evaluation item may fluctuate due to a state quantity that cannot be measured or predicted. For example, the amount of NOx emission, which is an evaluation item, fluctuates depending on the oxygen concentration during combustion and a time during which the high-frequency combustion gas is stagnant, but these values ​​cannot be measured or predicted. Hereinafter, a state quantity that cannot be measured or predicted is referred to as an unknown state quantity.

[0049] A management device according to a third embodiment sets a value of a target state quantity to be used for managing a target device 10 in view of an unknown state quantity. Meanwhile, a configuration of the management device 40 is the same as that of the second embodiment. However, a probability distribution storage unit 405 according to the third embodiment stores a probability distribution table in which a value of an evaluation item and a probability of accepting the value are related for each value of an unknown state quantity. That is, the probability distribution table according to the third embodiment is a table showing a conditional probability distribution with a value of an unknown state quantity as a precondition.In the third embodiment, the probability distribution storage unit 405 stores a probability distribution table in a case where a value of an unknown state quantity falls within a first range (a comparatively large value), a probability distribution table in a case where a value of an unknown state quantity falls within a second range (medium value), and a probability distribution table in a case where a value of an unknown state quantity falls within a third range (a comparatively small value). Operations of the management device

[0050] Fig. 8 is a flowchart showing operations of the management device according to the third embodiment.

[0051] When the management device 40 starts managing the target device 10, the measurement value acquisition unit 401 acquires a measurement value of a state variable measured by the measuring instrument 20 from the communication device 30 (step S201). Next, an omission detection unit 402 detects the omission of a measurement value acquired by the measurement value acquisition unit 401 (step S202). An estimation unit 404 applies the measurement value acquired by the measurement value acquisition unit 401 to each of a plurality of models to obtain an estimated value of a state variable for which the omission was detected (step S203).

[0052] Next, the management device 40 selects values ​​(a first range, a second range, and a third range) of the unknown state quantity sequentially (step S204) and executes processes of steps S205 and S206. That is, an evaluation value calculation unit 409 calculates values ​​of a plurality of evaluation items based on each of a plurality of estimated values ​​estimated by the estimation unit 404, the values ​​of the unknown state quantities selected in step S204, and the measured values ​​acquired by step S101 for each of the estimated values ​​(step S205). A probability specification unit 406 specifies an occurrence probability for each of the evaluation values ​​based on a probability distribution table associated with values ​​of the unknown state quantity selected in step S204 (step S206).

[0053] When the management device 40 calculates a plurality of occurrence probabilities for each value of a lower state quantity, a management value specifying unit 407 calculates the sum of the occurrence probabilities for each evaluation value calculated for the same state quantity (step S207).That is, the management value specification unit 407 calculates the sum of an appearance probability of an evaluation value assuming, as a precondition, that a value of an unknown state variable falls within the first range as a precondition, an appearance probability of an evaluation value assuming, as a precondition, that a value of an unknown state variable falls within the second range, and an appearance probability of an evaluation value assuming, as a precondition, that a value of the unknown state variable falls within the third range, for each evaluation value calculated from the same state variable.

[0054] In addition, the management value specifying unit 407 specifies the highest probability in the sum of the probabilities calculated by the probability specifying unit 406 and selects an estimated value used to calculate an evaluation value related to the probability of specifying a value of a state variable for which the omission was detected (step S208). In addition, the management unit 408 manages the target device 10 based on the measured value acquired by the measured value acquiring unit 401 and the value specified by the management value specifying unit 407 (step S209). Specific example of the operation

[0055] Here, a method for specifying a management value according to the third embodiment will be described using a specific example.

[0056] Fig. 9 is a diagram showing a specific example of a method for specifying a management value according to the third embodiment.

[0057] A case in which a probability distribution of the values ​​to be evaluated is as in Fig. 9 as a function of a value of an unknown state variable, and in which an estimation unit 404 outputs an estimated value G1 based on a first model and an estimated value e2 based on a second model, will be described.

[0058] All curves G2-1, G2-2 and G2-3, which are in Fig. 9 are curves in which a vertical axis shows a probability density and a horizontal axis shows a value of a probability distribution. Curve G2-1 shows a distribution of the probabilities of occurrence of the values ​​to be evaluated when a value of an unknown state variable is in the first range. Curve G2-2 shows a distribution of the probabilities of occurrence of the values ​​to be evaluated when a value of an unknown state variable is in the second range. Curve G2-3 shows a distribution of the probabilities of occurrence of the values ​​to be evaluated when a value of an unknown state variable is in the third range.

[0059] In step S204, the management device selects a value in the first range as a value of the unknown state quantity. The evaluation value calculation unit 409 calculates an evaluation value f1(e1), which is a value of an evaluation item, in a case where a value of an unknown state quantity is in the first range, based on the estimated value e1. In addition, the evaluation value calculation unit 409 calculates an evaluation value f1(e2), which is a value of an evaluation item, in a case where a value of an unknown state quantity is in the first range, based on the estimated value e2.

[0060] Next, the probability specifying unit 406 obtains an occurrence probability of the evaluation value f1(e1) based on the probability distribution shown in curve G2-1. Fig. 9, the probability density of the occurrence probability of the evaluation value f1(e1)=0.30. In addition, the probability specification unit 406 obtains an occurrence probability of the evaluation value f1(e2) based on the probability distribution shown in the curve G2-1. In the example shown in Fig. 9, the probability density of the probability of occurrence of the evaluation value is f1(e2)=0.35.

[0061] Next, the management device 40 selects a value in the second range as a value for an unknown state variable. The evaluation value calculation unit 409 calculates an evaluation value f2(e1), which is a value of an evaluation item, in a case where a value of an unknown state variable is in the second range, based on the estimated value e1. In addition, the evaluation value calculation unit 409 calculates an evaluation value f2(e2), which is a value of an evaluation item, in a case where a value of an unknown state variable is in the second range, based on the estimated value e2. Next, the probability specification unit 406 obtains an occurrence probability of the evaluation value f2(e1) based on the probability distribution shown in the curve G2-2. In the Fig. 9, the probability density of the occurrence probability of the evaluation value f2(e1)=0.5. In addition, the probability specification unit 406 obtains an occurrence probability of the evaluation value f2(e2) based on the probability distribution shown in the curve G2-2. In the example shown in Fig. In the example shown in Figure 9, the probability density of the probability of occurrence of the evaluation value f2(e2)=0.10.

[0062] Next, the management device 40 selects a value in the third range as a value of the unknown state variable. The evaluation value calculation unit 409 calculates an evaluation value f3(e2), which is a value of an evaluation item in a case where a value of an unknown state variable is in the third range, based on the estimated value e1. In addition, the evaluation value calculation unit calculates an evaluation value f3(e2), which is a value of an evaluation item in a case where a value of an unknown state variable is in the third range, based on the estimated value e2. Next, the probability specification unit 406 obtains an occurrence probability of the evaluation value f3(e1) based on the probability distribution shown in the curve G2-3. In the Fig. 9, the probability density is the probability of occurrence of the evaluation value f3(e1)=0.18. In addition, the probability specification unit 406 obtains a probability of occurrence of the evaluation value f3(e2) based on the probability distribution shown in the curve G2-3. In the Fig. In the example shown in Figure 9, the probability density of the probability of occurrence of the evaluation value f3(e2)=0.15.

[0063] In addition, the management value specification unit 407 calculates the sum (0.30+0.50+0.18=0.98) for the occurrence probabilities with respect to the evaluation value calculated based on the estimated value e1. In addition, the management value specification unit 407 calculates the sum (0.35+0.10+0.15=0.60) of the occurrence probabilities with respect to the evaluation value calculated based on the estimated value e2. In addition, the management value specification unit 407 specifies an estimated value to be used for calculating a higher occurrence probability in the sum of the specified occurrence probabilities as the management value. Since in the Fig. 9, if the appearance probability of the evaluation value calculated based on the estimated value e1 is higher than the appearance probability of the evaluation value calculated based on the estimated value e2, the management value specifying unit 407 will set the estimated value e1 to be a management value. Operations and effects

[0064] In this way, according to the third embodiment, the management device 40 specifies an occurrence probability of an evaluation value based on a plurality of values ​​that can be assumed by an unknown state quantity, and specifies a value to be used for management of the target device 10 based on the sum of probabilities corresponding to a plurality of evaluation values ​​for each estimated value. This allows the management device 40 to appropriately set a value of a target state quantity to be used for management of the target device 10, even in a case where an unknown value that cannot be measured or predicted exists in the calculation of the value to be evaluated. Fourth embodiment

[0065] According to the second and third embodiments, the management device 40 specifies a value of a target state quantity based on the occurrence probability of a value of a specific evaluation item. On the other hand, a management device 40 according to a fourth embodiment specifies a value of a target state quantity based on the values ​​of a plurality of evaluation items. For example, the management device 40 specifies a value of a target state quantity based on the value of the NOx emission amount, the value of the electric sales revenue, and a value of the exhaust gas temperature. Meanwhile, the configuration of the management device is the same as that of the second embodiment. Operations of the management device

[0066] Fig. 10 is a flowchart showing operations of the management apparatus according to the fourth embodiment.

[0067] When the management device 40 starts managing a target device 10, a measurement value acquisition unit 401 acquires a measurement value of a state quantity to be measured by a measuring instrument 20 from a communication device 30 (step S301). Next, a selection acquisition unit 402 detects the omission of the measurement value acquired by the measurement value acquisition unit 401 (step S302). An estimation unit 404 applies the measurement value acquired by the measurement value acquisition unit 401 to each of a plurality of models and estimates state quantity values ​​for which the omission was detected (step S303).

[0068] Next, the management device 40 selects the types of evaluation items one by one to execute the following processes up to S207 (step S204).

[0069] First, the evaluation value calculation unit 409 calculates values ​​of the evaluation items relating to the types selected in step S204 based on each of the plurality of estimated values ​​estimated by the estimation unit 404 (step S205). Next, the probability specification unit 406 specifies an occurrence probability for each evaluation value based on a probability distribution stored in a probability distribution storage unit 405 (step S206). Next, a management value specification unit 407 determines whether or not the specified occurrence probability is equal to or greater than a predetermined threshold (for example, a probability distribution of 0.3) (step S207).

[0070] When determining whether or not an occurrence probability is equal to or greater than a predetermined threshold for an evaluation item of each type, the management value specifying unit 407 specifies a value of a state variable for which omission has been detected by selecting an estimated value with the largest number of items for which an occurrence probability is equal to or greater than the predetermined threshold (step S208). Additionally, the management unit 408 manages the target device 10 based on the measured value acquired by the measured value acquiring unit 401 and the value specified by the management value specifying unit 407 (step S209). Specific example of the operation

[0071] Here, a method for specifying a management value according to the fourth embodiment will be described using a specific example.

[0072] Fig. 11 is a diagram showing a specific example of a method for specifying a management value according to the fourth embodiment.

[0073] A case where the estimator 404 outputs an estimator e1 based on a first model and an estimator e2 based on a second model, and types of evaluation items to be calculated are an amount of NOx emission, electric sales revenue, and the temperature of the exhaust gas, will be described.

[0074] The evaluation value calculation unit 409 calculates an evaluation value of the amount of NOx emission, an evaluation value of the electric sales revenue, and an evaluation value of the exhaust gas temperature based on the estimated value e1. In addition, the evaluation value calculation unit 409 calculates an evaluation value of the amount of NOx emission, an evaluation value of the electric sales revenue, and an evaluation value of the exhaust gas temperature based on the estimated value e2. Next, the probability specification unit 406 obtains an occurrence probability for each of the evaluation values ​​of the amount of NOx emission, the evaluation value of the electric sales revenue, and the evaluation value of the exhaust gas temperature obtained from the estimated value e1.Similarly, the probability specifying unit 406 obtains an occurrence probability for each of the evaluation values ​​of the amount of NOx emission, the evaluation value of electric sales revenue, and the evaluation value of the temperature of the exhaust gas obtained with the estimated value e2.

[0075] Here, the management value specification unit 407 determines whether or not the occurrence probability of the evaluation value for the amount of NOx emission obtained by the estimated value e1, the occurrence probability of the evaluation value for the electric sales revenue obtained by the estimated value e1, the occurrence probability of the evaluation value for the temperature of the exhaust gas obtained by the estimated value e1, the occurrence probability of the evaluation value for the amount of NOx emission obtained by the estimated value e2, the occurrence probability of the evaluation value for the electric sales revenue obtained by the estimated value e2, and the occurrence probability of the evaluation value for the temperature of the exhaust gas obtained by the estimated value e2 are equal to or greater than a predetermined threshold. Here, as shown in Fig. 11, it is assumed that the occurrence probability of the evaluation value of the amount of NOx emission obtained by the estimated value e1, the occurrence probability of the evaluation value of the amount of NOx emission obtained by the estimated value e2 and the occurrence probability of the evaluation value of the temperature of the exhaust gas obtained by the estimated value e2 are equal to or greater than a threshold value (indicated by “O” in Fig. 11), and that the others are smaller than the threshold (indicated by “X” in Fig. 11).

[0076] In addition, the management value specification unit 407 specifies an estimate with the largest number of items for which an occurrence probability is equal to or greater than the threshold as a management value. Since in the example shown in Fig. 11, the number of items for which the appearance probability is equal to or greater than the threshold value among the evaluation items obtained from the estimated value e1 is 1, and the number of items for which the appearance probability is equal to or greater than the threshold value among the evaluation items obtained from the estimated value e2 is 2, the management value specifying unit 407 sets the estimated value e2 to be a management value. Operations and effects

[0077] In this way, according to the fourth embodiment, the management device 40 specifies a value to be used for managing the target device based on a plurality of evaluation items. Here, the management device 40 can appropriately set a value of a target state quantity to be used for managing the target device 10 so that an evaluation item to be used for managing the target device 10 has an appropriate value.

[0078] While in the fourth embodiment, the management device 40 specifies a management value based on the number of items for which an appearance probability is equal to or greater, it is not limited to this. For example, in other embodiments, the management device 40 may determine a management value based on the sum of the appearance probabilities and a weighted average for each evaluation item, or may specify a management value based on the number of items with the highest appearance probability. Fifth embodiment

[0079] According to the fourth embodiment, the management device 40 generates an estimated value of a state variable based on a plurality of models. In the fifth embodiment, operations will be described in a case where one of the plurality of models is a statistical model. Configuration of the management device

[0080] Fig. 12 is a schematic block diagram showing a configuration of a management apparatus according to the fifth embodiment.

[0081] A management device 40 according to the fifth embodiment further includes, in addition to the components of the first embodiment, a state quantity storage unit 410 and a model update unit 411. The model update unit 411 updates a statistical model among a plurality of models stored in the model storage unit 413 based on values ​​of the past state quantities stored in the state quantity storage unit 410. Operations of the management device

[0082] Fig. 13 is a flowchart showing operations of the management device according to the fifth embodiment.

[0083] When the management device 40 starts managing the target device 10, the measurement value acquisition unit 401 acquires a measurement value of a state variable by a measuring instrument 20 from a communication device 30 (step S401). Next, the selection acquisition unit 402 detects an omission of the measurement value acquired by the measurement value acquisition unit 401 (step S402). An estimation unit 404 applies the measurement value acquired by the measurement value acquisition unit 401 to each of a plurality of models, including a statistical model, to obtain an estimated state variable value (on target state variables) for each of which the omission was detected (step S403).

[0084] Next, a probability specifying unit 406 specifies an occurrence probability for each estimated value based on a probability distribution stored in a probability distribution storage unit 405 (step S404). In addition, a management specifying unit 407 specifies the highest probability among the probabilities specified by the probability specifying unit 406 and selects an estimated value with respect to the probability to specify a value of a state quantity at which the omission was detected (step S405). A management unit 408 manages the target device 10 based on the measured value acquired by the measured value acquiring unit 401 and the value specified by the management value specifying unit 407 (step S406).The measurement value acquisition unit 401 and the management value specification unit 407 store values ​​to be used for managing the target device 10 in a state quantity storage unit 410 (step S407). In addition, the model update unit 411 updates a statistical model stored in the model storage unit 403 based on the values ​​stored in the state quantity storage unit 410 (step S408). Operating procedures and effects

[0085] In this way, according to the fifth embodiment, the estimation unit 404 can estimate a value of a state variable using a statistical model updated with a timing of the previous management at each management timing. The probability specification unit 406 specifies an occurrence probability of an estimated value using the updated statistical model. That is, according to the fifth embodiment, it is possible to estimate a statistical estimated value with higher accuracy by updating not only statistical data but also the statistical model itself at each management timing.

[0086] Meanwhile, according to the fifth embodiment, the management device 40 updates a statistical model based on values ​​of the past state variable, but this is not limited to this. For example, in other embodiments, the management device 40 may not update a statistical model while storing a state variable in the state variable storage unit 410. In this case, too, the accuracy of estimation using a statistical model can be improved by storing the values ​​of the past target state variables. For example, by storing data, an estimated value of a mean value approaches a true one according to the law of large numbers, and the range of dispersion is narrowed, so an improvement in estimation accuracy can be expected. Other embodiments

[0087] Although an embodiment has been described in detail with reference to the accompanying drawings, a specific configuration is not limited to the configurations described above, and various design changes and the like may be made.

[0088] For example, in the management system 1 according to the above-described embodiment, the management device 40 has a function of extracting and specifying a value to be used for managing the target device, but it is not limited to this. For example, in the management system 1 according to the other embodiment, an information processing device that extracts and specifies a value to be used for managing the target device 10 may be provided separately from the management device 40, and the corresponding device 40 can control the target device 10 using the value specified by the information processing device.

[0089] In addition, for example, the management device 40 of the above-described embodiments acquires a measured value via a network N, but this is not limited to this. For example, the management device 40 according to the other embodiments may acquire a measured value directly from the measuring instrument. In this case, the management system 1 does not need to include a communication device 30.

[0090] In addition to the embodiments described above, the management device 40 selects one of a plurality of estimated values ​​to set the selected estimated value to be a value of a state quantity to be used for managing the target device 10, but it is not limited to this. For example, in other embodiments, the management device 40 may obtain a weighted average of estimated values, using weighted coefficients corresponding to the occurrence probabilities to set the obtained weighted average as a value of a state quantity to be used for managing the target device 10. The weighted coefficient of each estimated value increases monotonically with respect to the occurrence probability. For example, the management device may use an appearance probability of an estimated value as it is as the weighting coefficient.

[0091] In addition to the embodiments described above, the management device 40 obtains a value at which the omission was detected through estimation, but this is not limited to this. For example, in other embodiments, the management device 40 may obtain a value of a state variable through estimation regardless of the presence or absence of an omission, and may specify a value to be used for managing the target device 10 based on a probability distribution for both the measured value and the estimated value. Computer configuration

[0092] Fig. 14 is a schematic block diagram showing a configuration of a computer according to at least one embodiment.

[0093] A computer 90 includes a CPU 91, a main storage device 92, an auxiliary storage device 93, and an interface 94.

[0094] The management device 40 mentioned above is attached to the computer 90. In addition, operations of the above-described processing units are stored in the auxiliary storage device 93 in a program format. The CPU 91 reads programs from the auxiliary storage device 93, develops the program in the main storage device 92, and executes the above-described processing according to a program. In addition, the CPU 91 saves a storage area corresponding to the model storage unit 403 and the probability distribution storage unit 405 in the main storage device 92 according to a program.

[0095] Examples of the auxiliary storage device 93 include a hard disk drive (HDD), a solid-state drive (SSD), a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, a semiconductor memory, and the like. The auxiliary storage device 93 may be an internal medium directly connected to a bus of the computer 90, or it may be an external medium connected to the computer 90 via the interface 94 or a communication line. Furthermore, in a case where the program is distributed to the computer 90 via a communication line, the computer can develop the program distributed to it in the main storage device 92 and can execute described processing. In at least one embodiment, the auxiliary storage device is a persistent tangible storage medium.

[0096] Additionally, the program may be a program for implementing some of the functions described above. For example, the program may be a so-called differential file (differential program) for implementing the functions described above by combining it with other programs stored in advance in the auxiliary storage device 93. Industrial applicability

[0097] The information processing apparatus according to the invention can appropriately set the value of a state quantity to be used for the management of a target device based on an estimated value of a target state quantity calculated using each model in view of a probability distribution of the state quantity values. [List of reference symbols] 1 Management system 10 Aiming device 20 measuring instrument 30 Communication device 40 Management device (information processing device) 401 Data acquisition unit 402 Outlet detection unit 403 Model storage unit 404 Estimation unit 405 Probability distribution storage unit 406 Probability Specification Unit 407 Management value specification unit 408 Management Unit 409 Valuation calculation unit 410 State variable storage unit 411 Model Update Unit

Claims

[1] Information processing device (40) comprising: a measured value acquisition unit (401) configured to receive measured values ​​of state variables measured by a plurality of measuring instruments (20); an omission detection unit (402) configured to detect, as a target state quantity, a state quantity whose value is a temporal or spatial omission among state quantities to be processed on the basis of the measured values ​​of a target device (10); an estimation unit (404) configured to estimate each of a plurality of estimated values ​​regarding the target state quantity in which the omission was detected using a plurality of models for explaining the target device (10) based on the measured values; a probability specification unit (406) configured to specify each of a plurality of probabilities of estimated values ​​based on a probability distribution of state variable values ​​with respect to the target state variable; and a management value specifying unit (407) configured to specify a management value to be used for management of the target device (10) based on the plurality of estimated values ​​and the plurality of probabilities. [2] Information processing apparatus (40) according to claim 1, further comprising: an evaluation value calculation unit (409) configured to calculate each of a plurality of evaluation values, which are values ​​of evaluation items of the management of the target device (10), on the basis of the plurality of estimated values, wherein the probability specifying unit (406) is configured to specify the plurality of probabilities each corresponding to the plurality of evaluation values ​​based on the probability distribution of the values ​​of the evaluation items. [3] Information processing device (40) according to claim 2, wherein the evaluation value calculation unit (409) is configured to calculate each of a plurality of evaluation values ​​for the plurality of estimated values ​​on the basis of a plurality of values ​​that can be assumed by an unknown state variable, which is a state variable with an unknown value, wherein the probability specification unit (406) is configured to specify the plurality of probabilities each corresponding to the plurality of evaluation values ​​for the plurality of estimated values, based on a conditional probability distribution of the evaluation items, with a value of the unknown state variable as a prerequisite; and wherein the management value specifying unit (407) is configured to specify a value to be used for the management of the target device (10) based on a sum of probabilities corresponding to the plurality of evaluation values ​​for the plurality of estimated values. [4] The information processing device (40) according to claim 2 or 3, wherein the evaluation value calculation unit (409) is configured to calculate a plurality of evaluation values ​​with respect to the evaluation items of a plurality of types based on the plurality of estimated values, and wherein the management value specification unit (407) is configured to specify a value to be used for the management of the target device (10) based on probabilities corresponding to the evaluation item of the plurality of types for each of the values ​​of the plurality of target state quantities. [5] The information processing device (40) according to any one of claims 1 to 4, wherein the management value specifying unit (407) is configured to determine the value of the target state quantity with respect to the highest probability as a value to be used for the management of the target device (10). [6] The information processing apparatus (40) according to claim 1 to 5, wherein the plurality of models includes at least one of a statistical model and a physical model. [7] Information processing apparatus (40) according to claim 6, further comprising: a model update unit (411) which updates the statistical model on the basis of values ​​of past state variables, wherein the probability specifying unit (406) is configured to specify a probability corresponding to an estimate value estimated using the updated statistical model. [8] Information processing apparatus (40) according to claim 1, further comprising: a management unit (408) configured to manage the target device (10) based on the measured values ​​and the specified management value. [9] The information processing device (40) according to claim 8, wherein the target device (10) is a gas turbine, and the management unit (408) manages the gas turbine by at least one action selected from the group consisting of changing a gas turbine output command value, changing the setting of an opening of an IGV, and changing a fuel flow rate, based on the measured values ​​and the management value. [10] Information processing method, comprising: Receiving measured values ​​of state variables measured by a plurality of measuring instruments (20); detecting, as a target state variable, a state variable whose value is a temporal or spatial omission among state variables to be processed on the basis of the measured values ​​of a target device (10); estimating each of a plurality of estimated values ​​regarding the target state quantity in which the omission was detected using a plurality of models for explaining the target device (10) based on the measured values; Specifying each of a plurality of probabilities corresponding to the plurality of estimates based on a probability distribution of the state variable values ​​with respect to the target state variable; and Specifying a management value to be used for managing the target device (10) based on the plurality of estimated values ​​and the plurality of probabilities. [11] Program that causes a computer to: to receive measured values ​​of state variables measured by a plurality of measuring instruments (20); detecting, as a target state quantity, a state quantity whose value is a temporal or spatial omission among state quantities to be processed on the basis of the measured values ​​of a target device (10); estimating each of a plurality of estimated values ​​regarding the target state quantity in which the omission was detected using a plurality of models for explaining the target device (10) on the basis of the measured values; Specifying each of a plurality of probabilities corresponding to the plurality of estimates based on a probability distribution of the state variable values ​​with respect to the target state variable; and Specifying a management value to be used for managing the target device (10) based on the plurality of estimated values ​​and the plurality of probabilities.

Images