Evaluation systems, evaluation methods, computer programs

Abstract

This system provides an evaluation system for changes resulting from parameter modifications and plant degradation. [Solution] The system includes: a plant data acquisition unit that acquires plant data including the manipulated quantities of the plant, sensor values ​​and parameters of sensors installed in the plant, and output data of the plant; a selection acquisition unit that acquires a plurality of selection lists defined by parameter combinations from the plant data; a model generation unit that generates a plurality of models using the parameter combinations defined in each of the plurality of selection lists; a model database capable of storing the plurality of models; a model comparison unit that compares inference data obtained by applying the corresponding parameter combination to each of the plurality of models with the plant output data corresponding to the applied parameter combination; a comparison database that stores the comparison results of the model comparison unit; and an output unit capable of outputting the comparison results corresponding to each of the plurality of models.

Description

Technical Field

[0005] , , ,

[0001] Embodiments of the present invention relate to an evaluation system, an evaluation method, and a computer program for evaluating, for example, a plant or the like.

Background Art

[0002] As an information processing system related to, for example, a plant that supplies electric power, a plant operation control system that controls the operation of a power generation plant and a plant operation plan formulation system that formulates an operation plan for a power generation plant are known. The plant operation plan formulation system formulates an operation plan for operating the power generation plant most efficiently while keeping the operation of the power generation plant stable, for example. Therefore, it is necessary to evaluate the plant efficiency from the plant operation data. <00​​​​​​​​​​​​​​​​​​​​​​​​​​​​Thus, conventional evaluation systems, evaluation methods, and computer programs cannot evaluate changes due to changes in various parameters or plant deterioration. The evaluation system in this embodiment was created to solve this problem, and aims to provide an evaluation system, evaluation method, and computer program that can evaluate changes due to changes in parameters or plant deterioration. [Means for solving the problem]

[0006] The evaluation system of the embodiment is an evaluation system for evaluating a plant. It comprises: a plant data acquisition unit that acquires plant data including parameters such as the manipulated quantities of the plant and sensor values ​​of sensors provided in the plant, and output data of the plant; a selection acquisition unit that acquires a plurality of selection lists that define combinations of parameters from the plant data; a model generation unit that generates a plurality of models using the combinations of parameters defined in each of the plurality of selection lists; a model database capable of storing the plurality of models; a model comparison unit that compares inference data obtained by applying the corresponding combination of parameters to each of the plurality of models with the output data of the plant corresponding to the applied combination of parameters; a comparison database that stores the comparison results of the model comparison unit; and an output unit capable of outputting the comparison results corresponding to each of the plurality of models. [Brief explanation of the drawing]

[0007] [Figure 1] This block diagram shows the functional configuration of the evaluation system according to the first embodiment. [Figure 2] This figure shows an example of a data selection list in the first embodiment. [Figure 3] This is a flowchart showing the model creation operation of the evaluation system according to the first embodiment. [Figure 4] This is a flowchart showing the evaluation operation of the evaluation system of the first embodiment. [Figure 5] This figure shows an example of the output of the evaluation system according to the first embodiment. [Figure 6] This figure shows an example of the comparison results in the evaluation system of the first embodiment. [Figure 7] This block diagram shows the functional configuration of the evaluation system of the second embodiment. [Modes for carrying out the invention]

[0008] (Configuration of the first embodiment) Embodiments of the present invention will be described below with reference to the drawings. In the following description, common elements will be denoted by the same reference numerals, and redundant descriptions will be omitted. Figure 1 is a block diagram showing the functional configuration of the plant evaluation system of the first embodiment. Figure 2 is a diagram showing an example of the data selection list of the first embodiment.

[0009] As shown in Figure 1, Plant 2 and the evaluation system 1 of this embodiment are configured to send and receive information from each other via a network or the like. The evaluation system 1 acquires plant data such as operational data of Plant 2, including power generation equipment, and sensor values ​​and calculated values ​​output by various sensors of Plant 2. Based on this acquired information, the evaluation system 1 creates a model related to the operation of Plant 2 and uses the model to output evaluation information that contributes to the evaluation of Plant 2.

[0010] The plant 2 illustrated in Figure 1 comprises a plant body 200, a plant control unit 210, and a display unit 220. The plant body 200 is, for example, equipment such as a power generation facility. The plant body 200 operates based on command signals, including an operandi a, provided by the plant control unit 210. The plant body 200 has sensors (not shown). The plant body 200 is configured to output measured values ​​(sensor values ​​b) output by the sensors. Examples of sensor values ​​b include actual measured values ​​such as various pressures, temperatures, and power outputs.

[0011] The plant control unit 210 is a functional element that controls the plant body 200 by providing command signals including an manipulated variable a to the plant body 200 based on commands received from the user via an input interface (not shown), such as a mouse or keyboard. The plant control unit 210 can be implemented by a computer consisting of, for example, a CPU, main memory, and auxiliary memory. The plant control unit 210 can acquire sensor values ​​b output by sensors provided in the plant body 200. Furthermore, the plant control unit 210 can perform predetermined calculations on the sensor values ​​b to acquire calculated values ​​c.

[0012] The display unit 220 is a display device that presents information to the user in plant 2. The plant control unit 210 can output the manipulated variable a, sensor value b, and calculated value c to the display unit 220 and present them to the user in plant 2.

[0013] The evaluation system 1 of the embodiment illustrated in Figure 1 includes a manipulated variable acquisition unit 100, a data acquisition unit 105, a data database (DB) 110, a data calculation unit 115, a calculated value database (DB) 120, a selection setting unit 125, a selection acquisition unit 130, a model generation unit 135, and a model database (DB) 140. The evaluation system 1 can be implemented by a computer consisting of, for example, a CPU, main memory, and auxiliary storage.

[0014] The manipulated variable acquisition unit 100 is a functional element that can acquire information from the plant 2 via a network or the like. For example, the manipulated variable acquisition unit 100 is an interface that can acquire the manipulated variable a of the plant body 200, which has been given to the plant body 200 by the plant control unit 210, from the plant control unit 210.

[0015] The data acquisition unit 105 is a functional element that can acquire information from the plant 2 via a network or the like. The data acquisition unit 105 is an interface that allows the plant control unit 210 to acquire sensor values ​​b acquired from the plant body 200 and calculated values ​​c calculated by the plant control unit 210.

[0016] The data DB110 is a database that stores the information acquired by the evaluation system 1 from the plant 2. The operation amount acquisition unit 100 stores the operation amount a acquired from the plant control unit 210 in the data DB110. Similarly, the data acquisition unit 105 stores the sensor value b and the calculated value c acquired from the plant control unit 210 in the data DB110.

[0017] The data calculation unit 115 is a functional element that executes a predetermined calculation on plant data d such as the sensor value and the calculated value acquired by the data acquisition unit 105 from the plant control unit 210. The data calculation unit 115 extracts the plant data d to be calculated from the data DB110 and executes the calculation.

[0018] The calculation value DB120 is a database that stores the calculation results of the data calculation unit 115. The data calculation unit 115 stores the executed calculation result as calculation data e in the calculation value DB120. The calculation value DB120 may store the plant data d used in the calculation.

[0019] The selection setting unit 125 is a functional element that sets the generation conditions of the model related to the operation of the plant 2 created by the evaluation system 1. The selection setting unit 125 receives a selection list f for model generation via an input device such as a mouse or a keyboard not shown. The selection list f is information for selecting the data used for model generation.

[0020] Figure 2 is an example of the selection list. The selection list f illustrated in Figure 2 includes the number of models to be generated (= N), "data selection list (output)" and "data selection list (input)" as parameters used for model creation. The "data selection list (output)" and the "data selection list (input)" may include the number of items for each. The data selection list (output) indicates whether to use the plant output for model creation. In the example shown in Figure 2, the plant output is used as one of the parameters for all N models.

[0021] The data selection list (input) specifies which of the various parameters, such as the manipulated variable a, sensor value b, calculated value c, and calculation data e of Plant 2, will be used to create the model. In the example shown in Figure 2, the parameters in the data selection list (input) are listed as heat source inlet temperature, heat source outlet temperature, turbine inlet pressure, main circulation pump outlet pressure, generator cooling water flow rate, turbine generator vibration, etc. Model number 1 uses all available parameters for model creation. Model number 2 uses all parameters except heat source outlet temperature. Model number 3 uses all parameters except turbine inlet pressure. On the other hand, model number N uses only heat source inlet temperature and heat source outlet temperature as parameters. Thus, each of the N models uses a different combination of parameters for model creation. The selection setting unit 125 accepts all parameter combinations for all N models from the user as a selection list f.

[0022] The selection and acquisition unit 130 is a functional element that extracts corresponding plant data d as parameters from the data DB 110 based on the selection list f received by the selection and setting unit 125. For example, for model number 1 shown in Figure 2, all parameters are used, so the selection and acquisition unit 130 extracts plant data d related to all items in the data selection list from the data DB 110 as parameters for creating the model for model number 1. The extracted parameters are associated with each model number. The selection and acquisition unit 130 sends the selection list f and the corresponding plant data d for all N models to the model generation unit 135.

[0023] The model generation unit 135 is a functional element that generates N models related to plant 2 based on N sets of selection lists f and corresponding plant data d sent from the selection acquisition unit 130. Each of the N models is created based on a different combination of parameters (plant data d). The model database 140 is a database that stores the models generated by the model generation unit 135. The model generation unit 135 stores all the models it has generated in the model database 140.

[0024] Furthermore, the evaluation system 1 of the embodiment illustrated in Figure 1 includes a model comparison unit 145, a comparison database (DB) 150, and an output unit 155.

[0025] The model comparison unit 145 is a functional element that compares the inference result obtained by applying plant data d to the generated model with the measured values ​​of the plant data d. Based on the selection list f acquired by the selection acquisition unit 130, the model comparison unit 145 reads the corresponding plant data d and calculation data e from the calculation value DB 120 (or data DB 110), reads the model corresponding to the selection list f from the model DB 140, and applies the plant data d and calculation data e. The model comparison unit 145 compares the inference result of the model obtained by applying the plant data d and calculation data e with the measured values ​​of the plant output included in the plant data d that correspond to the plant data d. In other words, the model comparison unit 145 outputs the error of the generated model as the comparison result h.

[0026] The comparison DB 150 is a database that stores the comparison results h from the model comparison unit 145. The model comparison unit 145 stores the comparison results h for N models in the comparison DB 150.

[0027] The output unit 155 is, for example, a display device such as a display screen. The output unit 155 outputs evaluation information based on the selection list f set by the selection setting unit 125 and the comparison result h. The output unit 155 may be configured to generate display information that displays the evaluation information and output it to a display device or the like.

[0028] (Model creation operation in the first embodiment) Next, the operation of the evaluation system 1 of the embodiment will be described with reference to Figure 3. Figure 3 is a flowchart showing the model creation operation of the evaluation system of the first embodiment.

[0029] The manipulated variable acquisition unit 100 acquires the manipulated variable a from the plant control unit 210. The data acquisition unit 105 also acquires the sensor value b and the calculated value c from the plant control unit 210 (S300). The manipulated variable acquisition unit 100 and the data acquisition unit 115 store the acquired manipulated variable a, sensor value b, and calculated value c as plant data d in the data DB 110 (S305). The data calculation unit 115c performs a predetermined calculation on the plant data d (S310) and stores the calculated data e in the calculated value DB 120 (S315). The plant data d stored in the data DB 110 and the calculated data e stored in the calculated value DB 120 are data that associates the parameters of the main plant dam 200 with the output performance.

[0030] The selection setting unit 125 accepts the number of models N to be generated by the user and the selection list f (S320). As shown in Figure 2, the selection list f contains N combinations of plant output as a data selection list (output) and various parameters as a data selection list (input). As shown in Figure 2, the plant output is required for all combinations. Examples of various parameters include those that include all parameters (model number 1 / N in the figure), those that lack one parameter (model number 2 / N or 3 / N), and those that include only some parameters (model number N / N), and all of these are different combinations.

[0031] The selection acquisition unit 130 sets the initial value to 1 (S325) and acquires the corresponding plant data d and calculation data e from the first selection list f, data DB 110, and calculation value DB 120, respectively, from the selection setting unit 125 (S330). In the example shown in Figure 2, since the first selection list utilizes all parameters, data corresponding to all parameters is acquired from the plant data d and calculation data e.

[0032] The model generation unit 135 generates a model related to the operation of the plant body 200 using the acquired data. A linear regression model is an example of the model. The model generation unit 135 stores the first model g it generates in the model DB 140 (S340).

[0033] The selection and acquisition unit 130 determines whether or not all N models have been generated (S345). Here, the first model has been generated (No. in S345), so next, the corresponding plant data d and calculation data e are obtained from the second (n=n+1) selection list f, data DB110, and calculation value DB120, respectively (S350, S330).

[0034] When N models have been generated (Yes in S345), the model generation operation terminates. In the operation shown in Figure 3, the generation of N models is performed sequentially, but this is not the only way. That is, the generation of the first N models may be performed in parallel.

[0035] (Evaluation operation of the first embodiment) Next, with reference to Figure 4, further operation of the evaluation system 1 of the embodiment will be described. Figure 4 is a flowchart of the evaluation operation of the evaluation system of the first embodiment.

[0036] The model comparison unit 145 sets an initial value of 1 (S400), obtains the corresponding first plant data d and calculation data e from the calculation value DB 120 based on the first selection list f (S410), and reads the corresponding first model g from the model DB 140 and applies it (S420).

[0037] The model comparison unit 145 compares the output of model g to which plant data d is applied with the actual value of the plant output of the plant body 200 corresponding to the plant data d (S430).

[0038] The model comparison unit 145 stores the comparison result h in the comparison DB 150 (S440).

[0039] The model comparison unit 145 determines whether or not a comparison has been performed for all N models (S450). At this stage, the first model has been compared (S450), so next, based on the second (n=n+1) selection list f, the corresponding second plant data d and calculation data e are obtained from the calculation value DB 120 (S410), and the corresponding second model g is read from the model DB 140 and applied (S460, S420).

[0040] When N models are compared (No. in S450), the output unit 155 reads N comparison results h from the comparison DB 150 and outputs them in a predetermined format (S470). In the operation shown in Figure 4, the N model comparisons are performed sequentially in series, but this is not the only way. That is, the first to N model comparisons may be performed in parallel.

[0041] (An example of a model evaluation method) In the sequence of operations shown in Figure 4, the first model (1 / N) through the Nth model (N / N) are provided with plant data d (manipulated variable a, sensor value b, calculated value c) and calculated data e from the calculated value DB120 based on the selection list f. As a result, the first model (1 / N) through the Nth model (N / N) each output inference results (results of applying parameters to the model).

[0042] The model comparison unit 145 compares each of these N inference results with the corresponding actual plant output data included in the plant data d. As a result, the N comparison results are stored in the comparison DB 150.

[0043] Figure 5 shows an example of the output of the evaluation system 1 according to the first embodiment. The comparison results of the evaluation system 1 illustrated in Figure 5 show that the number of models N is 127, and that the comparison results (absolute mean error) differ depending on the model.

[0044] Here, with reference to Figure 6, the model evaluation method will be explained in detail. Figure 6 shows an example of the comparison results in the evaluation system of the first embodiment. Figure 6 shows the progress of model creation and inference evaluation 1 to inference evaluation 3 in chronological order.

[0045] As shown in Figure 6, with respect to the time shown on the horizontal axis, the plant output is constant for the model creation data during model creation and for inference evaluation data 1 during inference evaluation 1. In contrast, the plant output differs between inference evaluation data 2 during inference evaluation 2 and inference evaluation data 3 during inference evaluation 3.

[0046] Here, data (combinations of parameters) within the range of inference evaluation data 1, inference evaluation data 2, and inference evaluation data 3 are applied (input) to N models appropriately created using the model creation data. At this time, the inference result (model output) using inference evaluation data 1 is expected to show a value close to the actual plant output of the plant data.

[0047] However, when data within the range of inference evaluation data 2 and inference evaluation data 3 is input to N models, it is expected that the inference results output from the models and the actual plant output of the plant data will show relatively distant values ​​(shifted values). Unless there is a sudden deterioration or failure of the plant, that is, unless the state of the plant changes in a short period of time, some change should be observed in the input parameters of the models during the period in which the distant values ​​are observed. In other words, if the discrepancy between the inference results using the models and the actual plant output of the plant data is relatively large, some change should occur in the input parameters of the models. The parameters that show a change in this case can be said to be parameters that may affect the evaluation results when evaluating plant 2 using the models.

[0048] On the other hand, the fact that the plant output is approximately constant with respect to time shown on the horizontal axis between the model creation data and the inference evaluation data 1 in inference evaluation 1 suggests that various parameters may be constant between model creation and inference evaluation 1, or that various parameters may be adjusted and changed to keep the plant output constant (for example, when the plant cooling water temperature changes). For example, while a change in plant cooling water temperature affects the plant output, as long as the plant output is controlled to be constant, it can be said that some parameter other than the plant cooling water temperature may affect the evaluation results.

[0049] Here, if the plant cooling water temperature is not included in the nth data point in the selection list f in Figure 1, the nth model generated using it will be a model that does not include the plant cooling water temperature as input. In other words, the selection list f used by the model generation unit 135 to generate the nth model is the same as the selection list f used by the model comparison unit 145 to compare the nth model. If the nth selection list f in the model creation in Figure 6 does not include the plant cooling water temperature, the comparison result of the nth model in inference evaluation 1 will also not include the plant cooling water temperature.

[0050] On the other hand, the selection list f used to generate the first model and the first model generated by the model generation unit 135 both include all input parameters. That is, both include the plant cooling water temperature as a parameter. In other words, the input parameters of the first model and the nth model differ in whether or not they include the plant cooling water temperature.

[0051] Therefore, during inference evaluation 1, by comparing the results of the first model and the nth model, it becomes possible to evaluate the difference between cases where the input does not include at least the plant cooling water temperature and cases where it does.

[0052] When the first model is fed the inference evaluation data 1 during the inference evaluation 1, its input parameters include the plant cooling water model. On the other hand, when the same data is fed to the nth model during the same period, its input parameters do not include the plant cooling water model. When comparing these, if they show different trends, a discrepancy occurs between the inference data and the actual data. That is, it can be seen that the plant cooling water temperature is a parameter that can have a significant influence. If there is no difference in the trends between the two, it can be seen that the plant cooling water temperature is a parameter that does not have a significant influence. In this way, the comparison value (inference error) between these model inference data and actual data can reveal the magnitude of the influence that parameters can have.

[0053] Next, we will explain in detail the model evaluation method using an example output of the evaluation system 1 of the embodiment shown in Figure 5.

[0054] The output example shown in Figure 5 is a graph with the number of input parameters on the horizontal axis and the absolute mean error of the inference error on the vertical axis. In the evaluation system 1 of the embodiment, all models take different input parameters from each other, so the horizontal axis is equivalent to plotting the inference error for each generated model. That is, the horizontal axis lists and represents the comparison results for each model with different parameters, from the first model which applies all 127 input parameters to the 127th model which has only one input parameter.

[0055] The output unit 155 of the evaluation system 1 outputs a plot diagram as shown in Figure 5. When any model from the 1st to the 127th is selected via an input means such as a mouse (not shown), the comparison results of the selected models are displayed in a pop-up window. In the example shown in Figure 5, the 1st model (1 / 127), the 5th model (5 / 127), and the 126th model (126 / 127) are displayed in the pop-up window. For example, if the 5th model (5 / 127), which has the smallest absolute mean error in the output results shown in Figure 5, is selected, the comparison results of the selected models are displayed. Here, the absolute mean error represents the discrepancy between the inference result and the actual data. In the example shown in Figure 5, the pop-up display shows the model creation data and the inference evaluation data 1 in chronological order. That is, the comparison results of the model creation data and the inference evaluation data 1 exemplified in Figure 6 are shown.

[0056] The first model (1 / 127) in Figure 5 applies all 127 input parameters. According to the pop-up display for the first model (1 / 127), the discrepancy between the inference results and the actual data is very small among the model creation data (Figure I), and large among the inference evaluation data 1 (Figure II).

[0057] The fifth model (5 / 127) in Figure 5 applies input parameters with four items excluded from all input parameters. According to the pop-up display for the fifth model (5 / 127), the discrepancy between the inference results of the fifth model (5 / 127) and the actual data is very small for the model creation data (III) and also small for the inference evaluation data 1 (IV).

[0058] Model 126 (126 / 127) in Figure 5 applies input parameters that retain only two items from all input parameters. According to the pop-up display for Model 126 (126 / 127), the discrepancy between the inference results of Model 126 (126 / 127) and the actual data is large for both the model creation data and the inference evaluation data 1 (V, VI).

[0059] Thus, according to the evaluation system 1 of the first embodiment, it is configured to output a comparison result between the inference result obtained by applying the combination of input parameters to multiple models generated based on different combinations of input parameters, and the corresponding actual plant output data. Furthermore, it is configured to output the comparison result for each of the multiple models. With this configuration, it becomes possible to provide the user with information to determine which of the input parameters may affect the operation of the plant.

[0060] Furthermore, according to the evaluation system 1 of the first embodiment, the combination of parameters corresponding to each of the multiple models and the corresponding comparison results are associated with each model and configured to be output visually, for example, in a pop-up display. With this configuration, it becomes possible for the user to easily evaluate the parameters that may affect the plant for each model.

[0061] (Configuration of the second embodiment) Next, with reference to Figure 7, the evaluation system according to the second embodiment will be described in detail. Figure 7 is a block diagram showing the functional configuration of the evaluation system according to the second embodiment. The evaluation system according to the second embodiment is configured to use differential data for the plant data d in the evaluation system according to the first embodiment. In the following description, components common to the evaluation system according to the first embodiment will be denoted by the same reference numerals, and redundant explanations will be omitted.

[0062] As shown in Figure 7, the evaluation system 1a of the second embodiment further includes a design value database (DB) 112 and a difference calculation unit 116 in addition to the configuration of the evaluation system 1 of the first embodiment. The design value DB 112 is a database that records the design value data k of the plant body 200. The difference calculation unit 116 is a functional element that calculates the difference value m between the plant data d stored in the data DB 110 and the design value data k.

[0063] (Differential calculation) Here, we define the actual data included in plant data d as shown in Equation 1.

[0064]

number

[0065] Similarly, the design value data k stored in the design value DB112 is defined as shown in Equation 2.

[0066]

number

[0067] Taking the difference between equation 1 and equation 2 yields the relationship shown in equation 3.

[0068]

number

[0069] Thus, the evaluation system 1a of the second embodiment uses the difference value m between the design value data and the plant data d. Therefore, the model generation unit 135 generates a model using the difference value m, and the model comparison unit 145 performs model comparison using the difference value m as a parameter. Consequently, the evaluation system 1a of this embodiment can reduce the amount of calculation data and improve calculation speed.

[0070] While several embodiments of the present invention have been described, these embodiments are presented as examples only and are not intended to limit the scope of the invention. These novel embodiments can be carried out in a variety of other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims of the invention and its equivalents. [Explanation of symbols]

[0071] 1,1a...Evaluation system, 100...Manipulated variable acquisition unit, 105...Data acquisition unit, 110...Data database, 115...Data calculation unit, 120...Calculated value database, 125...Selection setting unit, 130...Selection acquisition unit, 135...Model generation unit, 140...Model database, 145...Model comparison unit, 150...Comparison database, 155...Output unit, 112...Design value database, 116...Difference calculation unit 2...Plant, 200...Plant main unit, 210...Plant control unit, 220...Display unit a... Manipulated variable, b... Sensor value, c... Calculated value, d... Plant data, e... Calculation data, f... Selection list, g... Model, h... Comparison result, k... Design value data, m... Difference value

Claims

1. An evaluation system for evaluating plants, A plant data acquisition unit that acquires plant data of the plant, including parameters such as at least the manipulated amount of the plant and the sensor values ​​of sensors provided in the plant, and output data of the plant. A selection acquisition unit that acquires a plurality of selection lists defined by the combination of parameters from the plant data, A model generation unit that generates multiple models using combinations of parameters defined in each of the multiple selection lists, A model database capable of storing the aforementioned multiple models, A model comparison unit compares inference data obtained by applying the corresponding combination of parameters to each of the plurality of models with output data of the plant corresponding to the applied combination of parameters. A comparison database that stores the comparison results of the aforementioned model comparison unit, An output unit capable of outputting the comparison results corresponding to each of the aforementioned multiple models, An evaluation system equipped with the following features.

2. The evaluation system according to claim 1, wherein the output unit is capable of outputting the combination of parameters corresponding to each of the plurality of models and the comparison results in association.

3. The evaluation system according to claim 1, wherein the selection acquisition unit acquires a plurality of selection lists that define a plurality of mutually different combinations of the parameters from the plant data.

4. The evaluation system according to claim 1, wherein the output unit is capable of outputting combinations of the parameters that may affect the comparison result.

5. A design value database that stores design value data showing the design values ​​of the aforementioned plant data, The system further comprises a difference calculation unit that calculates the difference between the plant data and the design value data, The evaluation system according to claim 1, wherein the model comparison unit compares inference data obtained by applying a combination of difference values ​​of the corresponding parameters to each of the plurality of models with the difference values ​​of the output data of the plant corresponding to the applied combination of difference values ​​of the parameters.

6. An evaluation method for evaluating a plant, At least the parameters including the manipulated quantities of the plant and the sensor values ​​of the sensors provided in the plant, and the output data of the plant are acquired. A plurality of selection lists are obtained from the plant data, each of which defines a combination of the parameters. Multiple models are generated using the combinations of parameters defined in each of the multiple selection lists. The aforementioned multiple models are stored in a model database, The inference data obtained by applying the corresponding combination of parameters to each of the above-mentioned models is compared with the output data of the plant corresponding to the applied combination of parameters. The comparison results from the aforementioned model comparison unit are stored in the comparison database. An evaluation method that outputs the comparison results corresponding to each of the aforementioned multiple models.

7. A computer program that operates a computer as an evaluation system for evaluating a plant, At least the parameters including the manipulated quantities of the plant and the sensor values ​​of the sensors provided in the plant, and the output data of the plant are acquired. A plurality of selection lists are obtained from the plant data, each of which defines a combination of the parameters. Multiple models are generated using the combinations of parameters defined in each of the multiple selection lists. The aforementioned multiple models are stored in a model database, The inference data obtained by applying the corresponding combination of parameters to each of the above-mentioned models is compared with the output data of the plant corresponding to the applied combination of parameters. The comparison results from the aforementioned model comparison unit are stored in the comparison database. A computer program that outputs the comparison results corresponding to each of the aforementioned multiple models.

Images