Plant diagnostic device and plant diagnostic method

Abstract

To provide a plant diagnosis device and a plant diagnosis method that target equipment constituting a plant, detect failure and deterioration of the equipment separately, and diagnose thereof.SOLUTION: The plant diagnosis device is a plant diagnostic device that detects and diagnoses failures and deterioration of equipment that constitutes a plant, takes plant state values as time series data, calculates a first plant status value of the equipment calculated from the time series data using internal state parameters of the equipment by a first prediction model, calculates an equipment failure determination result that determines failure of the equipment using time series data and the first plant status value as input, calculates a second plant status value of the equipment from the time series data using the internal state parameters of the equipment by a second prediction model, calculates an equipment deterioration determination result that determines equipment deterioration by updating the internal state parameters that serve as equipment deterioration indicators using time series data and the second plant status value as input, and determines equipment failure and deterioration based on the equipment failure determination result and the equipment deterioration determination result.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a plant diagnostic device and a plant diagnostic method. [Background technology]

[0002] In power generation, chemical plants, and other similar facilities, equipment failures and deterioration during plant operation can lead to decreased plant efficiency and even shutdowns. To avoid these issues, the introduction of diagnostic equipment to detect and diagnose equipment failures and deterioration is progressing.

[0003] Patent Document 1 discloses a system for determining equipment deterioration that periodically compares a predicted value of the change in process volume due to equipment deterioration with an actual measured value of the equipment's process volume, and determines that there are signs of deterioration if the difference between the predicted value corresponding to the age at the time of comparison and the actual measured value at the time of comparison is greater than a predetermined amount.

[0004] Patent Document 2 discloses a monitoring device that generates time-series prediction data based on time-series measurement data output from a sensor and a prediction model, visualizes the differences between time-series past data, measurement data, and prediction data, and presents information useful for determining plant equipment failures. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2013-21703 [Patent Document 2] Japanese Patent Publication No. 2021-179740 [Overview of the Initiative] [Problems that the invention aims to solve]

[0006] Generally, the failure of equipment that makes up a plant occurs on a short time scale of about several days, while the deterioration of equipment occurs on a long time scale of about several months to several years. Equipment failure is an event that leads to plant operation stoppage, and when a failure is detected, measures such as promptly repairing or replacing the equipment are necessary. On the other hand, equipment deterioration is an event that leads to a decrease in plant efficiency, and when deterioration is detected, the equipment is repaired or replaced, for example, at the timing of regular inspections. That is, regarding equipment failure and deterioration, since the countermeasures when these events are detected are different, it is required to clearly distinguish and detect and diagnose equipment failure and deterioration.

[0007] In the deterioration determination system described in Patent Document 1, the deterioration of equipment can be detected and diagnosed from the difference between the predicted value and the measured value. Also, in the monitoring device described in Patent Document 2, the failure of equipment can be detected and diagnosed from the differences among past data, measurement data, and prediction data.

[0008] However, since the diagnostic target of the technology disclosed in Patent Document 1 is equipment deterioration and the diagnostic target of the technology disclosed in Patent Document 2 is equipment failure, there is a problem that equipment failure and deterioration cannot be distinguished and detected and diagnosed.

[0009] The present invention has been made in view of the above circumstances, and an object thereof is to provide a plant diagnostic device and a plant diagnostic method capable of distinguishing and detecting and diagnosing equipment failure and deterioration for equipment that makes up a plant.

Means for Solving the Problem

[0010] [[ID=I8]] Based on the above, the present invention provides a plant diagnostic device for detecting and diagnosing failures and deterioration of equipment constituting a plant, comprising: a measurement value acquisition unit for acquiring plant status values ​​of the plant; a measurement value database for storing plant status values ​​and outputting them as time-series data; a first prediction model for calculating the plant status values ​​of equipment as first target status values ​​from time-series data using the internal status parameters of the equipment; an equipment failure determination unit for determining equipment failure and outputting equipment failure determination results, taking time-series data and the first target status values ​​as input; a second prediction model for calculating and outputting the plant status values ​​of equipment as second target status values ​​from time-series data using the internal status parameters of the equipment; and an equipment deterioration determination unit for determining equipment deterioration by updating internal status parameters that serve as indicators of equipment deterioration and outputting equipment deterioration determination results, taking time-series data and the second target status values ​​as input. Input the updated time-series data for the internal state parameters. The internal state parameters used in the second prediction model are update The second prediction model update unit, The second prediction model update unit inputs the internal state parameters used in the updated second prediction model. Internal state parameters used in the first prediction model Updated The plant diagnostic device is characterized by comprising a first prediction model update unit and a diagnostic result output unit that takes equipment failure judgment results and equipment deterioration judgment results as inputs and outputs the judgment results for equipment failure and deterioration.

[0011] Furthermore, the present invention provides a plant diagnostic method for detecting and diagnosing failures and deterioration of equipment constituting a plant, comprising: acquiring and storing time-series data of the plant's plant state values; calculating the plant state value of the equipment as a first target state value from the time-series data using the equipment's internal state parameters; determining equipment failure using the time-series data and the first target state value as input; calculating the plant state value of the equipment as a second target state value from the time-series data using the equipment's internal state parameters; and updating the internal state parameters, which serve as indicators of equipment deterioration, using the time-series data and the second target state value to determine equipment deterioration. Input the updated time-series data for the internal state parameters. Internal state parameters for determining the second target state value Updated death, Enter the internal state parameters to be used when calculating the updated second target state value. The internal state parameters for determining the first target state value are, updateThis is a plant diagnostic method characterized by the following: [Effects of the Invention]

[0012] According to the present invention, it is possible to distinguish between equipment failure and deterioration, and to detect and diagnose them for equipment that constitutes a plant. [Brief explanation of the drawing]

[0013] [Figure 1] A diagram showing a schematic configuration example of a plant diagnostic device according to an embodiment of the present invention. [Figure 2] This diagram shows a heat exchanger, a specific example of a target for plant diagnostic equipment. [Figure 3] This figure shows an example of an image output by the diagnostic result output unit 19 to the display device, using the heat exchanger shown in Figure 2 as an example. [Figure 4] A flowchart illustrating a method for detecting and diagnosing equipment failure and deterioration according to an embodiment of the present invention. [Modes for carrying out the invention]

[0014] Embodiments of the present invention will be described below with reference to the drawings. In the following drawings, components common to all figures are denoted by the same reference numerals, and redundant explanations are omitted. [Examples]

[0015] Figure 1 shows a schematic example of the configuration of a plant diagnostic device according to an embodiment of the present invention.

[0016] As shown in Figure 1, the plant diagnostic device 10 consists of a measurement value acquisition unit 11, a measurement value database DB, a first prediction model M1, an equipment failure determination unit 14, a second prediction model M2, an equipment deterioration determination unit 16, a second prediction model update unit 17, a first prediction model update unit 18, and a diagnostic result output unit 19.

[0017] The plant diagnostic device 10 is a device that detects and diagnoses equipment failures and deterioration in the equipment that makes up Plant 1. The plant diagnostic device 10 is composed of, for example, a computer equipped with a computing unit such as a CPU (Central Processing Unit), a storage device such as ROM (Read Only Memory) and RAM (Random Access Memory), an external storage device such as a magnetic disk, an optical disk, and non-volatile memory, and an input interface (input section), an output interface (output section), and transmission / reception functions. Some of the functions of the plant diagnostic device 10 may be executed by a program recorded on a recording medium readable by the computer or by other hardware.

[0018] In Figure 1, the measurement value acquisition unit 11 corresponds to the input interface (input unit), the diagnostic result output unit 19 corresponds to the output interface (output unit), and the measurement value database DB, the first prediction model M1, and the second prediction model M2 are configured in a storage device such as ROM (Read Only Memory). The other parts are shown as blocks representing functions processed by the arithmetic unit.

[0019] In Figure 1, the measurement value acquisition unit 11 is a measuring instrument such as a flow meter, pressure gauge, and thermometer installed on the equipment that makes up Plant 1. It takes the condition of the equipment and piping components, or the condition of the fluid flowing inside the equipment and piping, as input and outputs plant condition values ​​such as flow rate, pressure, and temperature.

[0020] The measurement value database DB takes the plant status values ​​output by the measurement value acquisition unit 11 as input, stores the status quantities in an external storage device such as a magnetic disk, and outputs them as time-series data.

[0021] Time-series data of plant status values ​​stored in the measurement database DB is provided to the first prediction model M1 and the second prediction model M2. These prediction models have essentially the same prediction function, but they differ in whether the input they handle is short-term or long-term, and also in how their outputs are used. The first prediction model M1, which handles short-term inputs, is used in the subsequent equipment failure determination unit 14 to determine equipment failures, while the second prediction model M2, which handles long-term inputs, is used in the subsequent equipment degradation determination unit 16 to determine equipment degradation. First, the first prediction model M1 will be explained.

[0022] The first prediction model M1 is a mathematical model described based on physical laws to predict the behavior of the equipment constituting Plant 1. It takes time-series data output from the measurement database DB as input and calculates and outputs the plant state values ​​of the equipment to be predicted. Hereinafter, what is calculated by the first prediction model M1 will be referred to as the target state value. The time-series data that is the input to the first prediction model M1 is the state value for the plant state at past time t-Δt. On the other hand, the target state value that is the output of the first prediction model M1 is the state value for the plant state at the current time t, after Δt has elapsed since t-Δt.

[0023] The first prediction model M1 uses data from Δt hours ago to predict data for the current time t, while the second prediction model M2, described later, uses data from mΔt hours ago to predict data for the current time t. Therefore, the former handles short-term inputs, and the latter handles long-term inputs.

[0024] In the first prediction model M1, the mathematical model used to calculate the target state value includes parameters corresponding to the internal state of the plant equipment (hereinafter referred to as internal state parameters). These internal state parameters represent, for example, the state of material degradation, the state of heat transfer efficiency degradation, the state of increased friction of parts, and so on.

[0025] As a specific example of the first prediction model M1, a model for predicting the behavior of the heat exchanger 2 shown in FIG. 2 will be described. In the heat exchanger 2, a high-temperature fluid QH and a low-temperature fluid QL flow in. After heat is transferred from the high-temperature fluid QH to the low-temperature fluid QL through the heat transfer tubes 3 in the heat exchanger 2, the high-temperature fluid QH is cooled and flows out, and the low-temperature fluid QL is heated and flows out. The mathematical model of the first prediction model M1 for heat transfer is defined by equations (1), (2), and (3).

[0026] [Number]

[0027] [Number]

[0028] [Number] <00001​​​​​​​​​​​​​​​​​​​​​​​​​​is the temperature of the heat transfer pipe 3, and these values are the values for the plant state at the current time t. Also, the plant state values at the past time t - Δt, which are the input values of the first prediction model M1, are Gh *n-1 Gl *n-1 Th_in *n-1 Tl_in *n-1 and the target state values at the current time t, which are the output values, are Th_out n Tl_out n .

[0031] The equipment failure determination unit 14 takes as inputs the plant state value at the current time output by the measurement value acquisition unit 11 and the target state value at the current time output by the first prediction model M1, and determines that there is a sign of equipment failure if the difference between the plant state value and the target state value is greater than or equal to a predetermined amount, and outputs the determination result.

[0032] As a specific example of the equipment failure determination unit 14, the processing for the heat exchanger 2 shown in FIG. 2 will be described. The equipment failure determination unit 14 inputs the high-temperature fluid temperature Th_out *n flowing out from the heat exchanger 2 and the low-temperature fluid temperature Tl_out *n output by the measurement value acquisition unit 11, and the high-temperature fluid temperature Th_out n and the low-temperature fluid temperature Tl_out n output by the first prediction model M1. Th_out *n Tl_out *n Th_out n Tl_out n are all plant state values at the current time t. Next, the equipment failure determination unit 14 determines that there is a sign of failure in the heat exchange if any of the conditions shown in equations (4) and (5) is satisfied, and outputs the determination result.

[0033]

Equation

[0034]

Equation

[0035] Here, εh_out is the equipment failure threshold for high-temperature outflow fluid, and εl_out is the equipment failure threshold for low-temperature outflow fluid.

[0036] The second prediction model M2, like the first prediction model M1, is a mathematical model described based on physical laws to predict the behavior of equipment constituting the plant. It takes time-series data output from the measurement database DB as input and calculates and outputs the plant state value (target state value) of the equipment to be predicted.

[0037] The input values ​​for the second prediction model M2 are time series data for time tm×Δt, t-(m-1)×Δt, ..., t-2×Δt, t-Δt (time interval Δt, number of data points m). On the other hand, the output values ​​for the second prediction model M2 are time series data for time t-(m-1)×Δt, t-(m-2)×Δt, ..., t-Δt, t (time interval Δt, number of data points m).

[0038] Similar to the first prediction model M1, the second prediction model M2 includes internal state parameters in the mathematical model used to calculate the target state values. As a concrete example of the second prediction model M2, we will explain the model that predicts the behavior of the heat exchanger shown in Figure 2.

[0039] Similar to the first prediction model M1, the mathematical model of the second prediction model M2 for heat exchange is defined by equations (1), (2), and (3). Gh *n-1 , Gl *n-1 ,Th_in *n-1 ,Tl_in *n-1 , Tm n-1 , αhm n-1 , αml n-1 This is time series data for time tm×Δt, t-(m-1)×Δt, ..., t-2×Δt, t-Δt (time interval Δt, number of data points m). On the other hand, Th_out n ,Tl_out n , Tm nis time series data for times t-(m-1)×Δt, t-(m-2)×Δt, …, t-Δt, t (time interval Δt, number of data points m). The input values of the second prediction model M2 are Gh *n-1 and Gl *n-1 , Th_in *n-1 , Tl_in *n-1 , and the output values are Th_out n , Tl_out n .

[0040] As described above, it can be said that the time series data used in the first prediction model is past data close to the current time point, and the time series data used in the second prediction model is past data far from the current time point.

[0041] The equipment deterioration determination unit 16 takes as inputs the time series data of the plant state values output by the measurement value database DB and the time series data of the target state values output by the second prediction model M2, and by data assimilation, corrects the internal state parameters included in the calculation model of the second prediction model M2, outputs them as time series data, and determines that there is a sign of equipment deterioration if the value of the internal state parameter is outside a predetermined range, and outputs the determination result. Specific examples of correcting the internal state parameters by data assimilation can be appropriately designed by those skilled in the art based on known techniques and the like.

[0042] As a specific example of the equipment deterioration determination unit 16, the processing for the heat exchanger shown in FIG. 2 will be described. The equipment deterioration determination unit 16 inputs the high-temperature fluid outlet temperature Th_out *n , the low-temperature fluid outlet temperature Tl_out *n output by the measurement value database DB, and the high-temperature fluid outlet temperature Th_out n , the low-temperature fluid outlet temperature Tl_out n output by the second prediction model M2. Th_out *n , Tl_out *n , Th_out n , Tl_out n are all time series data for times t-(m-1)×Δt, t-(m-2)×Δt, …, t-Δt, t (time interval Δt, number of data points m).

[0043] Next, the equipment degradation determination unit 16 performs data assimilation to determine the internal state parameter αhm included in the calculation model of the second prediction model M2. n , αml n The data is corrected and output as time-series data. Here, αhm n , αml n These are time series data for time t-(m-1)×Δt, t-(m-2)×Δt, ..., t-Δt, and t (time interval Δt, number of data points m).

[0044] Next, the equipment degradation determination unit 16 determines that there are signs of degradation in the heat exchange if any of the conditions shown in equation (6) or (7) are met, and outputs the determination result.

[0045]

number

[0046]

number

[0047] Here, ε_αhm is the threshold value for determining equipment degradation based on the heat transfer efficiency of the amount of heat transferred from the high-temperature fluid to the heat transfer tube, and ε_αml is the threshold value for determining equipment degradation based on the heat transfer efficiency of the amount of heat transferred from the heat transfer tube to the low-temperature fluid.

[0048] The second prediction model update unit 17 receives time-series data of internal state parameters output by the equipment degradation determination unit 16 and overwrites and updates the internal state parameters included in the calculation model of the second prediction model M2.

[0049] The first prediction model update unit 18 receives the internal state parameters included in the calculation model of the second prediction model M2 and overwrites and updates the internal state parameters included in the calculation model of the first prediction model M1.

[0050] When overwriting and updating the internal state parameters, this update modifies the model so that the estimated output of the model approaches the current state variables of Plant 1. This is because, at the start of initial operation of Plant 1, the model's internal parameters are adjusted so that the model output matches the plant output. However, over time, the model output generally deviates from the plant output due to the performance degradation of Plant 1. The model's internal parameters are adjusted to match the degraded characteristics of Plant 1.

[0051] The diagnostic result output unit 19 has a display device. Examples of display devices include liquid crystal displays, plasma displays, organic EL displays, cathode ray tubes, etc. The diagnostic result output unit 19 takes the equipment failure determination result output by the equipment failure determination unit 14 and the equipment degradation determination result output by the equipment degradation determination unit 16 as inputs and outputs the determination results for equipment failure and degradation.

[0052] Figure 3 shows an example of an image output by the diagnostic result output unit 19 to the display device for the heat exchanger 2 shown in Figure 2. In this figure, the display device 120 displays the equipment failure diagnosis result 130 and the equipment deterioration diagnosis result 140. The equipment failure diagnosis result 130 displays a time-series image 131 and a numerical image 132. Similarly, the equipment deterioration diagnosis result 140 displays a time-series image 141 and a numerical image 142.

[0053] First, let's explain the display contents of the equipment failure diagnosis result 130. The time-series image 131 allows you to select the time-series data to display using the display target selection list 133. In Figure 3, high-temperature fluid outflow temperature is selected as the display target. The time-series image 131 displays the plant status value 134 output by the measurement value acquisition unit 11, the target status value 135, upper limit value 136, and lower limit value 137 output by the first prediction model M1. Here, the upper limit value 136 and lower limit value 137 are calculated from equations (8) and (9), respectively. Here, x is the target status value 135, ε is the equipment failure judgment threshold, and x max is the upper limit, x min This is the lower limit.

[0054]

number

[0055]

number

[0056] Numerical image 132 displays the target and measured values ​​as plant status values ​​134, and the predicted values ​​as target status values ​​135, upper limit 136, and lower limit 137, all presented as a table.

[0057] Next, the display contents of the equipment degradation diagnosis result 140 will be explained. The time-series image 141 allows you to select the time-series data to be displayed in the time-series image 141 using the display target selection list 143. In Figure 3, the heat transfer efficiency degradation parameter for the amount of heat transferred from the high-temperature fluid to the heat transfer tube is selected as the display target. The time-series image 141 displays the time-series data 144 of the internal state parameters output by the equipment degradation determination unit 16 and the equipment degradation determination threshold 145.

[0058] The numerical image 142 displays the target, the maximum value of the time-series data 144 of the internal state parameters, and the equipment degradation judgment threshold 145 as a table.

[0059] Next, we will explain in detail the process by which the plant diagnostic device 10 detects and diagnoses equipment failures and deterioration in the equipment that makes up the plant.

[0060] Figure 4 is a flowchart showing a method for detecting and diagnosing equipment failure and deterioration according to this embodiment. In this figure, first, in step S151 (measurement value acquisition step), the measurement value acquisition unit 11 takes the condition of the components of the equipment and piping, or the condition of the fluid flowing inside the equipment and piping, as input and outputs plant condition values ​​such as flow rate, pressure, and temperature. The measurement value database DB takes the plant condition values ​​output by the measurement value acquisition unit 11 as input and stores the condition values ​​in an external storage device such as a magnetic disk.

[0061] In step S152 (the process of calculating target state values ​​using the first prediction model), the first prediction model M1 takes time-series data output from the measurement value database DB as input and calculates and outputs the plant state values ​​of the equipment to be predicted.

[0062] In step S153 (equipment failure diagnosis process), the equipment failure determination unit 14 takes the plant status value at the current time output by the measurement value acquisition unit 11 and the target status value at the current time output by the first prediction model M1 as input. If the difference between the plant status value and the target status value is greater than or equal to a predetermined amount, it determines that there is a sign of equipment failure and outputs the determination result.

[0063] In step S154 (device failure diagnosis result display step), the diagnostic result output unit 19 takes the device failure determination result output by the device failure determination unit 14 as input and outputs the device failure determination result. In the example output image of the diagnostic result output unit 19 shown in Figure 3, the device failure diagnosis result 130 of the display device 130 is updated.

[0064] In step S155 (Equipment Degradation Diagnosis Start Determination Step), the second prediction model M2 determines whether the conditions for starting equipment degradation diagnosis are met. If the determination result is YES, the process proceeds to step S156 (Target State Value Calculation Process by First Prediction Model); otherwise, the process proceeds to step S151 (Measurement Value Acquisition Step), and the subsequent processing is repeated. Here, it is determined that equipment degradation diagnosis should be started if the condition in equation (10) is met. Here, t is the current time, t0 is the time when the equipment degradation diagnosis was last performed, and ΔT is the time interval during which equipment degradation diagnosis is performed.

[0065]

number

[0066] In step S156 (the process of calculating target state values ​​using the second prediction model), the second prediction model M2 takes time-series data output from the measurement database DB as input and calculates and outputs the plant state values ​​of the equipment to be predicted.

[0067] In step S157 (equipment degradation diagnosis process), the equipment degradation determination unit 16 takes the time-series data of plant state values ​​output by the measurement value database DB and the time-series data of target state values ​​output by the second prediction model M2 as input, modifies the internal state parameters included in the calculation model of the second prediction model M2 by data assimilation, and outputs it as time-series data. If the value of the internal state parameter is outside a predetermined range, it determines that there is a sign of equipment degradation and outputs the determination result.

[0068] In step S158 (device degradation diagnosis result display step), the diagnosis result output unit 19 takes the device degradation judgment result output by the device degradation determination unit 16 as input and outputs the device degradation judgment result. In the example of the output image of the diagnosis result output unit 19 shown in Figure 3, the device degradation diagnosis result 140 on the display device 130 is updated.

[0069] In step S159 (updating the second prediction model), the second prediction model update unit 17 receives time-series data of internal state parameters output by the equipment degradation determination unit 16 and overwrites and updates the internal state parameters included in the calculation model of the second prediction model M2.

[0070] In step S160 (updating the first prediction model), the first prediction model update unit 18 inputs the internal state parameters included in the calculation model of the second prediction model M2 and overwrites and updates the internal state parameters included in the calculation model of the first prediction model M1.

[0071] In step S161 (Equipment failure / deterioration detection and diagnosis process completion determination process), the plant diagnostic device 10 determines whether the process for detecting and diagnosing equipment failure / deterioration has reached its completion time. If the determination result is NO, the process proceeds to step SM21 (measurement value acquisition process) and the subsequent processes are repeated. On the other hand, if the determination result is YES, the process for detecting and diagnosing equipment failure / deterioration is terminated.

[0072] In this embodiment, during the process of diagnosing equipment deterioration, internal state parameters that serve as indicators of equipment deterioration are updated, and the internal state parameters included in the calculation models of the second prediction model M2 and the first prediction model M1 are overwritten and updated. As a result, the first prediction model M1 becomes a prediction model that reflects equipment deterioration. If the first prediction model M1 does not reflect equipment deterioration, it is difficult to distinguish whether the difference between the plant state value output by the measurement value acquisition unit 11 and the target state value output by the first prediction model M1 is due to equipment failure or equipment deterioration.

[0073] In this embodiment, since the first prediction model M1 is a prediction model that reflects equipment degradation, the difference between the plant status value output by the measurement value acquisition unit 11 and the target status value output by the first prediction model M1 is clearly due to equipment failure, and equipment failure can be appropriately detected and diagnosed. Furthermore, since internal status parameters that serve as indicators of equipment degradation are used in the process of diagnosing equipment degradation, equipment degradation can be appropriately detected and diagnosed.

[0074] As described above, in this embodiment, it is possible to distinguish between equipment failure and deterioration, and to detect and diagnose them for the equipment that makes up the plant. [Explanation of symbols]

[0075] 10: Plant diagnostic equipment 11: Measurement value acquisition unit 14: Equipment failure determination section 16: Equipment deterioration determination section 17: Second Prediction Model Update Section 18: First Prediction Model Update Section 19: Diagnostic Result Output Unit DB: Measurement Database M1: First Prediction Model M2: Second Prediction Model

Claims

1. A plant diagnostic device that detects and diagnoses failures and deterioration of equipment that make up a plant, A measurement value acquisition unit that acquires plant status values ​​of the aforementioned plant, A measurement value database that stores the aforementioned plant status values ​​and outputs them as time-series data, A first prediction model that uses the internal state parameters of the equipment to calculate the plant state value of the equipment as a first target state value from the time series data, A device failure determination unit takes the aforementioned time-series data and the first target state value as input, determines whether the device has failed, and outputs a device failure determination result. A second prediction model that uses the internal state parameters of the equipment to calculate and output the plant state value of the equipment as a second target state value from the time series data, A device degradation determination unit takes the aforementioned time-series data and the aforementioned second target state value as input, updates the aforementioned internal state parameter which serves as an indicator of device degradation, determines device degradation, and outputs a device degradation determination result. A second prediction model update unit inputs the time-series data of the updated internal state parameters and updates the internal state parameters used in the second prediction model, A first prediction model update unit inputs the internal state parameters used in the second prediction model updated by the second prediction model update unit and updates the internal state parameters used in the first prediction model, A diagnostic result output unit takes the equipment failure determination result and the equipment deterioration determination result as inputs and outputs the determination results for equipment failure and deterioration. A plant diagnostic device characterized by being equipped with the following features.

2. A plant diagnostic device according to Claim 1, The plant diagnostic device is characterized in that the equipment degradation determination unit updates the internal state parameters by data assimilation.

3. A plant diagnostic device according to claim 1, The plant diagnostic device is characterized in that the diagnostic result output unit displays the equipment failure determination result and the equipment deterioration determination result as a time-series image or a numerical image.

4. A plant diagnostic device according to claim 1, The plant diagnostic device is characterized in that the equipment deterioration determination unit determines the internal state parameter such that the second target state value obtained by the second prediction model is close to the time series data.

5. A plant diagnostic method for detecting and diagnosing failures and deterioration of equipment constituting a plant, The time-series data of the plant status values ​​of the aforementioned plant is acquired and stored. Using the internal state parameters of the equipment, the plant state value of the equipment is calculated from the time-series data as the first target state value. The time-series data and the first target state value are used as input to determine if the equipment has failed. Using the internal state parameters of the equipment, the plant state value of the equipment is calculated from the time-series data as the second target state value. Using the aforementioned time-series data and the second target state value, the internal state parameter, which serves as an indicator of equipment degradation, is updated to determine equipment degradation. The time-series data of the updated internal state parameters is input, and the internal state parameters for determining the second target state value are updated. A plant diagnostic method characterized by inputting the internal state parameters used when calculating the updated second target state value, and updating the internal state parameters for determining the first target state value.

6. A plant diagnostic method according to claim 5, A plant diagnostic method characterized by updating the internal state parameters by data assimilation.

7. A plant diagnostic method according to claim 5, A plant diagnostic method characterized by displaying equipment failure diagnosis results and equipment degradation diagnosis results as time-series images or numerical images.

8. A plant diagnostic method according to claim 5, A plant diagnostic method characterized by determining the internal state parameter such that the second target state value is close to the time-series data.

Images