WO2019159883A1 - Model creation method, plant operation support method, model creating device, model, program, and recording medium having program recorded thereon - Google Patents

Abstract

The objective of the present invention is to provide a model creation method, a plant operation support method, a model creating device, a model, a program, and recording medium having program recorded thereon, which efficiently and accurately create or utilize an operation simulation model of a target plant by making effective use of actual results from a preceding plant, even if the plant specifications differ from those of the preceding plant. This model creation method which utilizes actual results of a preceding plant to create a model representing a relationship between an input parameter and a process value of a target plant is characterized by including a reading step of reading an existing model of the preceding plant, and a model creating step of adding a physical parameter relating to the plant specification of the target plant to input parameters of the existing model of the preceding plant, to create a new model.

Description

Model creation method, plant operation support method, model creation device, model, program, and recording medium recording program

The present invention creates a model of operation simulation of a target plant by utilizing the results of a preceding plant, and relates to the use of the model, a model creation method, a plant operation support method, a model creation device, a model, a program, And a recording medium on which the program is recorded.

Conventionally, in various plants, an operation simulator that simulates the plant is configured and used for operation control, or operators are trained.

For example, Patent Document 1 is known as a simulation technique incorporated into plant operation control. In Patent Document 1, a plant control device uses a statistical model that estimates the value of a measurement signal acquired when a control signal is given to the plant. If the data used to construct the statistical model is modified, the plant is corrected. The model is updated using the collected data.

Japanese Patent No. 5378288

When creating / updating a model, more data can be used to simulate the characteristics of the target plant with high accuracy. For this reason, when the data of the target plant is insufficient, it is effective to increase the number of data by referring to the data of the preceding plant and the model creation history (hereinafter simply referred to as the results).

When using the results of the preceding plant, it is assumed that the specifications will change from the current plant. However, regarding this point, Patent Document 1 describes only a point where data is corrected in the same plant, and does not describe a response when plant specifications or fuel changes.

Accordingly, in the present invention, even if the preceding plant and the plant specifications are different, a model for operating simulation of the target plant is created efficiently or accurately by effectively utilizing the results of the preceding plant, or a model using this It is an object to provide a creation method, a plant operation support method, a model creation device, and a model, a program, and a recording medium on which the program is recorded.

From the above, in the present invention, “a model creation method for creating a model that shows the relationship between the input parameter of the target plant and the process value by utilizing the results of the preceding plant, A model creation method comprising: a reading step to read; and a model creation step of creating a new model by adding a physical parameter related to the plant specification of the target plant to an input parameter of the existing model in the preceding plant ".

Further, in the present invention, “a plant operation support method using a new model generated by a model creation method, a simulation step of calculating a process value using new operation data and a new model of a target plant, and a predetermined condition” The plant operation support method further includes an operation instruction step of calculating an operation instruction value of the target plant based on a process value satisfying the above.

Further, in the present invention, “a model creation device that creates a model indicating the relationship between the input parameter of the target plant and the process value by utilizing the results of the preceding plant, and reads the existing model in the preceding plant. And a model creating apparatus characterized in that a new model is created by adding physical parameters related to the plant specifications of the target plant to the input parameters of the existing model at the preceding plant.

Further, in the present invention, “a model indicating the relationship between the input parameter of the target plant and the process value, created by utilizing the results of the preceding plant, the input parameter of the existing model in the preceding plant, "A model created by adding physical parameters related to the plant specifications of the target plant so that they can be identified."

Further, in the present invention, “a new model is obtained by acquiring an existing model indicating a relationship between an input parameter of a preceding plant and a process value, and adding a physical parameter related to the plant specification of the target plant to the input parameter of the existing model. Is a program that causes a computer to execute the process of creating the "."

In the present invention, “a recording medium on which a program is recorded” is used.

Even if the plant specifications of the target plant are different from those of the preceding plant, the operation simulation model of the target plant can be created efficiently and accurately by effectively utilizing the results of the preceding plant.

The schematic block diagram of the typical boiler plant used as the object of modeling. The flowchart which shows the process flow about a model preparation method. The flowchart which showed the driving | operation instruction | indication flow which gives the result of the simulation using a model as a driving | operation instruction | indication with respect to an operation control apparatus. The figure which shows the example of whole structure of the operation control apparatus of the boiler plant incorporating the model. The figure which shows the detailed structural example of a model production apparatus. The figure which shows an example of all the operation data for model creation. Schematic diagram showing the relationship between model inputs and outputs. The figure which shows the data format example in a past driving | operation database. The figure which shows the data format example in the existing model database. The figure which shows the data format example in a plant specification database. The figure which shows the data format example in a fuel property database. The figure which shows the data format example in an additional parameter candidate database. The figure which shows the data format example in a sensitivity database. The figure which shows the example of a display of the confirmation screen of sensitivity. The figure which shows the example of whole structure of the driving assistance system which assists model creation. The figure which shows the detailed structural example of the driving assistance system which assists model creation. The figure which shows the hardware constitutions of an operation control apparatus. The figure which shows the hardware constitutions of a driving assistance device.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

In the embodiment of the present invention, an example in which a model simulating the operation when a plant is a boiler plant will be described. First, in a first embodiment, a configuration example of a typical boiler plant and a model creation method will be described. The configuration of the driving support device and the driving control device to which the created model is applied will be described. In the second embodiment, a description will be given of configuring a driving support system by a plurality of driving support devices.

First, a configuration example of a boiler plant and a model configuration method that simulates its operation will be described.

FIG. 1 is a schematic configuration example diagram of a typical boiler plant to be modeled.

A boiler plant 100 shown in FIG. 1 is equipment used for power generation and heat supply, and uses pulverized coal obtained by pulverizing coal as pulverized fuel (solid fuel) for burning solid fuel. The pulverized coal is used as a furnace 11. This is a coal fired boiler capable of generating steam by exchanging heat generated by this combustion with feed water and steam. The fuel is not limited to coal, and may be other solid fuel that can be burned in a boiler, such as biomass. Further, various kinds of solid fuels may be mixed and used.

The boiler plant 100 includes a furnace 11, a combustion device 12, and a flue 13.

Among these, the furnace 11 is installed along the vertical direction, for example, in a hollow shape of a square tube. As for the furnace 11, the wall surface is comprised by the fin which connects an evaporation pipe (heat-transfer pipe) and an evaporation pipe, and is suppressing the temperature rise of a furnace wall by heat-exchanging with water supply or a vapor | steam. Specifically, on the side wall surface of the furnace 11, a plurality of evaporator tubes are arranged, for example, along the vertical direction and arranged side by side in the horizontal direction. The fin closes between the evaporation pipe and the evaporation pipe. The furnace 11 is provided with an inclined surface 62 on the furnace bottom, and a furnace bottom evaporation pipe 70 is provided on the inclined surface 62 to become a bottom surface.

The combustion device 12 is provided on the vertical lower side of the furnace wall constituting the furnace 11. In the embodiment of FIG. 1, the combustion device 12 has a plurality of burners (eg, 21, 22, 23, 24, 25) mounted on the furnace wall. For example, a plurality of the burners 21, 22, 23, 24, 25 are arranged at equal intervals along the circumferential direction of the furnace 11. However, the shape of the furnace, the arrangement of the burners, the number of burners in one stage, and the number of stages are not limited to this embodiment.

The burners 21, 22, 23, 24, 25 are connected to pulverizers (pulverized coal machines / mills) 31, 32, 33, 34, 35 through pulverized coal supply pipes 26, 27, 28, 29, 30. Has been. When coal is transported by a transport system (not shown) and put into the pulverizers 31, 32, 33, 34, and 35, it is pulverized into a predetermined fine powder size and together with transport air (primary air). The pulverized coal (pulverized coal) can be supplied from the pulverized coal supply pipes 26, 27, 28, 29, 30 to the burners 21, 22, 23, 24, 25.

Further, the furnace 11 is provided with a wind box 36 at the mounting position of each burner 21, 22, 23, 24, 25. One end of an air duct 37b is connected to the wind box 36, and the other end is It is connected to an air duct 37a for supplying air at a connection point 37d. As a result, the conveying air (primary air) and the combustion air (secondary air) from the air duct 37b are introduced into the furnace 11.

Further, a flue 13 is connected vertically above the furnace 11, and a plurality of heat exchangers (41, 42, 43, 44, 45, 46, 47) for generating steam in the flue 13 are provided. Is arranged. Therefore, a flame is formed by the burners 21, 22, 23, 24, 25 injecting a mixture of pulverized coal fuel and combustion air into the furnace 11, combustion gas is generated, and flows into the flue 13. Then, the heated water or steam that flows through the furnace wall and the heat exchanger (41, 42, 43, 44, 45, 46, 47) is heated or superheated by the combustion gas to generate superheated steam, and the generated superheated steam is supplied. Then, a steam turbine (not shown) is driven to rotate, and a generator (not shown) connected to the rotating shaft of the steam turbine is driven to rotate to generate power.

Further, an exhaust gas passage 48 is connected to the flue 13, a denitration device 50 for purifying the combustion gas, air sent from the forced blower 38 a to the air duct 37 a, and exhaust gas sent through the exhaust passage 48 An air heater 49, a dust processing device 51, an induction blower 52, etc. are provided for heat exchange between them, and a chimney 53 is provided at the downstream end. The denitration device 50 may not be provided as long as the exhaust gas standard is satisfied.

Also, the air for conveying pulverized coal (primary air) is blown from the primary air blower 38b by an air duct 37g in which an air duct 37e passing through the air heater 49 and an air duct 37f bypassing the air heater 49 are coupled. Yes. After the air flow rate of both the air ducts 37e and 37f is adjusted, they are merged and sent to the pulverizers (pulverized coal machine / mill) 31, 32, 33, 34, and 35 via the air duct 37g. The air for conveying charcoal (primary air) is adjusted to a predetermined temperature or the like.

In the furnace 11 of the first embodiment, the combustion air (primary air) and the combustion air (secondary air) supplied from the wind box 36 to the furnace 11 are excessively combusted and then newly burned air ( This is a so-called two-stage combustion furnace in which after-air is introduced to perform lean fuel combustion. Therefore, the furnace 11 is provided with an after air port 39, one end of an air duct 37c is connected to the after air port 39, and the other end is connected to an air duct 37a for supplying air at a connecting point 37d. If the two-stage combustion method is not adopted, the after-air port 39 may not be provided.

The air sent from the primary air blower 38b to the air duct 37a is heated by the air heater 49 by heat exchange with the combustion gas, and the secondary air led to the wind box 36 through the air duct 37b at the connection point 37d. Then, the air branches to the after air led to the after air port 39 via the air duct 37c.

A typical boiler plant 100 is as shown in FIG. 1, and in the following, the construction of a model of the boiler plant 100 will be described.

Here, a model creation method for creating a model that shows the relationship between the input parameters of the target plant and the process value using the results and experience of the preceding plant will be described. FIG. 2 illustrates a processing flow for the model creation method.

In this model creation method, the operation data of the preceding plant and the target plant, the reading step S1 for reading the existing model in the preceding plant, and adding the physical parameters related to the plant specifications to the input parameters of the existing model, It comprises a model creation step S2 to be created, a verification step S3 for verifying the accuracy of the created new model using operation data, and an output step S4 for outputting the new model whose accuracy has been verified.

By using this process flow, an operation model that can take into account differences in plant specifications can be created.

The model creation method in Fig. 2 will be described in more detail. First, the reading step S1 is processed according to the following concept.

In reading step S1, first, all operation data of the preceding plant and the target plant and the existing model in the preceding plant are read.

Here, a boiler plant is illustrated as an example of a power plant, and the following description is based on the assumption of a boiler plant. However, the plant is not limited to this, and is widely applied to plants that produce industrial products and materials. Needless to say, it is applicable. For example, as a combustion plant for burning fuel, a steam supply plant and an iron manufacturing plant are exemplified in addition to a power generation plant. Other than the combustion plant, chemical and papermaking plants are exemplified.

The preceding plant is an existing plant with a track record of creating a model, and the existing model is a model created at the preceding plant.

Also, the model indicates the relationship between plant input parameters (inputs) and process values (outputs). Input parameters are input to the model and used to predict (simulate) process values. In principle, a model is created for each process value, but the present invention is not limited to this, and a plurality of process values may be output by one model.

Next, in the model creation step S2, each step shown in detail below is sequentially executed.

In step S21, 1 is first added to the number N of model creations (initial value is 0).

Next, in step S22, model creation conditions and additional parameter candidates are read. When the number of times N is 2 or more, these changes are made. Here, the model creation conditions are a model creation target (process value), a method (function formula), an allowable error, and the like. Further, the additional parameter candidate is an input parameter addition candidate to be described later.

Next, in step S23, the difference in plant specifications and fuel properties between the preceding plant and the target plant is confirmed.

If the plant specifications are different as a result of the confirmation in step S23, the process proceeds to step S24, and physical parameters related to the plant specifications are added to the input parameters of the existing model. Here, the physical parameter is a parameter related to the structure, performance and / or design conditions of the plant. By adding physical parameters, appropriate parameters representing plant specifications can be selected.

As a result of the confirmation in step S23, if the fuel properties are different, the process proceeds to step S25, and the fuel parameters related to the fuel properties are added to the input parameters of the existing model. Here, the fuel parameter is a parameter related to any of fuel adjustment, combustion, environmental load, and moisture.

If, as a result of the confirmation in step S23, there is no difference in either the plant specifications or the fuel properties, the process proceeds to step S26, and no input parameters are added.

In addition, the case where the plant specifications and the fuel properties are different in step S23 indicates that the input parameters of the existing model are different to the extent that it is necessary to add physical parameters and fuel parameters, respectively. For example, when the boiler dimensions are different among the plant specifications and the physical parameters related to the boiler dimensions are already included in the input parameters of the existing model, it is not necessary to add new input parameters. In such a case, it is determined in step S23 that there is no difference in plant specifications.

In the verification step S3, the steps shown in detail below are sequentially executed.

In step S31, the accuracy of the created new model is verified using all operation data. Next, in step S32, the validity of the new model is determined. For example, an actual value (actual process value) of a process value in all operation data is compared with a simulation value (virtual process value) of a process value calculated using a model, and an error is confirmed. Incidentally, if the error is within the allowable error, the model is judged to be appropriate.

If it is determined in step S32 that the error exceeds the allowable error, it is confirmed in step S33 whether the number N is equal to or less than the allowable number Nth. If the number N is less than or equal to the allowable number Nth, the process returns to the model creation step S2 again, and the model creation conditions and additional parameter candidates are changed to correct the new model.

In the verification step S3, the relationship between the input and output (trend) of the theory or the operator's experience is verified by verifying the relationship between the typical input parameter of the new model and the process value based on a predetermined criterion. By verifying the validity from the viewpoint, the accuracy of the model can be further increased.

In the output step S4, the steps shown in detail below are sequentially executed.

In the process in step S32, when it is determined that the model is valid, in step S41, the new model is output to an output or a database to be described later.

In the process in step S33, if the above-mentioned number N exceeds the allowable number Nth, a model creation error is output to the input / output unit 309. In this way, the model creation flow ends after outputting either a new model or a model creation error.

In the output step S4, by further outputting the relationship between the representative input parameters of the new model and the process value, the operator can confirm again the relationship (trend) between the input and the output.

In FIG. 2, the plant model creation method has been described, but the model created in this way is used by being incorporated in, for example, a plant operation control device. The concept of configuring the plant operation control device will be described with reference to FIGS.

First, FIG. 3 is a flowchart showing an operation instruction flow for giving a result of simulation using the model created in FIG. 2 as an operation instruction to the operation control device.

In this flowchart, first, in step S5, new operation data and a new model of the target plant are read.

Next, simulation is performed in simulation step S6. First, in step S61, simulation conditions are set. A simulation condition is a set of input parameters.

In step S62, the input parameter set is input to the new model created in the process of FIG. As a result of the simulation, a virtual process value can be obtained.

In operation instruction step S7, an operation instruction value for the operation control device is created. First, in step S71, the simulation result is evaluated. Next, in step S72, it is determined whether or not the virtual process value obtained by the simulation is optimal (predetermined conditions are satisfied). If it is not optimal, the process returns to step S61 to reset the simulation conditions and instruct to calculate a new virtual process value.

Here, evaluation may be conversion of each virtual process value into a score (non-dimensional) with a predetermined conversion coefficient. Also, “optimal” may be a case where the total value of the calculated scores is equal to or greater than a predetermined value. Alternatively, a simulation may be performed in a plurality of cases (simulation conditions), and the result may be the highest score or the operator may determine that the highest number of cases is optimal. Furthermore, a case with a higher score may be automatically searched using a genetic algorithm or particle swarm optimization technique, and it may be determined whether or not the result is optimal.

In step S72, a driving instruction value is calculated based on the simulation conditions and results determined to be optimal, and the results are output to an output screen or the like to be described later.

FIG. 4 is a diagram showing an example of the overall configuration of an operation control apparatus for a boiler plant incorporating a model. FIG. 4 shows a boiler plant 100 that is a control target when roughly classified, and an operation control device 200 that controls the boiler plant 100.

Among these, the boiler plant 100 has a system configuration as shown in FIG. 1 in detail, but the sensor SR and the operation end OP are typically described therein. The operation end OP refers to a valve or a damper. The sensor SR detects operation data such as process values of each part of the boiler plant 100.

On the other hand, the operation control apparatus 200 obtains the operation data from the sensor SR installed in the boiler plant 100 as an input, and finally gives an operation amount at each operation end OP in the boiler plant 100 as an output. Is.

The operation control device 200 is based on the concept of automatic control that consistently processes from input to output, but the operation support device 300 can be configured by omitting the operation control unit 201. In the case of the driving support device 300, the parameter that the driving support device 300 presents is determined based on the operator M or a predetermined rule base and then a parameter value that is considered appropriate is transmitted to the driving control unit 201. Thus, operation control of the boiler plant 100 is performed. The operation control unit 201 obtains an appropriate feedback signal for the set parameter, and executes automatic control by so-called feedback control.

In the following description of the present invention, the configuration of the driving support device 300 will be described. However, since this concept can be easily developed in the driving control device 200, a detailed description thereof will be omitted.

The driving support apparatus 300 according to the present invention shown in FIG. 4 handles a lot of data, and therefore has various databases DB. The database DB adopted by the driving support device 300 and the stored contents are as follows.

The past operation database DB1 stores operation data in the preceding plant. A data configuration example of the past operation database DB1 is shown in FIG. 8 and will be described later.

The new operation database DB2 stores new operation data acquired in the target plant. The data configuration example of the new operation database DB2 is basically the same as the data configuration example of the past operation database DB1.

The existing model database DB3 stores the existing model created in the preceding plant. A data configuration example of the existing model database DB3 is shown in FIG. 9 and will be described later.

The plant specification database DB4 stores plant specifications of the preceding plant and the target plant. A data configuration example of the plant specification database DB4 is shown in FIG. 10 and will be described later.

The fuel property database DB5 stores the properties of the fuel used in the preceding plant and the target plant. A data configuration example of the fuel property database DB5 is shown in FIG. 11 and will be described later.

The new model database DB6 stores the created new model.

The driving support apparatus 300 shown in FIG. 4 works as follows using the data stored in the database.

The data acquisition unit 301 acquires new operation data from the boiler plant (target plant) 100 and stores it in the new operation database DB2.

The data extraction / conversion unit 302 extracts data necessary for model creation and operation control (new operation data, past operation data) from the new operation database DB2 and the past operation database DB1, and complements and converts the format as necessary. An example of the conversion here is a process of estimating and identifying operation data that cannot be directly measured by the sensor SR from other data. Since the estimation process is executed in software using a computer, the estimated value is called a soft sensor value.

The model creation device 303 will be described in detail with reference to FIG. 5. The outline of the model creation device 303 is new operation data from the data extraction / conversion unit 302, past operation data, existing model data from the existing model database DB3, and plant from the plant specification database DB4. A model (new model) indicating the input / output relationship of the boiler plant 100 is created using the specification data and the fuel property data from the fuel property database DB5. The created new model is stored in the new model database DB6.

The simulation unit 306 calculates a virtual process value using the new operation data output from the data extraction conversion unit 302 and the new model output from the new model database DB6, and outputs the calculation result to the optimization unit 307.

The optimization unit 307 determines whether or not the virtual process value is optimal. If it is determined to be optimal, the optimization unit 307 outputs the virtual process value to the operation instruction unit 308. If it is determined that the virtual process value is not optimal, the simulation condition is reset. It outputs to the simulation part 306 so that it may simulate again. The driving instruction unit 308 calculates a driving instruction value based on the simulation conditions and results determined to be optimal, and outputs them to the driving control unit 201. Further, the driving instruction unit 308 outputs the simulation result output from the optimization unit 307 and the calculated driving instruction value to the input / output unit 309. The details here are the same as those described in the operation instruction step S7 of FIG.

The input / output unit 309 displays a new model creation / verification result, a simulation evaluation result, and a driving instruction proposal screen, and receives an instruction from the operator M for each. Further, if there is input of additional information for the past operation database DB1, the existing model database DB3, the plant specification database DB4, and the fuel property database DB5, the input result is output to each.

Note that the operation control device can be configured by providing the operation control value given by the operation support device 300 to the operation control unit 201. In this case, the operation control unit 201 controls the operation (valve opening degree, etc.) of each operation end OP of the boiler based on the operation instruction value. The operation control may be performed automatically based on the operation instruction value or may be performed after the operator M's consent at the input / output unit 309. Further, even if the operation instruction value from the operation support apparatus 300 is added as a bias value to the operation instruction value from the existing boiler plant control apparatus (not shown), the final operation instruction value is instructed. Good.

In the operation support apparatus 300 described above, a simulation step for calculating a process value using the new operation data of the target plant in addition to the model generated by the model generation apparatus 303, and the target plant so that the process value satisfies a predetermined condition. It can be said that the operation instruction step for calculating the operation instruction value is further provided so that the operation support of the plant can be performed using a general-purpose model.

Next, a detailed configuration example of the model creation device 303 will be described with reference to FIG. In the model creation device 303, new operation data from the data extraction conversion unit 302, past operation data, existing model data from the existing model database DB3, plant specification data from the plant specification database DB4, and fuel properties from the fuel property database DB5. A model (new model) indicating the input / output relationship of the boiler plant is created using the data, additional parameter candidate data from the additional parameter candidate database DB7, and sensitivity data from the sensitivity database DB8. The created new model is stored in the new model database DB6. The additional parameter candidate data and sensitivity data are data set and input via the input / output unit 309, such as existing model data, plant specification data, and fuel property data.

Note that an additional parameter candidate database DB7 and a sensitivity database DB8 are newly added to the model creation device 303. Of these, the additional parameter candidate database DB7 stores additional candidates for input parameters, and is a sensitivity database. The DB 8 stores criteria for verifying the relationship (change tendency) between typical input parameters and process values. An example of the configuration of the additional parameter candidate database DB7 is illustrated in FIG. 12, and an example of the configuration of the sensitivity database DB8 is illustrated in FIG.

The model creation device 303 creates a model indicating the relationship between the input parameters of the target plant and the process value by utilizing the results of the preceding plant, and includes all the operation data of the preceding plant and the target plant, and the preceding plant. The data reading unit 3031 for reading the existing model in the model, the model correcting unit 3032 for creating a new model by adding physical parameters related to the plant specifications to the input parameters of the existing model, and the created new model for the entire operation The model verification unit 3033 performs accuracy verification using data, and the model output unit 3034 outputs a new model whose accuracy has been verified.

Each part of the model creation device 303 in FIG. 5 functions in detail as follows.

First, the data reading unit 3031 reads all the operation data (past operation data and new operation data) extracted by the data extraction / conversion unit 302 and the existing model stored in the existing model database DB3.

The model correction unit 3032 corrects the existing model read from the existing model database DB3 based on the model creation conditions stored inside. At that time, the difference between the preceding plant and the target plant is confirmed for the plant specifications read from the plant specification database DB4 and the fuel properties read from the fuel property database DB5. Further, input parameter addition candidates are read from the additional parameter candidate database DB7. From these addition candidates, a physical parameter is selected if the plant specification is different, and a fuel parameter is selected if the fuel property parameter is different, and added to the input parameter. This detail is shown in the model creation step S3 of FIG.

The model verification unit 3033 verifies the accuracy of the new model read from the model correction unit 3032. For accuracy verification, the following two items are implemented.

Item 1 of accuracy verification is to perform accuracy verification of the created new model using all operation data. Here, the actual measurement value (actual process value) of the process value in all the operation data is compared with the simulation value (virtual process value) of the process value calculated using the model, and the prediction error is calculated. Then, accuracy is verified by comparing the calculated prediction error with the allowable error stored in the model verification unit 3033 in advance.

Item 2 of accuracy verification verifies the relationship between typical input parameters and process values of the new model based on the criteria read from the sensitivity database DB8.

By the above-described accuracy verification in the model verification unit 3033, the accuracy of the model can be further improved by verifying the validity of the relationship (trend) between input and output from the viewpoint of the theory or the experience of the operator M.

The model output unit 3034 reads the verified new model from the model verification unit 3033 and outputs it to the new model database DB 6 and the input / output unit 309.

Next, an example of all operation data for creating a model will be described with reference to FIG. Here, the total operation data is past operation data and new operation data obtained from the data extraction / conversion unit 302. In the example of FIG. 6, the past operation data is the operation data in the plants A and B that are the preceding plants. The new operation data is operation data in the plant C that is the target plant. The operation data includes test operation data and actual operation data. By including both the test operation data and the actual operation data during actual operation as the total operation data, it is possible to improve the accuracy of the model by using all the data of the preceding plant.

Thus, in the example of FIG. 6, there are three plants (A, B, C), the plant A and the plant B are the preceding plants, and the plant C is the target plant. To create a model for plant C, which is a new target plant, let's create a new model for plant C by referring to the operation data in plants A and B, which are the preceding plants, and the models created in these preceding plants. It is said.

The relationship between operation and model creation at each plant will be explained in turn below. First, since the model of the plant A is necessary before the plant A is actually operated, it is created using the trial operation data of the plant A (operation data acquired by changing parameters and fuel experimentally). Here, it is assumed that the model creation process for the plant A has been completed, the operation is started, and the actual operation data is acquired.

The plant B model is created using both the plant A test operation data and the actual operation data (data acquired during the operation control period) and the plant B test operation data. Here, it is assumed that the model creation process for the plant B has been completed, the operation is started, and the actual operation data is acquired. During the model creation process for plant B, a model suitable for plant B is created by adding the physical and fuel parameters of plant B.

The new plant C model will be created now. In this creation, a model suitable for the plant C is created by adding the physical and fuel parameters of the plant C using the trial operation data and actual operation data of the plants A and B and the trial operation data of the plant C. Become.

In this way, when creating a model for a new target plant, the accuracy of the model can be improved by using all the data of the preceding plant. In addition, by creating a model in stages, it is possible to examine input parameters to be added and carefully select appropriate input parameters. Here, once added input parameters are not deleted, but are succeeded by the next plant. This is to maximize the results of existing plants and provide continuity.

Fig. 7 is a schematic diagram showing the relationship between model inputs and outputs. Here, the input parameters (inputs) of the model include physical parameters and / or fuel parameters in addition to parameters for the operation end. The parameter for the operation end is a parameter indicating an instruction value (valve opening degree or the like) to the operation end OP.

A model is created for each process value. Here, the process value is an output obtained in the boiler plant 100 as a result of the control in the operation control unit 201 in FIG. 4, and most of them are monitoring items (or monitoring control items) for monitoring and controlling the boiler plant 100. Is shown. These are, for example, the NOx value of exhaust gas discharged into the atmosphere for environmental monitoring, the temperature and pressure of steam given to a steam turbine, etc., and the control target amounts to be controlled by various devices and auxiliary equipment, and the sensor SR Is detected. Therefore, the model defines a relationship such as correlation established between each of a plurality of process values that are the output of the boiler plant 100 and an input parameter that affects the process value.

Below, an example of the configuration of the main database DB will be described and explained.

FIG. 8 shows an example of a data format in the past operation database DB1. In this example, the horizontal axis is divided by plant name, fuel used, number of operation cases, and the like. The vertical axis is divided by data relating to each of the input parameter and the process value. The new operation database DB2 basically has the same configuration as the data format in the past operation database DB1.

FIG. 9 shows an example of the data format in the existing model database DB3. The horizontal axis is divided by plant by creation date, function, input parameters, and model details. The function describes the method of modeling. Examples of the modeling method include, but are not limited to, stepwise method, random forest, k-nearest neighbor method (KNN), neural network method, deep learning, reinforcement learning, and the like. In the model details, each item of the model formula F shown below is described. Or you may describe the quotation destination of another DB in which each item was described.

The model formula F is, for example, the formula (1), where f is a modeling method (function), x is an input parameter, ω is weighted, λ is an intercept, and n is the number of input parameters.
[Equation 1]
F = f (x, ω, λ, n) (1)
The vertical axis is divided by the model creation unit. When creating a model for each process value, list the process values to be modeled.

When updating the model in the same plant, all versions of the model may be stored so that the update history can be understood. For example, in FIG. 9, the model PR1 is updated and represented as a model PR1 ′.

Since the creation date of each model is stored, it is possible to identify at which stage the input parameter was added at which plant. This is because the input parameter addition history can be followed later.

FIG. 10 shows an example of a data format in the plant specification database DB4. The horizontal axis is divided by plant name. List plant names that cover all preceding and target plants.

縦 軸 The vertical axis is divided by items representing plant specifications. Here, a distinction is made between structural specifications and performance specifications. The structural specifications are dimensions, and boiler dimensions are exemplified. The performance specification is a value representing the performance of the plant, and examples thereof include exhaust gas temperature and steam temperature. The structure specification and performance specification may describe not only the representative value of the measurement result but also the value of the design condition.

FIG. 11 shows an example of a data format in the fuel property database DB5. The horizontal axis is divided by fuel. List the fuels that cover all the preceding and target plants.

縦 軸 The vertical axis is divided by items representing fuel properties. Examples of the fuel properties include industrial analysis (fuel ratio, etc.) and elemental analysis (carbon amount, etc.). Here, the fuel ratio is the ratio of fixed carbon to volatile matter.

FIG. 12 shows an example of the data format in the additional parameter candidate database DB7. The horizontal axis is divided according to the data acquisition method and applicable conditions. In the data acquisition method, not only measured values but also calculated values (soft sensor values) calculated by combining a plurality of measured values may be used. If there is no measured value of an appropriate parameter that represents the plant specification, a calculated value can be substituted.

The vertical axis is divided by physical parameters and fuel parameters. Some physical parameters are obtained from boiler specifications such as structural dimensions, and others are obtained from measured values or calculated values such as gas temperature and steam temperature. The latter may be set with reference to the design value from the boiler specification. The physical parameter may be a soft sensor value calculated by combining a plurality of measurement values. By treating the soft sensor value as a physical parameter, a calculated value can be substituted when there is no measurement value of an appropriate parameter representing the plant specification.

The fuel parameter is an item used when evaluating the characteristics of the fuel, and is the motor current value, hydraulic pressure, differential pressure, etc. of the pulverizer (pulverized coal machine / mil) for coal pulverization, and fuel consumption for coal combustion. Quantity (call flow), absorbed heat quantity of the heat transfer surface, boiler output, etc., NOx value in exhaust gas with respect to environmental load, SO2 value, etc., and inlet air temperature of pulverizer, etc. with respect to moisture. By adding a fuel parameter as an additional parameter, it is possible to create an operation model that takes into account differences in fuel properties. Further, by using a parameter relating to any of fuel adjustment, combustion, environmental load, and moisture as the fuel parameter, an appropriate parameter representing the fuel property can be selected.

In addition, about the operation data regarding a grinder, when the boiler plant 100 is equipped with two or more grinders, it is preferable to create a new model using operation data of two or more grinders as input parameters. In this case, when calculating the virtual process value, the operation data of one pulverizer is used as a representative, and when the pulverizer stops, the operation data is switched to another pulverizer operation data. This is because even if one pulverizer stops due to maintenance or the like, operation support can be continued using operation data of another pulverizer. What is necessary is just to select a grinder according to the burner position which supplies an operation rate and pulverized coal. In particular, it is preferable to select from a pulverizer that supplies pulverized coal to the middle burner as much as possible. This is because the average behavior in the boiler can be reflected.

FIG. 13 shows an example of the data format in the sensitivity database DB8. The horizontal axis is divided by typical process values. Although steam temperature, metal temperature, and NOx value are illustrated, it is not restricted to this.

】 The vertical axis is classified by typical input parameters. Although a burner angle and an air damper opening degree are illustrated, it is not restricted to this. The sensitivity confirmation criterion is indicated by a linear change tendency such as proportional, inversely proportional, or constant, or a non-linear change tendency such as upward or downward, with respect to the relationship between the input parameter and the process value.

FIG. 14 is a display example of the sensitivity confirmation screen. The vertical axis represents a typical process value, and the horizontal axis represents a typical input parameter. The relationship between the two data is displayed. The model verification unit 3033 compares this result with the confirmation criteria stored in the sensitivity database DB8 to verify the validity of the new model. Alternatively, the operator M confirms the validity on the confirmation screen of the input / output unit.

Note that some or all of the functions of the operation control device 200 may be remotely or on the cloud and connected to the boiler plant 100 via the Internet line.

FIG. 17 is a diagram illustrating an example of a hardware configuration of the operation control apparatus 200. The operation control device 200 includes a CPU (CPU (Central Processing Unit) 601, a RAM (Random Access Memory) 602, a ROM (Read Only Memory) 603, an HDD (Hard Disk Drive) 604, an input I / F 605, and an output I / F 605, 6. These are configured using computers connected to each other via a bus 607. The hardware configuration of the operation control device 200 is not limited to the above, and is configured by a combination of a control circuit and a storage device. The operation control apparatus 200 is configured by a computer executing programs that realize the functions of the operation control apparatus 200, and these programs are stored in the cloud 1601 or the recording medium 1602. .

The program stored in the recording medium 1602 is, for example, a program having the function of the flowchart shown in FIG. 2, acquires an existing model indicating the relationship between the input parameter of the preceding plant and the process value, and sets the input parameter of the existing model as It is good also as a program (new model creation program) which adds the physical parameter relevant to the plant specification of a target plant, and makes a computer perform the process which creates a new model.

Further, the program stored in the recording medium 1602 is a program having the function of the flowchart shown in FIG. 3, for example, acquires a new model generated by the new model creation program, acquires new operation data of the target plant, It is good also as a program (driving support program) which calculates a process value using new operation data and a new model, and makes a computer perform the process which calculates the operation instruction value of a target plant based on the process value which satisfy | fills predetermined conditions.

The operation control apparatus 200 also includes an external communication device 608, for example, a wireless LAN communication device such as a 4G or 5G line communication device or Wi-Fi (registered trademark), and the CPU 601 executes a program from the cloud 1601 via the external communication device 608. May be read and loaded into the RAM 602 for execution. Alternatively, the operation control apparatus 200 may include a driver 609 for reading data on the recording medium 1602, and the CPU 601 may read a program from the recording medium 1602, load it into the RAM 602, and execute it. Regardless of the type of the recording medium 1602, various recording media such as an SD card, a USB memory, an external HDD, and the like depending on the capacity of the program can be used.

FIG. 18 is a diagram illustrating an example of a hardware configuration of the driving support device 300. While taking the same configuration as the operation control device 200 described above, from the output I / F 606 to an output unit (input / output unit 309) such as a monitor or a printer, a new model creation / verification result, a simulation evaluation result, It is configured to output (display) a driving instruction proposal screen.

In Example 2, it will be described that a driving support system that supports model creation is configured by collectively combining a plurality of boiler plants.

FIG. 15 is a diagram illustrating an example of the overall configuration of a driving support system 500 that supports model creation. In this figure, a driving support system 500 that supports model creation includes, for example, a plurality of local support systems 300A, 300B, and 300C provided for each of a plurality of boiler plants 100A, 100B, and 100C, and local support systems 300A, 300B, and 300C. And a remote support system 400 capable of communicating via the network N.

FIG. 16 is a diagram illustrating a detailed configuration example of the driving support system 500 that supports model creation. For example, the configuration of the local support system 300A is illustrated as a representative configuration example. The local support systems 300B and 300C have the same configuration as the local support system 300A.

The local support system 300A includes a driving support device 300 and a first transmission / reception unit 301. The driving control device 200 may be used instead of the driving support device 300. The remote support system 400 includes a second transmission / reception unit 401 and a common model database DB10.

The first transmission / reception unit 301 sends the update result of the new model and the new driving data created by the model creation device 303 in the driving support device 300 to the second transmission / reception unit 401 at regular intervals or by an instruction from the second transmission / reception unit 401. Send.

The second transmission / reception unit 401 receives new operation data and new model update results transmitted from the local support systems 300A, 300B, and 300C. When a new update result is received, it is transmitted to the first transmission / reception units 301 of all other local support systems 300A, 300B, and 300C as update results of all the operation data and the existing model at any time or at a constant cycle.

When the first transmission / reception unit 301 receives new operation data and an update result of the existing model, the first transmission / reception unit 301 transmits the update result of the entire operation data to the past operation database DB1 and the update result of the existing model to the existing model database DB3. .

Thus, the driving support system that supports model creation illustrated in FIGS. 15 and 16 includes a plurality of local support systems 300A, 300B, and 300C including the model creation device 303, the local support systems 300A, 300B, and 300C, and the network N. The local support systems 300A, 300B, and 300C transmit new driving data and new model update results to the remote support system 400 and are transmitted from the remote support system 400. The remote support system 400 includes a first transmission / reception unit 301 that receives all the driving data in other local support systems and the update result of the existing model. The remote support system 400 includes new driving data transmitted from the respective local systems 300A, 300B, and 300C. New model Which receives the result of the update, all local system 300A the update result of the other, 300B, is to have a second transceiver 401 that transmits a result of updating existing models and all operation data to 300C.

With this configuration, operation data and model update results in other local support systems (preceding plants) can be shared.

In the above description, the local systems 300A, 300B, and 300C are configured to transmit signals via the remote support system 400. However, this is for direct transmission / reception between the local systems 300A, 300B, and 300C. May be.

The present invention can be widely applied not only to coal-fired power plants but also to general plants.

31, 32, 33, 34, 35: Crusher, 100, 100A, 100B, 100C: Boiler plant, 200: Operation control device, 201: Operation control unit, 300: Operation support device, 300A, 300B, 300C: Local support System 301: first transmission / reception unit 303: model creation device 306: simulation unit 308: driving instruction unit 309: input / output unit (output unit) 400: remote support system 401: second transmission / reception unit 500 : Operation support system, DB1: Past operation database, DB2: New operation database, DB3: Existing model database, DB4: Plant specification database, DB5: Fuel property database, DB6: New model database, DB7: Additional parameter candidate database, DB8: Sensitivity database , N: network

Claims (14)

1. A model creation method for creating a model that shows the relationship between the input parameters of a target plant and a process value by utilizing the results of a preceding plant,
   A reading step of reading an existing model in the preceding plant, and a model creating step of adding a physical parameter related to the plant specification of the target plant to an input parameter of the existing model in the preceding plant to create a new model A model creation method characterized by comprising:
2. The model creation method according to claim 1,
   The model creation method characterized in that the physical parameter is a parameter related to a structure, performance and / or design conditions of a plant.
3. The model creation method according to claim 1 or 2,
   The model creation method characterized in that the physical parameter includes a soft sensor value calculated by combining a plurality of measurement values.
4. The model creation method according to any one of claims 1 to 3,
   The preceding plant and the target plant are fuel combustion plants,
   The model creation step is characterized in that a new model is created by adding a fuel parameter related to a property of the fuel to an input parameter of the existing model.
5. A model creation method according to claim 4,
   The model creation method, wherein the fuel parameter is a parameter relating to any of fuel adjustment, combustion, environmental load, and moisture.
6. A model creation method according to any one of claims 1 to 5,
   A verification step of verifying the accuracy of the new model using all operation data of the preceding plant and the target plant;
   In the verification step, the relationship between a representative input parameter of the new model and a process value is verified based on a predetermined criterion.
7. The model creation method according to claim 6,
   The total operation data includes both test operation data and actual operation data during actual operation.
8. A model creation method according to any one of claims 1 to 7,
   An output step of outputting the new model;
   The output step further outputs a relationship between a representative input parameter of the new model and a process value.
9. A plant operation support method using a new model generated by the model creation method according to any one of claims 1 to 8,
   A simulation step of calculating the process value using the new operation data of the target plant and the new model;
   A plant operation support method, further comprising an operation instruction step of calculating an operation instruction value of the target plant based on the process value satisfying a predetermined condition.
10. It is a model that shows the relationship between the input parameters of the target plant and the process value, created by utilizing the results of the preceding plant,
    A model created by adding, in an identifiable manner, physical parameters related to the plant specifications of the target plant to the input parameters of the existing model in the preceding plant.
11. A model creation device that creates a model that shows the relationship between the input parameters of the target plant and the process value by utilizing the results of the preceding plant,
    A data reading unit for reading an existing model in the preceding plant, and adding a physical parameter related to the plant specification of the target plant to an input parameter of the existing model in the preceding plant to create a new model, Model creation device.
12. The computer acquires the existing model indicating the relationship between the input parameter of the preceding plant and the process value, adds the physical parameter related to the plant specification of the target plant to the input parameter of the existing model, and creates a new model A program to be executed.
13. A new model generated by the program according to claim 12 is acquired,
    Obtain new operation data of the target plant,
    Calculate the process value using the new operation data and the new model,
    A program that causes a computer to execute a process of calculating an operation instruction value of the target plant based on the process value that satisfies a predetermined condition.
14. A recording medium on which the program according to claim 12 or 13 is recorded.

Images