WO2021171654A1 - Operation improvement assistance system, operation improvement assistance method, and operation improvement assistance program - Google Patents

Abstract

The present invention addresses the problem of automatically formulating a switching condition of a driving condition in a real-time manner. An operation improvement assistance system 200 that, by using driving data about equipment such as a coal-fired plant which is an application target, imparts a manipulation condition of the application target by using driving data from the application target in order to formulate a switching condition of a driving condition, is provided with a driving condition switching means 300 comprising: a switching condition generation means 310 that creates an initial value of a switching condition based on the driving data and knowledge data about driving of the application target; a switching condition evaluation means 320 that evaluates the initial value of the switching condition; and a switching condition correction means 330 that corrects the initial value of the switching condition on the basis of the result of evaluation by the switching condition evaluation means.

Description

Operation improvement support system, operation improvement support method and operation improvement support program

The present invention relates to an operation improvement support system, an operation improvement support method, and an operation improvement support program that support operation improvement of various devices and systems (hereinafter, simply devices).

In recent years, with the technological innovation of ICT (Information and Communication Technology) and IoT (Internet of Things), an environment where high-speed computers, network communications, and large-capacity data storage devices can be used is being prepared. While attention is focused on the utilization of a large amount of accumulated data in many industrial fields, integration of data collected on local sites such as equipment measurement data and inspection / maintenance data with systems that manage corporate management and asset information. Therefore, it is required to improve the important performance evaluation index.

For example, in the field of power generation business, it is important to use a thermal power plant as a backup power source because there is a concern that fluctuations in the amount of power generated due to increased use of renewable energy such as wind power generation and solar power generation will reduce the stability of the power system. The sex is increasing. In addition, a thermal power plant plays a role not only as a load adjustment but also as a base load power source, and is required to be operated in consideration of operational performance such as efficiency, environmental performance, and operating rate.

In order to improve the operational performance of a thermal power plant, Patent Document 1 and Patent Document 2 disclose a control device for reducing the nitrogen oxide concentration and carbon monoxide concentration, which are environmental performances.

Japanese Unexamined Patent Publication No. 2012-141862 Japanese Unexamined Patent Publication No. 2009-244933

In the technique described in the prior art document, an operation signal is generated by combining a model that simulates the characteristics of a plant and a learning means that learns an optimum operation method for this model. By using this technique, the operating conditions can be moved to the optimum values. Here, the operation condition is a value used for generating an operation signal.

Combustion state changes when operating conditions such as load and mill pattern change. Therefore, it is necessary to prepare a model for each operating condition and switch the model according to the change of the operating condition to learn the optimum operation method. However, if the operating conditions to be considered in advance cannot be grasped, or if the plant operation method is changed, for example, from the base load to the load adjustment, it is necessary to cope with the new operating conditions.

In order to solve the above problems, the present invention formulates the switching conditions of the operating conditions by using the operating data of the device to which the application is applied. The present invention also includes the following aspects. The formulation of switching conditions includes at least one of the creation and modification of switching conditions.

An operation improvement support system that gives operating conditions for the application target using operation data from the application target, a model that simulates the characteristics of the application target using the operation data and outputs an estimated value, and the estimation. An evaluation value calculation means that calculates an evaluation value based on a value, a learning means that learns an operation condition of the application target based on the evaluation value, and an operation condition of the application target are generated according to a learning result by the learning means. The operating condition switching means is provided with an operating condition signal generating means for performing the operation condition signal, a changeover switch for switching the data for constructing the model according to the operating condition, and an operating condition switching means for determining the switching condition for the operating condition. Is a switching condition generating means that creates knowledge data about the operation to be applied and the initial value of the switching condition based on the operation data, a switching condition evaluation means that evaluates the initial value of the switching condition, and the switching condition. It is an operation improvement support system having a switching condition correction means for correcting the initial value of the switching condition based on the evaluation result by the evaluation means.

The present invention includes an operation improvement support method using an operation improvement support system and a program product for executing this method. The program product includes a program and a medium in which the program is stored.

According to the present invention, it is possible to automatically formulate switching conditions for operating conditions in real time.

It is a block diagram explaining the configuration example of the operation improvement support system which concerns on embodiment of this invention. It is a flowchart explaining the operation in the operation improvement support system. It is a figure explaining the model 500 and learning means 600. It is a figure explaining the mode of the data stored in the database provided in the operation improvement support system 200. It is a flowchart explaining the operation of the operation condition switching means 300. It is a figure explaining the operation of the switching condition evaluation means 320. It is a figure explaining the operation of the switching condition correction means 330. It is a figure explaining the Example of the screen displayed on the image display device 940. It is a schematic diagram which shows the structure of the coal-fired power plant which is an Example of application target 100. It is a figure explaining the mode of the data stored in the knowledge database 240 when the application target is a coal-fired plant.

Hereinafter, each embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram illustrating a configuration example of the operation improvement support system 200 according to the first embodiment. In this embodiment, the operation improvement support system 200 is connected to the application target 100 of this system and the external device 900.

The operation improvement support system 200 of FIG. 1 is generally composed of a computer device. That is, an arithmetic unit such as a CPU executes various processing functions according to a program. If the processing function in the arithmetic unit is schematically shown, the control means 270, the operation condition switching means 300, the changeover switch 400, the model 500, the learning means 600, the evaluation value calculation means 700, and the operation condition signal generation means 800 are provided. It can be said that it is a thing. Further, the operating condition switching means 300 is provided with a switching condition generating means 310, a switching condition evaluation means 320, and a switching condition correcting means 330. The contents of the processing function, which is the operation of each part in the operation improvement support system 200, will be described with reference to FIGS. 2 and 2. The model 500 refers to a neural network model or the like as described later, and executes processing by itself.

The operation improvement support system 200 has an operation data database 230 (operation data DB), a knowledge database 240 (knowledge DB), and an operation data database 250 for each operation condition (operation data DB for each operation condition) as databases (described as DB in the figure). ), The processing result database 260 (writing result DB) is provided. Digitized information is stored in each database, and the information is usually stored in a form called an electronic file (electronic data). It should be noted that these DBs may be provided outside the operation improvement support system 200 and can be connected via a network.

Further, the operation improvement support system 200 includes an external input interface 210 and an external output interface 220 as interfaces with the outside, and is connected to the application target 100 of this system and the external device 900 via the interfaces.

With the interface configuration, the external input signal 1 created by operating the external input device 910 (keyboard 920 and mouse 930) provided in the external device 900 and the application target 100 via the external input interface 210 are collected. The existing operation data 2 is taken into the operation improvement support system 200.

The application target 100 is composed of a control device 180 and a device 190, a measurement signal 70 is transmitted from the device 190 to the control device 180, and an operation signal 80 is transmitted from the control device 180 to the device 190. The operation data 2 described above is data including the measurement signal 70 and the operation signal 80.

The operation data 2 taken into the operation improvement support system 200 is stored in the operation data database 230 as the operation data 3, and the knowledge data 4 included in the external input signal 1 is stored in the knowledge database 240.

Further, the operation improvement support system 200 outputs the operation condition signal 17 to one or both of the control device 180 in the application target 100 and the image display device 940 in the application target 100 via the external output interface 220. The control device 180 can control in real time by generating an operation signal 80 according to the received operation condition signal 17 and transmitting it to the device 190. Further, the content of the operation condition signal 17 is displayed on the image display device 940 as an operation guidance, and the operator of the device 190 changes the set value of the control device 180 to adjust the operation signal 80 at the operation end of the device 190. Can be adjusted directly. It should be noted that the control in real time is not limited to simultaneous processing as long as it can be controlled by time constraints. Time constraints include controlling within a preset time range.

In the operation improvement support system 200 of this embodiment, an example is shown in which the arithmetic unit constituting the computer device and the database DB are provided inside the operation improvement support system 200. However, some of these devices may be arranged outside the operation improvement support system 200 so that only data is communicated between the devices.

Further, the signal database information 50, which is a signal stored in each database DB, can be displayed on the image display device 940 via the external output interface 220. Further, these information can be corrected by the external input signal 1 generated by operating the external input device 910.

In this embodiment, the external input device 910 is composed of a keyboard 920 and a mouse 930, but any device such as a microphone for voice input and a touch panel for inputting data may be used.

Needless to say, this embodiment also includes a method using the operation improvement support system 200. Further, in this embodiment, the application target of the operation improvement support system 200 is a plant, but it goes without saying that the application target can be implemented as equipment other than the plant. Specific examples include, but are not limited to, engines, turbines, and the like. Further, the plant may be any equipment that executes operation (operation) in response to some operation such as a chemical plant in addition to a power plant such as a coal-fired power plant.

FIG. 2 is a flowchart of the operation improvement support system 200 when learning the operation improvement method.

Next, the operation of each part in the operation improvement support system 200 in this embodiment will be described with reference to FIG. In the first processing step 1000 in FIG. 2, it is determined whether or not the model needs to be updated. As a result of this determination, if the model update is required, the process proceeds to the process step 1100, and if the model update is rejected, the process ends. Here, as an example of the criterion for determining the necessity of model update, it is conceivable to compare the model output and the operation data, and if the error exceeds the threshold value, the model update is required. In the present invention, there is no limitation on the determination criteria for the necessity of updating the model, and other determination criteria may be used, or may be based on an instruction input from the user via the external device 900.

Next, in the processing step 1100, it is determined whether or not the operating condition switching condition needs to be changed. If it is necessary to change the operating condition switching condition, the process proceeds to process step 1020, and if the operation condition switching condition is not changed, the process proceeds to process step 1030. Here, as an example of the criteria for determining whether or not the operating condition switching condition needs to be changed, the operating condition switching condition needs to be changed when the operation method of the plant changes and a request for changing the switching condition is input from the operator. Can be considered. The present invention is not limited to the criteria for determining whether or not the operating condition switching condition needs to be changed, other criteria may be used, or the invention may be based on an instruction input from the user via the external device 900. good.

Note that the processing of the processing steps 1000 and 1100 may use the input from the external device 900 in addition to the control means 270. In the processing step 1020, the operating condition switching means 300 is operated. The operation content of the operating condition switching means will be described later with reference to FIG.

Next, in the processing step 1030, the operating condition switching means 300 divides the operating data 5 stored in the operating data database 230 for each operating condition determined in the processing step 1020, and divides the operating data database 250 for each operating condition. Save to. Note that FIG. 1 describes an example in which the number of operating conditions is n.

Next, in the processing step 1040, the data for constructing the model 500 is switched by switching the changeover switch 400 according to the operation of the operation condition switching means 300.
The changeover switch 400 first selects the operation database 250a for each operation condition of the operation condition 1. After that, the processing step 1050 and subsequent steps are performed, and when the process returns to the processing step 1040, the operation data database 250b for each operation condition of the operation condition 2 is selected.
By repeating this process, the model 500 is constructed and learned for each operating condition (for example, all operating conditions). The operating conditions are conditions related to the operation of the applicable target 100.

Next, in the processing step 1050, the operating condition switching means 300 sets the initial value of the operating condition 11 generated by the learning means 600. Then, the learning means 600 transmits the operation condition 11 to the model 500.

Next, in the processing step 1060, the model 500 applies the input of the operating condition 11 to itself to calculate the model output 12, and transmits the model output 12 to the evaluation value calculating means 700.

In the processing step 1070, the evaluation value calculation means 700 calculates the evaluation value 13 based on the input of the model output 12, and transmits the evaluation value 13 to the learning means 600.

In the processing step 1080, the learning means 600 learns how to generate the operating condition 11 so that the evaluation value 13 is maximized. Then, the learning means 600 generates the next operation condition 11, and transmits the operation condition 11 to the model 500. The learning means 600 learns a method of calculating an operation amount so that the operability of the application target 100 becomes a desired characteristic. The learning means 600 can be implemented by using an optimization algorithm such as reinforcement learning, a genetic algorithm, or a nonlinear programming method, but the present invention does not limit the implementation method of the learning means 600.

Next, in the processing step 1090, the control means 270 determines whether or not the number of operations of the processing steps 1060-1080 exceeds a predetermined threshold value. As a result, if it is below the threshold value, the process returns to the processing step 1060, and if it exceeds the threshold value, the process proceeds to the processing step 1100.

Next, in the processing step 1100, the control means 270 determines whether or not the number of operations of the processing step 1050 exceeds a predetermined threshold value, and if it does not exceed, the process returns to the processing step 1050, and if it exceeds, the processing step 1110 Proceed to.

In the processing step 1110, the control means 270 determines whether the number of learned operating conditions, that is, the number of operations of the processing step 1040 is equal to or greater than the number of operating conditions, and if it is less than the number of operating conditions, the process step 1040 is performed. Return, and if the number of operating conditions is exceeded, proceed to the end.

The result of operating the arithmetic unit in each processing step is transmitted to the processing result database 260 as the processing result 14 and saved. Further, when operating the arithmetic unit in each processing step, the information stored in the processing result database 260 can be used as needed.

After learning the operation improvement method according to the flowchart shown in FIG. 2, the process for operating the application target 100 is performed. The operation condition signal generation means 800 calculates the operation condition signal 16 in real time based on the input of the processing result 15, and transmits it to the external output interface 220. After that, the external output interface 220 transmits the operation condition signal 17 to the control device 180 and the image display device 940. As described above, it is possible to operate the application target 100 using the operation condition signal 17 and to display the value of the operation condition signal 17 on the image display device 940 as the operation guidance.

Next, the operation of the model 500 and the learning means 600 will be described with reference to FIG.

3 (a) and 3 (b) are diagrams for explaining the details of the model 500. The model 500 is constructed by a neural network model as shown in FIG. 3A, and outputs an index for evaluating operability such as efficiency and environmentally hazardous substances in response to input of operating conditions such as an air flow rate set value. do.

FIG. 3B is a diagram showing the relationship between the input and the output of the neural network model. According to the neural network model, the value of an index for interpolating the input operation data and evaluating the operability for an arbitrary operation condition. Can be sought.

FIG. 3C is an example showing the result of operating the learning means 600. In this example, the result of learning the relationship between the operating condition and the change width of the operating condition when executing each process of FIG. 2 is shown at present. In the example of FIG. 3C, the operating condition is increased when the current manipulated variable is in the area A, and the operating condition is decreased when the current manipulated variable is in the region B. By changing the operating conditions in this way, the index for evaluating the operability in FIG. 3B becomes a minimum value, and the operating performance can be improved.

In the above description, the operations of the model 500 and the learning means 600 have been described with reference to FIGS. 3 (a), 3 (b), and 3 (c), but the model 500 and the learning means 600 are used as a neural network model. Needless to say, it can be constructed by other technologies.

Next, each database (DB) will be described with reference to FIG. FIG. 4A and FIG. 4B both show the contents of the data stored in the database. Specifically, FIG. 4A shows the data stored in the operation data database 230. As shown in FIG. 4A, operation data such as items A, B, and C measured by the sensor are stored for each sampling cycle. The trend graph of the operation data can be displayed on the image display device 940. FIG. 4B shows data stored in the knowledge database 240, that is, data related to operation (operation). As shown in FIG. 4B, the knowledge database 240 stores the PIDs of operation data and their classification conditions. Operation data includes analog signals and digital signals. The analog signal is a signal that continuously changes, for example, temperature and pressure, and the correspondence between the condition and the range of the analog signal value is stored. The digital signal is, for example, a signal indicating ON / OFF of the device (1 when ON, 0 when OFF), and the correspondence between the condition and the digital signal value is stored.

Note that the knowledge database 240 may store classification conditions related to values calculated using operation data, for example, values such as signal change rate and moving average value. Although not shown, the processing result database 260 is required to obtain the weighting coefficients of the neural network model shown in FIG. 3 (a) and the results shown in FIGS. 3 (b) and 3 (c). Information etc. are saved.

Next, the operation of each configuration requirement of the operating condition switching means 300 will be described with reference to FIG. FIG. 5 is a flowchart for explaining the operation of the operating condition switching means 300, and is a diagram for explaining the details of the processing step 1020 in FIG.

In the processing step 1021, the switching condition generating means 310 generates the initial value of the switching condition. The initial value of the switching condition is created based on the knowledge data 6 stored in the knowledge database 240, that is, the information shown in FIG. 4 (b).

Next, in the processing step 1022, the switching condition evaluation means 320 evaluates the switching condition. The evaluation method in the switching condition evaluation means 320 will be described later with reference to FIG.

Next, in the processing step 1023, the switching condition correcting means 330 corrects the switching condition based on the result of the switching condition evaluation means 320. The correction method in the switching condition correction means 330 will be described later with reference to FIG. 7.

Next, in the processing step 1024, the switching condition correction means 330 executes the end determination of the switching condition correction, and if it ends, the process proceeds to the process step 1030, and if it does not end, the process returns to the process step 1022. As a result, the operation data 8 for each operation condition is output to the operation data database 250 for each operation condition. Then, the operation data 9 for each operation condition is output to the changeover switch 400 from the operation data database 250 for each operation condition.

By operating the operating condition switching means 300 as shown in FIG. 5 described above, it is possible to modify the switching conditions in real time according to the operating status of the application target 100.

Next, the operation of the switching condition evaluation means 320 will be described with reference to FIG. As shown in FIG. 6, in the switching condition evaluation means 320, the smaller the variation in the operation data values with respect to the same instrumental variable value, the higher the evaluation. That is, as shown in FIG. 6A, when the variation is large, it is assumed that the relationship between the manipulated variable and the operation data cannot be estimated accurately, so the evaluation is low. On the contrary, as shown in FIG. 6B, the higher the number of data for each operating condition is, the higher the evaluation is. Further, when the number of data for each operating condition is small, it is assumed that the number of data is not required to create the model 500, so the evaluation is low.

The switching condition evaluation means 320 calculates the evaluation value E of the switching condition using the equation 1.

E = E1 + E2 ... (number 1)
Here, E1 is an evaluation value of the degree of variation in the operation data, E2 is an evaluation value of the number of data for each operation condition, and is calculated by the equations 2 and 3.

E1 = Σ (∫ V dt)… (Equation 2)
E2 = ΣC ... (number 3)
Here, Σ means to add for each operating condition (for example, all operating conditions). Further, V is the variation of the operation data value for the same operation variable value, C is zero when the number of data for each operation condition is smaller than the threshold value, and is a value which becomes higher as the number of data is larger than the threshold value.

Next, the operation of the switching condition correction means 330 will be described with reference to FIG. 7. The switching condition correcting means 330 modifies the switching condition so that the evaluation value of the switching condition evaluation means 320 becomes high. That is, if the variation in the operation data values for the same instrumental variable value is larger than the predetermined condition, the classification is made finer (more subdivided). In addition, if the number of data for each operating condition is less than a certain number, the classification is roughened (more centralized). It should be noted that the large variation is determined, for example, by comparison with a predetermined threshold value. FIG. 7A shows a switching condition classification method before modification of the analog signal A001. The results of classifying the values from 0 to 100 in increments of 10 are shown. Further, FIG. 7B is a result when the classification is roughly modified. That is, the new condition 1 (80 to 100) is obtained by combining the condition 1 (90 to 100) and the condition 2 (80 to 90) in FIG. 7 (a). Further, FIG. 7 (c) shows the result when the classification is finely modified. That is, the condition 1 (90 to 100) in FIG. 7 (a) is changed to the new condition 1 (95 to 100) and the condition 2 (90 to 100). Further, FIG. 7D is an example in which the operating conditions are defined by the combination of the digital signal D001 and the signal D002. The digital signal D001 and the signal D002 are values indicating ON / OFF of the device (0 means OFF and 1 means ON), and there are four combinations as shown in FIG. 7D. Further, FIG. 7 (e) shows the result when the classification is roughly modified. By integrating the operating conditions in which the number of devices that are turned on is the same (the sum of the digital signals D001 and D002 is the same), the classification is roughly corrected.

It is also possible to increase the number of data items used for operating condition classification as a method of making the classification more detailed. It is also possible to reduce the number of data items used for operating condition classification as a method of roughening the classification. The switching condition correction means 330 has a function of combining the above-mentioned methods to make the classification finer or coarser.

As described above, the operation improvement support system 200 of the present embodiment can reduce the variation in the operation data values with respect to the same operation variable value, and can increase the number of data for each operation condition to a certain number or more, according to the model 500. The estimation accuracy can be improved.

Next, FIG. 8 shows an example of a screen displayed on the image display device 940. Here, FIG. 8A is a screen for selecting a signal to be used for classifying operating conditions. Based on the selected signal, the items of data stored in the knowledge database 240 are determined. Further, FIG. 8B is a screen displayed in real time when the system is being used, and the current operating conditions and the definition contents of the operating conditions are displayed.

From the display screen shown above, it is possible to check the current operating conditions and the definition contents of the operating conditions in real time.

Next, in Example 2, a case where a coal-fired power plant is applied will be described.

FIG. 9 shows the configuration of a coal-fired power plant to which this embodiment is applied. First, the mechanism of power generation by a coal-fired plant will be briefly described with reference to FIG.

In FIG. 9, the boiler 101 constituting the coal-fired power plant is supplied with pulverized coal, which is a fuel obtained by finely crushing coal with a mill 134, primary air for transporting pulverized coal, and secondary air for combustion adjustment. A plurality of burners 102 are provided. Then, the pulverized coal supplied through the burner 102 is burned inside the boiler 101. As shown in the figure, the structure of the burner 102 is arranged in a plurality of stages in front of and behind the boiler 101, and a plurality of burners are arranged in a row in each stage. According to the burner structure and arrangement shown in FIG. 9, pulverized coal is burned inside the boiler 101 from the front surface (hereinafter referred to as the can front) and the back surface (hereinafter referred to as the can rear) of the boiler. By improving the burner combustion balance before and after the can, the heat recovery effect of the boiler is improved and the thermal efficiency of the plant is also improved.

The pulverized coal and the primary air are guided from the pipe 139, and the secondary air is guided from the pipe 141 to the burner 102, respectively. The primary air is guided from the fan 120 to the pipe 130 and branches into a pipe 132 that passes through the air heater 104 installed on the downstream side of the boiler 101 on the way and a pipe 131 that bypasses the air heater 104 without passing through the air heater 104. do. However, it becomes a pipe 133 arranged on the downstream side of the air heater 104, joins again, and is guided to the mill 134 for producing pulverized coal installed on the upstream side of the burner 102. The primary air passing through the air heater 104 is heated by exchanging heat with the combustion gas flowing down the boiler 101. Along with this heated primary air, the primary air bypassing the air heater 104 conveys the differential coal crushed in the mill 134 to the burner 102.

Mills 134 are arranged so as to correspond to each burner stage (4 units in FIG. 9), and pulverized coal and primary air are supplied to the burners constituting each stage. That is, when the coal supply amount is reduced, such as when the power generation output is reduced, the mill can be stopped and the burner can be stopped for each burner stage. In the mill 134, in consideration of the combustibility of the boiler 101, the rotation speed of the mill is adjusted so that pulverized coal having a desired particle size can be obtained according to the properties of the coal used. Further, the coal stored in the coal bunker 136 is guided to the coal feeder 135 via the coal conveyor 137, and the supply amount is adjusted by the coal feeder 135. After that, it is supplied to the mill 134 via the coal conveyor 138.

Further, the boiler 101 is provided with an after-airport 103 for injecting air for staged combustion into the boiler 101. The air for staged combustion is guided from the pipe 142 to the after airport 103. In the boiler 101 shown in FIG. 9, the air introduced from the pipe 140 using the fan 121 is similarly heated by the air heater 104. After that, the pipes 141 for the secondary air and the pipes 142 for the after-airport are branched, and are guided to the burner 102 and the after-airport 103 of the boiler 101, respectively. The air flow rate supplied to the burner 102 and the after airport 103 can be adjusted by operating air dampers (not shown) installed in the pipes 141 and 142, respectively.

The high-temperature combustion gas generated by burning pulverized coal inside the boiler 101 flows down to the downstream side along the internal path of the boiler 101, and is supplied with water by the heat exchanger 106 arranged inside the boiler 101. Heat exchanges with and generates steam. After that, it becomes exhaust gas and flows into the air heater 104 installed on the downstream side of the boiler 101, and heat is exchanged by the air heater 104 to raise the temperature of the air supplied to the boiler 101.

Then, the exhaust gas that has passed through the air heater 104 is discharged from the chimney to the atmosphere after being treated with an exhaust gas (not shown).

The water supply circulating in the heat exchanger 106 of the boiler 101 is supplied to the heat exchanger 106 via the water supply pump 105, and is overheated by the combustion gas flowing down the boiler 101 in the heat exchanger 106 to become high-temperature and high-pressure steam. Although the number of heat exchangers is one in this embodiment, a plurality of heat exchangers may be arranged.

The high-temperature and high-pressure steam generated in the heat exchanger 106 is guided to the steam turbine 108 via the turbine governor 107, and the energy of the steam drives the steam turbine 108 to generate power in the generator 109.

In the coal-fired power plant of this embodiment, various measuring instruments for detecting the state quantity indicating the operating state are arranged. The measurement signal of the coal-fired power plant acquired from the measuring instrument arranged in the coal-fired power plant is stored in the operation data database 230 shown in FIG. As the measuring instrument, for example, there is one shown in FIG. That is, the temperature measuring device 151 for measuring the temperature of the high-temperature and high-pressure steam supplied from the heat exchanger 106 to the steam turbine 108, the pressure measuring device 152 for measuring the steam pressure, and the amount of power generated by the generator 109 are measured. There is a power generation output measuring instrument 153.

Further, the water supply generated by cooling the steam by the condenser of the steam turbine 108 (not shown) is supplied to the heat exchanger 106 of the boiler 101 by the water supply pump 105, and the flow rate of this supply water is the flow rate measuring device. Measured by 150. The components contained in the exhaust gas is a combustion gas discharged from the boiler 101 (oxides of nitrogen (NOx), carbon monoxide (CO), and hydrogen sulfide (H 2 _S), etc.) of the state quantity relating to the concentration of The measurement signal is measured by the concentration measuring device 154 provided on the downstream side of the boiler 101.

In addition, there are the following measuring instruments related to the coal supply system. Measure the flow rate of the primary air supplied to the mill 134 through the pipe 133. Measure the amount of coal supplied to the mill 134 through the coal conveyor 138 from the primary air flow meter 155 and the coal feeder 135. There is a coal supply meter 156 and a rotation meter 157 that measures the rotation speed of the mill 134. These are configured to be able to measure the above information for each mill and coal feeder.

That is, in the operation data database 230 in this embodiment, the coal flow rate supplied to the boiler 101, which is the state quantity of the applicable target 100 which is the coal-fired power plant measured by each of the above measuring instruments, the rotation speed of the mill 134, and the boiler 101 The primary and secondary air flow rates supplied to the boiler 101, the water supply flow rate supplied to the heat exchanger 106 of the boiler 101, the steam temperature generated by the heat exchanger 106 of the boiler 101 and supplied to the steam turbine 108, and the boiler 101. The supply pressure of the supply water supplied to the heat exchanger 106, the gas temperature of the exhaust gas discharged from the boiler 101, the gas concentration of the exhaust gas, and the exhaust gas recirculation that recirculates a part of the exhaust gas discharged from the boiler 101 to the boiler 101. Circulation flow rate, etc. are included.

In general, a large number of measuring instruments other than those shown in FIG. 9 are arranged in a coal-fired power plant, but the illustration is omitted here.

Next, the data stored in the knowledge database 240 in this embodiment will be described with reference to FIG. As shown in FIG. 10, the knowledge database 240 stores candidate conditions for switching the model for each item of load, load change rate, mill ON / OFF state, and fuel calorific value. In addition to the items shown in FIG. 10, the operating conditions are determined according to the fuel composition values such as water content, hydrogen content, distribution, volatile content, carbon content, sulfur content, and nitrogen content contained in the fuel. Is also good.

The total number of operating conditions is calculated by multiplying the number of conditions of each item, but when the performance is evaluated for all operating conditions, it takes time to carry out the process shown in FIG. Therefore, the processing time can be reduced by equipping the switching condition generating means 310 and the switching condition correcting means 330 with the functions described below. Generally, in a coal-fired plant, the number of mills to be used is determined according to the load, so it is possible to first grasp the relationship between the load and the number of mills from the operation data or based on the knowledge about plant operation. .. Using this result, the time required for the processing of FIG. 5 can be reduced by excluding in advance the operating conditions of the combination of the load and the number of mills, which are not expected in the switching condition generating means 310 and the switching condition correcting means 330.

When the operation improvement support system 200 is applied to a thermal power plant, the value used for air damper control can be considered as the operating condition. For example, the excess air ratio, the burner air ratio, the front-rear ratio of the after-airport air amount, the left-right ratio of the after-airport air amount, and the like. Further, in the index for evaluating operability, that is, the model output, it is desirable to use at least one of the following state quantities. That is, the amount of coal consumed in the thermal power plant, the unburned content in the ash emitted from the thermal power plant, carbon monoxide, nitrogen oxides, sulfide oxides, mercury, fluorine, fine particles consisting of soot and mist, It is defined as one of the volatile organic compounds. By constructing the model 500 that estimates the state quantity from these operating conditions, the operation of the thermal power plant can be improved.

In addition, when operating conditions such as load, load change rate, mill ON / OFF state, and fuel calorific value change, the combustion state changes and the relationship between the operating variable and the state amount discharged from the thermal power plant also changes.

By using the operation improvement support system 200 of the present invention, it is possible to modify the operation condition switching method in real time according to the change of the plant operation method, and reduce the value of the state quantity emitted from the thermal power plant. It becomes possible to do.

The present invention can be widely used as an operation improvement support system for various devices.

1 ... External input signal, 2 ... Operation data, 3 ... Operation data, 4 ... Knowledge data, 5 ... Operation data, 6 ... Knowledge data, 7 ... Changeover switch signal , 8 ... Operating data for each operating condition, 9 ... Operating data for each operating condition, 10 ... Model construction data, 11 ... Operating conditions, 12 ... Model output, 13 ... Evaluation value, 14 ... Processing result, 15 ... Processing result, 16 ... Operation condition signal, 17 ... Operation condition signal, 70 ... Measurement signal, 80 ... Operation signal, 100 ...・ Applicable target, 180 ・ ・ ・ Control device, 190 ・ ・ ・ Equipment, 200 ・ ・ ・ Operation improvement support system, 210 ・ ・ ・ External input interface, 220 ・ ・ ・ External output interface, 230 ・ ・ ・ Operation data database, 240 ... Knowledge database, 250 ... Operation data database for each operating condition, 260 ... Processing result database, 270 ... Control means, 300 ... Operating condition switching means, 310 ... Switching condition generating means, 320 ... Switching condition evaluation means, 330 ... Switching condition correction means, 400 ... Changeover switch, 500 ... Model, 600 ... Learning means, 700 ... Evaluation value calculation means, 800 ... -Operating condition signal generation means, 900 ... external device, 910 ... external input device, 920 ... keyboard, 930 ... mouse, 940 ... image display device

Claims (15)

1. It is an operation improvement support system that gives the operating conditions of the application target using the operation data from the application target.
   A model that simulates the characteristics of the application target using the above operation data and outputs an estimated value,
   An evaluation value calculation means that calculates an evaluation value based on the estimated value, and
   A learning means for learning the operating conditions of the application target based on the evaluation value, and
   An operation condition signal generation means that generates an operation condition to be applied according to a learning result by the learning means, and an operation condition signal generation means.
   A changeover switch that switches the data for building the model according to the operating conditions,
   It is provided with an operating condition switching means for determining the switching condition of the operating condition.
   The operating condition switching means includes a switching condition generating means for creating knowledge data about the operation to be applied and an initial value of the switching condition based on the operating data, and a switching condition evaluation means for evaluating the initial value of the switching condition. An operation improvement support system characterized by having a switching condition correction means for correcting the initial value of the switching condition based on the evaluation result by the switching condition evaluation means.
2. The operation improvement support system according to claim 1.
   The switching condition evaluation means is an operation improvement support system characterized in that the initial value of the switching condition is evaluated based on the variation of a plurality of operation data related to the same operation variable value and the number of data for each operation condition.
3. The operation improvement support system according to claim 1.
   The switching condition correcting means subdivides the classification in the initial value of the switching condition when the variation of the operating data is larger than the predetermined condition, and when the number of data for each operating condition is a certain number or less, the switching condition. Operation improvement support system characterized by centralizing the classification of.
4. The operation improvement support system according to any one of claims 1 to 3.
   The applicable target is a coal-fired plant.
   Estimates of the model consist of coal flow consumed by the coal-fired plant, unburned ash emissions from the coal-fired plant, carbon monoxide, nitrogen oxides, sulfide oxides, mercury, fluorine, soot or mist. A state amount containing at least one of fine particles and volatile organic compounds.
   The switching condition generating means is an operation improvement support system characterized in that the load of the coal-fired plant, the load change rate, the ON / OFF state of the mill, the fuel composition, and the classification conditions of the fuel composition are used as the knowledge data.
5. The operation improvement support system according to claim 4.
   The relationship between the load and the number of mills is stored in the storage means, and the relationship is stored.
   The switching condition generating means and the switching condition correcting means are characterized in that they specify an operating condition including a combination of the load and the mill which is not considered as the operating condition, based on the relationship between the load and the number of the mills. Operation improvement support system.
6. It is an operation improvement support method using an operation improvement support system that gives the operation conditions of the application target using the operation data from the application target.
   The operation improvement support system is
   A model that simulates the characteristics of the application target using the above operation data and outputs an estimated value,
   An evaluation value calculation means that calculates an evaluation value based on the estimated value, and
   A learning means for learning the operating conditions of the application target based on the evaluation value, and
   An operation condition signal generation means that generates an operation condition to be applied according to a learning result by the learning means, and an operation condition signal generation means.
   A changeover switch that switches the data for building the model according to the operating conditions,
   It is provided with an operating condition switching means for determining the switching condition of the operating condition.
   The operation condition switching means creates knowledge data about the operation to be applied and an initial value of the switching condition based on the operation data, evaluates the initial value of the switching condition, and based on the evaluation result of the initial value. , An operation improvement support method characterized by modifying the initial value of the switching condition.
7. The operation improvement support method according to claim 6.
   The operating condition switching means evaluates the initial value of the switching condition based on the variation of a plurality of operating data related to the same instrumental variable value and the number of data for each operating condition. Operational improvement support method characterized by this.
8. The operation improvement support method according to claim 6.
   When the variation of the operation data is larger than the predetermined condition, the operation condition switching means subdivides the classification in the initial value of the switching condition, and when the number of data for each operation condition is a certain number or less, the switching condition. Operation improvement support method characterized by centralizing the classification of.
9. The operation improvement support method according to any one of claims 6 to 8.
   The applicable target is a coal-fired plant.
   Estimates of the model consist of coal flow consumed by the coal-fired plant, unburned ash emissions from the coal-fired plant, carbon monoxide, nitrogen oxides, sulfide oxides, mercury, fluorine, soot or mist. A state amount containing at least one of fine particles and volatile organic compounds.
   The operation condition switching means is an operation improvement support method characterized in that the load of the coal-fired plant, the load change rate, the ON / OFF state of the mill, the fuel composition, and the classification conditions of the fuel composition are used as the knowledge data.
10. The operational improvement support method according to claim 9.
    The relationship between the load and the number of mills is stored in the storage means, and the relationship is stored.
    The operation condition switching means is an operation improvement support method, characterized in that an operation condition including a combination of the load and the mill, which is not considered as the operation condition, is specified based on the relationship between the load and the number of the mills.
11. It is an operation improvement support program for making a computer function as an operation improvement support system that gives operating conditions of the application target by using operation data from the application target.
    The operation improvement support system
    A model that simulates the characteristics of the application target using the above operation data and outputs an estimated value,
    An evaluation value calculation means that calculates an evaluation value based on the estimated value, and
    A learning means for learning the operating conditions of the application target based on the evaluation value, and
    An operation condition signal generation means that generates an operation condition to be applied according to a learning result by the learning means, and an operation condition signal generation means.
    A changeover switch that switches the data for building the model according to the operating conditions,
    It functions as an operating condition switching means for determining the switching condition of the operating condition.
    The operating condition switching means is a switching condition generating means for creating knowledge data about the operation to be applied and an initial value of the switching condition based on the operating data, and a switching condition evaluation means for evaluating the initial value of the switching condition. An operation improvement support program characterized by functioning as a switching condition correction means for correcting the initial value of the switching condition based on the evaluation result by the switching condition evaluation means.
12. The operational improvement support program according to claim 11.
    The switching condition evaluation means is characterized in that the initial value of the switching condition is evaluated based on the variation of a plurality of operating data related to the same instrumental variable value and the number of data for each operating condition. Operational improvement support program.
13. The operational improvement support program according to claim 11.
    When the variation of the operation data is larger than the predetermined condition by the switching condition correcting means, the classification in the initial value of the switching condition is subdivided, and when the number of data for each operation condition is less than a certain number, the switching is performed. An operation improvement support program characterized by the realization of centralizing the classification of conditions.
14. The operation improvement support program according to any one of claims 11 to 13.
    The applicable target is a coal-fired plant.
    Estimates of the model consist of coal flow consumed by the coal-fired plant, unburned ash emissions from the coal-fired plant, carbon monoxide, nitrogen oxides, sulfide oxides, mercury, fluorine, soot or mist. A state amount containing at least one of fine particles and volatile organic compounds.
    The switching condition generating means is characterized in that it is possible to use the load of the coal-fired plant, the load change rate, the ON / OFF state of the mill, the fuel composition, and the classification conditions of the fuel composition as the knowledge data. Operation improvement support program.
15. The operational improvement support program according to claim 14.
    The relationship between the load and the number of mills is stored in the storage means, and the relationship is stored.
    It is realized that the switching condition generating means and the switching condition correcting means specify an operating condition including a combination of the load and the mill which is not considered as the operating condition based on the relationship between the load and the number of the mills. Operation improvement support program characterized by doing.

Images