Incinerator control device and incinerator control method

Abstract

Provided is an incinerator control device that appropriately controls the operating state of an incinerator. [Solution] The incinerator control device 100 comprises a process data acquisition unit 51 that acquires process data of the incinerator, an operating state information acquisition unit 53 that acquires operating state data of the incinerator, a memory unit 55 that stores the process data and operating state data, a prediction value calculation unit 57 that calculates a short-term fluctuation prediction value over a predetermined period based on the process data and operating state data in the memory unit 55 as a prediction value of the predetermined numerical value after a predetermined first hour, for a predetermined numerical value related to the operation of the incinerator, and calculates a long-term fluctuation prediction value over a period longer than the predetermined period based on the process data and operating state data in the memory unit 55 as a prediction value of the predetermined numerical value after a predetermined second hour, a short-term control unit 59 that controls a short-term control target quantity based on the short-term fluctuation prediction value, and a long-term control unit 61 that controls a long-term control target quantity based on the long-term fluctuation prediction value.

Description

[Technical Field]

[0001] The present invention relates to an incinerator control device that operates a boiler using heat from an incinerator that burns materials to be incinerated, and a method for controlling an incinerator. [Background technology]

[0002] Conventionally, power generation systems have been used in which heat generated when materials to be incinerated (e.g., garbage) are burned in an incinerator to generate steam in a boiler, and this steam is used to rotate a steam turbine generator to generate electricity. Various numerical values ​​related to the operation of the incinerator, such as the amount of steam generated per unit time in the boiler (hereinafter also referred to as "steam flow rate"), vary depending on the amount of materials to be incinerated supplied and the quality of the materials to be incinerated. In a power generation system, it is necessary to suppress fluctuations in the amount of power generated. One example of a technology for suppressing fluctuations in these numerical values ​​is described in Patent Document 1, the source of which is shown below.

[0003] Patent Document 1 describes a control device that controls the combustion of waste. This control device includes a steam flow rate prediction unit and a control unit. The steam flow rate prediction unit uses multiple prediction models to predict the steam flow rate 60 seconds, 120 seconds, and 180 seconds from now. The control unit controls combustion to promote combustion when the future predicted value of the steam flow rate falls below a preset lower threshold, and controls combustion to suppress combustion when the predicted value exceeds a preset upper threshold. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 2023-59470 Summary of the Invention [Problem to be solved by the invention]

[0005] The steam flow rate, which is an example of a numerical value related to the operation of an incinerator, may deviate from its set value due to an excessive or insufficient supply of air to the combustion space. An excessive or insufficient supply of fuel (garbage in Patent Document 1) may also cause the steam flow rate to deviate from its set value. Because the combustion state in a combustion furnace changes from moment to moment, a control delay occurs, making it impossible to perform optimal control according to the combustion state. This may result in the steam flow rate undershooting below the lower threshold or overshooting above the upper threshold. In other words, there is room for improvement in the accuracy of incinerator operation control.

[0006] Therefore, there is a need for technology that can appropriately control the operation of incinerators. [Means for solving the problem]

[0007] A characteristic configuration of an incinerator control device according to the present invention is a control device for an incinerator that operates a boiler using heat from an incinerator that burns materials to be incinerated, and includes: a process data acquisition unit that acquires process data of the incinerator; an operating state information acquisition unit that acquires operating state data of the incinerator; a memory unit that sequentially stores the process data and the operating state data; a prediction value calculation unit that calculates a short-term fluctuation prediction value over a predetermined period based on the process data and the operating state data stored in the memory unit as a prediction value of a predetermined numerical value related to the operation of the incinerator that is not the amount of steam discharged from the boiler after a predetermined first hour, and calculates a long-term fluctuation prediction value over a period longer than the predetermined period based on the process data and the operating state data stored in the memory unit as a prediction value of the predetermined numerical value after a predetermined second hour; a short-term control unit that controls a short-term control target quantity based on the short-term fluctuation prediction value; and a long-term control unit that controls a long-term control target quantity based on the long-term fluctuation prediction value.

[0008] With this characteristic configuration, it is possible to reduce fluctuations in incinerator operation by predicting predetermined values ​​related to the incinerator's operation after a predetermined time based on the process data and operating status data stored in the memory unit and controlling the incinerator based on this predicted result. Furthermore, since this configuration calculates short-term fluctuation prediction values ​​and long-term fluctuation prediction values ​​separately, optimal control can be achieved depending on the cause of deviations in incinerator operation from the settings. Therefore, it is possible to appropriately control the incinerator's operation.

[0009] It is also preferable that the predetermined value is a temperature measured at a predetermined portion of the incinerator.

[0010] It is also preferable that the specified numerical value is at least one numerical value selected from the group consisting of the temperature of the secondary combustion chamber of the incinerator, the temperature of the post-combustion chamber of the incinerator, and the temperature of steam discharged from the boiler.

[0011] It is also preferable that the predetermined value is the concentration of a predetermined gas measured in a predetermined portion of the incinerator.

[0012] It is also preferable that the predetermined value is the concentration of oxygen inside the incinerator.

[0013] According to these preferred configurations, control is performed based on predictions of indicators related to the operation of the incinerator (temperature measured in a specified part of the incinerator, concentration of a specified gas measured in a specified part of the incinerator), so that optimal control can be achieved depending on the cause of the deviation of the indicator from the setting.

[0014] a short-term control step of controlling a short-term controlled variable based on the short-term controlled variable; and a long-term control step of controlling a long-term controlled variable based on the long-term controlled variable.

[0015] With this characteristic configuration, it is possible to reduce fluctuations in incinerator operation by predicting predetermined values ​​related to the incinerator's operation after a predetermined time based on process data and operating status data, and controlling the incinerator based on the predicted results. Furthermore, since this method calculates short-term fluctuation prediction values ​​and long-term fluctuation prediction values ​​separately, optimal control can be achieved depending on the cause of deviations in incinerator operation from the settings. Therefore, it is possible to appropriately control the incinerator's operation. [Brief explanation of the drawings]

[0016] [Figure 1] FIG. 1 is a diagram illustrating a configuration of a power generation system. [Figure 2] FIG. 2 is a block diagram showing the configuration of a control device. [Figure 3] 4 is an example of process data stored in a storage unit. [Figure 4] FIG. 10 is a diagram illustrating preprocessing performed before calculating a short-term fluctuation prediction value. [Figure 5] FIG. 10 is an explanatory diagram of calculation of a short-term fluctuation prediction value. [Figure 6] FIG. 10 is an explanatory diagram of calculation of a long-term fluctuation prediction value. [Figure 7] FIG. 10 is an explanatory diagram of calculation of a short-term fluctuation predicted value and a long-term fluctuation predicted value. [Figure 8] FIG. 10 is a diagram illustrating control by a short-term control unit and a long-term control unit. DETAILED DESCRIPTION OF THE INVENTION

[0017] The incinerator control device according to the present invention is configured to be able to control an incinerator that operates a boiler using heat from the incinerator that burns materials to be incinerated. The incinerator control device (hereinafter referred to as the "control device") 100 of this embodiment will be described below. However, the control device 100 is not limited to the following embodiment, and various modifications are possible within the scope of the gist thereof.

[0018] Fig. 1 shows a power generation system 1 that generates power using steam from a boiler 20 that operates based on heat from an incinerator 10 controlled by a control device 100 shown in Fig. 2. As shown in Fig. 1, the power generation system 1 includes the incinerator 10, the boiler 20, a steam turbine generator 30, and a chimney 40.

[0019] In this embodiment, the incinerator 10 is configured as a rotary stoker-type incinerator in which the furnace body 11 is formed in a cylindrical shape and rotates around its axis. The furnace body 11 has an inlet 11a for the material to be incinerated W on the upstream side and an outlet 11b for the material to be incinerated W on the downstream side. The axis of the furnace body 11 is inclined relative to the horizontal so that the inlet 11a is higher than the outlet 11b. The furnace body 11 is formed using a metal such as carbon steel. The material to be incinerated W may be, for example, sludge or garbage.

[0020] The furnace body 11 is housed in a cover casing 12. A plurality of water pipes 13 extending along the axial direction are provided in the furnace body 11 at predetermined intervals in the circumferential direction. Metallic fins 14 extending along the axial direction are provided between two water pipes 13 adjacent to each other along the circumferential direction. Therefore, the water pipes 13 and the fins 14 are arranged alternately along the circumferential direction. The fins 14 have a plurality of air holes 15 penetrating in the radial direction. These water pipes 13 and fins 14 form a fire grate.

[0021] The furnace body 11 is provided with rotation transmission members (not shown) on the inlet 11a side and outlet 11b side of the furnace body 11, and these rotation transmission members are configured to be rotatable around their axis by a driving device (not shown). The furnace body 11 rotates in response to this rotation.

[0022] A hopper 5 is provided on the inlet 11a side of the furnace body 11. The materials to be incinerated W fed into the hopper 5 are supplied to the furnace body 11 by a dust feeder 6. In this embodiment, the dust feeder 6 is of a pusher type.

[0023] A plurality of wind boxes 25 are provided below the furnace body 11, and communicate with the lower end of the cover casing 12. The primary gas (e.g., air) supplied to the wind boxes 25 is introduced into the furnace body 11 from the bottom of the furnace body 11 through the air holes 15. The amount of primary gas supplied to the furnace body 11 and the proportion of primary gas supplied to each combustion area in the furnace body 11 can be adjusted using the rotation speed of the forced draft fan and a flow control device (not shown) such as a damper. The amount of primary gas supplied and the proportion of primary gas supplied to each combustion area in the furnace body 11 can be changed depending on, for example, the composition, amount, and distribution of the material to be incinerated W in the furnace body 11.

[0024] The incineration material W is supplied into the furnace body 11 while the incinerator 10 is rotating at a low speed. The incineration material W supplied to the furnace body 11 is stirred by the rotation of the furnace body 11 and gradually moves downstream. Furthermore, while the incineration material W moves through the furnace body 11, primary gas is introduced into the furnace body 11 from the wind box 25. The amount of primary gas supplied is preferably set to an amount sufficient to maintain slow combustion of the incineration material W. During slow combustion, unburned gas is generated, and this unburned gas is introduced into the secondary combustion chamber 27 located downstream of the furnace body 11. A secondary gas, such as air, is supplied to the secondary combustion chamber 27 together with the unburned gas. This combusts the unburned gas. Furthermore, unburned components contained in the ash of the incineration material W discharged from the furnace body 11 are combusted in the post-combustion stoker 28.

[0025] Boiler 20 is provided above secondary combustion chamber 27 and connected to it, and generates steam by utilizing the heat of the exhaust gas discharged from the furnace, post-combustion chamber 29, and secondary combustion chamber 27. Water is supplied to boiler 20 from water supply device 8, and steam is generated by heat exchange with the exhaust gas.

[0026] The steam turbine generator 30 generates electricity by rotating a turbine using steam supplied from the boiler 20. The exhaust gas from which heat has been recovered in the boiler 20 is cooled and subjected to dust removal treatment by a dust removal device installed on the exhaust gas route between the boiler 20 and the chimney 40, and then discharged to the outside of the system via the chimney 40.

[0027] The water supply system 8 includes a condenser that cools and condenses the low-pressure wet steam discharged from the turbine outlet of the steam turbine generator 30, thereby returning it to saturated water and storing it, and a deaerator that degasses the saturated water that is returned from the condenser to the boiler 20 by a pump.

[0028] The operation of such an incinerator 10 is controlled by a control device 100. Figure 2 is a block diagram showing a schematic configuration of the control device 100. As shown in Figure 2, the control device 100 is configured with a process data acquisition unit 51, an operating state information acquisition unit 53, a memory unit 55, a predicted value calculation unit 57, a short-term control unit 59, and a long-term control unit 61, and each functional unit is constructed with hardware or software, or both, with a CPU as a core component, in order to perform processing related to the control of the incinerator 10.

[0029] The process data acquisition unit 51 acquires process data of the incinerator 10. The process data is data indicating the state of the power generation system 1 when the incinerator 10 combusts the materials to be incinerated W. Therefore, the process data does not only include data on the incinerator 10, but also includes data indicating the state of the boiler 20, for example. Such process data corresponds to, for example, the flow rate of forced air introduced into the furnace body 11, the pressure from the multiple wind boxes 25 provided below the incinerator 10, the flow rate of secondary air, the amount of gas generated when the materials to be incinerated W are burned, the flow rate of gas (exhaust gas) discharged from the chimney 40, the oxygen concentration in the secondary combustion chamber 27, the oxygen concentration at the dust collector outlet, the inlet temperature within the furnace body 11, the temperature of the post-combustion stoker 28, the exhaust gas temperature at the dust collector inlet, the temperature of gas generated when the materials to be incinerated W are burned, the amount of materials to be incinerated W fed into the hopper 5, the pressure of the boiler 20, the amount of steam from the boiler 20 (hereinafter also referred to as the "steam flow rate"), and the steam temperature. These are just examples, and data other than these can be used as process data.

[0030] The process of acquiring such process data of the incinerator 10 is referred to as a process data acquisition step in the control method for the incinerator 10, which operates the boiler 20 using heat from the incinerator 10 that combusts the materials W to be incinerated.

[0031] The operating status information acquisition unit 53 acquires operating status data of the incinerator 10. The operating status data is data indicating the status of devices related to the operation of the incinerator 10 (power generation system 1). Such data corresponds to, for example, device setting values ​​that set the operating status of the devices and measurement values ​​that indicate the operating status of the devices. These values ​​do not need to be absolute values, but may be, for example, deviations or relative values ​​relative to a predetermined value. Such operating status data corresponds to, for example, data related to the movement of the incinerated materials W within the incinerator 10, such as the pushing amount of the materials W pushed into the furnace body 11, the dust feeding cycle of the dust feeding device 6, the dust feeding stroke length of the dust feeding device 6, the rotation speed (grate speed) of the furnace body 11, the period setting value of the post-combustion stoker 28, the rotation speed of the forced draft fan that pushes air from the wind box 25 or the motor driving the fan, and the damper opening. It may also include the steam flow rate setting value of the boiler 20 and a soot blower signal indicating that the soot blower is operating in the boiler 20. These are just examples, and data other than these can be used as the operating condition data.

[0032] Such a process of acquiring the operating state data of the incinerator 10 is referred to as an operating state information acquisition step in the incinerator 10 control method.

[0033] The memory unit 55 sequentially stores the process data acquired by the process data acquisition unit 51 and the operating state data acquired by the operating state information acquisition unit 53. As will be described in detail later, the control device 100 uses the process data and operating state data in controlling the incinerator 10. For this reason, the process data and operating state data stored in the memory unit 55 are associated with a timestamp indicating the time at which they were acquired and are continuously stored.

[0034] FIG. 3 shows an example of process data stored in the storage unit 55. FIG. 3 shows process data for processes A, B, and C. These process data are stored in association with timestamps indicating the times at which they were acquired. Of course, it is also possible to configure the storage unit 55 to store process data for processes other than processes A, B, and C, and in fact, operating state data is also stored. When predicting the temperature of the secondary combustion chamber 27 (described later), such process data and operating state data can be configured to select relevant variable information from domain knowledge, select based on the accuracy of the prediction results, or select data determined to be important by machine learning (multiple regression analysis, feature analysis, etc.). In this way, the process data and operating state data are accumulated in the storage unit 55.

[0035] The process of sequentially storing such process data and operating state data in the memory unit 55 is referred to as a storage step in the control method for the incinerator 10.

[0036] The predicted value calculation unit 57 predicts the temperature of the secondary combustion chamber 27 (one example of a numerical value related to the operation of the incinerator). The predicted value calculation unit 57 calculates a short-term fluctuation predicted value (one example of a short-term fluctuation predicted value). The short-term fluctuation predicted value is a predicted value of the temperature of the secondary combustion chamber 27 after a predetermined first hour. The first hour may be one hour from the present, or one hour from a predetermined point in time. The predicted value calculation unit 57 calculates a short-term fluctuation predicted value over a predetermined period as such a predicted value. The short-term fluctuation predicted value over a predetermined period is a predicted fluctuation in the temperature of the secondary combustion chamber 27 from a certain point in time to the prediction point in time, and in this embodiment, it corresponds to a moving average value from the certain point in time to the prediction point in time. Therefore, the short-term fluctuation predicted value over a predetermined period corresponds to a short-term moving average value of the temperature of the secondary combustion chamber 27. The predicted value calculation unit 57 continuously calculates the short-term fluctuation predicted value consisting of such a moving average value. Therefore, the predicted value calculation unit 57 continuously calculates a short-term fluctuation predicted value consisting of a moving average value from a certain point in time to a prediction point in time as a predicted value of the temperature of the secondary combustion chamber 27 after a predetermined first hour.

[0037] The short-term fluctuation predicted value composed of such a moving average value is calculated based on the process data and the operating state data stored in the storage unit 55. When calculating the short-term fluctuation predicted value, the predicted value calculation unit 57 performs preprocessing using the process data and the operating state data stored in the storage unit 55. In this embodiment, as the preprocessing using the process data, calculation of a lag feature amount of the process data, calculation of a moving average value of the process data, and calculation of a lag feature amount for the moving average value of the process data are performed. The number of lags used in calculating the lag feature amount of the process data, the number of process data used in calculating the moving average value of the process data (times at which the process data was acquired), and the number of lags used in calculating the lag feature amount for the moving average value of the process data differ depending on the scale and configuration of the power generation system 1, and can be changed as appropriate.

[0038] The predicted value calculation unit 57 calculates the lag feature of the process data, the moving average value of the process data, and the lag feature for the moving average value of the process data as preprocessing for calculating the short-term fluctuation predicted value.

[0039] FIG. 4 shows the preprocessing performed before the calculation of the short-term fluctuation prediction value. As shown in FIG. 4, the process data of process A from "0:00:00" to "0:00:10" is assumed to be "A1" to "A11." In the example of FIG. 4, the lag features of the process data of process A are calculated for s seconds and 2 seconds before. Specifically, "A1_Lags" is calculated as the lag feature s seconds before the time point "0:00:00," and "A1_Lag2s" is calculated as the lag feature 2 seconds before the time point "0:00:00." Similarly, "A2_Lags" is calculated as the lag feature s seconds before the time point "0:00:01," and "A2_Lag2s" is calculated as the lag feature 2 seconds before the time point "0:00:01."

[0040] Furthermore, the moving average value is calculated for the process data of process A. Specifically, "A1_ma" is calculated as the moving average value at the time point "0:00:00", and "A2_ma" is calculated as the moving average value at the time point "0:00:01".

[0041] Next, the lag feature for the moving average value of the process data of Process A is calculated. Specifically, "A1_ma_Lags" is calculated as the lag feature for the moving average value of the process data of Process A s seconds before the time point "0:00:00", and "A1_ma_Lag2s" is calculated as the lag feature for the moving average value of the process data of Process A 2s seconds before the time point "0:00:00". Similarly, "A2_ma_Lags" is calculated as the lag feature for the moving average value of the process data of Process A s seconds before the time point "0:00:01", and "A2_ma_Lag2s" is calculated as the lag feature for the moving average value of the process data of Process A 2s seconds before the time point "0:00:01".

[0042] Similarly, the process data of process B from "0:00:00" to "0:00:10" is "B1" to "B11", and lag features s seconds and 2 seconds before are calculated as the process data of process B. Specifically, "B1_Lags" is calculated as the lag feature s seconds before the time point "0:00:00", and "B1_Lag2s" is calculated as the lag feature 2 seconds before the time point "0:00:00". Similarly, "B2_Lags" is calculated as the lag feature s seconds before the time point "0:00:01", and "B2_Lag2s" is calculated as the lag feature 2 seconds before the time point "0:00:01". Although not shown, the moving average value of the process data of process B and the lag feature for the moving average value of the process data of process B are also calculated.

[0043] Furthermore, as preprocessing using the operating state data, calculation of a lag feature of the operating state data, calculation of a moving average value of the operating state data, and calculation of a lag feature for the moving average value of the operating state data are performed. The number of lags used in calculating the lag feature of the operating state data, the number of operating state data used in calculating the moving average value of the operating state data (times at which the operating state data was acquired), and the number of lags used in calculating the lag feature for the moving average value of the operating state data differ depending on the scale and configuration of the power generation system 1, and can therefore be changed as appropriate.

[0044] Although not shown in the figure, the predicted value calculation unit 57 calculates the lag feature of the driving state data, the moving average value of the driving state data, and the lag feature for the moving average value of the driving state data as preprocessing for calculating the short-term fluctuation predicted value.

[0045] As shown in FIG. 5 , the prediction value calculation unit 57 continuously calculates a short-term fluctuation prediction value (see FIG. 7 ) consisting of a moving average value from a certain point in time to the prediction point in time as a prediction value (objective variable) of the temperature of the secondary combustion chamber 27 after a predetermined first hour by machine learning (regression model) using the lag feature of the process data, the moving average value of the process data, and the lag feature for the moving average value of the process data obtained by preprocessing, the process data, the lag feature of the operating state data, the moving average value of the operating state data, and the lag feature for the moving average value of the operating state data, and the operating state data as explanatory variables. Specifically, a prediction model of the temperature of the secondary combustion chamber 27 is first generated by machine learning based on a combination of a preprocessed data set and correct data of the temperature of the secondary combustion chamber 27 (in this embodiment, the correct data is the moving average value from a certain point in time discharged one hour after the time of the preprocessed data set to the prediction point in time). Various prediction models are available, and can be implemented using various known machine learning algorithms. The temperature of the secondary combustion chamber 27 is then predicted using the preprocessed data set at the current time or a certain point in time and the machine learning model. As input for machine learning, multiple moving average patterns may be used for one piece of data, or a weighted average value may be used instead of a moving average.

[0046] The window width and number of steps for calculating the moving average value, and the number of lags and number of steps for calculating the lag feature value may be determined so as to maximize accuracy while comparing with the prediction accuracy of the temperature of the secondary combustion chamber 27. In addition, the prediction accuracy can be evaluated using known prediction evaluation indices such as RMSE, MSE, MAE, coefficient of determination R2, and AIC.

[0047] For example, the predicted value calculation unit 57 can predict a 60-second moving average value of the temperature of the secondary combustion chamber 27 as a short-term fluctuation prediction for 300 seconds from now. Of course, these 60 seconds and 300 seconds are just examples and can be changed as appropriate.

[0048] The response variable may be the difference between the temperature of the secondary combustion chamber 27 after the first hour and the current temperature, the deviation of the temperature of the secondary combustion chamber 27 after the first hour, or the class (classification model) of the temperature of the secondary combustion chamber 27 after the first hour. The deviation of the temperature of the secondary combustion chamber 27 can be calculated by (actual measurement value - set value) / set value. The optimal value of the first hour varies depending on the scale of the power generation system 1. Therefore, it is advisable to use the maximum time for which prediction accuracy is guaranteed.

[0049] Here, for example, if the amount of combustion air is inappropriate for the amount of materials W to be incinerated in the incinerator 10, or if the amount of materials W to be incinerated in the incinerator 10 is insufficient or excessive, the temperature of the secondary combustion chamber 27 will deviate from the set value. In this case, if dust supply control is performed based on the short-term fluctuation prediction value, the temperature of the secondary combustion chamber 27 may overshoot the set value, particularly if the amount of combustion air is inappropriate for the amount of materials W to be incinerated in the incinerator 10. Therefore, the control device 100 uses the long-term fluctuation prediction value described below.

[0050] The predicted value calculation unit 57 calculates a long-term fluctuation predicted value (one example of a long-term fluctuation predicted value). The long-term fluctuation predicted value is a predicted value of the temperature of the secondary combustion chamber 27 after a predetermined second hour. The second hour may be two hours from the present time or two hours from a predetermined time point. The second hour may be the same length as the first hour, or may be a different length. The "predetermined time point" may be the same as or different from the "predetermined time" that starts at the first hour. The predicted value calculation unit 57 calculates a long-term fluctuation predicted value over a period longer than the predetermined period as such a predicted value. A long-term fluctuation predicted value over a period longer than the predetermined period is a predicted fluctuation in the temperature of the secondary combustion chamber 27 from a certain time point to a prediction time point, which is a period longer than the period used to calculate the short-term fluctuation predicted value. In this embodiment, this corresponds to a moving average value from a certain time point to a prediction time point. Therefore, the long-term fluctuation predicted value over a predetermined period corresponds to a long-term moving average value of the temperature of the secondary combustion chamber 27. Therefore, the long-term fluctuation predicted value is a moving average value over a longer period than the short-term fluctuation predicted value. The predicted value calculation unit 57 continuously calculates the long-term fluctuation predicted value consisting of such a moving average value. Therefore, the predicted value calculation unit 57 continuously calculates the long-term fluctuation predicted value consisting of a moving average value from a certain point in time to a prediction point in time, which is a longer period than the short-term fluctuation predicted value, as a predicted value of the temperature of the secondary combustion chamber 27 after the predetermined second hour.

[0051] The long-term fluctuation predicted value, which is made up of such moving average values, is calculated based on the process data and operating state data stored in the storage unit 55. When calculating the long-term fluctuation predicted value, the predicted value calculation unit 57 performs preprocessing using the process data and operating state data stored in the storage unit 55. The preprocessing using the process data and operating state data has been explained in the calculation of the short-term fluctuation predicted value described above, and therefore will not be explained here.

[0052] As shown in FIG. 6 , the prediction value calculation unit 57 continuously calculates a long-term fluctuation prediction value (see FIG. 7 ) consisting of a moving average value from a certain point in time to a prediction point in time as a predicted value (objective variable) of the temperature of the secondary combustion chamber 27 after a predetermined second hour by machine learning (regression model) using the lag feature value of the process data, the moving average value of the process data, and the lag feature value for the moving average value of the process data obtained by preprocessing, the process data, the lag feature value of the operating state data, the moving average value of the operating state data, and the lag feature value for the moving average value of the operating state data, and the operating state data as explanatory variables. Specifically, as in the case of the short-term fluctuation prediction value described above, a prediction model of the temperature of the secondary combustion chamber 27 is first generated by machine learning based on a combination of a preprocessed data set and correct data of the temperature of the secondary combustion chamber 27. Various prediction models are available, and can be implemented using various known machine learning algorithms. The temperature of the secondary combustion chamber 27 is then predicted using the preprocessed data set and the machine learning model. Note that multiple moving average patterns may be used for one data set, or a weighted average may be used instead of a moving average as an input for machine learning.

[0053] Also in this case, the window width and number of steps for calculating the moving average value, and the number of lags and number of steps for calculating the lag feature value may be determined so as to maximize accuracy while comparing with the prediction accuracy of the temperature of the secondary combustion chamber 27. Furthermore, the prediction accuracy can be evaluated using known prediction evaluation indices such as RMSE, MSE, MAE, coefficient of determination R2, and AIC.

[0054] For example, the predicted value calculation unit 57 can predict the 1800-second moving average value of the temperature of the secondary combustion chamber 27 as a long-term fluctuation prediction for 900 seconds from now. Of course, 1800 seconds and 900 seconds are just examples and can be changed as appropriate.

[0055] The response variable may be the difference between the temperature of the secondary combustion chamber 27 after the second hour and the current temperature, the deviation of the temperature of the secondary combustion chamber 27 after the second hour, or the class (classification model) of the temperature of the secondary combustion chamber 27 after the second hour. The deviation of the temperature of the secondary combustion chamber 27 can be calculated by (actual measurement value - set value) / set value. The optimal value of the second hour varies depending on the scale of the power generation system 1. Therefore, it is advisable to use the maximum time for which prediction accuracy is guaranteed.

[0056] The process of calculating a short-term fluctuation prediction value over a predetermined period based on the process data and operating state data stored in the memory unit 55 as a prediction value for the temperature of the secondary combustion chamber 27 after such a predetermined first hour, and calculating a long-term fluctuation prediction value over a period longer than the predetermined period based on the process data and operating state data stored in the memory unit 55 as a prediction value for the temperature of the secondary combustion chamber 27 after such a predetermined second hour, is referred to as the prediction value calculation step in the control method for the incinerator 10.

[0057] The process data and operating state data used by the prediction value calculation unit 57 to calculate the short-term fluctuation prediction value may be the same as the process data and operating state data used to calculate the long-term fluctuation prediction value. Alternatively, the prediction value calculation unit 57 may appropriately select and use optimal parameters when creating the machine learning models used to calculate the short-term fluctuation prediction value and the long-term fluctuation prediction value. When selectively using parameters, the process data and operating state data used to calculate the short-term fluctuation prediction value and the process data and operating state data used to calculate the long-term fluctuation prediction value may all be different from each other, or only some of them may be different from each other. Furthermore, even when selectively using parameters, the process data and operating state data used to calculate the short-term fluctuation prediction value and the process data and operating state data used to calculate the long-term fluctuation prediction value may be the same as each other.

[0058] Returning to FIG. 2 , the short-term control unit 59 controls the short-term control target quantity based on the short-term fluctuation prediction value. The short-term fluctuation prediction value is calculated and transmitted by the prediction value calculation unit 57 as described above. The short-term control target quantity is a control quantity performed using the short-term fluctuation prediction value to combust the materials to be incinerated W in the incinerator 10. In this embodiment, this corresponds to the amount of air supplied to the incinerator 10. The air supplied to the incinerator 10 refers to combustion air, and may be, for example, secondary air or forced air introduced into the incinerator. Therefore, the short-term control unit 59 controls the amount of combustion air according to the short-term fluctuation prediction value calculated by the prediction value calculation unit 57. Here, the total amount of combustion air may be the control target, or a specific amount of combustion air, such as the amount of secondary air or forced air, may be the control target. The amount of combustion air can be controlled by controlling the fan of the forced draft fan that pushes air from the wind box 25, the rotation speed of the motor driving the fan, the damper opening, etc.

[0059] In this embodiment, the short-term control unit 59 corrects the amount of combustion air (such as the flow rate of secondary air or forced air) calculated based on the actual measurement value of the temperature of the secondary combustion chamber 27 and the set value of the temperature of the secondary combustion chamber 27, based on the short-term fluctuation prediction value. That is, as shown in Fig. 8, the short-term control unit 59 calculates the control amount of the amount of combustion air by PID control based on the actual measurement value and the set value of the temperature of the secondary combustion chamber 27, and then corrects this control amount by PID control with the short-term fluctuation prediction value to control the amount of combustion air. Note that in practice, PID control is not performed based only on the temperature of the secondary combustion chamber 27, but the control amount of PID control is calculated based on, for example, the temperature of the furnace body 11 and the amount of material to be incinerated W.

[0060] The short-term control unit 59 may set a correction amount in response to deviation of the short-term control target variable from a predetermined range set based on a preset threshold value, and may perform correction based on this correction amount. The threshold value and correction amount may be set according to the scale and configuration of the power generation system 1.

[0061] In addition, the short-term control unit 59 can be configured to calculate the deviation between the temperature of the secondary combustion chamber 27 and the short-term control target quantity, set a correction amount according to this deviation, and perform correction based on this correction amount.

[0062] The process of controlling the short-term control target quantity based on such a short-term fluctuation predicted value is called a short-term control step in the control method for the incinerator 10.

[0063] The long-term control unit 61 controls the long-term control target quantity based on the long-term fluctuation prediction value. The long-term fluctuation prediction value is calculated and transmitted by the prediction value calculation unit 57 as described above. The long-term control target quantity is a control quantity performed to combust the materials to be incinerated W in the incinerator 10 using the long-term fluctuation prediction value, and in this embodiment, corresponds to the amount of materials to be incinerated W supplied to the incinerator 10. The materials to be incinerated W supplied to the incinerator 10 are pushed into the furnace body 11 by the dust feeder 6. Therefore, the long-term control unit 61 controls the amount of materials to be incinerated W pushed into the furnace body 11 by the dust feeder 6 in accordance with the long-term fluctuation prediction value calculated by the prediction value calculation unit 57. The amount of materials to be incinerated W can be controlled by controlling the dust feeding cycle of the dust feeder 6, the dust feeding stroke length of the dust feeder 6, etc.

[0064] In this embodiment, the long-term control unit 61 corrects the dust feeding cycle of the dust feeder 6, which is calculated based on the actual measured value of the temperature of the secondary combustion chamber 27 and the set value of the temperature of the secondary combustion chamber 27, based on the long-term fluctuation predicted value. That is, as shown in Fig. 8, the long-term control unit 61 calculates the control variable (operation cycle) of the dust feeder 6 by PID control based on the actual measured value and set value of the temperature of the secondary combustion chamber 27, and corrects this control variable by PID control with the long-term fluctuation predicted value to control the dust feeder 6. Note that in practice, PID control is not performed based only on the temperature of the secondary combustion chamber 27, but the control variable of PID control is calculated based on, for example, the temperature of the furnace body 11 and the amount of materials to be incinerated W.

[0065] The long-term control unit 61 may set a correction amount in response to deviation of the long-term control target variable from a predetermined range set based on a preset threshold value, and may perform correction based on this correction amount. The threshold value and correction amount may be set according to the scale and configuration of the power generation system 1.

[0066] In addition, the long-term control unit 61 can be configured to calculate the deviation between the temperature of the secondary combustion chamber 27 and the short-term control target quantity, set a correction amount according to this deviation, and perform correction based on this correction amount.

[0067] The process of controlling the long-term control target quantity based on such a long-term fluctuation predicted value is called the long-term control step in the control method for the incinerator 10.

[0068] Here, factors that cause the temperature of the secondary combustion chamber 27 to deviate from the set value include an inappropriate amount of combustion air relative to the amount of incineration material W in the furnace body 11, and an insufficient or excessive amount of incineration material W in the furnace body 11. For example, if the amount of incineration material W forced into the furnace body 11 by the dust feeder 6 is controlled based on a short-term fluctuation prediction value, the temperature of the secondary combustion chamber 27 may overshoot the set value if the cause of the temperature of the secondary combustion chamber 27 deviating from the set value is an inappropriate amount of combustion air relative to the amount of incineration material W in the furnace body 11. Therefore, as in this embodiment, by controlling the amount of incineration material W forced into the furnace body 11 by the dust feeder 6 based on a long-term fluctuation prediction value, it is possible to prevent overshooting.

[0069] Other Embodiments In the above embodiment, the incinerator 10 was described as being a rotary stoker type incinerator, but the incinerator 10 is not limited to being a rotary stoker type incinerator, and may be an incinerator other than a rotary stoker type incinerator.

[0070] In the above embodiment, the configuration in which the prediction value calculation unit 57 predicts the temperature of the secondary combustion chamber 27 has been described as an example. However, in the present invention, the numerical value predicted by the prediction value calculation unit is not limited as long as it is a numerical value related to the operation of the incinerator that is not the amount of steam discharged from the boiler. The numerical value related to the operation of the incinerator may be a temperature measured in a predetermined portion of the incinerator. Here, the predetermined portion may be, for example, the secondary combustion chamber of the incinerator, the post-combustion chamber of the incinerator, the outlet of the boiler, etc. Therefore, the numerical value related to the operation of the incinerator may be at least one numerical value selected from the group consisting of the temperature of the secondary combustion chamber of the incinerator, the temperature of the post-combustion chamber of the incinerator, and the temperature of steam discharged from the boiler. Furthermore, the numerical value related to the operation of the incinerator may be the concentration of a predetermined gas measured in a predetermined portion of the incinerator. The predetermined portion may be, for example, the inside of the incinerator. The predetermined gas may be, for example, oxygen. Therefore, the numerical value related to the operation of the incinerator may be the concentration of oxygen in the incinerator. In this case, the location in the incinerator where the oxygen concentration is an issue may be, for example, the secondary combustion chamber. As is clear from the above examples, the numerical values ​​related to the operation of the incinerator may function as indicators for evaluating the soundness and efficiency of the operation of the incinerator.

[0071] In the above embodiment, the short-term control unit 59 controls the amount of combustion air as a control variable. However, the present invention does not limit the short-term control quantity. However, the short-term control quantity is preferably at least one selected from the group consisting of the flow rate of secondary air introduced into the incinerator and the flow rate of forced air introduced into the incinerator, which are examples of the amount of combustion air, the rotation speed of the incinerator body, and the amount of incinerated material supplied to the incinerator. If the short-term control quantity includes the amount of incinerated material supplied to the incinerator, the amount of incinerated material can be controlled by controlling the operation of the dust feeder, for example. For example, if the prediction value calculation unit predicts the temperature of the secondary combustion chamber of the incinerator, and if the short-term fluctuation prediction value of the temperature of the secondary combustion chamber of the incinerator falls below a predetermined lower threshold, the temperature of the secondary combustion chamber of the incinerator can be increased by controlling the dust feeder to forcibly feed dust regardless of the dust feeding cycle. Conversely, if the predicted short-term fluctuation value of the temperature in the secondary combustion chamber of the incinerator exceeds a predetermined upper threshold, the dust feeding device can be controlled to forcibly stop dust feeding regardless of the dust feeding cycle, thereby reducing the temperature in the secondary combustion chamber of the incinerator.

[0072] In the above embodiment, an example has been described in which the long-term control unit 61 controls the dust feeding cycle of the dust feeding device 6 as the control variable, but the long-term control variable is not limited to this invention. However, it is preferable that the long-term control variable is at least one selected from the group consisting of the dust feeding cycle to the incinerator and the flow rate of forced air introduced into the incinerator.

[0073] In the above embodiment, the short-term control unit 59 and the long-term control unit 61 are described as being configured as separate entities, but it is also possible to configure, for example, a single control unit that has the short-term control unit 59 and the long-term control unit 61.

[0074] In the above embodiment, the short-term control unit 59 was described as controlling the amount of combustion air by correcting the control amount by PID control based on the actual measured value and set value of the temperature of the secondary combustion chamber 27 with a short-term fluctuation prediction value, but it is also possible to calculate a correction amount by PID control based on the set value and short-term fluctuation prediction value of the temperature of the secondary combustion chamber 27, and correct the control amount by PID control based on the actual measured value and set value of the temperature of the secondary combustion chamber 27 with this correction amount to control the forced air.

[0075] In the above embodiment, the long-term control unit 61 was described as controlling the dust supply device 6 by correcting the control amount by PID control based on the actual measured value and set value of the temperature of the secondary combustion chamber 27 with a long-term fluctuation predicted value, but it is also possible to configure the dust supply device 6 to control by calculating a correction amount by PID control based on the set value and long-term fluctuation predicted value of the temperature of the secondary combustion chamber 27, and correcting the control amount by PID control based on the actual measured value and set value of the temperature of the secondary combustion chamber 27 with this correction amount.

[0076] In the above embodiment, the long-term control unit 61 and the short-term control unit 59 are described as controlling the short-term control target quantity and the long-term control target quantity so that the fluctuation amount of the temperature of the secondary combustion chamber 27 is equal to or less than a preset fluctuation amount. In this case, however, the "preset fluctuation amount" used in the short-term control target quantity and the "preset fluctuation amount" used in the long-term control target quantity may be the same value or different values. [Industrial Applicability]

[0077] The present invention can be used in an incinerator control device and an incinerator control method that operate a boiler using heat from an incinerator that burns materials to be incinerated. [Explanation of symbols]

[0078] 10: Incinerator 20: Boiler 51: Process data acquisition unit 53: Operation status information acquisition unit 55: Storage section 57: Prediction value calculation unit 59: Short-term control section 61: Long-term control section 100: Control device (incinerator control device) W: Incinerated material

Claims

1. A control device for an incinerator that operates a boiler using heat from an incinerator that burns materials to be incinerated, a process data acquisition unit that acquires process data of the incinerator; an operating status information acquisition unit that acquires operating status data of the incinerator; a storage unit that sequentially stores the process data and the operating state data; a prediction value calculation unit that calculates a short-term fluctuation prediction value over a predetermined period based on the process data and the operating state data stored in the storage unit as a prediction value of a predetermined numerical value after a predetermined first hour, the predetermined numerical value being other than the amount of steam discharged from the boiler, and that calculates a long-term fluctuation prediction value over a period longer than the predetermined period based on the process data and the operating state data stored in the storage unit as a prediction value of the predetermined numerical value after a predetermined second hour; a short-term control unit that controls a short-term control target quantity based on the short-term fluctuation prediction value; A control device for an incinerator comprising: a long-term control unit that controls a long-term control target quantity based on the long-term fluctuation predicted value.

2. 2. The incinerator control device according to claim 1, wherein the predetermined value is a temperature measured at a predetermined portion of the incinerator.

3. 3. The incinerator control device according to claim 2, wherein the predetermined value is at least one value selected from the group consisting of the temperature of the secondary combustion chamber of the incinerator, the temperature of the post-combustion chamber of the incinerator, and the temperature of steam discharged from the boiler.

4. 2. The incinerator control device according to claim 1, wherein said predetermined value is the concentration of a predetermined gas measured in a predetermined portion of said incinerator.

5. 5. The incinerator control device according to claim 4, wherein the predetermined value is the concentration of oxygen inside the incinerator.

6. A method for controlling an incinerator that operates a boiler using heat from an incinerator that burns materials to be incinerated, comprising: a process data acquisition step of acquiring process data of the incinerator; an operating status information acquisition step of acquiring operating status data of the incinerator; a storage step of sequentially storing the process data and the operating state data in a storage unit; a predicted value calculation step of calculating a short-term fluctuation predicted value over a predetermined period based on the process data and the operating state data stored in the storage unit as a predicted value of a predetermined numerical value after a predetermined first hour, the predetermined numerical value being other than the amount of steam discharged from the boiler, and calculating a long-term fluctuation predicted value over a period longer than the predetermined period based on the process data and the operating state data stored in the storage unit as a predicted value of the predetermined numerical value after a predetermined second hour; a short-term control step of controlling a short-term control target quantity based on the short-term fluctuation predicted value; A long-term control step of controlling a long-term control target quantity based on the long-term fluctuation predicted value.

Images