JP2021008991A - Automatic combustion control method and automatic combustion control device - Google Patents

Abstract

To reduce frequency of an intervening operation manually performed by an operator in an incinerator.SOLUTION: An automatic combustion control method for controlling combustion of waste in accordance with operation amount reference values set at a plurality of operation ends of an incinerator includes: a parameter acquisition step of acquiring sensor information on the inner state of the incinerator and control values at the operation ends obtained by an automatic combustion control device as input parameters; and a control step of correcting and controlling the operation amount reference values at the operation ends in accordance with correction values relative to types of the operation ends and the operation amount reference values, which are obtained by inputting the input parameters to an operation learned model. The operation learned model is a learned model generated by performing at least mechanical learning and a statistical method, by using learning sensor information on the inner state of the incinerator and learning control values at the operation ends to be controlled as learning input parameters and using the correction values relative to the types of the operation ends and the operation amount reference values, which correspond to the learning sensor information and learning control values as learning output parameters.SELECTED DRAWING: Figure 1

Description

The present invention relates to an automatic combustion control method and an automatic combustion control device for a waste incinerator.

Conventionally, various demands have been made in the field of waste treatment in order to realize a low-carbon society and a recycling-oriented society. Waste incinerators are required not only to efficiently recover heat from combustion exhaust gas, but also to suppress harmful substances such as dioxins and nitrogen oxides (NO _x ) released into the atmosphere. Therefore, in the waste incinerator, automatic combustion control of waste is adopted in order to stabilize the combustion of waste as waste (see Patent Documents 1 and 2).

In the automatic combustion control executed in the waste incinerator, the set value of the operation item is calculated by the automatic combustion control device that controls the combustion, and the control is performed by the monitoring control device based on the calculated set value. .. In a waste incinerator, the calories used for combustion fluctuate and the combustion state changes depending on the calories of the received waste and the stirring condition in the waste pit where the waste is received and agitated. Therefore, an operator or the like uses a monitoring control device in the central control room to constantly monitor data obtained from various sensors, alarms associated with the data, and images of the combustion state. On top of that, in order to perform operation management that stabilizes the combustion in the waste incinerator and maintains the operation control value at a predetermined value, the operator manually sets the setting values of the main operation items from the automatic combustion control device. A so-called driving intervention operation (hereinafter referred to as an intervention operation) is performed. In operation management, the factory operator performs manual intervention operations to correct the control values for safe and stable operation for combustion control by various control values several to several tens of times a day. ing. In addition to the control data and the current operation data, the operator also considers the combustion state to determine the manual intervention operation.

JP-A-2017-180971 JP-A-2015-114040

In the above-mentioned conventional waste incinerators, a manual intervention operation by an experienced operator is generally performed. In addition, it is necessary to secure the number of operators so that intervention operations can be performed at all times during operation. However, in recent years, waste incinerators have a shortage of experienced operators, and there is an increasing need to manage and operate them with a small number of people. From these facts, in the future, it may be difficult to properly perform the manual intervention operation even if the intervention operation is required. Even in this case, in the waste incinerator, it was necessary to strictly stabilize the combustion and maintain the operation control value. Therefore, it has been required to reduce the frequency of manual operation intervention.

The present invention has been made in view of the above, and an object of the present invention is to provide an automatic combustion control method and an automatic combustion control device capable of significantly reducing the frequency of manual intervention operations by an operator in a waste incinerator. To provide.

In order to solve the above-mentioned problems and achieve the object, the automatic combustion control method according to one aspect of the present invention sets the operation amount reference value of a plurality of operation ends of the waste incinerator to a preset set value. It is an automatic combustion control method executed by an automatic combustion control device that is set based on the above and controls the combustion of waste inside the waste incinerator based on the operation amount reference value, and is the inside of the waste incinerator. A parameter acquisition step for acquiring sensor information regarding the state of the waste incinerator and facilities related to the waste incinerator, and a control value of the operation end controlled by the automatic combustion control device as input parameters, and an operation-learned model for the input parameters. The operation learning includes a control step for correcting and controlling the operation amount reference value of the operation end, which is obtained based on the type of the operation end and the correction value of the operation amount reference value. The completed model is used for learning the internal state of the waste incinerator, the learning sensor information about the facility related to the waste incinerator, and the learning control value of the operation end controlled by the automatic combustion control device. As input parameters, the type of operation end and the correction value for the operation amount reference value corresponding to the learning sensor information and the learning control value are used as learning output parameters in advance by at least one of machine learning and statistical methods. It is characterized by being a generated trained model.

In the automatic combustion control method according to one aspect of the present invention, in the above invention, the learning sensor information and the learning control value of the learning input parameter are data used for determining a manual intervention operation by the operator. Yes, the correction value for the operation end type and the operation amount reference value included in the learning output parameter is a correction value for the operation end type and the operation amount reference value operated by the operator in the manual intervention operation. The operation-learned model is a trained model generated by cluster analysis using the learning input parameters and the learning output parameters. In the automatic combustion control method according to one aspect of the present invention, in this configuration, the data used for determining the manual intervention operation of the operator is the control value, operation value, or operation value of the operation end in the waste incinerator. It is characterized by including instantaneous value data, data derived based on a plurality of the sensor information, data on a fluctuation tendency of the sensor information over a predetermined time, and data on a one-hour moving average of control values of the operation end. To do.

In the above invention, the automatic combustion control method according to one aspect of the present invention uses the learning combustion image data as a learning input parameter and the numerical value corresponding to the learning combustion image data as a learning output parameter. It is characterized in that the combustion image trained model generated in advance by the machine learning includes the numerical value obtained by inputting the combustion image data obtained by imaging the combustion state in the waste incinerator.

The automatic combustion control method according to one aspect of the present invention is characterized in that, in the above invention, the learning output parameter includes data for determining the start of the control step and data for determining the end of the control step. And.

In the automatic combustion control device according to one aspect of the present invention, the operation amount reference value of a plurality of operation ends of a waste incinerator is set based on a preset set value, and the disposal is performed based on the operation amount reference value. An automatic combustion control device that controls the combustion of waste inside a product incinerator, and provides sensor information on the internal state of the waste incinerator, facilities related to the waste incinerator, and the automatic combustion control device. A storage unit that stores the control value of the operation end controlled by the above as an input parameter, and the input parameter read from the storage unit are input to the operation-learned model, and the type of the operation end and the operation amount reference value are corrected. The operation-learned model includes a control unit that corrects and controls the operation amount reference value of the operation end obtained based on the value, and the operation-learned model includes the internal state of the waste incinerator and the waste. The learning sensor information about the facility related to the incinerator and the learning control value of the operation end controlled by the automatic combustion control device are used as learning input parameters, and correspond to the learning sensor information and the learning control value. It is characterized in that it is a trained model generated in advance by at least one of machine learning and statistical methods, with the type of operation end and the correction value for the operation amount reference value as output parameters for training.

According to the automatic combustion control method and the automatic combustion control device according to the present invention, it is possible to significantly reduce the frequency of manual intervention operations by the operator in the waste incinerator.

FIG. 1 is an overall configuration diagram schematically showing an incineration facility to which an automatic combustion control device according to an embodiment of the present invention is applied. FIG. 2 is a block diagram showing a configuration of an automatic combustion control device according to an embodiment of the present invention. FIG. 3 is a flowchart for explaining an automatic combustion control method according to an embodiment of the present invention. FIG. 4 is a flowchart for explaining an intervention operation in the automatic combustion control method according to the embodiment of the present invention. FIG. 5 is a graph showing the conventional manual intervention frequency (a) and the manual intervention frequency (b) of the automatic combustion control method according to the embodiment of the present invention.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In all the drawings of the following embodiment, the same or corresponding parts are designated by the same reference numerals. Further, the present invention is not limited to one embodiment described below.

FIG. 1 shows a grate-type waste incinerator to which the automatic combustion control device according to the embodiment of the present invention is applied. As shown in FIG. 1, a waste incinerator, which is a waste incinerator, includes a furnace 1 for burning waste, a waste input port 2 for charging waste, and a boiler 9. The boiler 9 includes a heat exchanger 9a and a steam drum 9b installed on the downstream side of the furnace outlet 7 of the furnace 1.

The garbage thrown in from the garbage inlet 2 is conveyed to the grate 4 by the garbage supply device 3. The reciprocating motion of the grate 4 causes the dust to be agitated and moved. The dust on the grate 4 is burned while being dried by blowing the combustion air supplied by the combustion air blower 6 into the air box below the grate 4, and exhaust gas and ash are generated. The generated ash falls through the ash drop port 5 and is discharged to the outside of the furnace 1.

The total amount of combustion air supplied from under the grate 4 to the inside of the furnace 1 is adjusted by the combustion air damper 14 provided in the immediate vicinity of the combustion air blower 6. The flow rate of the combustion air supplied to each air box is adjusted by the sub-grate combustion air dampers 14a, 14b, 14c, 14d provided in the piping for supplying the combustion air to each air box. To. That is, the ratio of the flow rates of the combustion air supplied to each of the air boxes is adjusted by the combustion air dampers 14a to 14d under the grate. In FIG. 1, the lower part of the grate 4 is divided into four air boxes along the direction of transporting the waste, and the combustion air is supplied through each air box. However, the under-grate combustion air damper The number of 14a to 14d and the number of air boxes is not necessarily limited to four, and can be appropriately changed according to the scale and purpose of the waste incinerator.

Cooling air is blown into the furnace 1 by the cooling air blower 11 from the cooling air blowing port 10 provided on the furnace wall. When the cooling air is blown into the furnace 1, the unburned components in the combustion gas are further burned, and the temperature of the furnace wall is suppressed from rising excessively. The flow rate of the cooling air supplied from the cooling air inlet 10 into the furnace 1 is adjusted by the cooling air damper 15 provided in the immediate vicinity of the cooling air blower 11.

The flammable gas generated in the upstream waste drying process and the main combustion process and the combustion exhaust gas generated in the downstream post-combustion process along the waste transport direction in the grate 4 are the furnace outlet 7 of the furnace 1. It merges at the gas mixing section provided on the side. The combustible gas and the combustion exhaust gas merged in the gas mixing section are agitated and mixed again, and then secondary combustion is performed by supplying air for secondary combustion. The boiler 9 is installed on the downstream side along the waste transport direction with respect to the portion where the secondary combustion is performed (hereinafter, the secondary combustion portion). The combustion gas subjected to the secondary combustion is exhausted to the outside from the chimney 8 after the heat energy is recovered by the heat exchanger 9a of the boiler 9.

In the furnace 1, an intermediate ceiling 16 is provided at an upper position along the height direction of the furnace 1. The gas flowing in the furnace 1 is divided into a gas containing a large amount of flammable gas generated in the waste drying process and the main combustion process on the upstream side and a combustion exhaust gas generated in the post-combustion process on the downstream side by the intermediate ceiling 16. Can be divided and discharged. Specifically, the combustion exhaust gas flows through the flue (main flue) below the intermediate ceiling 16, while the gas containing a large amount of flammable gas flows through the flue (secondary flue) above the intermediate ceiling 16. .. The merging of the combustion exhaust gas and the gas containing a large amount of combustible gas in the gas mixing section further promotes the stirring and mixing of the gas in the gas mixing section. As a result, the combustion in the secondary combustion section becomes more stable, the generation of dioxins in the combustion process can be suppressed, and the generation of unburned waste can be suppressed. The intermediate ceiling 16 may not be provided in the furnace 1.

Thermometers as sensors for measuring the gas temperature in the furnace 1 are provided at a plurality of positions in the furnace 1. Specifically, a combustion chamber gas thermometer 17 is provided at an intermediate position between the grate 4 and the cooling air inlet 10 along the height direction of the furnace 1. A main flue gas thermometer 18 is provided at a position below the furnace outlet 7 along the height direction of the furnace 1. A furnace outlet lower gas thermometer 19 is provided at a lower position of the furnace outlet 7 along the height direction of the furnace 1. A furnace outlet central gas thermometer 20 is provided at a central position of the furnace outlet 7 along the height direction of the furnace 1. A furnace outlet gas thermometer 21 for measuring the combustion control temperature is provided at a position downstream of the furnace outlet 7 along the height direction of the furnace 1. The measured values of the temperature measured by the combustion chamber gas thermometer 17, the main flue gas thermometer 18, the furnace outlet lower gas thermometer 19, the furnace outlet middle gas thermometer 20, and the furnace outlet gas thermometer 21 are the combustion process. As a measured value, it is stored in the storage unit 34 (see FIG. 2) of the combustion control device 30. The measured temperature value stored in the storage unit 34 is transmitted from the combustion control device 30 to the monitoring center 40 and stored in the storage unit 44.

The boiler 9 is provided with a boiler outlet oxygen concentration meter 22 for measuring the concentration of oxygen (O ₂ ) in the exhaust gas on the outlet side. At the inlet of the chimney 8, a gas concentration meter 23 for measuring the concentrations of carbon monoxide (CO) and nitrogen oxides (NO _x ) in the exhaust gas is provided. An exhaust gas flow meter 24 for measuring the amount of exhaust gas is provided in the pipe connecting the outlet of the boiler 9 and the chimney 8. The measured values of the gas concentration and the flow rate measured by the boiler outlet oxygen concentration meter 22, the gas concentration meter 23, and the exhaust gas flow meter 24 are stored in the storage unit 34 of the combustion control device 30 as the combustion process measurement values. The measured values of the measured gas concentration and flow rate are transmitted from the combustion control device 30 to the monitoring center 40 and stored in the storage unit 44.

A combustion image capturing unit 25 is provided on the downstream side in the waste transport direction inside the furnace 1. The combustion image imaging unit 25 images the combustion state of the dust on the grate 4, and stores the captured combustion image data in the storage unit 34 of the combustion control device 30. The combustion image data stored in the storage unit 34 is transmitted from the combustion control device 30 to the monitoring center 40 at predetermined time intervals.

The combustion control device 30 includes, for example, a private line, a public communication network such as the Internet, for example, a LAN (Local Area Network), a WAN (Wide Area Network), a telephone communication network such as a mobile phone, a public line, or a VPN (Virtual Private Network). ) And the like (not shown), the monitoring center 40 is connected to the monitoring center 40 via a network (not shown).

FIG. 2 is a block diagram showing the configurations of the combustion control device 30 and the monitoring center 40. As shown in FIG. 2, the combustion control device 30 as an automatic combustion control device includes a waste quality calculation unit 31, an operation amount reference value adjustment unit 32, an operation amount reference value correction unit 33, a storage unit 34, and an operation amount adjustment unit. 35 is provided. The waste quality calculation unit 31, the operation amount reference value adjustment unit 32, the operation amount reference value correction unit 33, and the operation amount adjustment unit 35 specifically include a CPU (Central Processing Unit), a DSP (Digital Signal Processor), and an FPGA ( It is equipped with a processor such as a Field-Programmable Gate Array and a main storage unit such as a RAM (Random Access Memory) or a ROM (Read Only Memory) (neither is shown). The storage unit 34 is composed of a storage medium selected from volatile memory such as RAM, non-volatile memory such as ROM, EPROM (Erasable Programmable ROM), hard disk drive (HDD, Hard Disk Drive), and removable media. .. The removable media is, for example, a USB (Universal Serial Bus) memory or a disc recording medium such as a CD (Compact Disc), a DVD (Digital Versatile Disc), or a BD (Blu-ray (registered trademark) Disc). is there. Further, the storage unit 34 may be configured by using a computer-readable recording medium such as a memory card that can be mounted from the outside.

The storage unit 34 can store an operating system (OS), various programs, various tables, various databases, and the like for executing the operation of the combustion control device 30. Here, the various programs also include an automatic combustion control program that realizes processing based on the trained model according to the present embodiment. These various programs can also be recorded on a computer-readable recording medium such as a hard disk, flash memory, CD-ROM, DVD-ROM, or flexible disk and widely distributed.

The combustion control device 30 has, as the operation amount of each operation end, the combustion air amount, the cooling air amount, the waste supply device feed speed, and the operation amount based on the conventionally known predetermined operation amount reference value setting relational expression. Control the grate feed rate. The combustion control device 30 also controls the stop and operation of the waste supply device feed rate and the grate feed speed. The operation amount reference value setting relational expression is a relational expression between the waste incineration amount setting value or the waste quality setting value and the operation amount reference value (the target value of the operation amount), and includes a control parameter as a correction coefficient. The control parameter is adjusted by the operation amount reference value adjusting unit 32 so as to match the waste incineration amount set value and the waste quality set value. The adjusted control parameter is set by the operation amount reference value adjusting unit 32 in response to the changed setting value when at least one of the waste incineration amount setting value and the waste quality setting value is changed. Be changed. By changing the control parameter, the preset operation amount reference value is corrected.

The waste quality calculation unit 31 calculates the waste quality (lower calorific value of waste) according to the set value of the incineration amount of waste. The operation amount reference value adjusting unit 32 adjusts the operation amount reference value by adjusting the control parameters included in the operation amount reference value setting relational expression. The operation amount reference value correction unit 33 corrects the operation amount reference value adjusted by the operation amount reference value adjustment unit 32 based on a predetermined control algorithm (PID control, fuzzy calculation, etc.). The storage unit 34 stores the data referred to by the waste quality calculation unit 31, the operation amount reference value adjusting unit 32, and the operation amount reference value correction unit 33. The storage unit 34 stores the preset incineration amount set value and the combustion process measured value acquired as the combustion state amount in the furnace, in addition to the predetermined operation amount reference value setting relational expression and control algorithm. There is.

The operation amount adjusting unit 35 adjusts each operation amount of each operation end so as to follow the operation amount reference value. Specifically, the operation amount adjusting unit 35 includes a combustion air amount adjusting unit 36, an air amount ratio adjusting unit 36A, a cooling air amount adjusting unit 37, a waste supply device feed speed adjusting unit 38, and a grate feed speed adjusting unit 39. Has. The combustion air amount adjusting unit 36 adjusts the operation amount so that the combustion air amount follows the operation amount reference value (hereinafter, corrected operation amount reference value) corrected by the operation amount reference value correction unit 33. The air amount ratio adjusting unit 36A controls each of the subgrate combustion air dampers 14a to 14d to adjust the mutual ratio of the flow rates in the respective air boxes. The cooling air amount adjusting unit 37 adjusts the operation amount so that the cooling air amount follows the correction operation amount reference value. Here, the amount of combustion air and the amount of cooling air are adjusted by controlling the opening degrees of the combustion air damper 14, the subgrate combustion air dampers 14a to 14d, and the cooling air damper 15. .. The waste supply device feed rate adjusting unit 38 adjusts the operation amount so that the waste supply device feed rate follows the correction operation amount reference value. The grate feed speed adjusting unit 39 adjusts the operation amount so that the grate feed speed follows the correction operation amount reference value. When the operation amount reference value is not corrected by the operation amount reference value correction unit 33, the operation amount adjustment unit 35 adjusts each operation amount based on the uncorrected operation amount reference value.

The monitoring center 40 includes a control unit 41, an output unit 42, an input unit 43, and a storage unit 44. Specifically, the control unit 41 includes a processor such as a CPU, DSP, and FPGA, and a main storage unit (none of which is shown) such as RAM and ROM. The output unit 42 as an output means displays a combustion image in the furnace 1 on the display monitor, displays characters and figures on the screen of the touch panel display, and outputs sound from the speaker according to the control by the control unit 41. It is configured so that predetermined information can be notified to the outside by outputting it. The input unit 43 as an input means uses a user interface such as a keyboard, input buttons, levers, a touch panel for manual input provided superimposed on a display such as a liquid crystal display, or a microphone for voice recognition. It is composed of. By operating the input unit 43, a user or the like can input predetermined information to the control unit 41. The output unit 42 and the input unit 43 may be integrated into an input / output unit, and the input / output unit may be composed of a touch panel display, a speaker microphone, or the like.

The storage unit 44 is composed of a storage medium selected from volatile memory such as RAM, non-volatile memory such as ROM, EPROM, HDD, and removable media. The removable medium is, for example, a USB memory or a disc recording medium such as a CD, DVD, or BD. Further, the storage unit 44 may be configured by using a computer-readable recording medium such as a memory card that can be mounted from the outside. The storage unit 44 can store an OS, various programs, various tables, various databases, etc. for executing the operation of the monitoring center 40. Here, the various programs also include an automatic combustion control program that realizes control using the operation-learned model according to the present embodiment, and an automatic judgment processing program that realizes judgment processing using the combustion image-learned model. Specifically, the storage unit 44 stores the operation-learned model 44a and the combustion image-learned model 44b. The storage unit 44 may be provided in another server capable of communicating via various networks, or may be provided in the combustion control device 30. Further, these various programs can be recorded on a computer-readable recording medium such as a hard disk, a flash memory, a CD-ROM, a DVD-ROM, or a flexible disk and widely distributed.

The control unit 41 loads the program stored in the storage unit 44 into the work area of the main storage unit and executes it, and controls each component or the like through the execution of the program to realize a function that meets a predetermined purpose. it can. In the present embodiment, the function of the learning unit 41a is executed by the execution of the program by the control unit 41, and the processing of the operation-learned model 44a and the combustion image-learned model 44b, which are programs, is executed.

Here, a method of generating the operation-learned model 44a, which is a program stored in the storage unit 44, will be described. That is, the operation-learned model 44a returns to the automatic combustion control from the judgment condition for performing the intervention operation, the operation amount at the time of the intervention operation, and the intervention operation in the intervention operation control obtained from the experience of the operator in the operation management work of the incinerator. It is generated by deriving the return condition etc. at the time of performing by cluster analysis. For example, the points at which the speed of the grate 4 is increased or decreased by the operation of the operator can be extracted by cluster analysis from the relationship between the combustion control temperature and the combustion chamber temperature, or the combustion control temperature and the NO _x concentration can be defined. Extract from relationships by cluster analysis.

The data used to generate the operation-learned model 44a includes not only the speed of the grate 4, but also the process value obtained from the sensor related to combustion control as learning sensor information, and the operation management calculated from the process value. Includes values used in the control of and combustion quantification data. The combustion quantification data is combustion quantification data in which the combustion state obtained by the combustion image learned model 44b is quantified from the combustion image periodically obtained by the combustion image imaging unit 25. The data used to generate the operation-learned model 44a further includes a control value by the combustion control device 30 as a learning control value. These learning sensor information and learning control values serve as learning input parameters when the operation-learned model 44a is generated.

The learning output parameter when the operation-learned model 44a is generated is a correction amount obtained from the control value operated by the operator. Further, the learning output parameters are determined by the start timing of switching the control mode between the automatic combustion control by the combustion control device 30 and the control by the monitoring center 40 when the operator performs the intervention operation, and by the monitoring center 40 after the control mode is switched. It includes parameters for the duration of control and the end timing of control by the monitoring center 40.

Further, the learning unit 41a of the control unit 41 of the monitoring center derives the correlation of data when the operator performs the intervention operation from the above-mentioned learning input parameters and learning output parameters by machine learning such as cluster analysis. As a result, the driving operation pattern in which the operator performs the intervention operation is generated as the operation-learned model 44a. The generated operation-learned model 44a is stored in the storage unit 44. The control unit 41 executes control of various intervention operations according to the driving operation pattern by the operation-learned model 44a generated by the learning unit 41a. It should be noted that the operation-learned model 44a may be generated by deriving the correlation of the data when the operator performs the intervention operation by performing machine learning such as cluster analysis on another computer or the like. In this case, the generated operation-learned model 44a is supplied to the learning unit 41a from the computer that has performed machine learning, and is stored in the storage unit 44. In addition to machine learning, the correlation may be derived by a statistical method such as average, quartile, or correlation analysis to generate an operation-learned model 44a. That is, the points of judgment of the start and end of intervention operations performed by multiple operators by analyzing multiple conditions related to complex intervention operations in which judgment is different for each operator by at least one of machine learning and statistical methods, and It is possible to extract the correction amount of each operation end set by the operator. The trained model is also called a model.

Further, the combustion image learned model 44b described above uses a plurality of learning combustion image data captured by the combustion image imaging unit 25 as learning input parameters. In addition, the combustion image trained model 44b quantifies the strength and weakness of the combustion state as learning output parameters, the data quantifying the position of the combustion cutoff point, and the degree of generation of unburned matter. Use the data. The combustion image trained model 44b is generated by machine learning such as deep learning, using the above-mentioned learning input parameters and learning output parameters as training data. When the combustion image is input, the combustion image trained model 44b digitizes the combustion state based on the input combustion image. Specifically, the combustion image trained model 44b classifies the combustion state into a plurality of types, for example, eight types, determines the confidence level for each of the eight types, and quantifies the combustion state based on each type. I do. In addition, it may be classified into 7 types or less and more than 8 types.

Table 1 shows an example of the measurement target and the operation target used in the operation-learned model 44a described above and the parameters used.

In Table 1, the instantaneous value is a measured value at a predetermined time point, and the 1-hour average value is the value of the 1-hour moving average from the instantaneous value to 1 hour before. The difference from the target value is the difference from the instantaneous value with respect to the preset target value. A slope of a few minutes to a few tens of minutes means a slope of any of a few minutes to a few tens of minutes of an instantaneous value. The mean value of several minutes to several tens of minutes means the average value of the instantaneous value for several minutes to the average value for several tens of minutes. Here, the number of minutes may be any value of 1 to 9 minutes, and the number of tens of minutes may be any value of 10 minutes to 60 minutes. The one-hour average value, the slope of several minutes to several tens of minutes, and the average value of several minutes to several tens of minutes are indicators of the fluctuation tendency in a predetermined time. The control value is a control value for each type of operating end controlled by the combustion control device 30, and the operating value is an actual value for the type of operating end being operated.

Next, an automatic combustion control method according to an embodiment of the present invention will be described. FIG. 3 is a flowchart showing a control operation by the combustion control device 30 and the monitoring center 40 for explaining the automatic combustion control method according to the present embodiment. In the flowchart shown in FIG. 3, steps ST1 to ST3 and ST5 are executed by the combustion control device 30, and step ST4 is executed by the monitoring center 40.

As shown in FIG. 3, in step ST1 executed by the combustion control device 30, the operator sets the target waste incineration amount as the incineration amount set value. Instead of the incineration amount set value, the operator may set the target waste incineration amount as the evaporation amount set value. The set incineration amount set value is stored in the storage unit 34. As a result, automatic combustion control for the incinerator is started.

Next, in step ST2, the waste quality calculation unit 31 determines the input heat amount other than the waste such as the heat amount of the combustion heating air and the furnace 1 according to the waste incineration amount set value stored in the storage unit 34. Based on the planned value of the amount of heat discharged from the waste and the amount of heat recovered and recovered as steam, the waste quality, which is the lower calorific value of the waste, is calculated and set as the waste quality set value. The waste quality set value is stored in the storage unit 34.

Next, in step ST3, the operation amount reference value adjusting unit 32 is stored in the storage unit 34 based on the set value including the incineration amount set value and the waste quality set value stored in the storage unit 34. Calculate the reference value of each operation amount by referring to the operation amount reference value setting relational expression. In this way, after the operation amount reference value is set, the operation amount of each operation end is controlled so as to follow the operation amount reference value, and the waste incinerator is operated.

Next, the process proceeds to step ST4 executed by the monitoring center 40. In step ST4, the instantaneous value of the measurement target shown in Table 1 is input as sensor information to the monitoring center 40 while the furnace 1 is operating. The sensor information is sensor information measured in a state inside the furnace 1 and a facility related to the furnace 1, specifically, in a power generation facility for generating electric power, for example. The input sensor information is stored in the storage unit 44 as a parameter (parameter acquisition step). The control unit 41 of the monitoring center 40 monitors the fluctuation of the instantaneous value of the measurement target stored in the storage unit 44, and also from the instantaneous value of each measurement target input from the combustion control device 30 and read from the storage unit 44. , Calculate the parameters used. The learning unit 41a of the control unit 41 determines the necessity of the intervention operation by the monitoring center 40 based on the input instantaneous value of the measurement target and the calculated usage parameter by the operation-learned model 44a.

FIG. 4 is a flowchart for explaining the determination of the intervention operation by the monitoring center 40 and the derivation of the correction amount in step ST4 of FIG. The control step is executed according to the flowchart shown in FIG. As shown in FIG. 4, in step ST21, the learning unit 41a of the control unit 41 inputs the usage parameters shown in Table 1 to the operation-learned model 44a as input parameters. The learning unit 41a determines whether or not it is necessary to correct the manipulated variable reference value, that is, whether or not an intervention operation is necessary, based on the operation-learned model 44a. When the learning unit 41a determines in step ST21 that an intervention operation is necessary, the process proceeds to step ST22.

In step ST22, the control unit 41 of the monitoring center 40 switches the operation of the incinerator from the automatic combustion control (ACC) by the combustion control device 30 to the control based on the operation-learned model 44a by the learning unit 41a. The learning unit 41a derives a correction amount (correction operation amount reference value) for correcting the operation amount reference value based on the input usage parameter.

In step ST23, the control unit 41 of the monitoring center 40 outputs the correction operation amount reference value. In step ST24, as shown in FIG. 2, the corrected operation amount reference value derived from the learning unit 41a is used in the combustion air amount adjusting unit 36, the cooling air amount adjusting unit 37, and the waste supply device feed speed adjusting unit. It is supplied to the operation amount adjusting unit 35 including the 38 and the grate feed speed adjusting unit 39. As a result, the control unit 41 controls the combustion air damper 14, the cooling air damper 15, the sub-grate combustion air dampers 14a to 14d, the waste supply device 3, and the grate 4, which are the operation ends.

As shown in FIG. 4, when the learning unit 41a reaches the upper limit of the time for ending the intervention operation based on the operation-learned model 44a (step ST25), or the use parameter becomes a condition for ending the intervention. If (step ST26), the process proceeds to step ST27. In step ST27, the control unit 41 changes the operation control of the incinerator from the control based on the operation-learned model 44a by the learning unit 41a to the automatic combustion control (ACC) by the combustion control device 30 by ending the intervention operation. Switch.

When the processing of the monitoring center 40 is switched to the automatic combustion control by the combustion control device 30 in step ST27, the process returns to FIG. 3 and proceeds to step ST5. In step ST5, the waste quality calculation unit 31 of the combustion control device 30 determines whether or not the waste is charged. When the waste quality calculation unit 31 determines that there is waste input (step ST5: Yes), the process returns to step ST2. When the waste quality calculation unit 31 determines that no waste is charged (step ST5: No), the process returns to step ST4. As described above, the automatic combustion control of the incinerator using the trained model by machine learning is executed.

(Example)
Next, the effect when the automatic combustion control method by the combustion control device 30 and the monitoring center 40 configured as described above is executed will be described. FIG. 5 is a graph showing the frequency with which the operator manually intervenes in two different furnaces, the first furnace and the second furnace. FIG. 5A is a graph showing the average number of operations per day in which the operator performed the intervention operation in the first furnace and the second furnace in the case of the conventional automatic combustion control method. FIG. 5B shows one day when the operator performed an intervention operation in the first furnace and the second furnace when the automatic combustion control method according to the above-described embodiment was executed in the waste supply system, the ventilation system, and the like. It is a graph which shows the average number of operations of. The waste supply system is an intervention operation on the grate 4 by the waste supply device 3 by the waste supply device feed rate adjusting unit 38 and the grate feed speed adjusting unit 39. Further, the ventilation system or the like is an intervention operation on the combustion air damper 14 by the combustion air amount adjusting unit 36 and the cooling air damper 15 by the cooling air amount adjusting unit 37. In FIG. 5B, the average number of operations per day in which the control unit 41 by the monitoring center 40 performs the intervention operation based on the operation-learned model 44a is also shown.

As shown in FIG. 5, the operator's manual intervention operation on the waste supply system of the first furnace was 10.2 times a day on average in the conventional automatic combustion control method, but it depends on one embodiment. It can be seen that in the automatic combustion control method, the daily average is reduced to 0.3 times. Similarly, the operator's manual intervention operation on the waste supply system of the second furnace was 18.3 times a day on average in the conventional automatic combustion control method, whereas the automatic combustion control method according to one embodiment. It can be seen that the daily average is reduced to 0.2 times. In the automatic combustion control method according to one embodiment, the average number of operations per day when the monitoring center 40 intervenes in the waste supply system based on the operation-learned model 44a is 17.3 times in the first furnace. , 16.3 times in the second furnace.

Further, from FIG. 5, when the automatic combustion control method according to one embodiment is executed for the blower system or the like, the average number of manual intervention operations per day is 5.9 times in the first furnace. On the other hand, it was reduced to 0.5 times, and it can be seen that the number was reduced to 0.0 times from 12.2 times in the past in the second furnace. That is, it can be seen that the automatic combustion control method according to the above-described embodiment can reduce the number of intervention operations for the ventilation system as well as the waste supply system.

According to one embodiment described above, in a waste incinerator, an intervention operation performed by a skilled or experienced operator (hereinafter referred to as a skilled operator) at his / her own discretion is performed, for example, by cluster analysis on behalf of the skilled operator. By executing the learning unit 41a based on the operation-learned model 44a, which is a program generated by machine learning by the above, stable combustion control by the control unit 41 of the monitoring center 40 becomes possible, so that the operator can perform an intervention operation. It can be significantly reduced and the operation management of the waste incinerator can be automated.

That is, while the conventional automatic combustion control is performed based only on the numerical value of the process, in the automatic combustion control by one embodiment, the control incorporating the judgment of the combustion state similar to the judgment of the skilled operator is added. can do. As a result, it is possible to establish a technique in which the judgment of the combustion state by the human eye and skillful technique is replaced by artificial intelligence using a learned model. Therefore, combustion control including the combustion state in addition to the process signal becomes possible.

INDUSTRIAL APPLICABILITY The present invention can determine the start of a driving intervention, correct the manipulated variable reference value in the driving intervention, end the driving intervention, or set the average intervention time by analyzing a large amount of actual data.

Although one embodiment of the present invention has been specifically described above, the present invention is not limited to the above-mentioned one embodiment, and various modifications based on the technical idea of the present invention are possible. For example, the numerical values given in the above-described embodiment are merely examples, and different numerical values may be used if necessary. The description and drawings that form part of the disclosure of the present invention according to the present embodiment are used in the present invention. The invention is not limited.

For example, the input parameters and output parameters given in the above-described embodiment are merely examples, and different input parameters and output parameters may be used as necessary.

For example, the storage unit 44 stores the operation-learned model 44a and the combustion image-learned model 44b, but the combustion image-learned model 44b can communicate with the monitoring center 40 via a network such as a public network. It is also possible to store it in the storage unit of the combustion image determination server that determines the combustion image. In this case, the combustion image data captured by the combustion image imaging unit 25 is transmitted to the combustion image determination server via the network and stored in the storage unit. After that, the control unit of the combustion image judgment server discriminates and classifies the combustion state based on the combustion image learned model 44b, derives a numerical value according to the combustion state as a result of the discrimination and classification, and monitors the monitoring center. Send to 40. In the monitoring center 40, the received numerical value is input to the operation-learned model 44a of the control unit 41, and each operation end is controlled based on the obtained correction amount.

For example, in the above-described embodiment, deep learning (deep learning) using a neural network is used as an example of machine learning, but machine learning based on other methods may be performed. Other supervised learning, such as support vector machines, decision trees, naive Bayes, and k-nearest neighbors, may be used. Also, semi-supervised learning may be used instead of supervised learning.

1 Furnace 2 Input port 3 Supply device 4 Grate 5 Ash drop port 6 Combustion air blower 7 Combustion outlet 8 Chimney 9 Boiler 9a Heat exchanger 9b Steam drum 10 Cooling air blow port 11 Cooling air blower 14 Combustion air damper 14a, 14b, 14c, 14d Under-grate combustion air damper 15 Cooling air damper 16 Intermediate ceiling 17 Combustion chamber gas thermometer 18 Main flue gas thermometer 19 Furnace outlet lower gas thermometer 20 Furnace outlet middle gas thermometer 21 Furnace outlet gas thermometer 22 Boiler outlet oxygen concentration meter 23 Gas concentration meter 24 Exhaust gas flow meter 25 Combustion image imaging unit 30 Combustion control device 31 Waste quality calculation unit 32 Operation amount reference value adjustment unit 33 Operation amount reference value correction unit 34 Storage unit 35 Operation amount adjustment unit 36 Combustion air amount adjustment unit 37 Cooling air amount adjustment unit 38 Supply device feed speed adjustment unit 39 Grate feed speed adjustment unit 40 Monitoring center 41 Control unit 41a Learning unit 42 Output unit 43 Input unit 44 Storage Part 44a Operation learned model 44b Combustion image learned model

Claims (6)

The operation amount reference value of a plurality of operation ends of the waste incinerator is set based on a preset set value, and the combustion of waste inside the waste incinerator is controlled based on the operation amount reference value. It is an automatic combustion control method executed by the automatic combustion control device.
A parameter acquisition step for acquiring sensor information about the internal state of the waste incinerator and facilities related to the waste incinerator, and a control value of the operation end controlled by the automatic combustion control device as input parameters.
A control step in which the input parameters are input to the operation-learned model, and the operation amount reference value of the operation end is corrected and controlled based on the type of the operation end and the correction value of the operation amount reference value. And, including
The operation-learned model includes learning sensor information about the internal state of the waste incinerator and facilities related to the waste incinerator, and learning control values of the operation end controlled by the automatic combustion control device. Is used as the learning input parameter, and at least the machine learning and statistical methods are performed using the learning sensor information and the correction value for the operation amount reference value corresponding to the learning control value as the learning output parameters. An automatic combustion control method characterized by being a trained model pre-generated by one. The learning sensor information and the learning control value of the learning input parameter are data used for determining the manual intervention operation of the operator, and are the type of operation end and the operation amount reference included in the learning output parameter. The correction value for the value is a correction value for the type of operation end and the operation amount reference value operated by the operator in the manual intervention operation, and the operation-learned model is the learning input parameter and the learning output parameter. The automatic combustion control method according to claim 1, wherein the model is a trained model generated by cluster analysis using the above. The data used for determining the manual intervention operation of the operator is the data of the control value, the operating value, or the instantaneous value of the operating end in the waste incinerator, and the data derived based on the plurality of sensor information. The automatic combustion control method according to claim 2, further comprising data on a fluctuation tendency of the sensor information for a predetermined time and data on a one-hour moving average of control values of the operation end. The sensor information is applied to a combustion image trained model generated in advance by machine learning using the training combustion image data as a learning input parameter and the numerical value corresponding to the learning combustion image data as a learning output parameter, and the waste. The automatic combustion control method according to any one of claims 1 to 3, further comprising a numerical value obtained by inputting combustion image data that images a combustion state in an incinerator. The automatic combustion control according to any one of claims 1 to 4, wherein the learning output parameter includes data for determining the start of the control step and data for determining the end of the control step. Method. The operation amount reference value of a plurality of operation ends of the waste incinerator is set based on a preset set value, and the combustion of waste inside the waste incinerator is controlled based on the operation amount reference value. It is an automatic combustion control device
A storage unit that stores the internal state of the waste incinerator, sensor information about the facility related to the waste incinerator, and the control value of the operation end controlled by the automatic combustion control device as input parameters.
The input parameter read from the storage unit is input to the operation-learned model, and the operation amount reference value of the operation end obtained based on the type of the operation end and the correction value of the operation amount reference value is corrected. It is equipped with a control unit that controls
The operation-learned model includes learning sensor information about the internal state of the waste incinerator and facilities related to the waste incinerator, and learning control values of the operation end controlled by the automatic combustion control device. Is used as the learning input parameter, and at least the machine learning and statistical methods are performed using the learning sensor information and the correction value for the operation amount reference value corresponding to the learning control value as the learning output parameters. An automatic combustion control device characterized by being a trained model pre-generated by one.

Images