Incineration combustion control method and integrated cabinet device

Abstract

This application relates to the field of cremation technology, providing a cremation combustion control method and an integrated cabinet device. The method includes: acquiring multiple control variables under the current operating conditions of the cremation equipment; constructing multiple candidate variable combinations based on the multiple control variables; selecting the optimal variable combination from the multiple candidate variable combinations based on a trained classification model and a trained regression model; and controlling the oil valve and blower of the cremation equipment based on the optimal variable combination. This application, by employing full-process machine control and introducing multiple control variables, as well as a trained classification model and a trained regression model, can avoid excessively high fuel consumption per body and can reduce fuel consumption in a timely and precise manner.

Description

Technical Field

[0001] This application relates to the field of cremation technology, specifically to a cremation combustion control method and an integrated cabinet device. Background Technology

[0002] Cremation equipment is a type of high-temperature combustion equipment, typically consisting of a furnace body, burner, blower, induced draft fan, oil valves, flue gas pipelines, and electrical control system. Many cremation combustion control methods still rely heavily on manual operation, using the main combustion chamber temperature as the sole basis for combustion control.

[0003] However, the above methods have several significant shortcomings. First, manual operation tends to be conservative and inject more fuel. If the fuel supply is not reduced when the temperature of the main combustion chamber meets the conditions for auto-ignition, it will lead to high fuel consumption per body. Second, many factors affect combustion fuel consumption, and cremation combustion control based on the single variable of main combustion chamber temperature cannot accurately reduce fuel consumption. Third, manual operation is mostly based on visual observation of the flame, furnace pressure gauge, or experience, and the adjustment is relatively lagging and cannot respond to high fuel consumption in a timely manner.

[0004] In summary, existing cremation combustion control methods rely on manual operation and use the main combustion chamber temperature as the sole basis for control, which leads to high fuel consumption per body and makes it impossible to reduce fuel consumption in a timely and accurate manner. Summary of the Invention

[0005] This application provides a cremation combustion control method and an integrated cabinet device to solve the technical problem that existing cremation combustion control methods rely on manual operation and use the main combustion chamber temperature as the sole basis for cremation combustion control, which leads to high fuel consumption per body and makes it impossible to reduce fuel consumption in a timely and accurate manner.

[0006] In a first aspect, embodiments of this application provide a cremation combustion control method, including: Obtain multiple control variables under the current operating conditions of the cremation equipment; Based on the aforementioned multiple control variables, multiple combinations of candidate variables are constructed; Based on the trained classification model and the trained regression model, the optimal variable combination is selected from the multiple candidate variable combinations; Based on the optimal combination of variables, the oil valves and blowers of the cremation equipment are controlled.

[0007] In one embodiment, the plurality of control variables includes a plurality of continuous variables and a plurality of switching variables; The construction of multiple candidate variable combinations based on the multiple control variables includes: The continuous variable is perturbed to obtain the perturbed continuous variable; The switch variable is flipped to obtain the flipped switch variable; Multiple candidate variable combinations are constructed based on the continuous variables before and after the perturbation and the switching variables before and after the flip.

[0008] In one embodiment, the step of selecting the optimal variable combination from the plurality of candidate variable combinations based on the trained classification model and the trained regression model includes: The multiple candidate variables are combined and input into the trained classification model to obtain the fuel consumption status of the multiple candidate variable combinations output by the trained classification model. From the multiple candidate variable combinations, select multiple target variable combinations that indicate low fuel consumption. The multiple target variables are combined and input into the trained regression model to obtain the predicted flow rate of pollutant gas output by the multiple target variables combination of the trained regression model; Based on the predicted flow rate of the polluting gas, the fitness of the combination of the multiple target variables is obtained; Based on the fitness, the optimal combination of variables is selected from the multiple combinations of target variables.

[0009] In one embodiment, selecting the optimal combination of variables from the plurality of combinations of target variables based on the fitness includes: Several target variable combinations are extracted from the multiple target variable combinations according to the fitness probability to obtain multiple first variable combinations; By swapping the continuous variables of some of the multiple first variable combinations, multiple second variable combinations are obtained; Resample the continuous variables of some of the multiple second variable combinations, and flip the on / off variables of some of the multiple second variable combinations to obtain multiple third variable combinations; If the preset number of iterations has not been reached, the multiple combinations of third variables are used as new combinations of candidate variables, and the process of inputting the multiple combinations of candidate variables into the trained classification model is repeated until the preset number of iterations is reached. The combination of third variables with the highest fitness among the multiple combinations of third variables at this time is determined as the optimal variable combination.

[0010] In one embodiment, controlling the oil valves and blowers of the cremation equipment based on the optimal combination of variables includes: Under preset constraints, the oil valves and blowers of the cremation equipment are controlled based on the differences between the multiple control variables and the optimal combination of variables.

[0011] In one embodiment, the plurality of control variables includes stage variables; The preset constraints are based on the flame conditions, main combustion chamber temperature, combustion duration, furnace pressure, oxygen content, and the cremation stage corresponding to the stage variables of the cremation equipment, and are applied to the oil valves and blowers of the cremation equipment. The cremation stage adopts a one-way latching mechanism.

[0012] In one embodiment, after controlling the oil valves and blowers of the cremation equipment based on the optimal combination of variables, the process includes: Return to obtain multiple control variables under the current operating conditions of the cremation equipment, until the oil valve and blower of the cremation equipment are controlled again based on the optimal combination of variables.

[0013] Secondly, embodiments of this application provide an integrated cremation combustion control cabinet device, comprising: The communication and PLC area is used to: acquire multiple control variables under the current operating conditions of the cremation equipment; The industrial computing and service area is used for: constructing multiple candidate variable combinations based on the multiple control variables; and selecting the optimal variable combination from the multiple candidate variable combinations based on the trained classification model and the trained regression model. The communication and PLC area is also used to control the oil valves and blowers of the cremation equipment based on the optimal combination of variables.

[0014] In one embodiment, it also includes: The execution and power supply area is used to: supply power to the equipment inside the cabinet, and provide communication lines for the communication and PLC area to control the oil valves and fans of the cremation equipment; The display and interaction area is used to display various parameters for cremation combustion control.

[0015] In one embodiment, the communication and PLC area includes: The Modbus gateway is used to: acquire multiple control variables under the current operating conditions of the cremation equipment; PLC control hardware is used to control the oil valves and blowers of the cremation equipment based on the optimal combination of variables.

[0016] The cremation combustion control method and integrated cabinet device provided in this application acquire multiple control variables under the current operating conditions of the cremation equipment. Based on these control variables, multiple candidate variable combinations are constructed. Using a trained classification model and a trained regression model, the optimal variable combination is selected from these candidate combinations. Based on this optimal variable combination, the oil valves and fans of the cremation equipment are controlled. In this application, the entire process is machine-controlled, avoiding the problems of high fuel consumption per body and delayed response caused by manual operation. Furthermore, multiple control variables are introduced to construct candidate variable combinations, and the optimal variable combination is accurately and quickly selected using the trained classification model and the trained regression model to control the oil valves and fans of the cremation equipment. This avoids the problem of inaccurate fuel consumption reduction caused by relying solely on the main combustion chamber temperature. In summary, this application, by employing full machine control and introducing multiple control variables, as well as trained classification and regression models, can avoid high fuel consumption per body and can reduce fuel consumption in a timely and accurate manner. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is one of the flowcharts illustrating the cremation combustion control method provided in the embodiments of this application; Figure 2 This is a second schematic flowchart of the cremation combustion control method provided in the embodiments of this application; Figure 3 This is the third flowchart of the cremation combustion control method provided in the embodiments of this application; Figure 4 This is the fourth flowchart of the cremation combustion control method provided in the embodiments of this application; Figure 5 This is a schematic diagram of the integrated cremation combustion control cabinet device provided in the embodiments of this application; Figure 6 This is a schematic diagram of the overall control flow of the integrated cremation combustion control cabinet device provided in the embodiments of this application; Figure 7 This is a schematic diagram of the real-time data collection and standardized processing flow in the overall control flow chart provided in the embodiments of this application; Figure 8 This is a schematic diagram of the combustion session identification and stage latching process in the overall control flow chart provided in the embodiments of this application; Figure 9This is a schematic diagram of the intelligent decision-making and fuel consumption optimization process in the overall control flowchart provided in the embodiments of this application; Figure 10 This is a schematic diagram of the warning, preset constraints and manual permission flow in the overall control flow diagram provided in the embodiments of this application; Figure 11 This is a schematic diagram of the command processing and execution control flow in the overall control flow diagram provided in the embodiments of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] It should be noted that in the description of the embodiments of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. The terms "upper," "lower," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application. Unless otherwise expressly specified and limited, the terms "installed," "connected," and "linked" should be interpreted broadly, for example, they can be fixed connections, detachable connections, or integral connections; they can be mechanical connections or electrical connections; they can be direct connections or indirect connections through an intermediate medium; and they can be internal connections between two elements. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0021] The terms "first," "second," etc., used in this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class, without limiting the number of objects; for example, a first object can be one or more. Furthermore, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects have an "or" relationship.

[0022] Figure 1 This is one of the schematic flowcharts of the cremation combustion control method provided in the embodiments of this application. (Refer to...) Figure 1 This application provides a cremation combustion control method, which may include: Step 101: Obtain multiple control variables under the current operating conditions of the cremation equipment; Step 102: Construct multiple combinations of candidate variables based on multiple control variables; Step 103: Based on the trained classification model and the trained regression model, select the optimal variable combination from multiple candidate variable combinations; Step 104: Control the oil valves and blowers of the cremation equipment based on the optimal combination of variables.

[0023] In step 101, the current working status of the cremation equipment is comprehensively reflected through multiple control variables; the specific number and type of control variables can be determined according to actual needs, and are not limited here.

[0024] In step 102, multiple combinations of variable values ​​near the current working state of the cremation equipment can be constructed based on multiple control variables to obtain multiple candidate variable combinations.

[0025] In step 103, the trained classification model and the trained regression model can accurately and quickly select the optimal variable combination from multiple candidate variable combinations.

[0026] In step 104, based on the optimal combination of variables, the oil valves and blowers of the cremation equipment are controlled. This mainly includes adjusting the opening and closing of the oil valves, the operating parameters of the blower and the operating parameters of the induced draft fan, so that when the temperature of the main combustion chamber meets the conditions for auto-ignition, the window for fuel replenishment is promptly identified and reduced; when the negative pressure and oxygen content meet the constraints, unnecessary long-term opening of the oil valves is avoided, and the combustion is promoted to be complete through the adjustment of the blower.

[0027] Based on the traditional manual cremation fuel consumption per body, the method described in this embodiment reduces the fuel consumption per body by approximately 10%. The specific calculation method is as follows: (1) in, The rate of reduction in fuel consumption for cremation of a single body. This refers to the fuel consumption for cremation of a single body under traditional manual control. This represents the fuel consumption for cremation of a single body under the same conditions using the method described in this embodiment.

[0028] Furthermore, for remains with significant weight differences, the following formula can be used for a horizontal comparison of fuel consumption: (2) in, For the first Cremation fuel consumption per unit mass of remains. For the first Fuel consumption for cremation of remains. For the first The weight or equivalent load mass of the remains; for a single remains with a large weight difference, the cremation fuel consumption per unit mass can be calculated based on formula (2), and then substituted into formula (1) to calculate the cremation fuel consumption reduction rate.

[0029] Furthermore, after step 104, you can return to step 101 to obtain the latest control variables, form a closed-loop feedback, and continue to execute steps 102 to 104 to continuously optimize combustion control and reduce fuel consumption.

[0030] The cremation combustion control method provided in this embodiment acquires multiple control variables under the current operating conditions of the cremation equipment. Based on these control variables, multiple candidate variable combinations are constructed. Using a trained classification model and a trained regression model, the optimal variable combination is selected from these candidate combinations. Based on this optimal variable combination, the oil valves and fans of the cremation equipment are controlled. In this embodiment, the entire process is machine-controlled, avoiding the problems of high fuel consumption per body and delayed response caused by manual operation. Furthermore, by introducing multiple control variables and constructing candidate variable combinations, and using the trained classification and regression models, the optimal variable combination is accurately and quickly selected to control the oil valves and fans of the cremation equipment, avoiding the problem of inaccurate fuel consumption reduction caused by relying solely on the main combustion chamber temperature. In summary, this embodiment, by employing full machine control and introducing multiple control variables, as well as trained classification and regression models, can avoid high fuel consumption per body and can reduce fuel consumption in a timely and accurate manner.

[0031] Figure 2 This is the second schematic flowchart of the cremation combustion control method provided in the embodiments of this application. (Refer to...) Figure 2In one embodiment, the multiple control variables include multiple continuous variables and multiple switching variables; step 102 may include: Step 201: Perturb the continuous variable to obtain the perturbed continuous variable; Step 202: Invert the switch variable to obtain the inverted switch variable; Step 203: Construct multiple candidate variable combinations based on the continuous variables before and after the disturbance and the switching variables before and after the flip.

[0032] In step 201, continuous variables can be selected according to actual needs, and there is no limitation here; in this embodiment, continuous variables may include main combustion chamber temperature, top wind speed, left wind speed, right wind speed, left smoke exhaust wind speed and right smoke exhaust wind speed; by subjecting these continuous variables to limited perturbation, multiple variable values ​​near the current value of these continuous variables can be generated so as to construct multiple candidate variable combinations in the future.

[0033] In step 202, the switching variables can be selected according to actual needs, and there is no limitation here; in this embodiment, the switching variables may include the burner, blower, main oil valve and induced draft fan; these switching variables are kept at their current values ​​or flipped with a certain probability, that is, the values ​​of some switching variables are kept, such as keeping the burner on, and the values ​​of other switching variables are flipped, such as flipping the blower on to the blower off, thereby generating more switching variable values ​​so as to construct multiple candidate variable combinations in the future.

[0034] In step 203, through perturbation and flipping, the continuous variable values ​​and switch variable values ​​can be effectively expanded, so that the expanded continuous variable values ​​and switch variable values ​​can be combined to form multiple candidate variable combinations. The specific combination method can be set according to actual needs and is not limited here. In this embodiment, it can be set that each candidate variable combination includes all continuous variables and switch variables, and all variable values, including continuous variables and switch variables, are not completely the same between any two groups.

[0035] This embodiment introduces continuous variables and switch variables, applies limited perturbation to the continuous variables, and performs probability flipping on the switch variables to expand the original variable values, thereby forming diverse combinations of candidate variables, which is helpful for subsequent iterative optimization.

[0036] Figure 3 This is the third schematic flowchart of the cremation combustion control method provided in the embodiments of this application. (Refer to...) Figure 3 In one embodiment, step 103 may include: Step 301: Input multiple candidate variables into the trained classification model to obtain the fuel consumption status of multiple candidate variable combinations output by the trained classification model. Step 302: Select multiple combinations of target variables from multiple candidate variable combinations that indicate low fuel consumption. Step 303: Input the combination of multiple target variables into the trained regression model to obtain the predicted flow rate of pollutant gas output by the combination of multiple target variables from the trained regression model. Step 304: Based on the predicted flow rate of pollutant gases, obtain the fitness of multiple target variable combinations; Step 305: Based on fitness, select the optimal combination of variables from multiple combinations of target variables.

[0037] In step 301, the classification model can be a random forest, gradient boosting tree, XGBoost, support vector machine, or neural network, etc., which is not limited here, and is used to classify the fuel consumption state corresponding to each candidate variable combination.

[0038] Since the fuel consumption of a single body under traditional manual control is generally greater than 20 liters, a fuel consumption of less than 20 liters can be defined as a low fuel consumption state and a fuel consumption of more than 20 liters as a high fuel consumption state. By training the classification model with historical candidate variable combinations and their corresponding fuel consumption state labels, the fuel consumption state corresponding to the current candidate variable combination can be determined through the trained classification model.

[0039] In step 302, the candidate variable combination with low fuel consumption status among multiple candidate variable combinations is determined as the target variable combination.

[0040] In step 303, the regression model can be a tree model, a linear model, an ensemble learning model, or a lightweight neural network, etc., which is not limited here, and is used to predict the pollutant gas flow rate corresponding to each combination of target variables.

[0041] The type of pollutant gas can be selected according to actual needs and is not limited here. In this embodiment, carbon monoxide, sulfur dioxide, nitrogen dioxide and flue gas can be selected as pollutant gases. The regression model is trained by using historical target variable combinations and their corresponding flow rates of each pollutant gas to obtain a trained regression model for predicting the flow rate of each pollutant gas. The predicted flow rate of each pollutant gas corresponding to the current target variable combination can then be determined by the trained regression model.

[0042] In step 304, each combination of target variables can be calculated based on the following formula. fitness : ; in, Combination of target variables The corresponding predicted carbon monoxide flow rate, Combination of target variables The corresponding predicted sulfur dioxide flow rate, Combination of target variables The corresponding predicted nitrogen dioxide flow rate, Combination of target variables The corresponding predicted flue gas flow rate, , , and To correspond to the weights of the predicted traffic, Combination of target variables The corresponding fuel consumption status indicator, when the fuel consumption status is low fuel consumption, When the fuel consumption state is high, ; The larger the value, the better the combination of target variables. The better the flue gas flow rate, the better. Since flue gas flow rate is related to combustion intensity and fuel utilization, incorporating it into the fitness function can indirectly constrain energy consumption and emission trends.

[0043] In step 305, the optimal combination of target variables can be selected from multiple combinations of target variables based on its fitness.

[0044] This embodiment first uses a trained classification model to select low-fuel-consumption variable combinations from candidate variable combinations. Then, it uses a trained regression model to predict the pollutant gas flow rate corresponding to the low-fuel-consumption variable combinations. Based on the pollutant gas flow rate, it calculates the fitness to measure the adverse environmental impact of the low-fuel-consumption variable combinations. Finally, based on the fitness, it selects the optimal variable combination from the low-fuel-consumption variable combinations to achieve dual control of fuel consumption and pollutants, so that the selected optimal variable combination meets the dual objectives of low fuel consumption and low pollution.

[0045] Figure 4 This is the fourth schematic flowchart of the cremation combustion control method provided in the embodiments of this application. (Refer to...) Figure 4 In one embodiment, step 305 may include: Step 401: Extract several target variable combinations from multiple target variable combinations according to the fitness probability to obtain multiple first variable combinations; Step 402: Swap the continuous variables of some of the first variable combinations to obtain multiple second variable combinations; Step 403: Resample the continuous variables of some variable combinations in multiple second variable combinations, and flip the on / off variables of some variable combinations in multiple second variable combinations to obtain multiple third variable combinations. Step 404: If the preset number of iterations has not been reached, combine multiple third variables as new combinations of multiple candidate variables and return to step 301; Step 405: When the preset number of iterations has been reached, the combination of third variables with the highest fitness among the multiple combinations of third variables at this time is determined as the optimal combination of variables.

[0046] In step 401, the probability of fitness for each target variable combination is first calculated, that is, the proportion of fitness of each target variable combination in the total fitness of all target variable groups. Then, several target variable combinations with a larger proportion of fitness are selected as the first variable combination.

[0047] In step 402, select some of the first variable combinations and exchange the values ​​of their corresponding continuous variables. The number of continuous variables exchanged can be set according to actual needs and is not limited here. The exchanged first variable combinations and the first variable combinations that did not participate in the exchange are used together as the second variable combinations.

[0048] In step 403, a portion of the second variable combinations are selected, and the values ​​of their continuous variables are resampled; a portion of the second variable combinations are selected, and the values ​​of their switching variables are flipped; the resampled and flipped second variable combinations, together with the second variable combinations that did not participate in these two operations, are used as the third variable combinations.

[0049] It should be noted that resampling should be performed near the current value of the continuous variable. In addition, the combination of the second variable selected in the two tests can be exactly the same or not. This is not limited here.

[0050] In step 404, if the preset number of iterations has not been reached, the third variable combination after the above operations is used as a new candidate variable combination, and steps 301 to 305 are executed to continuously iterate and optimize.

[0051] In step 405, after reaching the preset number of iterations, the third variable combination has achieved the preset optimization effect. The third variable combination with the highest fitness is selected as the optimal variable combination.

[0052] In this embodiment, a first variable combination is obtained by selecting a target variable combination with high fitness from multiple target variable combinations. The continuous variable values ​​of some first variable combinations are swapped to obtain a second variable combination. The continuous variable values ​​of some second variable combinations are resampled, and the on / off variable values ​​of some second variable combinations are flipped to obtain a third variable combination. This allows for the mutation of high-fitness variable combinations within a reasonable range to update the candidate variable combinations. The trained classification model and the trained regression model can then be used to continuously classify and predict. The third variable combination is iteratively optimized based on fitness, and finally, the third variable combination with the highest fitness is selected as the optimal variable combination to achieve the dual goals of low fuel consumption and low pollution.

[0053] In one embodiment, step 104 may include: Under preset constraints, the oil valves and blowers of the cremation equipment are controlled based on the differences between multiple control variables and the optimal combination of variables. Specifically, the oil valves and blowers of the cremation equipment are controlled based on the differences between the current values ​​of continuous variables among the multiple control variables and the values ​​of continuous variables in the optimal combination of variables, as well as the differences between the current values ​​of on / off variables among the multiple control variables and the values ​​of on / off variables in the optimal combination of variables, in order to reduce these differences. The blowers may include blowers and induced draft fans. The specific control methods may be adjusting the operating parameters of the blowers, including increasing, decreasing, and keeping them unchanged, and / or adjusting the operating parameters of the induced draft fans, including increasing, decreasing, and keeping them unchanged, and / or adjusting the on / off state of the oil valves, including opening and closing.

[0054] Furthermore, multiple control variables include stage variables, whose values ​​include the minute of cremation, the third minute of cremation, etc., used to reflect whether the cremation stage is the pre-start stage, the ignition stage, the spontaneous combustion stage, or the final stage. The specific cremation stage can be determined according to the following formula: ; in, The value of the stage variable Pollutant gases The relative rate of change of flow rate The value of the stage variable Pollutant gases Traffic, The value of the stage variable Pollutant gases Traffic, The sampling interval is... Represents absolute value. To prevent extremely small positive numbers with a denominator of zero; we obtain Then, it can be based on The value of the variable in the determination stage The cremation stage, for example when If the fluctuation is within the preset fluctuation range of stage A, then the value of the stage variable can be determined. The cremation stage at that time was stage A.

[0055] Furthermore, the cremation stage employs a one-way latching mechanism, meaning that it only proceeds in the order of the initial stage → ignition stage → auto-ignition stage → final stage, and does not allow reverse transitions.

[0056] The preset constraints are based on the flame conditions, main combustion chamber temperature, combustion duration, furnace pressure, oxygen content, and the corresponding cremation stage of the cremation equipment. These constraints are applied to the oil valves and blowers of the cremation equipment. For example, if no flame is detected or the flame detection signal is 0, regardless of whether the algorithm suggests opening the oil valve or whether afterburning is required, the oil valve is prohibited from opening or is forced to close, and a manual check of the ignition status, burner, and flame sensor is prompted. The oil valve is forced to open when the main combustion chamber temperature is below 300 degrees Celsius; the oil valve is forced to open during the ignition stage; the oil valve is forced to open during the tail stage; the oil valve is only allowed to close when the combustion duration exceeds the upper limit, and a manual check is prompted; the induced draft fan operating parameters are adjusted when the furnace pressure is below or above the target range; the blower operating parameters are adjusted when the oxygen content deviates from the target range, etc.

[0057] This embodiment aims to minimize the gap between multiple control variables and the optimal combination of variables by controlling the oil valves and blowers of the cremation equipment. During the control process, mandatory engineering constraints must be met to ensure absolute safety, thus forming an intelligent control strategy that prioritizes model optimization, provides a safety net of preset constraints, and allows for manual takeover. Furthermore, traditional manual control of the cremation stages has unclear boundaries and lacks a stage latching mechanism, making it prone to jumping back and forth between the ignition stage, the auto-ignition stage, and the final stage, resulting in control jitter. The stage latching mechanism in this embodiment can avoid the above situation and will not frequently change the oil valve and blower strategies due to short-term fluctuations in control variables.

[0058] Figure 5 This is a schematic diagram of the integrated cremation combustion control cabinet device provided in an embodiment of this application. (Refer to...) Figure 5 This application provides an integrated cremation combustion control cabinet device, which may include: The communication and PLC area is used to: acquire multiple control variables under the current operating conditions of the cremation equipment; The industrial computing and service area is used for: constructing multiple candidate variable combinations based on multiple control variables; and selecting the optimal variable combination from multiple candidate variable combinations based on the trained classification model and the trained regression model. The communication and PLC area is also used for controlling the oil valves and fans of the cremation equipment based on the optimal combination of variables.

[0059] Furthermore, the device also includes: The execution and power supply area is used to: supply power to the equipment inside the cabinet, and provide communication lines for the communication and PLC area to control the oil valves and fans of the cremation equipment; The display and interaction area is used to display various parameters for cremation combustion control.

[0060] Furthermore, the communication and PLC area includes: The Modbus gateway is used to: acquire multiple control variables under the current operating conditions of the cremation equipment; PLC control hardware is used to control the oil valves and blowers of cremation equipment based on the optimal combination of variables.

[0061] Reference Figure 5 The integrated cabinet device in this embodiment adopts an upper and lower partition structure: The upper part of the cabinet is the display and interaction area, where a monitor is installed to display the cremation combustion control software interface. This software interface directly displays the current time, system operating status, connection status, cremation stage, combustion duration, main combustion chamber temperature, oxygen content, carbon monoxide flow rate, sulfur dioxide flow rate, nitrogen oxide flow rate, main combustion chamber furnace pressure, control suggestions, Modbus coding, alarm logs, and manual / intelligent modes, etc.

[0062] The central section of the cabinet houses the industrial computing and service area, where industrial computers are installed to deploy the software runtime environment, model files, control services, log queues, REST interfaces, SSE (Server-Sent Events) log stream interfaces, and data archiving logic. The control service receives start, stop, mode switching, parameter saving, cleanup mode settings, and debug mode settings requests from the front end, and organizes the acquisition thread, heartbeat thread, and algorithm thread to run collaboratively.

[0063] The lower part of the cabinet is divided into the communication and PLC area and the execution and power supply area. The communication and PLC area houses the PLC control hardware, Modbus gateway, relays, and terminal blocks. The execution and power supply area houses the power supply, protective switches, and output terminals. The Modbus gateway connects to the field PLC or remote I / O module. The PLC control hardware processes coil control signals from the software and drives actuators such as blowers, induced draft fans, and oil valves. Relays and terminal blocks are used for signal isolation, termination, and maintenance testing. The power supply and protective switches ensure stable operation of the equipment within the cabinet. The output terminals connect to actuators such as blowers, induced draft fans, and oil valves, facilitating PLC control hardware control of these actuators.

[0064] and Figure 5 Correspondingly, the cabinet area of ​​the device, the main components of each area, and the functions of each area are shown in the table below: Table 1. Description of different areas of the integrated cabinet unit

[0065] As can be seen from the above, to enhance the system's robustness and ease of maintenance, the lower part of the cabinet features strict electrical isolation and standardized design. Relays not only achieve physical signal isolation between the control-side low-voltage and execution-side high-voltage circuits, effectively preventing surge currents from entering and burning out the core motherboard, but also provide clear node mapping. Dense terminal blocks standardize complex external wiring, providing intuitive and easy-to-use test contacts for daily inspections and troubleshooting. Furthermore, comprehensive protective switches (such as miniature circuit breakers and overload protectors) construct a robust power defense system, effectively resisting short circuits, overloads, and grid fluctuations, comprehensively ensuring the long-term, continuous, and stable operation of the precision components within the cabinet in harsh industrial environments.

[0066] It should be noted that the software interface described above can also be deployed on industrial tablets or embedded touchscreens, and the control services can also be deployed on embedded hosts or edge computing modules; there are no limitations on this. Furthermore, the display format of the control parameters can be a web frontend, a desktop client, or an HMI (Local Human Machine Interface); there are also no limitations on this.

[0067] Existing control cabinets or boxes typically emphasize electrical safety and PLC execution, but lack a structured design that integrates the host computer software interface, algorithm services, PLC gateway, data archiving, alarm logs, manual / intelligent dual-mode switching, and intelligent optimization results into a single cabinet. For cremation sites, operators usually need to directly observe the monitor, switch modes, view warnings, handle anomalies, and confirm control recommendations. Therefore, integrating display, computing, communication, and control hardware into a single cabinet can significantly improve deployment convenience and on-site maintainability.

[0068] Furthermore, cremation equipment operation typically involves strong on-site monitoring and continuous operation. Once the cremation process begins, furnace temperature, negative pressure, flue gas composition, and fan status will continuously change. The control system needs to simultaneously possess capabilities for rapid data reading, robust caching, periodic calculation, anomaly alerts, manual intervention, and log traceability. If the software, PLC, monitors, and communication equipment are deployed separately, on-site maintenance personnel need to check multiple nodes separately, resulting in high troubleshooting costs. In this embodiment, the integrated cabinet device forms a unified electrical, communication, and software structure, which can significantly reduce deployment and maintenance difficulties.

[0069] Reference Figure 5 and Figure 6 In one embodiment, the overall control flow of the device can be briefly described as follows: S1. Multi-source sensor signal acquisition: Temperature, furnace pressure, wind speed, oxygen content, flue gas flow rate, carbon monoxide flow rate, sulfur dioxide flow rate, nitrogen oxide flow rate, and on / off variable sensors are installed at the locations of the cremation equipment, its flue gas pipelines, ducts, and actuators. Signals include continuous variables, on / off variables, and stage variables; furthermore, signals may also include on-site status variables of the cremation equipment, its flue gas pipelines, ducts, and actuators.

[0070] S2. Connection to the integrated cremation combustion control cabinet: Connect the field signals obtained from S1 to the integrated cabinet. The integrated cabinet serves as a centralized carrier for multi-source data, intelligent algorithms, manual operation, and PLC command processing.

[0071] S3. Cremation Combustion Control Software Interface Enabled: The operator starts the cremation combustion control software on the display screen on the upper part of the integrated cabinet, and sets parameters such as body weight, freezing time, target negative pressure, manual / intelligent mode, cleaning mode, and debugging mode. The software interface provides start, stop, mode switching, real-time display of key control parameters, log stream, warning prompts, and final control result display.

[0072] S4. Real-time data collection and standardization: The industrial computer reads field data through the Modbus gateway and completes field mapping, unit unification, timestamp marking, memory caching, CSV (Comma-Separated Values) archiving, and persistence of running status, forming a data snapshot that can be read by the algorithm thread; among them, continuous variables and stage variables are read through Modbus registers, and switch variables and field status variables are read through Modbus coils or discrete points.

[0073] S5. Combustion Session Identification and Stage Latching: The industrial computer executes start-up interlocks based on the blower, induced draft fan, and main combustion chamber pressure; once the conditions are met, a combustion session is established, and the combustion stage is determined based on the main combustion chamber temperature and the relative change rate of pollutant gases. A one-way latching mechanism is used during the combustion stage to avoid strategy backtracking.

[0074] S6. Intelligent Decision Making and Fuel Consumption Optimization: The industrial computer constructs candidate variable combinations based on multiple control variables under the current operating conditions, uses a trained classification model to screen low-energy-consumption variable combinations, uses a trained regression model to predict pollutant gas flow, and then obtains the optimal variable combination through multi-objective fitness function and intelligent algorithm iteration, thereby forming control suggestions for oil valves, blowers and induced draft fans.

[0075] S7, Warnings, Preset Constraints and Manual Access: In intelligent mode, the industrial computer issues warnings for situations such as data disconnection, failure to meet start-up interlock conditions, no flame detected, excessively low main combustion chamber temperature, abnormal furnace pressure, oxygen deviation, model anomalies, cleanup mode conflicts, and PLC write failures. In manual mode, the operator has priority control access, and the algorithm only monitors and displays suggestions.

[0076] S8. Command processing and transmission to PLC control hardware: The industrial computer encodes the final control suggestions, derived from the fusion of S6 and S7, into PLC-recognizable control bits, writes them to the PLC coil via a Modbus gateway, and simultaneously maintains the heartbeat and mode bits. Fan regulation signals can be repeatedly transmitted in pulses to improve PLC recognition reliability.

[0077] S9. Execution Control and Generation of Control Results: The PLC control hardware sends control commands to actuators such as the induced draft fan, blower, and oil valves based on the control codes written into the coils, causing the cremation equipment to perform control operations. The execution results are transmitted back to the integrated cabinet device software interface for display, forming intelligent cremation combustion control results, including reduced fuel consumption, stage stability, emission trends, and log archiving; in addition, multi-source sensors re-collect the latest signals after control, forming a closed-loop feedback.

[0078] As can be seen from the above, in S6 to S8, control suggestions are not issued directly. Instead, warnings, logs, control suggestions, and manual confirmation status are first displayed on the software interface. Only after manual confirmation or permission from the intelligent mode is granted is the control suggestion encoded and written into the PLC coil.

[0079] The feedback lines in the overall process are used to represent continuous control cycles, rather than turning the process into a chaotic loop structure. Each control cycle first reads field data, then generates control suggestions and executes them, before entering the next cycle.

[0080] The above process abandons the scattered and disordered network interaction mode of traditional industrial control, and constructs a clear-bounded, top-down, closed-loop data flow system. It clearly tracks how data is collected in real time from complex field environments, converges to the integrated cabinet device, then commands are issued to the underlying PLC, which drives external actuators, and finally the physical execution results are fed back up to the software interface—the entire lifecycle path. This design transforms the black-box underlying control logic into a highly visualized white-box data flow.

[0081] It should be noted that in this embodiment: The number and type of multi-source sensors can be adjusted according to the structure of the on-site cremation equipment, as long as the control logic still takes the integrated cabinet device as the centralized processing carrier and retains the manual / intelligent dual mode and PLC command processing link. Furthermore, the PLC communication protocol can be Modbus TCP (Transmission Control Protocol), or it can be changed to Modbus RTU (Remote Terminal Unit), OPC UA (Open Platform Communications Unified Architecture), ProFiNet (Process Field Net), EtherNet / IP (Ethernet Industrial Protocol), or other industrial communication protocols depending on the field devices. The key is that the integrated cabinet device can read the field data and write the final control recommendations to the PLC or remote I / O devices.

[0082] Traditional manual control schemes focus on fixed parameter control and equipment-level interlocking, and do not integrate multi-source sensor data, cremation stage identification, machine learning prediction, intelligent algorithm optimization, manual takeover, and industrial cabinet integration into a unified, reproducible, and deployable integrated cabinet device.

[0083] This embodiment of the integrated cabinet device can be directly deployed at cremation sites. On one hand, it features both manual and intelligent modes. In manual mode, it retains the operator's priority control over start / stop, mode switching, cleaning modes, oil valves, and blower-related actions. In intelligent mode, it achieves closed-loop control through stage recognition, model selection, intelligent algorithms, multi-objective fitness functions, and preset constraint correction, and provides warnings for abnormal situations through a software interface. On the other hand, it integrates sensor readings, coil status, algorithm suggestions, manual confirmation, warning logs, PLC write results, and cremation task archive records into the integrated cabinet device for processing, ensuring the cabinet exhibits complete device attributes at both the physical structure and control logic levels. It can function as an independent field control device or connect to existing cremation equipment PLCs and multi-source sensors; it ensures the operator has significant control authority in manual mode while generating energy-saving control suggestions through stage recognition, model selection, and intelligent algorithms in intelligent mode; it can warn of abnormal operating conditions and reduce fuel consumption compared to traditional manual or fixed threshold control under the same cremation tasks and burnout requirements. Based on the objectives of this embodiment and the results of on-site commissioning, this device can reduce fuel consumption for cremation of a single body by approximately 10%.

[0084] Reference Figure 7 In one embodiment, S4 is briefly described as follows: S41, the Modbus register and coil read continuous variables, stage variables, switch variables, and field status variables. Continuous variables may include flue gas temperature, flue gas pressure, flue gas velocity, nitrogen dioxide meter reading, flue gas oxygen content, instantaneous flue gas flow rate, cumulative flue gas flow rate, carbon monoxide flow rate, carbon dioxide flow rate, sulfur dioxide flow rate, nitrogen monoxide flow rate, anemometer signal, main combustion chamber temperature, secondary combustion chamber temperature, heat exchanger temperature, main combustion chamber pressure, and secondary combustion chamber pressure. Switch variables may include oil valves, burners, blowers, and induced draft fans. Field status variables may include empty furnace signals, ignition signals, gate position signals, fan fault signals, negative pressure electric valve status signals, and flame detection signals. The main variables and their uses are shown in the table below. Table 2 Variable Examples Table 1

[0085] S42. The variables in S41 are mapped to fields and their dimensions are unified to convert sensor addresses into operational variables. For example, anemometers 1 to 5 can be mapped sequentially to top wind speed, left wind speed, right wind speed, left smoke exhaust wind speed, and right smoke exhaust wind speed; combustion chamber temperature can be mapped to main combustion chamber temperature; main combustion chamber pressure can be mapped to main combustion chamber furnace pressure; and nitric oxide flow rate and nitrogen dioxide flow rate can be combined into nitrogen oxide flow rate. By unifying field naming, the algorithm does not need to concern itself with the original addresses of the field instruments. Furthermore, the converted operational variables can be written into the daily CSV archive for easy comparison testing and traceability.

[0086] S43. Perform a quality check on the data processed in S42. The check includes identifying missing data, out-of-bounds data, device connection status, and timestamps. Furthermore, the quality-checked and processed data can be saved in runtime_state.json to achieve persistent runtime status. Runtime status can include whether the combustion session has started, the cremation start time, the current latching stage, the auto-ignition start time, the blower's continuous shutdown count, the last oil valve status, the manual mode status, and the empty furnace status. If the control service restarts, the system can restore critical contexts, thereby preventing stage identification errors caused by control session interruptions.

[0087] S44. Cache and snapshot the data processed by S43 in memory; Furthermore, refresh the front-end interface and log stream to display the latest key control parameters, warnings, and control suggestions on the software interface.

[0088] Furthermore, the algorithm thread reads the latest snapshot and forms combustion control every 20 seconds.

[0089] The above process employs three parallel cycles. First, the acquisition thread reads Modbus data every second and writes the data to the memory cache and the daily CSV file. Second, the heartbeat thread sends a signal to the PLC heartbeat coil every second to determine if the communication link is still online. Third, the algorithm thread reads the most recently cached data and executes combustion control every 20 seconds. This separation of time scales balances real-time interface updates with actuator stability, avoiding mechanical motion jitter caused by repeated high-frequency algorithm updates.

[0090] Reference Figure 8 In one embodiment, S5 is briefly described as follows: S51. Read the status of the blower, induced draft fan, main combustion chamber furnace pressure, and main combustion chamber temperature to identify whether a combustion session exists.

[0091] S52. If the number of times the blower is continuously shut down reaches the confirmation number, the current cremation combustion session is determined to be over. The combustion start time, current cremation stage, auto-ignition start time and blower continuous shutdown count are reset. An end suggestion of closing the oil valve or keeping it manually adjusted is output, and the process returns to S51 to execute the next cremation task.

[0092] S53. If the number of consecutive shutdowns of the blower does not reach the required number of confirmations, it is determined that the current cremation combustion session has not ended. An interlocking start-up determination is then executed. When the blower and induced draft fan are both on, and the main combustion chamber pressure is not higher than the main combustion chamber pressure threshold, the interlocking start-up condition is met. A combustion session is established or maintained, and the start time and session number are recorded. Otherwise, the interlocking start-up condition is not met. For example, in this embodiment, the main combustion chamber pressure threshold can be set to -10Pa. When the furnace pressure is greater than -10Pa, it is considered that the negative pressure is insufficient, and the interlocking start-up condition is not met. The induced draft fan operating parameters should be increased first, or a manual check should be performed before returning to S51, and timing and automatic control cannot begin.

[0093] S54. After establishing or maintaining a combustion session, the ignition stage is determined based on the relative change rate of the main combustion chamber temperature and the pollutant gas flow rate, and the combustion duration can be incorporated for joint determination. The ignition stage is usually characterized by a low main combustion chamber temperature or a generally low pollutant gas change rate; the auto-ignition stage is usually characterized by a main combustion chamber temperature higher than the auto-ignition threshold and a significant increase in the change rate of at least one pollutant gas; the tail stage is usually characterized by the current pollutant gas flow rate being lower than a certain percentage of the peak value of this session, and the combustion duration reaching the minimum tail stage cut-in duration.

[0094] S55. One-way latching during the execution phase prevents backtracking during the cremation phase. The cremation phase state is only allowed to proceed in the order of Non-started Phase → Ignition Phase → Auto-ignition Phase → Tail-end Phase. Once the tail-end phase is entered, there is no backtracking; if the tail-end phase pollutant gas flow ratio condition has been met but the combustion time has not yet reached the minimum tail-end cut-in time, the system remains in the ignition or auto-ignition phase, and the latching reason is recorded. This mechanism prevents the intelligent algorithm from frequently changing the oil valve and fan strategies due to short-term sensor fluctuations.

[0095] S56 outputs the cremation stage and control constraints for use by the intelligent algorithm and software interface, then returns to S54 to make the next cycle judgment.

[0096] Reference Figure 9 In one embodiment, S6 is briefly described as follows: S61. Read the latest data snapshot to obtain multiple control variables under the current operating conditions of the cremation equipment, including continuous variables, on / off variables, and stage variables; specifically, the main variables and their uses are shown in the table below: Table 3 Variable Examples Table 2

[0097] S62. Perturb the continuous variable and flip the switch variable to construct multiple candidate variable combinations.

[0098] S63. Based on the trained classification model, determine the fuel consumption status of the candidate variable combination.

[0099] S64. For candidate variable combinations with low fuel consumption, the fuel consumption state indicator in the fitness function is 1, so the fitness is not 0. Based on the trained regression model, the corresponding pollutant gas flow rate is predicted. For candidate variable combinations with high fuel consumption, the fuel consumption state indicator in the fitness function is 0, so the fitness is 0 and the combination is not selected, thus avoiding misleading optimization by subsequent regression results.

[0100] S64. Calculate the fitness of the low fuel consumption variable combination based on the predicted pollutant gas flow rate, and merge it with the aforementioned 0 fitness.

[0101] S65. Update candidate variable combinations based on fitness screening, continuous variable value exchange and resampling, and switch variable value flipping.

[0102] S66. If the preset number of iterations has not been reached, return to S63 to continue iterating; if the preset number of iterations has been reached, determine the optimal combination of variables and formulate control suggestions for the oil valve and the blower accordingly.

[0103] Furthermore, preset constraints and manual permissions can be superimposed on the control suggestions of S66 to output the final control suggestions.

[0104] As can be seen from the above, the fuel consumption reduction in this embodiment does not rely on a single mechanical modification, but rather on the synergy of multi-source sensing, stage latching, model selection, intelligent algorithms, mandatory constraints, and manual confirmation.

[0105] Reference Figure 10 In one embodiment, S7 is briefly described as follows: S71. Read the device connection status, cremation stage, multi-source sensor data, and log data to determine the current control mode.

[0106] S72. If the current control mode is manual mode, manual operation takes priority, and the algorithm only provides control suggestions.

[0107] S73. If the current control mode is intelligent mode, the control suggestion generated by the algorithm enters the control channel. The control suggestion is checked for warnings and preset constraints are superimposed. If the constraints are met and there are no warnings, the final control suggestion is formed. Furthermore, the final control suggestion can be issued for command processing and log recording, and the final control suggestion and control result are displayed. If the constraints are not met or there are warnings, alarm and protection actions are initiated, and security degradation is performed. Automatic writing is suspended, and manual confirmation or on-site takeover is required. The reason for the warning is manually recorded and displayed on the software interface.

[0108] S74, return to S71 and provide feedback on the current control mode.

[0109] This embodiment emphasizes two control modes: manual and intelligent. The manual mode is not an auxiliary function but a priority mode with greater authority. When the operator switches to manual mode, sensor data continues to be read and algorithm-generated control suggestions and log records are displayed, but the algorithm-generated control signals are no longer automatically written to the PLC coil. This retains the algorithm's auxiliary judgment while preventing the algorithm from directly overriding manual operation during maintenance, abnormal cremation, debugging, cleaning, or special remains conditions.

[0110] In intelligent mode, the algorithm is allowed to generate final control suggestions under preset constraints, which are then written to the PLC control hardware after command processing. Intelligent mode has a warning function; when field data or control conditions are abnormal, the software interface displays the reason in the log and warning areas. Warnings are not simply alarm lights, but are related to data, constraints, models, and PLC status, explaining why control is not allowed to begin, why the oil valve remains open, why the induced draft fan operating parameters are forcibly increased, or why automatic writing is paused.

[0111] Warning conditions in this embodiment include, but are not limited to: Modbus gateway or PLC connection failure; no sensor data read; empty cache file; blower or induced draft fan not turned on; main combustion chamber pressure higher than -10Pa and not meeting the start-up interlock; main combustion chamber temperature lower than 300℃ requiring forced opening of the oil valve; abnormal model or algorithm operation; target negative pressure setting higher than -10Pa; weight exceeding the range of 1kg to 200kg; freezing time exceeding the range of 0 days to 30 days; cleanup mode requested while the algorithm is running and not switched to manual mode; cleanup frequency not 15Hz, 20Hz, or 40Hz; PLC write failure; combustion time exceeding the maximum threshold requiring manual check of burnout status. Specific details are shown in the table below: Table 4. Example of a warning

[0112] The intelligent mode does not blindly shut off the oil valve upon receiving a warning. If the model malfunctions or data is insufficient, a safety degradation strategy is prioritized, such as keeping the oil valve open, increasing the operating parameters of the blower and induced draft fan, or prompting for manual inspection. This strategy aligns with the principle of prioritizing safety in cremation equipment.

[0113] Reference Figure 11 In one embodiment, S8 to S9 are briefly described as follows: S891. Command processing of the final control suggestion can be performed using PLC coding according to the following formula: in, To write the encoded value to the PLC control coil, This is the oil valve control position; 0 indicates closed, and 1 indicates open. Codes for adjusting the operating parameters of the induced draft fan; This indicates the code for adjusting the operating parameters of the blower. Among them, Indicates no change. This indicates an increase. This indicates a reduction. Similarly.

[0114] S892, The signal is written to the PLC coil via the Modbus gateway, while maintaining the heartbeat and mode bits.

[0115] S893. If an abnormality occurs, such as a write failure or communication failure, an alarm will be displayed and recorded on the software interface. Further, manual intervention or security downgrade will be implemented, and automatic writing will be paused. The process will return to step 891 to generate manual feedback. If there is no abnormality, the PLC control hardware will parse and interlock the written data to determine the heartbeat, mode, and interlock conditions. Further, the actuators such as the induced draft fan, blower, and oil valve will be driven to operate.

[0116] S894, the actuator changes the air supply, negative pressure and fuel replenishment in the cremation equipment, which further affects the temperature, pollutant gas concentration and flow rate. The multi-source sensor collects multiple control variables again and returns to step 891 to form the feedback for the next cycle.

[0117] In this embodiment, the final control suggestion is first determined in the software interface, and then encoded into PLC control bits through command processing. The lower two bits can be used for blower adjustment, the middle two bits for induced draft fan adjustment, and the higher bits for oil valve switching. The heartbeat bit, control mode bit, and cleaning mode frequency bit are not mixed with the control word to avoid different functions overlapping.

[0118] Furthermore, the heartbeat bit indicates that the integrated cabinet device control service is still online; the control mode bit indicates to the PLC control hardware that the current mode is manual or intelligent; the cleaning mode frequency bit can be used to select the maintenance operation frequency of 15Hz, 20Hz, or 40Hz. To avoid the fan adjustment signal being missed by the PLC control hardware, when the final control suggestion is to increase or decrease the operating parameters of the blower or induced draft fan, a pulse-type repeated transmission can be used three times; when the final control suggestion is to keep the operating parameters of the blower or induced draft fan unchanged, only the oil valve status and heartbeat need to be maintained.

[0119] This embodiment of the integrated cabinet device performs mode confirmation, manual permission judgment, warning check, safety constraint correction, coding, heartbeat maintenance, and log recording before sending control signals to the actuator. Therefore, it belongs to the cabinet-type device with intelligent processing capabilities. The following is a command processing example: Table 5 Example of Command Processing

[0120] This embodiment improves the reliability of field control signals and retains log traceability by using heartbeat bit, mode bit, control bit encoding, pulse repetition transmission and PLC coil processing mechanism.

[0121] In one embodiment, a typical operation process description of a cremation task is provided: The operator connects the cremation equipment and multi-source sensors to the integrated cabinet. After confirming that the cabinet's power supply and communication are normal, the operator starts the cremation combustion control software on the upper display. The operator enters the body weight, freezing time, and target negative pressure, and selects manual or intelligent mode according to the site conditions. In manual mode, the operator can first complete the cremation equipment check, ignition preparation, and fan pre-adjustment; the system only displays data and warnings.

[0122] Once the site conditions are met, the operator switches to intelligent mode or confirms entry into intelligent control. The system checks whether the blower, induced draft fan, and furnace pressure meet the start-up interlock requirements. If the start-up interlock conditions are not met, such as the blower or induced draft fan not being turned on, or the furnace pressure being higher than -10Pa, the software interface immediately issues an alarm, does not establish a combustion session, and prompts the user to increase the induced draft fan operating parameters or perform a manual check. If the start-up interlock conditions are met, the system establishes a combustion session, records the start time, and begins executing intelligent decisions in 20-second cycles.

[0123] During the ignition phase, the main combustion chamber temperature is low, the fuel valve is forcibly opened, and the blower adjustment is optimized by algorithm within a safe range. During the auto-ignition phase, the system allows intelligent algorithms to find a better combination among the fuel valve, blower, and induced draft fan. When the prediction results show that a certain variable combination is better in terms of pollutant gas flow, that variable combination obtains higher fitness. In the tail end phase, the system pays attention to the drop in peak pollutant gas flow and combustion duration, maintaining necessary afterburning and avoiding premature fuel cut-off that would lead to incomplete combustion.

[0124] During execution, the software interface continuously displays control suggestions and warning logs. In the event of data loss, model anomalies, or PLC write failure, the system prioritizes conservative handling and prompts for manual intervention. If the operator deems special handling necessary, they can switch to manual mode at any time; in manual mode, the algorithm thread continues to run and provide suggestions, but automatic PLC writing is suppressed. At the end of the task, after the blower has been continuously shut down a certain number of times, the system resets the session and archives the logs.

[0125] In cleaning or maintenance scenarios, the integrated cabinet device retains data monitoring and log recording functions, but the automatic control logic should be constrained by the manual mode. Operators can select the cleaning frequency according to the cremation equipment maintenance procedures, and the system checks whether the frequency is valid, whether there are mode conflicts, and whether the PLC is in a writable state. This mechanism prevents intelligent algorithms and manual maintenance actions from overlapping.

[0126] In communication interruption scenarios, the software interface first displays the connection error and the last valid sampling time; the algorithm thread no longer generates automatic commands based on expired data. If the PLC write fails, the system records the reason for the failure and the control word value, and prompts manual inspection of the gateway, cables, coil addresses, and PLC operating status. This enables on-site maintenance personnel to quickly locate the fault.

[0127] The above process can be illustrated in the table below: Table 6 Control Table for Cremation Task Operation Phase

[0128] In one embodiment, the effectiveness of the method of this application is compared with that of the conventional method: Traditional methods typically rely on operators manually adjusting oil valves, blowers, and induced draft fans based on main combustion chamber temperature, flame condition, experience, time, and smoke extraction. The advantage of this method is its flexibility, but its disadvantages include significant differences between operators and a tendency to adopt a conservative oil injection strategy to avoid incomplete cremation. While fixed-threshold PLC control is more stable than purely manual control, the threshold values ​​often fail to cover a wide range of factors, including body weight, freezing time, preheating status, and furnace pressure fluctuations.

[0129] This application integrates data acquisition, model screening, intelligent algorithms, PLC control, and manual verification into a single integrated cabinet, improving control quality on the same hardware basis. Especially during the auto-ignition stage, if sensors indicate that temperature and pollutant gas changes meet the conditions for complete combustion, the system can reduce unnecessary oil valve maintenance; if furnace pressure and oxygen levels deviate from targets, the system prioritizes correction via the blower rather than blindly increasing fuel oil. Thus, it achieves a fuel consumption reduction target of approximately 10% while ensuring safety and complete combustion.

[0130] The reduction in fuel consumption can be verified by conducting a controlled experiment using the same model of cremator, similar body weight, similar freezing time, and similar environmental conditions. The control group used traditional manual or fixed threshold control, while the experimental group used the intelligent mode control of the integrated cabinet device described in this application. Fuel consumption, body weight, combustion time, flow rate and concentration of key pollutants, and logs were recorded for each mission. The fuel consumption reduction rate was calculated using the aforementioned formula and statistically compared to verify the fuel-saving effect of this application.

[0131] In practical verification, simple comparisons of absolute fuel consumption for a single mission should be avoided. Instead, body mass, freezing time, furnace preheating status, ambient temperature, cremation time, and maintenance status should be used as screening criteria. Missions with obvious anomalies should be marked in the log and analyzed as separate samples to avoid individual outliers misleading the overall conclusions.

[0132] If the sample size is sufficient, the tasks can be grouped by weight range, freezing time range, and furnace preheating state, and the average fuel consumption, fuel consumption per unit mass, peak value of major pollutants, and combustion time can be calculated separately. This grouping method helps to demonstrate that this application does not depend on individual special operating conditions, but rather exhibits stable energy saving and control consistency across multiple types of field tasks.

[0133] The specific effects can be compared in the table below: Table 7 Comparison of the effects of the method in this application and traditional methods

[0134] In one embodiment, to ensure repeatability of this application during field deployment, a set of preferred commissioning parameters is provided. These parameters are for illustrative purposes only, and their specific values ​​can be adjusted according to the cremation equipment model, sensor range, PLC address table, and on-site safety regulations.

[0135] Table 8 Preferred Debugging Parameters

[0136] During on-site commissioning, first confirm that the sensor address and range mapping are correct, then confirm that the address written to the PLC coil is consistent with the actuator's action. For fan height adjustment, fan height adjustment, and oil valve switching actions, single-point tests should be performed under manual control conditions to confirm that the response direction is correct before allowing intelligent mode to participate in control.

[0137] Before deploying the model, offline validation of the classification and regression models should be performed using historical cremation data, with a focus on abnormal samples, extreme weight samples, sensor drift samples, and samples from the final stage. If significant deviations are found in the model predictions, the model should first be run using safety rules and manual verification mode, and the intelligent control ratio should be increased only after the model is updated.

[0138] After the system is put into use, the CSV archive file, runtime_state.json status file, and warning log should be checked regularly. If a certain type of warning occurs frequently, the corresponding sensor, PLC address, communication gateway, or field operation procedure should be investigated first, rather than simply disabling the warning function.

[0139] In one embodiment, to further illustrate the completeness of this application as an integrated cabinet device, the data archiving, model updating, and field maintenance mechanisms are further described. Although the above mechanisms do not directly change the immediate operation of the oil valve, blower, or induced draft fan, they affect the long-term stability, traceability, and model effectiveness of the system, and therefore can be considered as a preferred component of this application.

[0140] For data archiving, the system saves the sensor values, coil status, algorithm suggestions, final control words, manual / intelligent modes, warning content, and PLC write results obtained from each sampling to a CSV file or equivalent database for the day. For cremation tasks, a task summary can also be created according to the session number, recording the task start time, end time, stage switching time, spontaneous combustion start time, cause of tail-end determination, cumulative fuel consumption, peak values ​​of major pollutants, and anomaly logs.

[0141] For model updates, the system can employ a combination of offline training and on-site validation. During offline training, historical task data is used to train classification and regression models. During on-site validation, the system initially runs in manual or suggestion mode, displaying suggestions without automatically writing them to the PLC. Once the model output meets the consistency requirements with operator experience and preset constraints, the automatic control ratio in intelligent mode is gradually increased.

[0142] For on-site maintenance, the integrated cabinet device should retain functions such as sensor calibration, communication testing, coil single-point testing, model file version viewing, log export, and operation status cleanup. After replacing sensors, adjusting PLC addresses, or updating model files, maintenance personnel should verify the functionality through debug mode to avoid old addresses, old models, or old status files affecting new tasks.

[0143] Table 9. Equipment Maintenance Examples

[0144] During long-term operation, the system can also generate monthly or quarterly statistics based on archived data, such as average fuel consumption per body, fuel consumption per unit mass, duration of each stage, frequency of warnings, number of PLC write failures, and number of sensor offline occurrences. This statistical analysis does not alter the real-time control logic of the cremation process, but it can provide a basis for equipment maintenance, model retraining, and operational procedure optimization.

[0145] The aforementioned data archiving, model updating, and maintenance mechanisms further illustrate that the integrated cabinet device of this application is not an ordinary control box, but a field intelligent device with display, calculation, communication, control, feedback, and traceability capabilities. Even if CSV is replaced with a database, the Web interface is replaced with an HMI, or Modbus is replaced with other industrial communication protocols in future implementations, as long as the core structure of centralized processing and closed-loop control of the integrated cabinet is retained, it still constitutes a reasonable extension of the concept of this application.

[0146] In one embodiment, the technical terms used in this application are uniformly interpreted as follows: Table 10 Glossary of Terms

[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for controlling combustion in cremation, characterized in that, include: Obtain multiple control variables under the current operating conditions of the cremation equipment; Based on the aforementioned multiple control variables, multiple combinations of candidate variables are constructed; Based on the trained classification model and the trained regression model, the optimal variable combination is selected from the multiple candidate variable combinations; Based on the optimal combination of variables, the oil valves and blowers of the cremation equipment are controlled.

2. The cremation combustion control method according to claim 1, characterized in that, The multiple control variables include multiple continuous variables and multiple switching variables; The construction of multiple candidate variable combinations based on the multiple control variables includes: The continuous variable is perturbed to obtain the perturbed continuous variable; The switch variable is flipped to obtain the flipped switch variable; Multiple candidate variable combinations are constructed based on the continuous variables before and after the perturbation and the switching variables before and after the flip.

3. The cremation combustion control method according to claim 2, characterized in that, The optimal variable combination is selected from the multiple candidate variable combinations based on the trained classification model and the trained regression model, including: The multiple candidate variables are combined and input into the trained classification model to obtain the fuel consumption status of the multiple candidate variable combinations output by the trained classification model. From the multiple candidate variable combinations, select multiple target variable combinations that indicate low fuel consumption. The multiple target variables are combined and input into the trained regression model to obtain the predicted flow rate of pollutant gas output by the multiple target variables combination of the trained regression model; Based on the predicted flow rate of the polluting gas, the fitness of the combination of the multiple target variables is obtained; Based on the fitness, the optimal combination of variables is selected from the multiple combinations of target variables.

4. The cremation combustion control method according to claim 3, characterized in that, The step of selecting the optimal variable combination from the multiple combinations of target variables based on the fitness includes: Several target variable combinations are extracted from the multiple target variable combinations according to the fitness probability to obtain multiple first variable combinations; By swapping the continuous variables of some of the multiple first variable combinations, multiple second variable combinations are obtained; Resample the continuous variables of some of the multiple second variable combinations, and flip the on / off variables of some of the multiple second variable combinations to obtain multiple third variable combinations; If the preset number of iterations has not been reached, the multiple combinations of third variables are used as new combinations of candidate variables, and the process of inputting the multiple combinations of candidate variables into the trained classification model is repeated until the preset number of iterations is reached. The combination of third variables with the highest fitness among the multiple combinations of third variables at this time is determined as the optimal variable combination.

5. The cremation combustion control method according to claim 1, characterized in that, The control of the oil valves and blowers of the cremation equipment based on the optimal variable combination includes: Under preset constraints, the oil valves and blowers of the cremation equipment are controlled based on the differences between the multiple control variables and the optimal combination of variables.

6. The cremation combustion control method according to claim 5, characterized in that, The multiple control variables include stage variables; The preset constraints are based on the flame conditions, main combustion chamber temperature, combustion duration, furnace pressure, oxygen content, and the cremation stage corresponding to the stage variables of the cremation equipment, and are applied to the oil valves and blowers of the cremation equipment. The cremation stage adopts a one-way latching mechanism.

7. The cremation combustion control method according to claim 1, characterized in that, After controlling the oil valves and blowers of the cremation equipment based on the optimal variable combination, the process includes: Return to obtain multiple control variables under the current operating conditions of the cremation equipment, until the oil valve and blower of the cremation equipment are controlled again based on the optimal combination of variables.

8. A cremation combustion control integrated cabinet device, characterized in that, include: The communication and PLC area is used to: acquire multiple control variables under the current operating conditions of the cremation equipment; The industrial computing and service area is used for: constructing multiple candidate variable combinations based on the multiple control variables; and selecting the optimal variable combination from the multiple candidate variable combinations based on the trained classification model and the trained regression model. The communication and PLC area is also used to control the oil valves and blowers of the cremation equipment based on the optimal combination of variables.

9. The integrated cremation combustion control cabinet device according to claim 8, characterized in that, Also includes: The execution and power supply area is used to: supply power to the equipment inside the cabinet, and provide communication lines for the communication and PLC area to control the oil valves and fans of the cremation equipment; The display and interaction area is used to display various parameters for cremation combustion control.

10. The integrated cremation combustion control cabinet device according to claim 8, characterized in that, The communication and PLC area includes: The Modbus gateway is used to: acquire multiple control variables under the current operating conditions of the cremation equipment; PLC control hardware is used to control the oil valves and blowers of the cremation equipment based on the optimal combination of variables.

Images