Automatic garbage combustion control method and system based on multi-mode perception

Abstract

The invention discloses an automatic garbage combustion control method and system based on multi-mode perception, and relates to the technical field of combustion control. The method comprises the following steps: collecting multi-mode operation state data of the garbage incinerator in real time through a multi-mode sensing module; inputting the multi-mode operation state data into an intelligent decision-making model for garbage combustion state prediction, and generating combustion trend information; comparing the combustion trend information with a preset combustion control target, and generating an optimization instruction according to a comparison result; based on the optimization instruction, optimizing the combustion air amount, the fuel supply amount and the fire grate motion parameters according to the combustion trend information, and generating an optimal combustion parameter group; and controlling a combustion fan, a fuel supply device and a fire grate driving mechanism in the garbage incinerator according to the optimal combustion parameter group. The technical problem that in the prior art, an incinerator is not accurate in control, adjustment response lags behind, and consequently the garbage combustion efficiency is low is solved, and the technical effect of improving the garbage combustion efficiency is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of combustion control, and in particular to a method and system for automatically controlling garbage combustion based on multimodal perception. Background Art

[0002] As an effective method of waste disposal, waste incineration not only significantly reduces waste volume but also achieves resource utilization through energy recovery. However, the waste incineration process is affected by a variety of factors, including the complex and diverse composition of waste, large fluctuations in feed volume, and frequent changes in the furnace combustion environment, which poses significant challenges to the stability and efficiency of the combustion process. Traditional waste incinerators often rely on single sensor data or manual experience for combustion regulation. The control methods are simple and the response is delayed, making it difficult to achieve comprehensive perception and refined control of the combustion state. This results in low combustion efficiency, high energy consumption, and difficult to effectively control harmful gas emissions. Summary of the Invention

[0003] The present application provides a method and system for automatic garbage combustion control based on multimodal perception, which solves the technical problems in the prior art of incinerator control being inaccurate and regulation response being delayed, resulting in low garbage combustion efficiency.

[0004] The first aspect of the present application provides a method for controlling automatic garbage combustion based on multimodal perception, the method comprising: The multi-mode operating status data of the waste incinerator are collected in real time through the multi-modal sensing module, including at least temperature detection data, material layer thickness detection data, flue gas detection data and feed amount detection data; the multi-mode operating status data are input into the intelligent decision-making model to predict the waste combustion state and generate combustion trend information; the combustion trend information is compared with the preset combustion control target, and an optimization instruction is generated according to the comparison result; based on the optimization instruction, the combustion air amount, fuel supply amount and grate movement parameters are optimized according to the combustion trend information to generate an optimal combustion parameter group; the combustion fan, fuel supply device and grate drive mechanism in the waste incinerator are controlled by the optimal combustion parameter group.

[0005] The second aspect of the present application provides a garbage automatic combustion control system based on multimodal perception, the system comprising: Data acquisition component: collects multi-mode operating status data of the waste incinerator in real time through the multi-modal perception module, including at least temperature detection data, material layer thickness detection data, flue gas detection data and feed amount detection data; prediction component: inputs the multi-mode operating status data into the intelligent decision-making model to predict the waste combustion state and generate combustion trend information; comparison component: compares the combustion trend information with the preset combustion control target and generates optimization instructions according to the comparison result; optimization component: based on the optimization instructions, optimizes the combustion air amount, fuel supply amount and grate movement parameters according to the combustion trend information to generate an optimal combustion parameter group; control component: controls the combustion fan, fuel supply device and grate drive mechanism in the waste incinerator with the optimal combustion parameter group.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: First, a multimodal sensing module is used to collect multimodal operating status data of the waste incinerator in real time, including at least temperature detection data, material layer thickness detection data, flue gas detection data, and feed rate detection data. Next, the multimodal operating status data is input into an intelligent decision-making model to predict the waste combustion state and generate combustion trend information. Furthermore, the combustion trend information is compared with the preset combustion control target, and an optimization instruction is generated based on the comparison result. Then, based on the optimization instruction, the combustion air volume, fuel supply volume, and grate movement parameters are optimized according to the combustion trend information to generate an optimal combustion parameter group. Finally, the combustion fan, fuel supply device, and grate drive mechanism in the waste incinerator are controlled with the optimal combustion parameter group. This solves the technical problem of inaccurate incinerator control and delayed adjustment response in the prior art, which leads to low waste combustion efficiency, and achieves the technical effect of improving waste combustion efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0008] Figure 1 A schematic diagram of the flow chart of the automatic garbage combustion control method based on multimodal perception provided in an embodiment of the present application; Figure 2 Schematic diagram of the structure of the automatic garbage combustion control system based on multimodal perception provided in the embodiment of the present application.

[0009] Description of reference numerals: data collection component 11 , prediction component 12 , comparison component 13 , optimization component 14 , control component 15 . DETAILED DESCRIPTION

[0010] This application solves the technical problems in the prior art of incinerator control being inaccurate and regulating response being delayed, resulting in low garbage combustion efficiency, by providing a garbage automatic combustion control method and system based on multimodal perception.

[0011] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only some of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0012] It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products or devices.

[0013] Example 1, as Figure 1 As shown, the present application provides a method for controlling automatic garbage combustion based on multimodal perception, wherein the method includes: The multimodal sensing module collects multimodal operating status data of the waste incinerator in real time, including at least temperature detection data, material layer thickness detection data, flue gas detection data and feed quantity detection data.

[0014] Multimodal sensing modules deployed in different locations of the waste incinerator collect real-time data on the incinerator's operating status, generating multimodal operating status data including temperature, bed thickness, flue gas, and feed rate data. Temperature data reflects the heat distribution in different combustion zones; bed thickness data reflects the current bed distribution characteristics; flue gas data reflects the current combustion completeness and pollutant emissions; and feed rate data reflects the waste supply rate.

[0015] Furthermore, the multimodal perception module includes at least a temperature sensor array, a pressure and differential pressure sensor array, a gas composition detector and a weight sensor; wherein, the temperature sensor array is distributed at various positions of the furnace of the waste incinerator, and collects the temperature distribution of each area of ​​the furnace in real time to generate temperature detection data; the pressure and differential pressure sensor array is distributed on the combustion grate surface, and is used to collect the pressure difference above and below the grate surface and the air volume on the grate surface, judge the material layer thickness on the grate surface, and generate material layer thickness detection data; the gas composition detector is distributed at the flue gas exhaust port, and is used to detect the gas component content in the flue gas and generate flue gas detection data; the weight sensor is distributed at the feed port, and is used to detect the amount of garbage feed and generate feed amount detection data.

[0016] The multimodal sensing module includes at least a temperature sensor array, a pressure and differential pressure sensor array, a gas composition detector and a weight sensor, which are used to sense and collect multi-dimensional operating status information in the waste incinerator in real time and comprehensively. Specifically, the temperature sensor array is evenly distributed in multiple key positions of the incinerator furnace, including the upper layer, the middle layer and the area near the grate, to continuously monitor the temperature changes in different combustion areas in the furnace, and generate temperature detection data that characterizes the heat distribution state of the furnace. This data can be used to determine whether the local combustion is sufficient and whether the temperature meets the standard; the pressure and differential pressure sensor array is installed above and below the combustion grate, and by collecting the static pressure difference above and below the grate surface and the combustion air flow rate through the grate, the degree of obstruction of the material to the exhaust gas flow is calculated, and then combined with the set material layer thickness calculation model, the thickness distribution of the current material layer is inferred. The gas composition detector is arranged at the flue gas emission pipe or chimney outlet of the incineration system to monitor the concentrations of carbon monoxide (CO), carbon dioxide (CO2), oxygen (O2), nitrogen oxides (NOx), sulfur dioxide (SO2) and other gas components in the flue gas in real time, and output flue gas detection data to evaluate the degree of combustion and the level of pollutant emissions; the weight sensor is arranged at the bottom of the feed channel or feed chute to detect the change in garbage weight per unit time in real time during the garbage placement or feeding process, thereby generating feed quantity detection data.

[0017] Furthermore, the pressure and differential pressure sensor array is connected to a material layer thickness calculation unit, and the material layer thickness calculation unit is embedded with a material layer thickness calculation formula. The material layer thickness is calculated by reading the real-time sensing data of the pressure and differential pressure sensor array to generate the material layer thickness detection data.

[0018] The pressure and differential pressure sensor array is electrically connected to the material layer thickness calculation unit to realize real-time calculation and evaluation of the thickness of the garbage material layer on the grate surface.

[0019] The pressure and differential pressure sensor array includes multiple high-sensitivity pressure sensors arranged above and below the grate, which are used to collect the static pressure of the airflow above the grate surface, the air supply pressure of the wind chamber below the grate, and the upper and lower pressure difference data respectively; the material layer thickness calculation unit is an embedded edge computing module, which has a preset material layer thickness calculation formula. Specifically, the material layer thickness calculation unit periodically reads the original sensor data of the pressure and differential pressure sensor array, and performs thickness estimation operations based on this, outputs the thickness information of the garbage accumulation layer in each area of ​​the current grate, and then generates spatially distributed continuous material layer thickness detection data. The material layer thickness can be calculated based on the resistance encountered by the primary air passing through the combustion. The material layer thickness calculation formula is as follows: ;in, is the thickness of the material layer, A is the grate area, is the primary wind density, is the air chamber pressure, furnace pressure, PAL is the primary air flow rate in the wind chamber, PAT is the primary air temperature in the wind chamber, and k is a constant.

[0020] The multi-mode operation status data is input into an intelligent decision-making model to predict the garbage combustion status and generate combustion trend information.

[0021] Multi-source operational status data, including temperature, bed thickness, flue gas, and feed rate data, collected and preprocessed in real time by the multimodal sensing module, is fed into a constructed intelligent decision-making model. This model is designed based on a deep learning architecture, preferably employing neural network models with spatiotemporal modeling capabilities, such as graph neural networks (GNNs), convolutional neural networks (CNNs), or long short-term memory networks (LSTMs). A graph-based model, based on the location of sensor node locations and their distribution within the incinerator's spatial structure, is constructed. Multiple rounds of node state propagation and updates are performed via the graph neural network to extract high-dimensional feature representations of each region, enabling dynamic learning and trend modeling of the waste combustion state. After supervised training using multiple rounds of historical combustion data, the model possesses intelligent perception and trend prediction capabilities for the current combustion state. The intelligent decision-making model outputs combustion trend information corresponding to the current moment, including indicators such as whether combustion is unstable, whether there are drastic temperature fluctuations, whether the bed thickness is abnormal, and whether flue gas composition exceeds limits. This provides a quantitative and dynamic trend reference for subsequent control optimization.

[0022] Furthermore, the multi-mode operation status data is input into an intelligent decision-making model to predict the waste combustion status and generate combustion trend information, including: A graph structure is constructed based on the data collection positions corresponding to the multi-mode operating status data; based on the graph structure, iterative training of the graph neural network is performed to generate the intelligent decision model; the intelligent decision model is called to analyze the multi-mode operating status data and output the combustion trend information.

[0023] Firstly, according to the physical installation locations of various sensors in the multi-mode operation status data and the corresponding collection data types, a graph structure reflecting the operation space structure of the waste incinerator is constructed. Each node of the graph structure corresponds to a sensor unit or an operation parameter collection point. The node feature is its corresponding operation status data (such as temperature value, pressure difference value, gas concentration value, feed amount, etc.). The connection relationship of the edge is defined according to the proximity of the physical location of the sensor, the functional relevance or the airflow path logic, forming a graph topology that can reflect the spatiotemporal interaction characteristics in the furnace; then, based on the constructed graph structure model, the graph neural network is used as the core architecture to perform multiple rounds of node state propagation and An iterative training process of feature fusion, in which the training process uses historical multi-mode operating status data and its corresponding combustion trend labels as training samples, continuously optimizes model parameters through the node feature update mechanism, and finally generates an intelligent decision-making model with trend prediction capabilities; in the actual operation process, the current multi-mode operating status data collected in real time is used as the graph structure node input, and the trained intelligent decision-making model is called for forward reasoning to output the prediction results of the current incineration status, including the combustion temperature change trend, material layer stability change trend, pollutant emission change trend, etc. in the short future time, collectively referred to as combustion trend information, which provides a dynamic adjustment basis for combustion control.

[0024] Furthermore, based on the graph structure, iterative training of the graph neural network is performed to generate the intelligent decision model, including: Collect a first historical state prediction sample set and a second historical state prediction sample set; construct a graph neural network model based on the graph structure, use the first historical state prediction sample set to perform node feature update training, and generate a node feature update mechanism, wherein node feature update means that each node obtains information from neighboring nodes, and then updates its own state in combination with the node's own information, and the training parameters are learning parameters based on the neighboring nodes updating their own states; based on the node feature update mechanism and the graph neural network model, use the second historical state prediction sample set to perform combustion trend prediction training after performing node feature update, and use the node update data to generate the intelligent decision-making model.

[0025] First, historical multi-modal operating status data for training is collected to construct two sample sets, namely the first historical state prediction sample set and the second historical state prediction sample set. The first historical state prediction sample set is mainly used for learning the node feature propagation mechanism, and the second historical state prediction sample set is mainly used for training and optimization of the combustion trend output layer. Then, based on the previously constructed graph structure, a graph neural network (GNN) model is established. Each node in the model represents a state collection point, and its initial feature is the corresponding historical state data (such as temperature, material layer thickness, flue gas concentration, etc. at a certain moment). The first historical state prediction sample set is used to train the node feature propagation process, and each node is enabled to obtain state information from its adjacent nodes through the information aggregation mechanism. , and updates its features in combination with its original state. This process is called node feature update. The training goal is to optimize the state representation of each node so that it better reflects the joint dynamic characteristics of the local area and its neighborhood. The core parameters of node feature update include the neighbor node weight distribution function, information aggregation function, and activation function. Its update strategy is determined based on the adjacent edge weight, inter-node distance, or functional relevance. After the node feature update mechanism is constructed, the second historical state prediction sample set is further used for trend prediction training. That is, after the node state update is completed, the updated node features are used as input, and supervised training of trend labels is performed through the output layer. The labels may include combustion trend indicators such as furnace temperature trend changes (increasing, decreasing, or stable), material layer fluctuations, and flue gas emission changes. Finally, through multiple rounds of iterative optimization, the training of the intelligent decision-making model is completed, and the model is used to predict the current combustion trend information in real time, providing a dynamic decision-making basis for the optimization of combustion control strategies.

[0026] The combustion trend information is compared with a preset combustion control target, and an optimization instruction is generated according to the comparison result.

[0027] The combustion control targets preset in the system include control indicators in multiple dimensions, such as the stable range of furnace temperature, the reasonable range of material layer thickness, the safe concentration threshold of pollutants such as CO or NOx in the flue gas, the matching relationship between feed and combustion air, and other control benchmark data; the combustion trend information output by the intelligent decision-making model is used as a prediction reference for the current state, and is compared and analyzed item by item with each indicator in the combustion control target to determine whether the current predicted trend meets or deviates from the preset target. If the combustion trend information fully meets all control targets, the existing control parameters remain unchanged; if any trend prediction result exceeds the preset tolerance range of the control target, then according to the direction and degree of the deviation, combined with the priority setting of the control target, a corresponding optimization instruction is generated to prompt or drive the system to perform parameter adjustment.

[0028] Furthermore, the combustion trend information is compared with a preset combustion control target, and an optimization instruction is generated according to the comparison result, including: Determine whether the combustion trend information meets the preset combustion control target; if not, generate the optimization instruction.

[0029] The system determines the combustion trend information output by the intelligent decision-making model at the current moment. The combustion trend information includes at least the direction of furnace temperature change, the fluctuation trend of the material layer thickness, the change trend of key pollutant concentrations in the flue gas, and the dynamic evolution characteristics of the feed and combustion adaptation relationship in the future period. The system matches and judges the above trend information item by item according to the preset combustion control objectives. The combustion control objectives include but are not limited to: the stability of the furnace temperature within the target temperature range, the duration of the material layer thickness in the optimal combustion range, the continuous probability of pollutant emissions not exceeding the set threshold, and the target lower limit of the thermal efficiency of the combustion reaction. If the combustion trend information fully meets all the preset combustion control objectives, the system maintains the current operating parameters unchanged. If any trend indicator deviates from the corresponding control target (for example, it is predicted that the temperature will continue to drop or CO emissions will exceed the threshold), the system triggers the target deviation judgment mechanism, determines the deviation type and correction direction, and generates optimization instructions accordingly. The optimization instructions are used to drive the subsequent parameter optimization process, indicate the category of combustion variables to be adjusted and the direction of the adjustment strategy, provide a decision basis for the subsequent generation of the optimal combustion parameter group, and realize active adjustment and closed-loop optimization control of the incinerator operation status.

[0030] Based on the optimization instruction, the combustion air quantity, fuel supply quantity and grate movement parameters are optimized according to the combustion trend information to generate an optimal combustion parameter group.

[0031] After receiving the optimization instruction generated by the combustion trend deviation analysis result, the system first extracts the predicted dynamic data related to the combustion air supply, garbage feeding acceleration rate and grate operation status in the current combustion trend information; then, it calls the historical combustion control data that matches the trend information time window to construct a historical combustion parameter sample set, which includes the combustion air volume, fuel supply rate, grate operation cycle and beat under different combinations and the corresponding combustion response effect; then, based on the deviation direction and adjustment requirements specified in the optimization instruction, the system uses a multi-objective optimization algorithm (such as a multi-objective optimization algorithm based on non-dominated sorting) to optimize the combustion parameters. The NSGA-II (optimization algorithm) was developed. The fitness evaluation was based on satisfying control objectives (such as combustion temperature stability, material layer distribution uniformity, and flue gas emission compliance rate). A fitness function group containing multiple sub-objective evaluators was constructed, and multiple rounds of evolutionary optimization iterations were performed on the historical combustion parameter sample set. Through fitness evaluation and non-dominated sorting of each parameter combination, inferior solutions were eliminated and superior solutions were retained. The results of multiple rounds of optimization were continuously screened and integrated, and finally a parameter solution set at the non-dominated optimal frontier was extracted. The optimal combustion parameter group was selected from it, including the combination of the optimal combustion air supply, the optimal fuel supply rate, and the optimal grate motion parameters.

[0032] Furthermore, based on the optimization instruction, the combustion air quantity, fuel supply quantity and grate movement parameters are optimized according to the combustion trend information to generate an optimal combustion parameter set, including: Collect a historical combustion parameter set corresponding to the combustion trend information; construct a fitness evaluation mechanism based on the preset combustion control target, including several sub-target evaluators; use the several sub-target evaluators to perform iterative optimization based on non-dominated sorting on the historical combustion parameter set to generate the optimal combustion parameter group.

[0033] Based on the current combustion trend information, the system calls the multimodal perception data and control parameter records in the historical operation cycle associated with it, and collects the corresponding historical combustion parameter set. The historical combustion parameter set covers the combustion air flow, fuel supply rate, grate operation rhythm, primary / secondary air distribution ratio, etc. under different combustion environments in the historical period, and is accompanied by corresponding combustion response results (such as furnace temperature stability, material layer thickness uniformity, pollutant concentration in flue gas, etc.); a fitness evaluation mechanism is constructed based on the preset combustion control targets, which generally include but are not limited to: furnace temperature stability target, thermal efficiency target, emission compliance target and garbage burnout rate target. Several sub-target evaluators are set in the fitness evaluation mechanism, and each evaluator evaluates and scores a certain sub-target. , forming a multi-objective fitness function family; the system uses several sub-objective evaluators to calculate the fitness of each group of parameters in the historical combustion parameter set, and sorts and screens the multi-objective fitness through non-dominated sorting (such as NSGA-II algorithm) in each round of iteration to determine the parameter combination on the Pareto front in the current solution space, that is, the solution that performs better in multiple evaluation dimensions at the same time, and eliminates the dominated solutions degraded by multiple objectives; performs genetic crossover and mutation operations on the retained non-dominated solutions, expands the candidate parameter space, and performs sub-objective evaluation and non-dominated screening again; iterates until the preset convergence condition or iterative rounds are reached; finally, selects the parameter group with the best comprehensive fitness from the obtained global non-dominated front solution set as the optimal combustion parameter group under the current incineration state.

[0034] Furthermore, the fitness evaluation mechanism is used to perform iterative optimization based on non-dominated sorting on the historical combustion parameter set to generate the optimal combustion parameter group, including: Step 1: Use the several sub-target evaluators to perform fitness evaluation on each solution in the historical combustion parameter set to generate each fitness set; Step 2: Based on the each fitness set, calculate the fitness difference of any two solutions using the same sub-target evaluator to generate a fitness difference distribution; Step 3: Call the non-dominated stratification mechanism to perform dominance stratification on each solution according to the fitness difference distribution and construct a dominance relationship structure diagram; perform screening of the dominance relationship structure diagram according to a preset non-dominance level threshold to generate a screening population, repeat steps 1 to 3 after parameter expansion of the screening population, continue to iterate until a preset number of iterations is reached, globally fuse the dominance relationship structure diagram obtained by the iteration, extract the non-dominated optimal solution from the fused dominance relationship structure diagram to generate the optimal combustion parameter group.

[0035] Based on multiple preset sub-target evaluators (such as furnace temperature fluctuation amplitude evaluator, burnout efficiency evaluator, NO x Emission evaluator, thermal energy utilization efficiency evaluator, etc.) performs fitness evaluation on each set of candidate parameters (i.e., a solution) in the historical combustion parameter set, and obtains multiple fitness scores to form the fitness set of the solution. Each fitness score represents the performance of the parameter group under a specific evaluation target. The fitness sets of all candidate solutions are compared pairwise, and the fitness difference between any two solutions is calculated under the same sub-target evaluation dimension to obtain a fitness difference distribution matrix to measure the degree of dominance and relative performance differences between the solutions; based on the above difference distribution, the non-dominated hierarchical mechanism is called, and according to the Pareto optimal principle, the dominance relationship between the candidate solutions is judged, and the candidate solutions are stratified according to the number of times they are dominated, and a dominance relationship structure diagram is constructed. The first layer is the optimal solution set that is not dominated by any other solution, and the second layer, third layer, and other inferior solution sets are constructed in this way. The dominance relationship structure diagram is screened based on a preset non-dominated level threshold (for example, retaining the first two or three layers) to generate a screening population for this round of optimization; then, parameter expansion operations are performed on the candidate solutions in the screening population, and the crossover mutation mechanism in the genetic algorithm (such as simulated binary crossover SBX, non-uniform mutation, etc.) is used to generate a new set of candidate solutions, which is used as the input for the next round of iteration, and steps one to three are repeated; after each round of iteration, the system fuses the current dominance relationship structure diagram with the historical structure diagrams of the previous rounds, retaining the global dominance relationship, and continuously optimizing the search space until the preset number of iterations or convergence criterion is met; finally, the optimal solution of the non-dominated frontier is extracted from the fused global dominance relationship structure diagram as the optimal combustion parameter group.

[0036] Furthermore, the non-dominated hierarchical mechanism includes complete dominance of the several sub-goal evaluators and advanced dominance of at least one sub-goal evaluator, wherein advanced dominance means that the fitness difference between any two solutions on at least one sub-goal evaluator is higher than a preset threshold.

[0037] Complete dominance means that one solution is superior to another solution in all sub-goal evaluator dimensions; while high-level dominance means that in at least one sub-goal evaluator dimension, its fitness value has a significant advantage over the fitness value of another solution that exceeds a preset numerical threshold.

[0038] Specifically, if there are n sub-goal evaluators, for any two candidate solutions A and B, let the fitness function of the i-th evaluator be f i , then: when , (assuming all objectives are minimization problems), then solution A is said to completely dominate solution B; when there is at least one evaluator , making ,(in is the preset threshold of the j-th sub-goal evaluator), then solution A is said to have a high-level dominance over solution B.

[0039] In each round of optimization, the system prioritizes solutions that satisfy both complete dominance and high-level dominance. When constructing the dominance relationship structure, it prioritizes solutions with higher non-dominance levels. By integrating complete dominance and high-level dominance rules, the system not only accelerates the convergence of the solution space but also enhances the accuracy and reliability of finding balanced performance solutions within a multi-objective evaluation system.

[0040] The combustion-supporting fan, fuel supply device and grate drive mechanism in the waste incinerator are controlled using the optimal combustion parameter group.

[0041] The optimal combustion parameter set includes combustion air supply parameters, fuel dosing parameters, and grate operating parameters, which control the combustion fan, fuel supply device, and grate drive mechanism, respectively. The combustion fan adjusts the air volume and pressure based on the combustion air supply parameters to ensure sufficient and uniform oxygen supply during different combustion stages. The fuel supply device controls the waste feed rate and feeding interval based on the fuel dosing parameters to achieve a fuel supply that matches the actual calorific value required. The grate drive mechanism adjusts the grate movement frequency, propulsion speed, and propulsion mode based on the grate operating parameters to ensure uniform distribution and sufficient combustion of the waste within the incinerator.

[0042] Through precise linkage control of the above three types of key execution devices, the system makes the combustion conditions stable, the furnace temperature reasonably controlled, and the emission concentration controlled, thereby realizing intelligent and dynamic regulation of the garbage combustion process, effectively improving the incineration efficiency and reducing the environmental impact.

[0043] In summary, the embodiments of the present application have at least the following technical effects: First, a multimodal sensing module is used to collect multimodal operating status data of the waste incinerator in real time, including at least temperature detection data, material layer thickness detection data, flue gas detection data, and feed rate detection data. Next, the multimodal operating status data is input into an intelligent decision-making model to predict the waste combustion state and generate combustion trend information. Furthermore, the combustion trend information is compared with the preset combustion control target, and an optimization instruction is generated based on the comparison result. Then, based on the optimization instruction, the combustion air volume, fuel supply volume, and grate movement parameters are optimized according to the combustion trend information to generate an optimal combustion parameter group. Finally, the combustion fan, fuel supply device, and grate drive mechanism in the waste incinerator are controlled with the optimal combustion parameter group. This solves the technical problem of inaccurate incinerator control and delayed adjustment response in the prior art, which leads to low waste combustion efficiency, and achieves the technical effect of improving waste combustion efficiency.

[0044] Example 2 is based on the same inventive concept as the method for controlling automatic garbage combustion based on multimodal perception in the previous embodiment. Figure 2 As shown, the present application provides a garbage automatic combustion control system based on multimodal perception, wherein the system includes: Data acquisition component 11: collects multi-mode operating status data of the waste incinerator in real time through the multi-modal sensing module, including at least temperature detection data, material layer thickness detection data, flue gas detection data and feed amount detection data; prediction component 12: inputs the multi-mode operating status data into the intelligent decision-making model to predict the waste combustion state and generate combustion trend information; comparison component 13: compares the combustion trend information with the preset combustion control target, and generates optimization instructions according to the comparison result; optimization component 14: based on the optimization instruction, optimizes the combustion air amount, fuel supply amount and grate movement parameters according to the combustion trend information to generate an optimal combustion parameter group; control component 15: controls the combustion fan, fuel supply device and grate drive mechanism in the waste incinerator with the optimal combustion parameter group.

[0045] Furthermore, the data acquisition component 11 is used to perform the following method: The multimodal perception module includes at least a temperature sensor array, a pressure and differential pressure sensor array, a gas composition detector and a weight sensor; wherein the temperature sensor array is distributed at various positions in the furnace of the waste incinerator, and collects the temperature distribution of each area of ​​the furnace in real time to generate temperature detection data; the pressure and differential pressure sensor array is distributed on the combustion grate surface, and is used to collect the pressure difference above and below the grate surface and the air volume on the grate surface, judge the material layer thickness on the grate surface, and generate material layer thickness detection data; the gas composition detector is distributed at the flue gas exhaust port, and is used to detect the gas component content in the flue gas and generate flue gas detection data; the weight sensor is distributed at the feed port, and is used to detect the amount of garbage fed and generate feed amount detection data.

[0046] Furthermore, the data acquisition component 11 is used to perform the following method: The pressure and differential pressure sensor array is connected to a material layer thickness calculation unit, and the material layer thickness calculation unit is embedded with a material layer thickness calculation formula. The material layer thickness is calculated by reading the real-time sensing data of the pressure and differential pressure sensor array to generate the material layer thickness detection data.

[0047] Furthermore, the prediction component 12 is used to perform the following method: A graph structure is constructed based on the data collection positions corresponding to the multi-mode operating status data; based on the graph structure, iterative training of the graph neural network is performed to generate the intelligent decision model; the intelligent decision model is called to analyze the multi-mode operating status data and output the combustion trend information.

[0048] Furthermore, the prediction component 12 is used to perform the following method: Collect a first historical state prediction sample set and a second historical state prediction sample set; construct a graph neural network model based on the graph structure, use the first historical state prediction sample set to perform node feature update training, and generate a node feature update mechanism, wherein node feature update means that each node obtains information from neighboring nodes, and then updates its own state in combination with the node's own information, and the training parameters are learning parameters based on the neighboring nodes updating their own states; based on the node feature update mechanism and the graph neural network model, use the second historical state prediction sample set to perform combustion trend prediction training after performing node feature update, and use the node update data to generate the intelligent decision-making model.

[0049] Furthermore, the optimization component 14 is used to perform the following method: Collect a historical combustion parameter set corresponding to the combustion trend information; construct a fitness evaluation mechanism based on the preset combustion control target, including several sub-target evaluators; use the several sub-target evaluators to perform iterative optimization based on non-dominated sorting on the historical combustion parameter set to generate the optimal combustion parameter group.

[0050] Furthermore, the optimization component 14 is used to perform the following method: Step 1: Use the several sub-target evaluators to perform fitness evaluation on each solution in the historical combustion parameter set to generate each fitness set; Step 2: Based on the each fitness set, calculate the fitness difference of any two solutions using the same sub-target evaluator to generate a fitness difference distribution; Step 3: Call the non-dominated stratification mechanism to perform dominance stratification on each solution according to the fitness difference distribution and construct a dominance relationship structure diagram; perform screening of the dominance relationship structure diagram according to a preset non-dominance level threshold to generate a screening population, repeat steps 1 to 3 after parameter expansion of the screening population, continue to iterate until a preset number of iterations is reached, globally fuse the dominance relationship structure diagram obtained by the iteration, extract the non-dominated optimal solution from the fused dominance relationship structure diagram to generate the optimal combustion parameter group.

[0051] Furthermore, the optimization component 14 is used to perform the following method: The non-dominated hierarchical mechanism includes complete dominance of the several sub-goal evaluators and advanced dominance of at least one sub-goal evaluator, wherein advanced dominance means that the fitness difference between any two solutions on at least one sub-goal evaluator is higher than a preset threshold.

[0052] Furthermore, the comparison component 13 is used to perform the following method: Determine whether the combustion trend information meets the preset combustion control target; if not, generate the optimization instruction.

[0053] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0054] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

[0055] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. The automatic garbage combustion control method based on multimodal perception is characterized by: The method comprises: The multimodal sensing module collects multimodal operating status data of the waste incinerator in real time, including at least temperature detection data, material layer thickness detection data, flue gas detection data, and feed amount detection data; Inputting the multi-mode operation status data into an intelligent decision-making model to predict the garbage combustion status and generate combustion trend information; comparing the combustion trend information with a preset combustion control target, and generating an optimization instruction based on the comparison result; Based on the optimization instruction, optimizing the combustion air quantity, fuel supply quantity and grate movement parameters according to the combustion trend information to generate an optimal combustion parameter set; The combustion-supporting fan, fuel supply device and grate drive mechanism in the waste incinerator are controlled using the optimal combustion parameter group.

2. The method for controlling automatic garbage combustion based on multimodal sensing according to claim 1, characterized in that: The multimodal sensing module includes at least a temperature sensor array, a pressure and differential pressure sensor array, a gas composition detector and a weight sensor; Among them, the temperature sensor array is distributed at various positions of the furnace of the waste incinerator, and collects the temperature distribution of each area of ​​the furnace in real time to generate temperature detection data; the pressure and pressure difference sensor array is distributed on the combustion grate surface, and is used to collect the pressure difference between the upper and lower parts of the grate surface and the air volume on the grate surface, judge the thickness of the material layer on the grate surface, and generate material layer thickness detection data; the gas composition detector is distributed at the flue gas exhaust port, and is used to detect the gas component content in the flue gas and generate flue gas detection data; the weight sensor is distributed at the feed port, and is used to detect the amount of garbage fed and generate feed amount detection data.

3. The method for controlling automatic garbage combustion based on multimodal sensing according to claim 2, characterized in that: The pressure and differential pressure sensor array is connected to a material layer thickness calculation unit, and the material layer thickness calculation unit is embedded with a material layer thickness calculation formula. The material layer thickness is calculated by reading the real-time sensing data of the pressure and differential pressure sensor array to generate the material layer thickness detection data.

4. The method for controlling automatic garbage combustion based on multimodal sensing according to claim 1, characterized in that: The multi-mode operation status data is input into the intelligent decision model to predict the garbage combustion status and generate combustion trend information, including: Building a graph structure based on data collection locations corresponding to the multi-mode operating status data; Based on the graph structure, iterative training of the graph neural network is performed to generate the intelligent decision model; The intelligent decision-making model is called to analyze the multi-mode operating status data and output the combustion trend information.

5. The method for controlling automatic garbage combustion based on multimodal sensing according to claim 4, characterized in that: Based on the graph structure, iterative training of the graph neural network is performed to generate the intelligent decision model, including: Collect a first historical state prediction sample set and a second historical state prediction sample set; Based on the graph structure, a graph neural network model is constructed, and node feature update training is performed using the first historical state prediction sample set to generate a node feature update mechanism, wherein node feature update refers to each node obtaining information from neighboring nodes and then updating its own state based on the node's own information. The training parameters are learning parameters based on the neighboring nodes updating their own states. Based on the node feature update mechanism and the graph neural network model, the second historical state prediction sample set is used to perform combustion trend prediction training using node update data after executing node feature update to generate the intelligent decision-making model.

6. The method for controlling automatic garbage combustion based on multimodal sensing according to claim 1, characterized in that: Based on the optimization instruction, the combustion air quantity, fuel supply quantity and grate movement parameters are optimized according to the combustion trend information to generate an optimal combustion parameter group, including: collecting a historical combustion parameter set corresponding to the combustion trend information; Constructing a fitness evaluation mechanism based on the preset combustion control target, including a plurality of sub-target evaluators; The plurality of sub-objective evaluators are used to perform iterative optimization based on non-dominated sorting on the historical combustion parameter set to generate the optimal combustion parameter group.

7. The method for controlling automatic garbage combustion based on multimodal sensing according to claim 6, characterized in that: The fitness evaluation mechanism is used to perform iterative optimization based on non-dominated sorting on the historical combustion parameter set to generate the optimal combustion parameter group, including: Step 1: using the plurality of sub-objective evaluators to perform fitness evaluation on each solution in the historical combustion parameter set to generate each fitness set; Step 2: Based on each fitness set, calculate the fitness difference of the same sub-goal evaluator for any two solutions to generate a fitness difference distribution; Step 3: Call the non-dominated stratification mechanism to perform dominance stratification on each solution according to the fitness difference distribution, and construct a dominance relationship structure diagram; The dominance relationship structure diagram is screened by a preset non-dominance level threshold to generate a screened population. After parameter expansion of the screened population, steps one to three are repeated, and iteration is continued until a preset number of iterations is reached. The dominance relationship structure diagrams obtained by the iterations are globally fused, and the non-dominance optimal solution is extracted from the fused dominance relationship structure diagram to generate the optimal combustion parameter group.

8. The method for controlling automatic garbage combustion based on multimodal sensing according to claim 7, characterized in that: The non-dominated hierarchical mechanism includes complete dominance of the several sub-goal evaluators and advanced dominance of at least one sub-goal evaluator, wherein advanced dominance means that the fitness difference between any two solutions on at least one sub-goal evaluator is higher than a preset threshold.

9. The method for controlling automatic garbage combustion based on multimodal sensing according to claim 1, characterized in that: Comparing the combustion trend information with a preset combustion control target and generating an optimization instruction based on the comparison result includes: determining whether the combustion trend information satisfies the preset combustion control target; If not, generate the optimization instruction.

10. The automatic garbage combustion control system based on multimodal perception is characterized by: A system for implementing the method for controlling automatic garbage combustion based on multimodal sensing according to any one of claims 1 to 9, the system comprising: Data acquisition component: The multi-modal sensing module collects multi-modal operating status data of the waste incinerator in real time, including at least temperature detection data, material layer thickness detection data, flue gas detection data, and feed amount detection data; Prediction component: inputs the multi-mode operation status data into the intelligent decision model to predict the garbage combustion status and generate combustion trend information; Comparison component: compares the combustion trend information with a preset combustion control target and generates an optimization instruction based on the comparison result; Optimization component: Based on the optimization instruction, the combustion air quantity, fuel supply quantity and grate movement parameters are optimized according to the combustion trend information to generate an optimal combustion parameter group; Control component: controls the combustion-supporting fan, fuel supply device and grate drive mechanism in the waste incinerator with the optimal combustion parameter group.

Images