Combustible waste gas heat balance method, device, electronic equipment and storage medium

Abstract

The present application provides a method, device, electronic device and storage medium for heat balance of combustible waste gas. The method includes: first constructing a simplified model of combustible waste gas, reducing the computational complexity by simplifying the components and processes; then, using regression screening technology to determine a set of key parameters that have a significant impact on the heat balance results; based on the set, a multilayer perceptron neural network is used to establish a nonlinear correlation between active and passive quantities, and a polynomial expression is obtained through genetic programming optimization; finally, the polynomial expression is embedded in a programmable logic controller to realize real-time heat balance and control of the combustible waste gas treatment process; the present application can realize real-time, lightweight and high-precision control of the heat balance of combustible waste gas, significantly improving the efficiency of industrial waste gas treatment and energy utilization.

Description

Technical Field

[0001] The present application relates to the field of industrial automation technology, and in particular to a combustible waste gas heat balance method, device, electronic equipment and storage medium. Background Art

[0002] Heat balance is an essential step in clarifying energy distribution in industrial processes. In the field of combustible waste gas treatment, accurate heat balance plays a decisive role in designing efficient treatment equipment and optimizing operational processes. Heat balance provides a deep understanding of the energy conversion and transfer patterns within the combustible waste gas treatment process, enabling the rational planning of treatment equipment heat exchange structures, precise control of reaction temperatures, and optimization of key parameters such as gas flow rates. This not only helps improve waste gas treatment efficiency and ensure compliance with emission standards, but also optimizes heat treatment and reduces treatment costs, which is of great significance for promoting sustainable industrial development.

[0003] In the treatment of combustible waste gas, simplifying heat calculations is essential for designing efficient treatment equipment and optimizing operational processes. Traditional heat balance methods often involve complex thermodynamic formulas, extensive data collection and processing, and tedious calculations. This not only increases the workload but can also lead to significant errors that affect the accuracy of the results. Therefore, developing a heat balance method that simplifies calculations while maintaining relatively high accuracy is crucial. Summary of the Invention

[0004] In view of this, the embodiments of the present application provide a combustible waste gas heat balance method, device, electronic device and storage medium, which can realize real-time, lightweight and high-precision control of combustible waste gas heat balance, and significantly improve industrial waste gas treatment efficiency and energy utilization.

[0005] The technical solution of the embodiment of the present application is implemented as follows:

[0006] In a first aspect, an embodiment of the present application provides a method for calculating heat balance of combustible waste gas, the method comprising:

[0007] Establishing a simplified model of combustible waste gas; wherein the simplified model is constructed based on component simplification and process simplification;

[0008] Based on regression screening, determine the parameters whose influence on the heat balance result is greater than the influence threshold, and generate a set of key parameters;

[0009] Based on the key parameter set, a nonlinear correlation between the active quantity and the driven quantity is established through a multilayer perceptron neural network, and a polynomial expression is generated through genetic programming optimization;

[0010] The polynomial expression is embedded in a programmable logic controller, and combustible waste heat balance is performed based on the programmable logic controller; wherein the programmable logic controller is used to control a combustible waste treatment device.

[0011] In a second aspect, an embodiment of the present application further provides a combustible waste gas heat balance device, the device comprising:

[0012] A construction module for establishing a simplified model of combustible waste gas; wherein the simplified model is constructed based on component simplification and process simplification;

[0013] A determination module is used to determine parameters whose influence on the heat balance result is greater than an influence threshold based on regression screening, and generate a key parameter set;

[0014] A generation module is used to establish a nonlinear correlation between the active quantity and the driven quantity through a multilayer perceptron neural network based on the key parameter set, and to generate a polynomial expression through genetic programming optimization;

[0015] A balancing module is used to embed the polynomial expression into a programmable logic controller and perform combustible waste heat balancing based on the programmable logic controller; wherein the programmable logic controller is used to control the combustible waste treatment device.

[0016] In a third aspect, an embodiment of the present application further provides an electronic device comprising: a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to execute the combustible waste gas heat balancing method described in any one of the first aspects.

[0017] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the combustible waste gas heat balance method described in any one of the first aspects is executed.

[0018] The embodiments of the present application have the following beneficial effects:

[0019] By establishing a simplified model (segmented simplification of components and processes) and the synergy of intelligent algorithms (multi-layer perceptron neural networks), traditional offline calculations that rely on complex thermodynamic formulas are transformed into polynomial correlation calculations based on key parameters. Combined with PLC real-time dynamic control, the heat balance response time is greatly shortened, while adapting to fluctuations in different operating conditions, ensuring efficient and precise control of industrial sites. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 10 is a flow chart of steps S101-S104 provided in an embodiment of the present application;

[0022] Figure 2 1 is a flow chart of steps S201-S204 provided in an embodiment of the present application;

[0023] Figure 3 Schematic diagram of the process of steps S301-S302 provided in an embodiment of the present application;

[0024] Figure 4 This is a simplified calculation model flow chart of the combustible waste gas treatment system provided in an embodiment of the present application;

[0025] Figure 5 is a data processing flow chart of the neural network algorithm provided in an embodiment of the present application;

[0026] Figure 6 This is a diagram of the training and optimization process of the heat balance prediction model provided in the embodiment of the present application;

[0027] Figure 7 It is a structural schematic diagram of a combustible waste gas heat balance device provided in an embodiment of the present application;

[0028] Figure 8 It is a schematic diagram of the composition structure of the electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the drawings in the present application only serve the purpose of illustration and description and are not used to limit the scope of protection of the present application. In addition, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate the operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowcharts can be implemented out of sequence, and steps without logical context can be reversed or implemented simultaneously. In addition, those skilled in the art, under the guidance of the contents of this application, can add one or more other operations to the flowchart, or remove one or more operations from the flowchart.

[0030] In the following description, reference is made to “some embodiments”, which describes a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0031] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present application.

[0032] In the following description, the terms "first\second\third" involved are merely used to distinguish similar objects and do not represent a specific ordering of the objects. It can be understood that "first\second\third" can be interchanged with a specific order or sequence where permitted, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0033] It should be noted that the term "comprising" will be used in the embodiments of the present application to indicate the existence of the features declared thereafter, but does not exclude the addition of other features.

[0034] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application belongs. The terms used herein are for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0035] See also Figure 1 , Figure 1 This is a flow chart of steps S101-S104 of the combustible waste gas heat balance method provided in the embodiment of the present application, which will be combined with Figure 1 Steps S101-S104 are shown for explanation.

[0036] In step S101 , a simplified model of combustible exhaust gas is established; wherein the simplified model is constructed based on component simplification and process simplification.

[0037] Here, we first establish a simplified model of combustible exhaust gas. Combustible exhaust gas typically contains multiple components. To simplify the calculation, the exhaust gas composition needs to be simplified. This can include omitting components that have a minor impact on the heat balance or combining components with similar thermophysical properties. The exhaust gas treatment process may involve multiple complex steps, such as combustion, cooling, and purification. To simplify the heat balance, these steps need to be simplified. For example, a complex combustion process can be reduced to one or a few key heat exchange steps.

[0038] In some embodiments, the component simplification is used to simplify the combustible exhaust gas into a mixed model consisting of preset key combustible components and non-combustible components, and ignore minor components whose content is lower than a content threshold and whose heat contribution is lower than a contribution threshold;

[0039] The process simplification is used to divide the combustible waste gas treatment process into multiple energy transfer stages, model each stage independently, and extract the main energy transfer mode.

[0040] For example, for component simplification, a combustible waste gas sample from a chemical plant (containing 60% methane, 25% CO, 10% H2, 4% N2, and 1% other trace components) was analyzed, and trace components with a content of less than 0.5% (such as SO2 and particulate matter) were filtered out to construct a mixed model consisting of methane, CO, and H2.

[0041] To simplify the process, the exhaust gas treatment process can be divided into a "combustion reaction stage" and a "water cooling stage." In the reaction stage, combustion heat and heat conduction are calculated; in the cooling stage, only convection heat transfer and cooling water heat absorption are considered, ignoring secondary factors such as radiation heat dissipation.

[0042] In some embodiments, the content threshold of the simplified component is set to a content lower than 0.5%, and the average content range of the main components is determined by analyzing actual combustible exhaust gas samples, and the average content range of the main components is used as the basic parameter of the simplified model.

[0043] For example, 10 sets of waste gas samples from different industrial scenarios were collected and analyzed. It was found that when the content of minor components (such as ethane and propane) was ≥0.5%, their contribution to the total calorific value exceeded 1.5%, while the contribution was negligible when the content was <0.5%. In the waste gas treatment of a steel plant, ethane (0.3%) and propane (0.2%) were filtered out, retaining only CO (28%), H2 (12%), and CH4 (58%) as model inputs, which significantly reduced the calculation error.

[0044] The above method can construct a physical model that can reflect the essence of the actual processing process and is relatively simple and easy to understand.

[0045] In step S102 , parameters whose influence on the heat balance result is greater than an influence threshold are determined based on regression screening, and a key parameter set is generated.

[0046] Regression analysis can be used to determine the degree of influence of each parameter on the heat balance results. Regression analysis can reveal linear or nonlinear relationships between parameters and the heat balance results. Based on the results of the regression analysis, parameters with an impact on the heat balance results exceeding a preset threshold can be selected. These parameters will constitute the key parameter set.

[0047] For example, Lasso regression (λ=0.05) can be used to screen out parameters that have a significant impact on heat balance, including exhaust gas flow rate (weight 0.6), cooling water inlet temperature (weight -0.4), and exhaust gas initial temperature (weight 0.3).

[0048] In some embodiments, the regression screening is Lasso regression screening, and the loss function is:

[0049]

[0050] in, is the response variable for the ith observation, is the jth predictor variable for the i-th observation, is the coefficient of the j-th predictor variable, and λ is the regularization parameter. By adjusting the regularization parameter λ, the parameters related to the heat balance results are screened out.

[0051] Here, in the Lasso regression model, λ=0.1 is set, and the exhaust gas flow rate (β=0.62), cooling water inlet temperature (β=-0.38), and exhaust gas initial temperature (β=0.28) are screened out, and the coefficients of other parameters (such as ambient humidity and air pressure) are compressed to 0.

[0052] By adjusting the λ value (0.05~0.2), it was verified that when λ=0.1, the model mean square error (MSE) was the smallest (0.023), and the parameter selection was optimal.

[0053] In step S103, based on the key parameter set, a nonlinear correlation between the active quantity and the passive quantity is established through a multilayer perceptron neural network, and a polynomial expression is generated through genetic programming optimization.

[0054] Here, a multilayer perceptron neural network is used to establish a nonlinear relationship between active variables (such as exhaust gas flow rate and temperature) and passive variables (such as heat change). Neural networks can capture complex relationships between parameters, improving the accuracy of heat balance calculations. Using a genetic programming algorithm, the nonlinear relationship established by the neural network is optimized to generate a polynomial expression. Polynomial expressions offer advantages such as simple computation and ease of integration into control systems.

[0055] In some embodiments, see Figure 2 , Figure 2 This is a flow chart of steps S201-S204 provided in an embodiment of the present application, wherein the multilayer perceptron neural network includes an input layer, multiple hidden layers, and an output layer; the input layer nodes include m principal components and s original parameters, each of the multiple hidden layers adopts a two-layer structure, the activation function is LeakyReLU, and the output layer is a single-node linear unit; the nonlinear correlation between the active quantity and the follower quantity is established through the multilayer perceptron neural network, and a polynomial expression is generated through genetic programming optimization, which can be achieved through steps S201-S203 and will be explained in conjunction with each step.

[0056] In step S201 , the MLP model is trained through cross-validation, and a heat balance prediction model is generated based on the Adam optimizer and the early stopping mechanism.

[0057] In step S202, a sensitivity analysis is performed on the trained MLP model to extract the input parameter gradient contribution, and a polynomial correlation is derived based on Taylor second-order expansion and symbolic regression.

[0058] In step S203, an optimal polynomial combination is searched through genetic programming, and a polynomial expression with a degree of fit greater than or equal to a degree of fit threshold is selected to express the relationship between the active variable and the driven variable.

[0059] Here, the input layer contains m principal components and s raw parameters. Principal components are typically derived by performing dimensionality reduction (e.g., PCA) on the raw parameters, preserving most information while reducing computational complexity. Raw parameters can be directly measured variables in the exhaust gas treatment process, such as temperature, pressure, and flow rate.

[0060] Each hidden layer uses a two-layer structure, and each hidden layer actually consists of two sub-layers. The activation function is LeakyReLU (Leaky Rectified Linear Unit), which is an improved ReLU function that avoids the neuron "death" problem by allowing small negative gradients.

[0061] The output layer is a single-node linear unit, so the output is a continuous value used to predict the result of heat balance.

[0062] When establishing nonlinear correlations and optimizing the generated polynomial expressions, using cross-validation to train the MLP model helps avoid overfitting and improve the model's generalization. The Adam optimizer is an adaptive learning rate optimization algorithm that combines the advantages of momentum and adaptive learning rates and is suitable for most deep learning tasks. Early stopping is a regularization technique that prevents model overfitting by stopping training when performance on the validation set no longer improves.

[0063] A sensitivity analysis is performed on the trained MLP model to assess the contribution of each input parameter to the output. By calculating the gradient contribution of the input parameters, the parameters with the greatest impact on the heat balance can be identified. Based on Taylor's second-order expansion, the model can be locally linearized to approximate nonlinear relationships. Symbolic regression, a method that uses evolutionary algorithms to find mathematical expressions that explain the data, can be used to derive explicit polynomial correlations.

[0064] Genetic programming is an algorithm that optimizes program structure by simulating natural selection and genetic mechanisms. In this embodiment, genetic programming searches for optimal polynomial combinations, screening out polynomial expressions with a fitness greater than or equal to a preset fitness threshold. These polynomial expressions are used to express the relationship between active variables (such as exhaust flow rate and temperature) and passive variables (such as heat change).

[0065] The above approach, combining MLP and genetic programming, effectively models the complex nonlinear relationships in the combustible waste gas treatment process. By optimizing the generated polynomial expressions, it provides a concise and interpretable model for practical control. The advantage of this approach lies in its flexibility and adaptability, enabling it to process complex industrial process data and generate practical control strategies.

[0066] In some embodiments, the multilayer perceptron neural network is optimized by:

[0067] Select activation function to improve the nonlinear expression ability of the network;

[0068] Alternatively, an optimization algorithm is used to speed up the training and avoid falling into a local optimal solution; wherein the optimization algorithm at least includes a stochastic gradient descent method;

[0069] Alternatively, determine the optimal number of hidden layers and neurons through experiments to balance computational complexity and model performance.

[0070] Optimizing the Multilayer Perceptron (MLP) neural network is a key step in improving its performance and efficiency. The activation function introduces nonlinearity to the neural network, enabling it to learn and represent complex patterns. Different activation functions have a significant impact on the network's training speed and ultimate performance.

[0071] The activation function of choice can be LeakyReLU. As mentioned earlier, LeakyReLU allows small negative gradients, avoiding the "neuron death" problem of ReLU in negative areas. Other commonly used activation functions can also be:

[0072] ReLU is simple and computationally efficient, but may cause gradients to vanish in negative regions.

[0073] Sigmoid: The output is between 0 and 1. It is suitable for binary classification problems, but it can easily cause gradient disappearance.

[0074] Tanh: The output is between -1 and 1, and has stronger gradient characteristics than Sigmoid.

[0075] Swish: A self-gating activation function that often outperforms ReLU in some situations.

[0076] In the embodiment of the present application, a suitable activation function can be selected according to specific tasks and data characteristics.

[0077] The role of the optimization algorithm is to adjust the network parameters to minimize the loss function. Different optimization algorithms perform differently in terms of convergence speed and avoiding local optimal solutions.

[0078] Common optimization algorithms include:

[0079] Stochastic Gradient Descent (SGD): Simple and easy to implement, but may converge slowly and easily fall into local optimal solutions.

[0080] Adam: combines the advantages of momentum and adaptive learning rate, and generally performs well in most tasks.

[0081] RMSprop: Adaptive learning rate, suitable for non-stationary objectives.

[0082] Adagrad: Adaptive learning rate, suitable for sparse data.

[0083] In the embodiments of the present application, a suitable optimization algorithm can be selected according to the task characteristics and data scale.

[0084] The number of hidden layers and neurons directly affects the complexity and expressiveness of the model. Too many layers and neurons can lead to overfitting, increasing computational complexity, while too few layers and neurons can lead to underfitting, failing to capture complex patterns in the data. Therefore, the optimal network structure can be determined through experimentation and cross-validation.

[0085] For example, you can start with a simple structure, gradually increase the complexity, observe the changes in model performance, and use regularization techniques (such as L1 / L2 regularization and early stopping mechanism) to prevent overfitting.

[0086] The above approach can effectively improve the performance and efficiency of MLP by selecting appropriate activation functions, optimization algorithms, and experimentally determining the optimal network structure. The optimization process requires combining theoretical knowledge with experimental verification to find the configuration that best suits a specific task.

[0087] In step S104, the polynomial expression is embedded in a programmable logic controller, and combustible waste heat balance is performed based on the programmable logic controller; wherein the programmable logic controller is used to control a combustible waste treatment device.

[0088] The optimized polynomial expression is embedded in a programmable logic controller (PLC), enabling it to perform real-time heat balance calculations. Based on the results of this heat balance, the PLC controls the combustible waste gas treatment device in real time, adjusting factors such as the fuel supply to the burner and the cooling water flow rate to ensure the stability and efficiency of the waste gas treatment process.

[0089] In some embodiments, see Figure 3 , Figure 3 It is a flow chart of steps S301-S302 provided in an embodiment of the present application. The method also includes steps S301-S302, which will be described in combination with each step.

[0090] In step S301, the deviation between the heat balance result and the measured value is monitored in real time.

[0091] In step S302, the weight of the follower parameter in the programmable logic controller is dynamically adjusted based on the monitoring result to correct the error.

[0092] By comparing the predicted values of the heat balance model with the measured values in real time, we can promptly detect any deviations between the model and the actual system. This feedback mechanism is the basis for dynamic adjustments. Deviations may arise from simplifying model assumptions, sensor errors, environmental disturbances, or dynamic process changes. Continuous monitoring helps identify the main sources of error.

[0093] The output of the heat balance model and the measured values of key parameters can be collected synchronously to ensure time synchronization, calculate the absolute or relative deviation between the predicted value and the measured value, set the deviation threshold to trigger the adjustment mechanism, record the deviation data, perform trend analysis, and identify the stability and potential problems of system operation.

[0094] In a programmable logic controller (PLC), the weights of slave parameters (such as temperature and flow) directly influence the results of heat balance calculations. Adjusting these weights can correct model deviations. Dynamic adjustment enables the system to adapt to changing operating conditions, such as fluctuating exhaust gas composition and varying processing loads, thereby maintaining heat balance accuracy. Adjustment strategies can be rule-based, such as threshold triggering, initiating an adjustment process when a deviation exceeds a preset threshold. Alternatively, incremental adjustment can be used, adjusting parameter weights in preset steps based on the magnitude and direction of the deviation. Alternatively, model-based adjustment can be implemented using online learning, updating model parameters using real-time data, such as through online parameter estimation using recursive least squares (RLS). Alternatively, adaptive filtering can be used, combining predicted and measured values using methods such as the Kalman filter to dynamically optimize parameter estimates.

[0095] The embodiment of the present application can be adjusted by the following adjustment process:

[0096] (1) Deviation assessment: Regularly or continuously assess the deviation between the predicted value and the measured value.

[0097] (2) Decision making: Based on the deviation assessment results, decide whether to adjust the parameter weights and the adjustment range.

[0098] (3) Parameter update: Implement parameter weight adjustment in PLC and verify the adjustment effect.

[0099] (4) Effect feedback: monitor the deviation changes after adjustment to form a closed-loop control.

[0100] The following is a comprehensive description of the core processes of system simplified modeling and real-time control. Figure 4 , Figure 4 This is a simplified calculation model flow chart of the combustible waste gas treatment system provided in the embodiment of the present application, such as Figure 4 As shown, it includes five stages:

[0101] Composition simplification: Exhaust gas is separated by a gas analyzer, retaining only core components (such as methane, CO, and H2) and filtering out trace components (content <0.5%). Based on actual sample analysis, the average content range of key components is determined and used as model input parameters.

[0102] Procedural abstraction:

[0103] The complex process is divided into two independent modules: the "reaction phase" and the "cooling phase." The reaction phase focuses on combustion heat and heat conduction, while the cooling phase focuses on convection heat transfer and ignores minor energy losses (such as radiation heat dissipation).

[0104] Parameter screening:

[0105] Lasso regression (L1 regularization) is used to filter significant parameters (e.g., exhaust gas flow rate, cooling water temperature, etc. when λ=0.1). Parameters with low contribution (e.g., nitrogen weight -0.2) are eliminated to reduce computational redundancy.

[0106] Introduction of intelligent algorithms:

[0107] Based on the MLP model training data, a polynomial correlation between the active and the passive momentum is generated.

[0108] Real-time control adjustments:

[0109] By embedding the correlation equation into the PLC controller, the driven variables such as cooling water flow rate can be adjusted dynamically in response to fluctuations in exhaust gas composition or sudden increases in flow rate.

[0110] Through the closed-loop process of "simplified modeling - parameter screening - intelligent association - real-time control", the embodiment of the present application can achieve efficient and accurate heat balance and process optimization.

[0111] Below, we will give a comprehensive description of the data processing flow based on the multi-layer perceptron neural network (MLP). Figure 5 , Figure 5 This is a data processing flow chart of the neural network algorithm provided in the embodiment of the present application, such as Figure 5 As shown, the input layer receives active quantity parameters related to heat balance (such as exhaust gas inlet temperature, exhaust gas flow, cooling water inlet temperature) and original parameters (such as exhaust gas component ratio). Each input node (X1~X n ) corresponds to an independent parameter and provides initial data for the neural network.

[0112] The hidden layer adopts a two-layer structure (e.g. 64→32 nodes), the first hidden layer (h1~h n ) and the second hidden layer (h2-h4) extract and transform the input data using nonlinear activation functions (such as LeakyReLU), capturing the complex relationships between parameters. The design of the hidden layer enhances the model's nonlinear mapping capabilities.

[0113] The output layer generates prediction results (Y1~Y m ). The output layer uses linear units to directly reflect the final target parameters of heat balance.

[0114] The above method simplifies the derivation of traditional thermodynamic formulas into nonlinear correlations between parameters through layer-by-layer calculations of multi-layer neural networks, providing a data basis for the subsequent generation of polynomial expressions.

[0115] Below, we will provide a comprehensive description of the entire process from predictive model construction to optimization. Figure 6 , Figure 6This is a diagram of the training and optimization process of the heat balance prediction model provided in the embodiment of the present application, such as Figure 6 As shown, during the model construction phase, active and passive variables are identified, with input parameters (active variables) and output targets (passive variables) defined. For example, exhaust gas flow is the active variable, and cooling water flow is the passive variable. The network structure is determined, with the input layer containing the primary parameters, a two-layer hidden layer, and the output layer corresponding to a single target variable. During the training and validation phase, cross-validation is used for training, employing the Adam optimizer (learning rate 0.001) and early stopping to avoid overfitting and ensure model generalization. Gradient contribution analysis is used to identify key input parameters (e.g., cooling water temperature weighted at 0.4).

[0116] For correlation expression generation and optimization, including Taylor expansion, symbolic regression and genetic programming (GP) screening, the black box model of the neural network is converted into an interpretable polynomial expression, and the optimal polynomial with a fitting degree ≥ 95% is screened from the candidate expressions.

[0117] The activation function and structure optimization select the ReLU function to improve the nonlinear expression ability, and determine the optimal number of hidden layer nodes through experiments.

[0118] The above method converts the complex neural network model into a lightweight polynomial correlation, which is convenient for embedding in industrial controllers to realize real-time calculation.

[0119] In summary, the embodiments of the present application have the following beneficial effects:

[0120] (1) Simplified calculation process: By establishing a simplified model and simplifying and associating parameters, the complex thermodynamic formula derivation and unnecessary data processing in traditional methods are discarded. The heat balance process is simplified to calculations based on key parameters and simple correlations, which greatly reduces the number of calculation steps and workload. Operators can easily use the methods of the application embodiment to perform heat balance without having deep thermodynamic expertise, thereby improving work efficiency.

[0121] (2) Improved calculation accuracy: On the one hand, the introduction of intelligent algorithms can automatically learn complex parameter relationships, compensating for the accuracy loss that may be caused by simplified models in traditional methods. On the other hand, the error correction mechanism adjusts the calculation results in a timely manner through real-time monitoring and dynamic correction, effectively reducing error accumulation and ensuring high accuracy of the heat balance results. Compared with traditional methods, the heat balance results of the embodiments of the present application can more accurately reflect the actual energy distribution in the combustible waste gas treatment process, providing more reliable data support for the design and operation of the treatment device.

[0122] (3) Improve industrial efficiency: The embodiments of the present application contribute to the design of more reasonable combustible waste gas treatment devices. For example, during the design phase, key parameters such as the heat exchange structure and reactor size of the device can be optimized based on accurate heat balance results, thereby improving energy efficiency and reducing equipment costs. During the operation phase, through real-time heat balance and error correction, operating process parameters can be adjusted in a timely manner to ensure that the treatment device is always in the best operating state, improve waste gas treatment efficiency, reduce pollutant emissions, and promote the development of the combustible waste gas treatment industry in the direction of high efficiency, energy saving, and environmental protection.

[0123] Based on the same inventive concept, an embodiment of the present application also provides a combustible waste gas heat balance device corresponding to the combustible waste gas heat balance method in the first embodiment. Since the principle of solving the problem by the device in the embodiment of the present application is similar to the above-mentioned combustible waste gas heat balance method, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0124] like Figure 7 As shown, Figure 7 Schematic diagram of the structure of the combustible waste gas heat balance device 700 provided in an embodiment of the present application. The combustible waste gas heat balance device 700 includes:

[0125] A construction module 701 is used to establish a simplified model of combustible waste gas; wherein the simplified model is constructed based on component simplification and process simplification;

[0126] A determination module 702 is configured to determine, based on regression screening, parameters whose impact on the heat balance result is greater than an impact threshold, and generate a key parameter set;

[0127] A generating module 703 is configured to establish a nonlinear correlation between the active quantity and the driven quantity based on the key parameter set through a multilayer perceptron neural network, and generate a polynomial expression through genetic programming optimization;

[0128] The balancing module 704 is used to embed the polynomial expression into a programmable logic controller and perform combustible waste heat balancing based on the programmable logic controller; wherein the programmable logic controller is used to control the combustible waste treatment device.

[0129] Those skilled in the art should understand that Figure 7 The functions of each unit in the combustible waste gas heat balance device 700 can be understood by referring to the relevant description of the combustible waste gas heat balance method mentioned above. Figure 7 The functions of the various units in the combustible waste gas heat balance device 700 shown can be implemented by a program running on a processor, or by a specific logic circuit.

[0130] In one possible embodiment, the component simplification is used to simplify the combustible exhaust gas into a mixed model consisting of preset key combustible components and non-combustible components, and ignore minor components whose content is lower than a content threshold and whose heat contribution is lower than a contribution threshold;

[0131] The process simplification is used to divide the combustible waste gas treatment process into multiple energy transfer stages, model each stage independently, and extract the main energy transfer mode.

[0132] In one possible embodiment, the content threshold of the simplified component is set to a content lower than 0.5%, and the average content range of the main components is determined by analyzing actual combustible exhaust gas samples, and the average content range of the main components is used as the basic parameter of the simplified model.

[0133] In one possible implementation, the regression screening is Lasso regression screening, and the loss function is:

[0134]

[0135] in, is the response variable for the ith observation, is the jth predictor variable for the i-th observation, is the coefficient of the j-th predictor variable, and λ is the regularization parameter. By adjusting the regularization parameter λ, the parameters related to the heat balance results are screened out.

[0136] In one possible implementation, the multilayer perceptron neural network includes an input layer, multiple hidden layers, and an output layer; the input layer nodes include m principal components and s original parameters, each of the multiple hidden layers adopts a two-layer structure, the activation function is LeakyReLU, and the output layer is a single-node linear unit;

[0137] The generation module 703 establishes a nonlinear correlation between the active quantity and the driven quantity through a multilayer perceptron neural network, and generates a polynomial expression through genetic programming optimization, including:

[0138] The MLP model is trained through cross-validation, and a heat balance prediction model is generated based on the Adam optimizer and early stopping mechanism;

[0139] Perform sensitivity analysis on the trained MLP model, extract the input parameter gradient contribution, and derive and display the polynomial correlation based on Taylor second-order expansion and symbolic regression;

[0140] The optimal polynomial combination is searched through genetic programming, and a polynomial expression with a degree of fit greater than or equal to a degree of fit threshold is screened to express the relationship between the active variable and the driven variable.

[0141] In one possible implementation, the multilayer perceptron neural network is optimized as follows:

[0142] Select activation function to improve the nonlinear expression ability of the network;

[0143] Alternatively, an optimization algorithm is used to speed up the training and avoid falling into a local optimal solution; wherein the optimization algorithm at least includes a stochastic gradient descent method;

[0144] Alternatively, determine the optimal number of hidden layers and neurons through experiments to balance computational complexity and model performance.

[0145] In one possible implementation, the accounting module 704 further includes:

[0146] Real-time monitoring of the deviation between heat balance results and measured values;

[0147] The weight of the follower parameter in the programmable logic controller is dynamically adjusted based on the monitoring result to correct the error.

[0148] The above-mentioned combustible waste gas heat balance device has the following beneficial effects:

[0149] (1) Simplified calculation process: By establishing a simplified model and simplifying and associating parameters, the complex thermodynamic formula derivation and unnecessary data processing in traditional methods are discarded. The heat balance process is simplified to calculations based on key parameters and simple correlations, which greatly reduces the number of calculation steps and workload. Operators can easily use the methods of the application embodiment to perform heat balance without having deep thermodynamic expertise, thereby improving work efficiency.

[0150] (2) Improved calculation accuracy: On the one hand, the introduction of intelligent algorithms can automatically learn complex parameter relationships, compensating for the accuracy loss that may be caused by simplified models in traditional methods. On the other hand, the error correction mechanism adjusts the calculation results in a timely manner through real-time monitoring and dynamic correction, effectively reducing error accumulation and ensuring high accuracy of the heat balance results. Compared with traditional methods, the heat balance results of the embodiments of the present application can more accurately reflect the actual energy distribution in the combustible waste gas treatment process, providing more reliable data support for the design and operation of the treatment device.

[0151] (3) Improve industrial efficiency: The embodiments of the present application contribute to the design of more reasonable combustible waste gas treatment devices. For example, during the design phase, key parameters such as the heat exchange structure and reactor size of the device can be optimized based on accurate heat balance results, thereby improving energy efficiency and reducing equipment costs. During the operation phase, through real-time heat balance and error correction, operating process parameters can be adjusted in a timely manner to ensure that the treatment device is always in the best operating state, improve waste gas treatment efficiency, reduce pollutant emissions, and promote the development of the combustible waste gas treatment industry in the direction of high efficiency, energy saving, and environmental protection.

[0152] like Figure 8 As shown, Figure 8 This is a schematic diagram of the structure of an electronic device 800 provided in an embodiment of the present application. The electronic device 800 includes:

[0153] A processor 801, a storage medium 802 and a bus 803, wherein the storage medium 802 stores machine-readable instructions executable by the processor 801. When the electronic device 800 is running, the processor 801 communicates with the storage medium 802 via the bus 803, and the processor 801 executes the machine-readable instructions to perform the steps of the combustible waste gas heat balance method described in the embodiment of the present application.

[0154] In actual application, the various components in the electronic device 800 are coupled together via a bus 803. It is understood that the bus 803 is used to achieve connection and communication between these components. In addition to the data bus, the bus 803 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clarity, Figure 8 Various buses are labeled as bus 803.

[0155] The electronic device has the following beneficial effects:

[0156] (1) Simplified calculation process: By establishing a simplified model and simplifying and associating parameters, the complex thermodynamic formula derivation and unnecessary data processing in traditional methods are discarded. The heat balance process is simplified to calculations based on key parameters and simple correlations, which greatly reduces the number of calculation steps and workload. Operators can easily use the methods of the application embodiment to perform heat balance without having deep thermodynamic expertise, thereby improving work efficiency.

[0157] (2) Improved calculation accuracy: On the one hand, the introduction of intelligent algorithms can automatically learn complex parameter relationships, compensating for the accuracy loss that may be caused by simplified models in traditional methods. On the other hand, the error correction mechanism adjusts the calculation results in a timely manner through real-time monitoring and dynamic correction, effectively reducing error accumulation and ensuring high accuracy of the heat balance results. Compared with traditional methods, the heat balance results of the embodiments of the present application can more accurately reflect the actual energy distribution in the combustible waste gas treatment process, providing more reliable data support for the design and operation of the treatment device.

[0158] (3) Improve industrial efficiency: The embodiments of the present application contribute to the design of more reasonable combustible waste gas treatment devices. For example, during the design phase, key parameters such as the heat exchange structure and reactor size of the device can be optimized based on accurate heat balance results, thereby improving energy efficiency and reducing equipment costs. During the operation phase, through real-time heat balance and error correction, operating process parameters can be adjusted in a timely manner to ensure that the treatment device is always in the best operating state, improve waste gas treatment efficiency, reduce pollutant emissions, and promote the development of the combustible waste gas treatment industry in the direction of high efficiency, energy saving, and environmental protection.

[0159] The embodiment of the present application further provides a computer-readable storage medium, which stores executable instructions. When the executable instructions are executed by at least one processor 801, the combustible waste gas heat balance method described in the embodiment of the present application is implemented.

[0160] In some embodiments, the storage medium can be a magnetic random access memory (FRAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a flash memory, a magnetic surface storage, an optical disc, or a compact disc read-only memory (CD-ROM); it can also be various devices including one or any combination of the above memories.

[0161] In some embodiments, executable instructions may be in the form of a program, software, software module, script, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0162] As an example, executable instructions may, but do not necessarily, correspond to a file in a file system, may be stored as part of a file that stores other programs or data, for example, in one or more scripts in a HyperText Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinated files (for example, files storing one or more modules, subroutines, or code portions).

[0163] By way of example, executable instructions may be deployed to be executed on one computing device, or on multiple computing devices at one site, or on multiple computing devices distributed across multiple sites and interconnected by a communication network.

[0164] The computer-readable storage medium has the following advantages:

[0165] (1) Simplified calculation process: By establishing a simplified model and simplifying and associating parameters, the complex thermodynamic formula derivation and unnecessary data processing in traditional methods are discarded. The heat balance process is simplified to calculations based on key parameters and simple correlations, which greatly reduces the number of calculation steps and workload. Operators can easily use the methods of the application embodiment to perform heat balance without having deep thermodynamic expertise, thereby improving work efficiency.

[0166] (2) Improved calculation accuracy: On the one hand, the introduction of intelligent algorithms can automatically learn complex parameter relationships, compensating for the accuracy loss that may be caused by simplified models in traditional methods. On the other hand, the error correction mechanism adjusts the calculation results in a timely manner through real-time monitoring and dynamic correction, effectively reducing error accumulation and ensuring high accuracy of the heat balance results. Compared with traditional methods, the heat balance results of the embodiments of the present application can more accurately reflect the actual energy distribution in the combustible waste gas treatment process, providing more reliable data support for the design and operation of the treatment device.

[0167] (3) Improve industrial efficiency: The embodiments of the present application contribute to the design of more reasonable combustible waste gas treatment devices. For example, during the design phase, key parameters such as the heat exchange structure and reactor size of the device can be optimized based on accurate heat balance results, thereby improving energy efficiency and reducing equipment costs. During the operation phase, through real-time heat balance and error correction, operating process parameters can be adjusted in a timely manner to ensure that the treatment device is always in the best operating state, improve waste gas treatment efficiency, reduce pollutant emissions, and promote the development of the combustible waste gas treatment industry in the direction of high efficiency, energy saving, and environmental protection.

[0168] In the several embodiments provided in this application, it should be understood that the disclosed methods and electronic devices can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0169] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed across multiple network elements. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0170] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0171] If the functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, platform server, or network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage media include various media that can store program code, such as USB flash drives, mobile hard drives, ROM, RAM, magnetic disks, or optical disks.

[0172] The above are only specific embodiments of the present application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A method for calculating heat balance of combustible waste gas, characterized in that: The method comprises: Establishing a simplified model of combustible waste gas; wherein the simplified model is constructed based on component simplification and process simplification; Based on regression screening, determine the parameters whose influence on the heat balance result is greater than the influence threshold, and generate a set of key parameters; Based on the key parameter set, a nonlinear correlation between the active quantity and the driven quantity is established through a multilayer perceptron neural network, and a polynomial expression is generated through genetic programming optimization; Embedding the polynomial expression into a programmable logic controller, and performing a combustible waste gas heat balance based on the programmable logic controller; wherein the programmable logic controller is used to control a combustible waste gas treatment device; The multilayer perceptron neural network includes an input layer, multiple hidden layers and an output layer; the input layer nodes include m principal components and s original parameters, each hidden layer in the multiple hidden layers adopts a two-layer structure, the activation function is LeakyReLU, and the output layer is a single-node linear unit; The method of establishing a nonlinear correlation between the active quantity and the driven quantity through a multilayer perceptron neural network and generating a polynomial expression through genetic programming optimization includes: The MLP model is trained through cross-validation, and a heat balance prediction model is generated based on the Adam optimizer and early stopping mechanism; Perform sensitivity analysis on the trained MLP model, extract the input parameter gradient contribution, and derive and display the polynomial correlation based on Taylor second-order expansion and symbolic regression; The optimal polynomial combination is searched through genetic programming, and a polynomial expression with a degree of fit greater than or equal to a degree of fit threshold is screened to express the relationship between the active variable and the driven variable.

2. The method according to claim 1, characterized in that The component simplification is used to simplify the combustible exhaust gas into a mixed model consisting of preset key combustible components and non-combustible components, and ignore minor components whose content is lower than a content threshold and whose heat contribution is lower than a contribution threshold; The process simplification is used to divide the combustible waste gas treatment process into multiple energy transfer stages, model each stage independently, and extract the main energy transfer mode.

3. The method according to claim 2, characterized in that The content threshold of the simplified component is set to less than 0.5%, and the average content range of the main components is determined by analyzing actual combustible exhaust gas samples, and the average content range of the main components is used as the basic parameter of the simplified model.

4. The method according to claim 1, wherein The regression screening is Lasso regression screening, and the loss function is: in, is the response variable for the ith observation, is the jth predictor variable for the ith observation, is the coefficient of the j-th predictor variable, and λ is the regularization parameter. By adjusting the regularization parameter λ, the parameters related to the heat balance results are screened out.

5. The method according to claim 1, wherein The multilayer perceptron neural network is optimized as follows: Select activation function to improve the nonlinear expression ability of the network; Alternatively, an optimization algorithm is used to speed up the training and avoid falling into a local optimal solution; wherein the optimization algorithm at least includes a stochastic gradient descent method; Alternatively, determine the optimal number of hidden layers and neurons through experiments to balance computational complexity and model performance.

6. The method according to claim 1, characterized in that The method further comprises: Real-time monitoring of the deviation between heat balance results and measured values; The weight of the follower parameter in the programmable logic controller is dynamically adjusted based on the monitoring result to correct the error.

7. A combustible waste gas heat balance device, characterized in that: The device comprises: A construction module for establishing a simplified model of combustible waste gas; wherein the simplified model is constructed based on component simplification and process simplification; A determination module is used to determine parameters whose influence on the heat balance result is greater than an influence threshold based on regression screening, and generate a key parameter set; A generation module is configured to establish a nonlinear correlation between the active quantity and the driven quantity through a multilayer perceptron neural network based on the key parameter set, and generate a polynomial expression through genetic programming optimization; the multilayer perceptron neural network includes an input layer, multiple hidden layers, and an output layer; the input layer nodes include m principal components and s original parameters, each of the multiple hidden layers adopts a two-layer structure, the activation function is LeakyReLU, and the output layer is a single-node linear unit; the nonlinear correlation between the active quantity and the driven quantity is established through the multilayer perceptron neural network, and the polynomial expression is generated through genetic programming optimization, including: training an MLP model through cross-validation, generating a heat balance prediction model based on an Adam optimizer and an early stopping mechanism; performing sensitivity analysis on the trained MLP model, extracting the gradient contribution of the input parameters, and displaying the polynomial correlation based on Taylor second-order expansion and symbolic regression derivation; searching for an optimal polynomial combination through genetic programming, and screening a polynomial expression with a fit greater than or equal to a fit threshold for expressing the relationship between the active quantity and the driven quantity; A balancing module is used to embed the polynomial expression into a programmable logic controller and perform combustible waste gas heat balance based on the programmable logic controller; wherein the programmable logic controller is used to control the combustible waste gas treatment device.

8. An electronic device, characterized in that: include: A processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the storage medium communicate via the bus, and the processor executes the machine-readable instructions to perform the combustible waste gas heat balance method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the combustible waste gas heat balance method according to any one of claims 1 to 6 is executed.

Images