Cremation combustion efficiency and emission multi-objective optimization method and system based on intelligent algorithm
By optimizing combustion efficiency and emissions of multiple pollutants during the cremation process through intelligent algorithms, the contradiction between efficiency and emissions in the traditional cremation process is resolved. This achieves synergistic optimization of maximizing combustion efficiency and minimizing pollutants, and provides intelligent and precise control technology support.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-30
- Publication Date
- 2026-04-07
AI Technical Summary
Existing technologies struggle to balance combustion efficiency with emissions of multiple pollutants during cremation. Traditional control methods present a trade-off between efficiency and emissions and lack customized optimization algorithms for the cremation process, resulting in low operational precision, unsuitable prediction models, and poor optimization effects.
A multi-objective optimization method based on intelligent algorithms is adopted, which achieves synergistic optimization of combustion efficiency and multi-pollutant emissions through data preprocessing, combustion efficiency pattern recognition, multi-objective prediction model establishment and improved optimization algorithm.
It maximizes combustion efficiency and minimizes emissions of multiple pollutants during cremation, forming an intelligent and precise control technology closed loop that supports the green transformation of the funeral industry.
Smart Images
Figure CN121809262A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of environmental protection and intelligent control technology for funeral cremation, and particularly relates to a multi-objective optimization method and system for cremation combustion efficiency and emissions based on intelligent algorithms. Background Technology
[0002] Against the backdrop of global demographic shifts and a deepening aging population, the volume of cremation services in the funeral industry has been increasing year by year. Cremation is essentially a high-temperature combustion reaction that simultaneously produces various pollutants and greenhouse gases, including smoke, carbon monoxide, nitrogen oxides, sulfur dioxide, and carbon dioxide. For the funeral industry, there is a significant contradiction between improving combustion efficiency to reduce operating costs and reducing pollutant emissions to meet environmental protection requirements. Traditional control methods struggle to balance these two aspects. For example, increasing the main combustion chamber temperature to improve efficiency leads to a significant increase in nitrogen oxide emissions, while increasing air supply to promote complete combustion may reduce furnace temperature and efficiency. This contradiction between efficiency and emissions has become a core bottleneck restricting the industry's green development. From the perspective of current technology, firstly, current research largely focuses on controlling single pollutants, lacking systematic research on the synergistic optimization of combustion efficiency and multiple pollutants, and many proposed strategies do not address the actual parameter control of the microscopic combustion process. Secondly, existing research methods largely rely on numerical simulations based on ideal operating conditions, failing to fully consider complex variables such as differences in body composition, equipment heat loss, and data noise in actual cremation, resulting in significant deviations between research results and actual operating conditions, making practical application difficult. Furthermore, while multi-objective optimization algorithms are widely used in other fields, they are all general-purpose designs and have not been customized for the special conditions of cremation combustion, such as dynamic stages, strong nonlinearity, and high-dimensional asynchronous data. Direct application of these algorithms can easily lead to slow convergence, unstable optimization schemes, and a disconnect from actual operation. From a practical application perspective, current cremator operation relies heavily on operator experience, resulting in crude and low-precision control methods. Existing prediction models struggle to adapt to the unique characteristics of cremation data, limiting their prediction accuracy. Simultaneously, optimization schemes lack staged designs to address the significant differences in parameter requirements at different stages of combustion, often employing a one-size-fits-all control strategy, leading to poor overall optimization results. Therefore, the key technical problems this invention aims to solve are: effectively addressing the deficiencies in cremation data quality, improving the synergistic prediction accuracy of combustion efficiency and multi-pollutant emissions, designing optimization algorithms adapted to the special conditions of cremation, and achieving precise staged control. Summary of the Invention
[0003] To address the aforementioned technical problems, this invention proposes a multi-objective optimization method and system for cremation combustion efficiency and emissions based on intelligent algorithms, achieving both improved combustion efficiency and synergistic emission reduction of pollutants.
[0004] To achieve the above objectives, this invention provides a multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms, comprising: S1. Systematically preprocess the raw operational and emission data collected during the cremation process to construct a modeling dataset; S2. Based on the preprocessed data, different combustion efficiency modes are objectively identified and classified using a data-driven approach. S3. Using the preprocessed modeling dataset, train a multi-objective prediction model to establish a mapping relationship between combustion efficiency and various pollutant emission indicators from cremator control parameters. S4. Based on the multi-objective prediction model, construct a bi-objective optimization problem and solve it using an improved optimization algorithm to output the optimal control parameter schemes corresponding to different combustion stages.
[0005] Optionally, in S1, the systematic preprocessing of the raw operational and emission data collected during the cremation process includes: addressing the data missing problem by using time-series linear interpolation to repair continuous parameters and forward filling to repair discrete parameters; addressing the asynchronous timestamp problem of multi-source data by using a time window-based fuzzy merging method for data alignment; and addressing the problem of inconsistent measurement units by converting pollutant concentration indicators of different units to a consistent unit according to standard formulas and performing normalization processing.
[0006] Optionally, in S2, the combustion efficiency modes are objectively identified and classified through a data-driven approach, including: selecting flue gas temperature, carbon monoxide concentration, carbon dioxide concentration, and flue gas oxygen content as core indicators for standardization and principal component analysis dimensionality reduction; using multiple clustering algorithms to perform cluster analysis on the dimensionality-reduced data, and combining the silhouette coefficient with actual operating conditions for verification, to finally classify the combustion efficiency into three modes: low efficiency, medium efficiency, and high efficiency.
[0007] Optionally, in S3, the multi-objective prediction model is constructed using an extreme gradient boosting algorithm; the input features of the multi-objective prediction model are the preprocessed operating parameters of the cremator, and the output targets are combustion efficiency, sulfur dioxide concentration, nitrogen oxide concentration, carbon monoxide concentration, and carbon dioxide concentration; the learning rate, maximum tree depth, and regularization parameters of the model are optimized through cross-validation.
[0008] Optionally, in S4, the improved optimization algorithm is the Jaya algorithm, which integrates a back-learning strategy and a simulated annealing mechanism; wherein the back-learning strategy is used to generate a diverse initial population, and the simulated annealing mechanism is used to accept inferior solutions with probability during the individual update process in order to escape local optima.
[0009] Optionally, in S4, the parameters of the improved optimization algorithm are optimized using the Taguchi experiment method, with the quality of the Pareto solution set as the evaluation index, to determine the optimal combination of parameters, including the number of iterations, population size, and initial temperature.
[0010] Optionally, in S4, the phased solution includes: dividing the cremation combustion process into five stages: combustion preparation, early combustion, middle combustion, late combustion, and combustion end; running the improved optimization algorithm for each stage to independently solve and obtain the Pareto optimal control parameter scheme corresponding to that stage.
[0011] On the other hand, to achieve the above objectives, the present invention also provides a multi-objective optimization system for cremation combustion efficiency and emissions based on intelligent algorithms, comprising: The data preprocessing module is used to systematically preprocess the raw operational and emission data collected during the cremation process in order to construct a modeling dataset. The combustion mode classification module is used to objectively identify and classify different combustion efficiency modes based on preprocessed data in a data-driven manner. A multi-objective prediction module is used to train a multi-objective prediction model using the preprocessed modeling dataset to establish a mapping relationship between combustion efficiency and various pollutant emission indicators from cremator control parameters. The optimization and control module is used to construct a bi-objective optimization problem based on the multi-objective prediction model, and solve it using an improved optimization algorithm to output the optimal control parameter scheme corresponding to different combustion stages.
[0012] An electronic device, the electronic device comprising: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, it implements the multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms.
[0013] A computer storage medium storing computer program instructions, which, when executed by a processor, implement the aforementioned multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms.
[0014] Technical Effects of this Invention: This invention discloses a multi-objective optimization method and system for cremation combustion efficiency and emissions based on intelligent algorithms. Through a systematic data preprocessing strategy, it effectively addresses the issues of missing, asynchronous, and inconsistent units in cremation data, providing a high-quality data foundation for subsequent modeling. The data-driven combustion efficiency classification method objectively identifies different combustion state modes, overcoming the subjectivity of human experience. The constructed multi-objective prediction model accurately captures the complex nonlinear relationship between cremation parameters and combustion performance, significantly improving prediction reliability. The optimized algorithm for cremation conditions combines reverse learning and simulated annealing mechanisms, and performs parameter optimization and staged solution, significantly enhancing the algorithm's optimization ability and the practicality of the solution in the specific scenario of cremation. Ultimately, an integrated technical closed loop is formed, achieving the synergistic optimization goal of maximizing combustion efficiency and minimizing multiple pollutant emissions, providing reliable technical support for the intelligent and precise control of the cremation process and the green transformation of the funeral industry. Attached Figure Description
[0015] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart illustrating a multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms, according to an embodiment of the present invention. Figure 2 This is a clustering profile coefficient diagram of an embodiment of the present invention; Figure 3 Visualization of K-Means clustering results in this embodiment of the invention; Figure 4 Visualization of hierarchical clustering results in an embodiment of the present invention; Figure 5 Visualization of DBSCAN clustering results in an embodiment of the present invention; Figure 6 Visualization of GMM clustering results in an embodiment of the present invention; Figure 7 This is a diagram of the XGBoost prediction algorithm model according to an embodiment of the present invention; Figure 8 The flowchart of the improved Jaya algorithm is shown in the embodiment of the present invention. Detailed Implementation
[0016] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.
[0017] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.
[0018] like Figure 1 As shown, this embodiment provides a multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms, including: S1. Systematically preprocess the raw operational and emission data collected during the cremation process to construct a modeling dataset; S2. Based on the preprocessed data, different combustion efficiency modes are objectively identified and classified using a data-driven approach. S3. Using the preprocessed modeling dataset, train a multi-objective prediction model to establish a mapping relationship between combustion efficiency and various pollutant emission indicators from cremator control parameters. S4. Based on the multi-objective prediction model, construct a bi-objective optimization problem and solve it using an improved optimization algorithm to output the optimal control parameter schemes corresponding to different combustion stages.
[0019] Specifically, such as Figure 1 As shown, this technical approach revolves around a prediction-optimization synergy mechanism. First, a theoretical framework is constructed through literature review. Then, a data foundation is established through data preprocessing. Subsequently, a multi-objective prediction model is established using the XGBoost algorithm. Next, a multi-objective optimization model is constructed with the objectives of maximizing combustion efficiency and minimizing pollutant emissions. This model is then solved using the Jaya algorithm, which is optimized by Taguchi orthogonal experiments. Finally, the effectiveness of the scheme is verified through comparative experiments, and the optimal combustion decision strategy is analyzed.
[0020] Furthermore, in S1, the systematic preprocessing of the raw operational and emission data collected during the cremation process includes: addressing the data missing problem by using time-series linear interpolation to repair continuous parameters and forward filling to repair discrete parameters; addressing the asynchronous timestamp problem of multi-source data by using a time window-based fuzzy merging method for data alignment; and addressing the problem of inconsistent measurement units by converting pollutant concentration indicators of different units to a consistent unit according to standard formulas and performing normalization processing.
[0021] Specifically, the implementation process of this embodiment includes: To address the issues of missing data, asynchronous timing, and inconsistent units, a three-step processing strategy is designed: Accurate missing value repair: The location of missing data is identified by the isnull() function of the Pandas library, and the missing rate of each feature is calculated; for continuous key parameters such as main combustion chamber temperature and flue gas oxygen content, the time-series linear interpolation method (pandas.DataFrame.interpolate(method='linear')) is used to repair short-term missing values within 5 minutes, and the time-series continuity of data is used to ensure repair accuracy; for discrete indicators such as valve opening and closing status, the forward filling method (ffill) is used to avoid introducing false signals.
[0022] Multi-source data time alignment: The timestamp fields of all data sources are uniformly converted to Pandas datetime type and sorted in ascending order of time; the timestamp-based fuzzy merging (pandas.merge_asof) method is adopted, and the maximum allowed time difference Δt=30 seconds is set to achieve accurate alignment of equipment operation data, flue gas data, and infrared temperature data, ensuring that the samples at the same time contain complete input-output features.
[0023] Standardized unit conversion: For concentration indicators such as CO and CO2, unit conversion was performed according to standard formulas to ensure consistent measurement standards for all indicators. Simultaneously, normalization was applied due to the different magnitudes of data for different variables. The formula for converting CO from mg / m³ to ppm is: (1); in, The volume concentration of carbon monoxide (CO) is expressed in ppm. The mass concentration of carbon monoxide (unit: mg / m³). The value represents the molar mass of carbon monoxide (in g / mol). The formula for converting CO2 from mg / m³ to volume fraction (%) is: (2); in, This represents the volume concentration of carbon monoxide (CO) (in %).
[0024] Furthermore, in S2, the objective identification and classification of combustion efficiency modes through a data-driven approach includes: selecting flue gas temperature, carbon monoxide concentration, carbon dioxide concentration, and flue gas oxygen content as core indicators for standardization and principal component analysis dimensionality reduction; employing multiple clustering algorithms to perform cluster analysis on the dimensionality-reduced data, and combining silhouette coefficients with actual operating conditions for verification, ultimately classifying combustion efficiency into three modes: low efficiency, medium efficiency, and high efficiency. For example... Figure 2As shown in the figure, the silhouette coefficient changes for different cluster numbers K: when the K value increases from 2 to 5, the silhouette coefficient first rises and then falls. The coefficient reaches its peak when K=3 and drops to its trough when K=5. This can help determine the optimal cluster K=3. Based on this, the present invention selects K=3 as the optimal cluster number for this clustering.
[0025] Specifically, the implementation process of this embodiment includes: Based on the preprocessed data, an objective classification of combustion efficiency patterns is achieved through a two-step method of "dimensionality reduction-clustering": Dimensionality reduction of high-dimensional data: Four core indicators, namely flue gas temperature, CO concentration, CO2 concentration and flue gas oxygen content, were selected. After standardization, principal component analysis (PCA) was used to reduce the dimensionality and extract two principal components (PCA1 represents "combustion enrichment", with high values corresponding to high CO2 and low oxygen content; PCA2 represents "combustion completeness", with high values corresponding to high CO and low flue gas temperature), so as to achieve data dimensionality compression and key information retention.
[0026] Multi-algorithm clustering validation: Four classic clustering algorithms—K-Means, hierarchical clustering, DBSCAN, and Gaussian mixture model—were used to cluster the dimensionality-reduced data. By calculating the silhouette coefficient for different K values and validating the results under actual combustion conditions, combustion efficiency was ultimately classified into three categories: "low efficiency (excessive ventilation leading to incomplete combustion), medium efficiency (insufficient fuel combustion), and high efficiency (ideal complete combustion)," providing a labeling basis for subsequent phased optimization. Figure 3 As shown in the figure, the distribution of K-Means clustering results is as follows: sample points of different colors form three independent clusters on the two-dimensional plane of "contribution of high CO emissions and low flue gas temperature" and "contribution of high CO2 emissions and low flue gas oxygen content". The samples within each cluster have high similarity in the two-dimensional features, while the features between clusters are significantly different, indicating that K-Means clustering effectively distinguishes samples according to these two-dimensional features. Figure 4 As shown in the figure, the distribution of the hierarchical clustering results is as follows: sample points of different colors form three independent clusters on the two-dimensional plane of "contribution of high CO emissions and low flue gas temperature" and "contribution of high CO2 emissions and low flue gas oxygen content". The samples within each cluster have high similarity in the two-dimensional features, while the features between clusters are significantly different, indicating that hierarchical clustering effectively distinguishes samples according to these two-dimensional features. Figure 5 As shown in the figure, the distribution of DBSCAN clustering results is as follows: sample points of different colors form three independent clusters on the two-dimensional plane of "contribution of high CO emissions and low flue gas temperature" and "contribution of high CO2 emissions and low flue gas oxygen content". The samples within each cluster have high similarity in the two-dimensional features, while the features between clusters are significantly different, indicating that DBSCAN clustering effectively distinguishes samples according to these two-dimensional features. Figure 6As shown in the figure, the distribution of the GMM clustering results is as follows: the sample points of different colors form three independent clusters on the two-dimensional plane of "contribution of high CO emission and low flue gas temperature" and "contribution of high CO2 emission and low flue gas oxygen content". The two-dimensional features of the samples within each cluster are highly similar, and the features between clusters are significantly different, indicating that GMM clustering effectively distinguishes the samples according to the features of these two dimensions.
[0027] Furthermore, in S3, the multi-objective prediction model is constructed using an extreme gradient boosting algorithm; the input features of the multi-objective prediction model are the preprocessed operating parameters of the cremator, and the output targets are combustion efficiency, sulfur dioxide concentration, nitrogen oxide concentration, carbon monoxide concentration, and carbon dioxide concentration; the learning rate, maximum tree depth, and regularization parameters of the model are optimized through cross-validation.
[0028] Specifically, the implementation process of this embodiment includes: The Extreme Gradient Boosting (XGBoost) algorithm is selected to construct a predictive model for combustion efficiency and multi-pollutant emissions, addressing the challenges of nonlinear and multivariate interactive prediction. Model structure design: The pre-processed equipment operating parameters (19 items including main / secondary combustion chamber temperature, top wind speed, and instantaneous flow rate) are used as input features, with combustion efficiency (based on flue gas analysis) as the input characteristic. The model uses CO, NO, NO2, SO2, and CO2 concentrations as output targets to construct a multi-output prediction model. The model adopts a gradient boosting framework and generates 150 decision trees (n_estimators=150) through iteration. Each tree is optimized based on the prediction residual of the previous round to achieve nonlinear relationship fitting.
[0029] Hyperparameter optimization configuration: Key parameters were adjusted using 5-fold cross-validation. The learning rate was set to 0.1 (to control the contribution of each tree and avoid model oscillation), the maximum tree depth was set to 8 (to balance the model's segmentation ability with the risk of overfitting), the subsample ratio and column sample ratio (colsample_bytree) were both set to 0.8 (to improve generalization ability through random sampling), the regularization parameter λ=1 (to control model complexity), and the mean squared error (MSE) loss function was used to ensure that the model maintains stable performance on both the training and validation sets.
[0030] Model performance validation: The dataset is split into training, validation, and test sets in a 70%:15%:15% ratio. The training set is used for model fitting, the validation set is used for hyperparameter tuning, and the test set is used for performance evaluation. The three metrics RMSE (root mean square error), R² (coefficient of determination), and MAE (mean absolute error) are used for evaluation. The following process is used to train and validate the model, as shown in Table 1.
[0031] Table 1 Validation results show that the model achieves an R² of 0.808 in combustion efficiency prediction on the test set, meeting the accuracy requirements for subsequent multi-objective optimization. Figure 7 As shown in the figure, the model flow of the XGBoost prediction algorithm is as follows: First, the training dataset D is processed to obtain multiple training subsets. Then, base models such as CART 1 to CART K are trained respectively. After the output of each base model is updated and optimized, the ensemble learning model is finally constructed by integrating the outputs of each base model.
[0032] Furthermore, in S4, the improved optimization algorithm is the Jaya algorithm, which integrates a back-learning strategy and a simulated annealing mechanism; wherein, the back-learning strategy is used to generate a diverse initial population, and the simulated annealing mechanism is used to accept inferior solutions with probability during the individual update process in order to escape local optima.
[0033] Furthermore, in S4, the parameters of the improved optimization algorithm are optimized using the Taguchi experiment method, with the quality of the Pareto solution set as the evaluation index, to determine the optimal parameter combination, including the number of iterations, population size, and initial temperature.
[0034] Specifically, the implementation process of this embodiment includes: Improved Jaya algorithm for finding the optimal solution: Based on the prediction model, a dual-objective optimization model is constructed. A phased optimization is achieved by improving the Jaya algorithm, thus addressing the problem of algorithm adaptability to combustion conditions. Construction of dual objective function: Objective 1 is to maximize combustion efficiency, using combustion efficiency based on flue gas analysis. .
[0035] (3); in, This represents the proportion of heat loss from flue gas exhaust. This represents the percentage of heat loss due to incomplete combustion. , ; The temperature of the exhaust gas; The ambient temperature.
[0036] Objective 2 is to minimize pollutant emissions, using a weighted summation formula: (4); in, The goal is to minimize pollutant emissions; The index number of the remains; Index of Remains A set; This is the index number of the cremator; Index for EDM machines A set; For the remains Arranged to crematorium The value is 1 if it is 1, otherwise it is 0. This represents the weighted value for pollutant CO; For the remains In the cremator CO emissions during cremation; This represents the weighted value for pollutant CO2. For the remains In the cremator CO2 emissions during cremation; This represents the weight value for pollutant SO2; For the remains In the cremator SO2 emissions during cremation; This represents the weighted value for pollutant NO; For the remains In the cremator NO emissions during cremation; This represents the weighted value for pollutant NO2. For the remains In the cremator NO2 emissions during cremation.
[0037] Constraint settings: Considering the safe operation and environmental protection requirements of the cremator, three types of constraints are set: equipment parameter constraints, oxygen content constraints, and control parameter constraints.
[0038] Improved Jaya algorithm design: To address the issues of insufficient initial population diversity and susceptibility to local optima in the traditional Jaya algorithm, three improvements were made based on the characteristics of cremation conditions. The specific process is shown in Table 2.
[0039] Table 2 The algorithm is described in detail below: Initial population optimization: Combined with the reverse learning strategy, an initial population and a reverse population are generated. The optimal individuals that meet the population size are selected as the initial population through non-dominated sorting to improve the diversity of optimization. The specific expression of the reverse learning strategy is shown in equation (5). Individual update mechanism: Incorporate simulated annealing acceptance mechanism. If the new individual dominates the original individual, it will be accepted directly. Otherwise, it will be accepted according to the probability of Equation (6) to avoid getting trapped in local optima. Parameter optimization configuration: The optimal parameters were determined through Taguchi experiments (4 factors and 4 levels). The target deviation sum (ODS) was used as the evaluation index. The smaller the ODS value, the better the Pareto solution set effect. Finally, the optimal parameter combination (150 iterations, 700 population size, 80 initial temperature, and 0.9 temperature decrease coefficient) was selected to ensure that the algorithm balances convergence speed and solution quality.
[0040] like Figure 8 As shown in the figure, the process of the Jaya algorithm is as follows: starting from the initialization of the population, the best and worst solutions in the population are identified, and new individuals are generated based on the two. It is then determined whether the new individuals are better to decide whether to accept or retain the original individuals. This process is repeated until the termination condition is met, and finally the population is output.
[0041] Furthermore, in S4, the phased solution includes: dividing the cremation combustion process into five stages: combustion preparation, early combustion, middle combustion, late combustion, and combustion end; running the improved optimization algorithm for each stage to independently solve and obtain the Pareto optimal control parameter scheme corresponding to that stage.
[0042] Specifically, the implementation process of this embodiment includes: The combustion process is divided into five stages: combustion preparation (before cremation starts), early combustion (0-25% of the duration), middle combustion (25%-75% of the duration), late combustion (75%-100% of the duration), and combustion end (after cremation ends). Algorithms are run for each stage to output the optimal control parameters for each stage.
[0043] (5); (6); (7); in, Index for the dimension of the individual vector; Index for the target; This represents the number of iterations. For individual The reverse element generated by the dimension; For the individual vector of the first The upper boundary of the range of values that a dimensional element can take; For the individual vector of the first The lower boundary of the range of values that a dimensional element can take; For the individual vector of the first Initial values for dimensional elements; To simulate the acceptance probability value of the annealing mechanism; For the first The change in each target value; The current temperature; The initial temperature; Indicates the coefficient of temperature decrease Power of 1.
[0044] This embodiment also provides a multi-objective optimization system for cremation combustion efficiency and emissions based on intelligent algorithms, including: The data preprocessing module is used to systematically preprocess the raw operational and emission data collected during the cremation process in order to construct a modeling dataset. The combustion mode classification module is used to objectively identify and classify different combustion efficiency modes based on preprocessed data in a data-driven manner. A multi-objective prediction module is used to train a multi-objective prediction model using the preprocessed modeling dataset to establish a mapping relationship between combustion efficiency and various pollutant emission indicators from cremator control parameters. The optimization and control module is used to construct a bi-objective optimization problem based on the multi-objective prediction model, and solve it using an improved optimization algorithm to output the optimal control parameter scheme corresponding to different combustion stages.
[0045] An electronic device, the electronic device comprising: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, it implements the multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms.
[0046] A computer storage medium storing computer program instructions, which, when executed by a processor, implement the aforementioned multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms.
[0047] This invention discloses a multi-objective optimization method and system for cremation combustion efficiency and emissions based on intelligent algorithms. Through a systematic data preprocessing strategy, it effectively addresses the issues of missing, asynchronous, and inconsistent units in cremation data, providing a high-quality data foundation for subsequent modeling. A data-driven combustion efficiency classification method objectively identifies different combustion state modes, overcoming the subjectivity of human experience. The constructed multi-objective prediction model accurately captures the complex nonlinear relationship between cremation parameters and combustion performance, significantly improving prediction reliability. The optimized algorithm, tailored to cremation conditions, combines back-learning and simulated annealing mechanisms, and incorporates parameter optimization and phased solutions, significantly enhancing the algorithm's optimization capability and the practicality of the solutions in the specific scenario of cremation. Ultimately, an integrated technical closed loop is formed, achieving the synergistic optimization goal of maximizing combustion efficiency and minimizing multiple pollutant emissions, providing reliable technical support for intelligent and precise control of the cremation process and the green transformation of the funeral industry.
[0048] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms, characterized in that, include: S1. Systematically preprocess the raw operational and emission data collected during the cremation process to construct a modeling dataset; S2. Based on the preprocessed data, different combustion efficiency modes are objectively identified and classified using a data-driven approach. S3. Using the preprocessed modeling dataset, train a multi-objective prediction model to establish a mapping relationship between combustion efficiency and various pollutant emission indicators from cremator control parameters. S4. Based on the multi-objective prediction model, construct a bi-objective optimization problem and solve it using an improved optimization algorithm to output the optimal control parameter schemes corresponding to different combustion stages.
2. The multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms as described in claim 1, characterized in that, In S1, the systematic preprocessing of the raw operational and emission data collected during the cremation process includes: addressing the data missing problem by using time-series linear interpolation to repair continuous parameters and forward filling to repair discrete parameters; addressing the asynchronous timestamp problem of multi-source data by using a time window-based fuzzy merging method for data alignment; and addressing the problem of inconsistent measurement units by converting pollutant concentration indicators of different units to consistent units according to standard formulas and performing normalization processing.
3. The multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms as described in claim 2, characterized in that, In S2, the combustion efficiency modes are objectively identified and classified through a data-driven approach, including: selecting flue gas temperature, carbon monoxide concentration, carbon dioxide concentration, and flue gas oxygen content as core indicators for standardization and principal component analysis dimensionality reduction; using multiple clustering algorithms to perform cluster analysis on the dimensionality-reduced data; and combining the silhouette coefficient with actual operating conditions for verification, ultimately classifying the combustion efficiency into three modes: low efficiency, medium efficiency, and high efficiency.
4. The multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms as described in claim 3, characterized in that, In S3, the multi-objective prediction model is constructed using an extreme gradient boosting algorithm. The input features of the multi-objective prediction model are the preprocessed operating parameters of the cremator, and the output targets are combustion efficiency, sulfur dioxide concentration, nitrogen oxide concentration, carbon monoxide concentration, and carbon dioxide concentration. The learning rate, maximum tree depth, and regularization parameters of the model are optimized through cross-validation.
5. The multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms as described in claim 4, characterized in that, In S4, the improved optimization algorithm is the Jaya algorithm, which integrates a back-learning strategy and a simulated annealing mechanism. The back-learning strategy is used to generate a diverse initial population, and the simulated annealing mechanism is used to accept inferior solutions with probability during individual update to escape local optima.
6. The multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms as described in claim 5, characterized in that, In S4, the parameters of the improved optimization algorithm are optimized using the Taguchi experiment method, with the quality of the Pareto solution set as the evaluation index, to determine the optimal combination of parameters, including the number of iterations, population size, and initial temperature.
7. The multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms as described in claim 6, characterized in that, In S4, the phased solution includes: dividing the cremation combustion process into five stages: combustion preparation, early combustion, middle combustion, late combustion, and combustion end; running the improved optimization algorithm for each stage to independently solve and obtain the Pareto optimal control parameter scheme corresponding to that stage.
8. A multi-objective optimization system for cremation combustion efficiency and emissions based on intelligent algorithms, characterized in that, For implementing the multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms as described in any one of claims 1-7, the system comprises: The data preprocessing module is used to systematically preprocess the raw operational and emission data collected during the cremation process in order to construct a modeling dataset. The combustion mode classification module is used to objectively identify and classify different combustion efficiency modes based on preprocessed data in a data-driven manner. A multi-objective prediction module is used to train a multi-objective prediction model using the preprocessed modeling dataset to establish a mapping relationship between combustion efficiency and various pollutant emission indicators from cremator control parameters. The optimization and control module is used to construct a bi-objective optimization problem based on the multi-objective prediction model, and solve it using an improved optimization algorithm to output the optimal control parameter scheme corresponding to different combustion stages.
9. An electronic device, characterized in that, The electronic device includes: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, it implements the multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms as described in any one of claims 1-7.
10. A computer storage medium, characterized in that, The computer storage medium stores computer program instructions, which, when executed by a processor, implement the multi-objective optimization method for cremation combustion efficiency and emissions based on intelligent algorithms as described in any one of claims 1-7.