XGBoost model-based sludge thermal hydrolysis temperature dynamic regulation and control method

Abstract

The invention provides a sludge thermal hydrolysis temperature dynamic regulation and control method based on an XGBoost model, which is applied to the technical field of sludge treatment.The method comprises the steps that detection physicochemical property parameters, preset thermal hydrolysis process parameters and preset anaerobic digestion parameters of sludge to be regulated and controlled are obtained; inputting the detected physicochemical property parameters, the preset thermal hydrolysis process parameters and the preset anaerobic digestion parameters into a trained target prediction model based on an XGBoost model to obtain predicted net energy output output by the target prediction model; based on the predicted net energy output, target thermal hydrolysis process parameters are determined, and the target thermal hydrolysis process parameters are preset thermal hydrolysis process parameters when the predicted net energy output is the maximum value. According to the invention, pyrohydrolysis temperature control can be optimized in real time according to different sludge characteristics and process parameters, and the efficiency and economical efficiency of sludge treatment are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of sludge treatment, and particularly to a method for dynamically regulating the temperature of sludge hydrothermal hydrolysis based on the XGBoost model. Background Art

[0002] For sludge with a high organic matter content, a relatively low hydrothermal hydrolysis intensity (low temperature, short time) may result in insufficient hydrolysis of the organic matter in the sludge, reducing the conversion degree of the sludge in subsequent anaerobic digestion; while for sludge with a low organic matter content, a relatively high hydrothermal hydrolysis intensity (high temperature, long time) may cause energy waste.

[0003] Regional, seasonal variations, and sewage treatment processes all have significant effects on sludge characteristics. However, the hydrothermal hydrolysis process parameters of most current sewage treatment plants always remain fixed, which will undoubtedly result in over - or under - hydrothermal hydrolysis. In order to select the optimal hydrothermal hydrolysis temperature under specific sludge characteristic conditions, traditional anaerobic digestion optimization experiments are difficult to achieve this optimization because they are time - consuming and laborious and not conducive to quickly obtaining the required experimental results.

[0004] Therefore, it can be seen that the hydrothermal hydrolysis temperature regulation method in the related technology has the technical problem of low efficiency. Summary of the Invention

[0005] The present invention provides a method for dynamically regulating the temperature of sludge hydrothermal hydrolysis based on the XGBoost model, which is used to solve the defect of low efficiency in the existing hydrothermal hydrolysis temperature regulation method, and realizes real - time optimization of the hydrothermal hydrolysis temperature control according to different sludge characteristics and process parameters, improving the efficiency and economy of sludge treatment.

[0006] The present invention provides a method for dynamically regulating the temperature of sludge hydrothermal hydrolysis based on the XGBoost model, including the following steps. Obtain the detected physical and chemical characteristic parameters of the sludge to be regulated, the preset hydrothermal hydrolysis process parameters, and the preset anaerobic digestion parameters; input the detected physical and chemical characteristic parameters, the preset hydrothermal hydrolysis process parameters, and the preset anaerobic digestion parameters into the trained target prediction model based on the XGBoost model to obtain the predicted net energy output output by the target prediction model; determine the target hydrothermal hydrolysis process parameters based on the predicted net energy output, where the target hydrothermal hydrolysis process parameters are the preset hydrothermal hydrolysis process parameters when the predicted net energy output is the maximum value.

[0007] A method for dynamically regulating the temperature of sludge hydrothermal hydrolysis based on the XGBoost model provided by the present invention. Before inputting the detected physicochemical characteristic parameters, the preset hydrothermal hydrolysis process parameters, and the preset anaerobic digestion parameters into the trained target prediction model based on the XGBoost model, the method further includes: obtaining the sample physicochemical characteristic parameters, sample hydrothermal hydrolysis process parameters, sample anaerobic digestion parameters, and sample net energy output of the sludge sample; using the sample physicochemical characteristic parameters, the sample hydrothermal hydrolysis process parameters, and the sample anaerobic digestion parameters as input features, and using the sample net energy output as the output feature; performing data preprocessing based on the input features and the output features to obtain a sludge sample data set; training a preset prediction model based on the XGBoot model using the sludge sample data set to obtain a trained target prediction model.

[0008] A method for dynamically regulating the temperature of sludge hydrothermal hydrolysis based on the XGBoost model provided by the present invention. The performing data preprocessing based on the input features and the output features to obtain a sludge sample data set includes: removing missing data from and performing statistical analysis on the input features and the output features to obtain statistical data; performing normalization processing on the statistical data to obtain a sludge sample data set, where the normalization processing includes: ; where represents the -dimensional data in the sludge sample data set, represents the -dimensional data in the statistical data, represents the mean of the statistical data, represents the standard deviation of the statistical data.

[0009] A method for dynamically regulating the temperature of sludge hydrothermal hydrolysis based on the XGBoost model provided by the present invention. The training a preset prediction model based on the XGBoot model using the sludge sample data set to obtain a trained target prediction model includes: inputting the sludge sample data set into a preset prediction model based on the XGBoot model to obtain the sample predicted net energy output output by the preset prediction model; determining the target loss function of the sample predicted net energy output and the sample true net energy output through the mean square error; performing supervised training on the preset prediction model based on the target loss function, and obtaining a trained target prediction model through iterative optimization.

[0010] A method for dynamically regulating the sludge hydrothermal hydrolysis temperature based on the XGBoost model provided by the present invention further includes: adjusting the hyperparameters of the target prediction model through cross-validation error to obtain a plurality of adjusted hyperparameters; and determining the target hyperparameters based on the plurality of adjusted hyperparameters through a genetic algorithm.

[0011] A method for dynamically regulating the sludge hydrothermal hydrolysis temperature based on the XGBoost model provided by the present invention, the target loss function includes: ; wherein, represents the target loss function, represents the number of samples in the sludge sample data set, represents the sample index of the sludge sample data set, represents the th sample's true net energy output, represents the th sample's predicted net energy output, represents the regularization coefficient, represents the sum of the absolute values of all weights.

[0012] The present invention also provides a system for dynamically regulating the sludge hydrothermal hydrolysis temperature based on the XGBoost model, including the following modules: an acquisition module, configured to acquire the detected physical and chemical characteristic parameters of the sludge to be regulated, the preset hydrothermal hydrolysis process parameters, and the preset anaerobic digestion parameters; a prediction module, configured to input the detected physical and chemical characteristic parameters, the preset hydrothermal hydrolysis process parameters, and the preset anaerobic digestion parameters into the trained target prediction model based on the XGBoost model to obtain the predicted net energy output output by the target prediction model; a determination module, configured to determine the target hydrothermal hydrolysis process parameters based on the predicted net energy output, wherein the target hydrothermal hydrolysis process parameters are the preset hydrothermal hydrolysis process parameters when the predicted net energy output is the maximum value.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the program, it implements the method for dynamically regulating the sludge hydrothermal hydrolysis temperature based on the XGBoost model as described in any one of the above.

[0014] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the method for dynamically regulating the sludge hydrothermal hydrolysis temperature based on the XGBoost model as described in any one of the above.

[0015] The present invention also provides a computer program product, including a computer program which, when executed by a processor, implements the method for dynamically regulating the temperature of sludge hydrothermal hydrolysis based on the XGBoost model as described in any one of the above.

[0016] The method for dynamically regulating the temperature of sludge hydrothermal hydrolysis based on the XGBoost model provided by the present invention provides a necessary data basis for subsequent model prediction by obtaining the detected physical and chemical characteristic parameters of the sludge to be regulated, the preset hydrothermal hydrolysis process parameters, and the preset anaerobic digestion parameters; through the target prediction model based on the XGBoost model, it can comprehensively consider the physical and chemical characteristic parameters of the sludge, the preset hydrothermal hydrolysis process parameters, and the preset anaerobic digestion parameters, so as to accurately predict the net energy output under different hydrothermal hydrolysis process conditions; based on the predicted net energy output, by comparing the prediction results under different hydrothermal hydrolysis process parameters, the target hydrothermal hydrolysis process parameters that maximize the net energy output can be determined. Thus, accurate regulation operations can be achieved based on the target hydrothermal hydrolysis process parameters, and at the same time, a large amount of experimental and time cost consumption is saved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required to be used in the description of the embodiments or the prior art one by one. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 is a flowchart of the method for dynamically regulating the temperature of sludge hydrothermal hydrolysis based on the XGBoost model provided by the present invention.

[0019] Figure 2 is a data graph of the experiment and evaluation results of the target prediction model provided by the present invention.

[0020] Figure 3 is a structural diagram of the method for dynamically regulating the temperature of sludge hydrothermal hydrolysis based on the XGBoost model provided by the present invention.

[0021] Figure 4 is an interface diagram of the input, output, and regulation of the online prediction website provided by the present invention.

[0022] Figure 5 is a data graph of the regulation results of the target prediction model provided by the present invention.

[0023] Figure 6 is a module diagram of the system for dynamically regulating the temperature of sludge hydrothermal hydrolysis based on the XGBoost model provided by the present invention.

[0024] Figure 7It is a schematic diagram of the physical structure of the electronic device provided by the present invention. Detailed implementation manners

[0025] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without making creative efforts shall fall within the protection scope of the present invention.

[0026] With the increase in the sludge production of sewage treatment plants, sludge treatment and resource utilization have become increasingly important. Anaerobic digestion is one of the main technologies for sludge treatment. It not only promotes sludge reduction and harmlessness, but also can produce biogas for providing heat and electricity for sewage treatment plants and surrounding residents, thereby reducing the operating costs of sewage treatment plants. However, since sludge (mainly excess activated sludge) has a large amount of extracellular polymers that adhere to each other to form a floc structure, it seriously hinders the hydrolysis and acidification of organic matter during the anaerobic digestion of sludge. The hydrothermal hydrolysis technology is an important means to break this barrier. It refers to the reaction of materials with saturated steam under specific temperature and pressure conditions. Through processes such as pulping, hydrothermal treatment, and flash evaporation, it promotes the disintegration of sludge flocs, the rupture of microbial cells, and the release of a large amount of organic matter in extracellular polymers and cells into the supernatant, thereby improving the anaerobic digestion performance and sludge dewatering performance of sludge and achieving complete elimination of pathogens. Although hydrothermal hydrolysis pretreatment can effectively promote methane production in the anaerobic digestion of sludge, due to the high specific heat of sludge (close to that of liquid water) and the high temperature conditions (150-180 °C) required in the hydrothermal hydrolysis process, a large amount of energy is often consumed. Therefore, balancing the energy input of the hydrothermal hydrolysis technology and the additional energy output for promoting methane production in sludge is crucial for the sludge hydrothermal hydrolysis-anaerobic digestion process.

[0027] The hydrothermal hydrolysis temperature is one of the important factors affecting the energy balance. Although there are currently a large number of studies on the influence of different hydrothermal hydrolysis process parameters on the anaerobic digestion performance of sludge, there has always been a controversy over the optimal conditions of the hydrothermal hydrolysis temperature. Some studies have found that the optimal hydrothermal hydrolysis temperature of sludge depends on sludge characteristics (TS, VS, glycoprotein lipid organic matter components, etc.). The maximum solubility of carbohydrates usually occurs at 140-150 °C, while the maximum solubility of proteins is around 180 °C. In addition, for sludge with a high organic matter content, a lower hydrothermal hydrolysis intensity (low temperature, short time) may cause insufficient hydrolysis of organic matter in the sludge, reducing the conversion degree of the sludge in subsequent anaerobic digestion; while for sludge with a low organic matter content, a higher hydrothermal hydrolysis intensity (high temperature, long time) may result in energy waste.

[0028] Geographical location, seasonal variations, and sewage treatment processes can all have a significant impact on sludge characteristics. However, the thermochemical hydrolysis process parameters in most current sewage treatment plants always remain fixed, which will undoubtedly result in over- or under-hydrolysis. To select the optimal thermochemical hydrolysis temperature under specific sludge characteristics, traditional anaerobic digestion optimization experiments are difficult to achieve this optimization because they are time-consuming and laborious and not conducive to quickly obtaining the required experimental results.

[0029] Therefore, the present invention provides an intelligent control method based on the XGBoost model, which can dynamically predict the energy output benefit of sludge anaerobic digestion and optimize the thermochemical hydrolysis temperature, thereby achieving the best net energy output. Through the present invention, a learning mode or mapping relationship between sludge characteristics - thermochemical hydrolysis and anaerobic digestion process parameters - net energy benefit can be established, and with the goal of maximizing the net energy output of the sludge thermochemical hydrolysis - anaerobic digestion process, dynamic control of the optimal thermochemical hydrolysis process parameters with changes in sludge characteristics is realized.

[0030] The present invention relates to the field of sludge treatment and energy recovery, especially the dynamic control management of the thermochemical hydrolysis temperature of sludge. Specifically, it is a method and system for controlling the thermochemical hydrolysis temperature of sludge based on the XGBoost model to maximize the net energy output of the sludge thermochemical hydrolysis - anaerobic digestion system. Among them, the above method takes the physical and chemical characteristics of sludge, thermochemical hydrolysis process parameters, and anaerobic digestion parameters as input features, predicts the net energy output of sludge anaerobic digestion through the XGBoost model, and takes maximizing the net energy output as the objective function to realize intelligent control of the thermochemical hydrolysis temperature. The above system integrates functions such as data collection, data processing, model prediction, temperature control, and feedback optimization, and can optimize the thermochemical hydrolysis temperature control in real time according to different sludge characteristics and process parameters, improving the efficiency and economy of sludge treatment.

[0031] Optionally, the method for dynamically controlling the thermochemical hydrolysis temperature of sludge based on the XGBoost model in the embodiments of the present application can be executed by a server, or by a terminal device, or jointly by a server and a terminal device. Taking the server executing the method for dynamically controlling the thermochemical hydrolysis temperature of sludge based on the XGBoost model in this embodiment as an example.

[0032] Reference Figure 1 , Figure 1 is a schematic flowchart of the method for dynamically controlling the thermochemical hydrolysis temperature of sludge based on the XGBoost model provided by the present invention. As Figure 1 shown, the method includes the following: Step 101, obtain the detected physical and chemical characteristic parameters, preset thermochemical hydrolysis process parameters, and preset anaerobic digestion parameters of the sludge to be regulated.

[0033] In an embodiment of the present invention, the physical and chemical properties of the sludge to be regulated are detected to determine the preset hydrothermal hydrolysis technical parameters of the sewage treatment plant and the preset anaerobic digestion conditions of the sludge biogas project.

[0034] Among them, the physical and chemical properties of the sludge include organic matter content (VS / TS), soluble COD ratio (SCOD / TCOD), proportion of carbohydrate organic matter, proportion of protein organic matter, and proportion of fat organic matter. The hydrothermal hydrolysis process parameters include hydrothermal hydrolysis temperature and hydrothermal hydrolysis duration. The anaerobic digestion parameters include fermentation temperature, fermentation concentration, and organic load.

[0035] Step 102, input the detected physical and chemical property parameters, preset hydrothermal hydrolysis process parameters, and preset anaerobic digestion parameters into the trained target prediction model based on the XGBoost model to obtain the predicted net energy output output by the target prediction model.

[0036] In an embodiment of the present invention, the detected sludge physical and chemical property parameters, preset hydrothermal hydrolysis process parameters, and preset anaerobic digestion parameters are used as input features, and the trained target prediction model based on the XGBoost model is used to predict the input features to obtain the predicted net energy output.

[0037] Among them, the format of the input data is consistent with the data format used during model training.

[0038] In some embodiments, the collected detected physical and chemical property parameters, preset hydrothermal hydrolysis process parameters, and preset anaerobic digestion parameters are cleaned to remove outliers and missing values; the numerical data is standardized or normalized to improve the convergence speed and prediction performance of the model; the categorical variables are encoded, such as using one-hot encoding or label encoding.

[0039] In some embodiments, a preset prediction model based on the XGBoost model is trained using a sludge sample data set (including sample physical and chemical property parameters, sample hydrothermal hydrolysis process parameters, sample anaerobic digestion parameters, and sample net energy output); the hyperparameters of the model are set, such as learning rate, maximum depth, subsample ratio, etc., and are optimized by methods such as cross-validation.

[0040] The performance of the preset prediction model is evaluated using the validation set data, such as calculating indicators such as mean squared error (MSE) and root mean squared error (RMSE), and the model parameters are adjusted according to the validation results until the target prediction performance is achieved to obtain the trained target prediction model.

[0041] Step 103: Determine the target thermohydrolysis process parameters based on the predicted net energy output, where the target thermohydrolysis process parameters are the preset thermohydrolysis process parameters when the predicted net energy output is at its maximum value.

[0042] In an embodiment of the present invention, taking the predicted net energy output of sludge anaerobic digestion as the objective function, calculate the thermohydrolysis process parameters (including thermohydrolysis temperature and thermohydrolysis duration) of the sludge to be regulated when the predicted net energy output is maximized.

[0043] In some embodiments, within the range of preset thermohydrolysis parameters, traverse all possible combinations at a certain step size to find the parameter combination that maximizes the predicted net energy output.

[0044] Through the above steps of the embodiment of the present invention, by obtaining the detected physical and chemical characteristic parameters, preset thermohydrolysis process parameters, and preset anaerobic digestion parameters of the sludge to be regulated, it provides a necessary data basis for subsequent model prediction; through the target prediction model based on the XGBoost model, it can comprehensively consider the physical and chemical characteristic parameters of the sludge, preset thermohydrolysis process parameters, and preset anaerobic digestion parameters, thereby accurately predicting the net energy output under different thermohydrolysis process conditions; based on the predicted net energy output, by comparing the prediction results under different thermohydrolysis process parameters, the target thermohydrolysis process parameters that maximize the net energy output can be determined. Thus, precise regulation operations can be achieved based on the target thermohydrolysis process parameters, and at the same time, a large amount of experimental and time cost consumption is saved.

[0045] According to a method for dynamically regulating the thermohydrolysis temperature of sludge based on the XGBoost model provided by the present invention, before inputting the detected physical and chemical characteristic parameters, preset thermohydrolysis process parameters, and preset anaerobic digestion parameters into the trained target prediction model based on the XGBoost model, the above method further includes: Obtain the sample physical and chemical characteristic parameters, sample thermohydrolysis process parameters, sample anaerobic digestion parameters, and sample net energy output of the sludge sample; Take the sample physical and chemical characteristic parameters, sample thermohydrolysis process parameters, and sample anaerobic digestion parameters as input features, and take the sample net energy output as the output feature; Perform data preprocessing based on the input features and output features to obtain the sludge sample data set; Train the preset prediction model based on the XGBoot model using the sludge sample data set to obtain the trained target prediction model.

[0046] In an embodiment of the present invention, consult relevant literature on sludge thermohydrolysis research, and collect the physical and chemical characteristic parameters, thermohydrolysis process parameters, anaerobic digestion parameters, and corresponding net energy output of the sludge sample.

[0047] Take the physical and chemical property parameters of the sludge, the hydrothermal hydrolysis process parameters, and the anaerobic digestion parameters as input features, take the net energy output as the output feature, and perform data preprocessing on the input features and output features to obtain a sample data set; divide the sample data set to obtain a training set, and train a target prediction model through the training set.

[0048] In some embodiments, remove data points with duplicates, missing values, or outliers, and based on correlation analysis or feature importance evaluation, select the physical and chemical property parameters, hydrothermal hydrolysis process parameters, and anaerobic digestion parameters that have a significant impact on the net energy output as input features; for features with non-linear relationships or different dimensions, data transformation is required, such as logarithmic transformation, standardization, or normalization; combine the input features (sample physical and chemical property parameters, sample hydrothermal hydrolysis process parameters, sample anaerobic digestion parameters) and the output feature (sample net energy output) into a complete sludge sample data set.

[0049] To evaluate the performance of the model, the data set (sludge sample data set) needs to be divided into a training set and a test set. The training set is used to train the model, while the test set is used to verify the prediction accuracy of the model.

[0050] Set the parameters of the preset prediction model based on the XGBoot model, such as the learning rate, the maximum number of iterations, the number and depth of the trees, etc. Use the training set data to train the preset prediction model based on the XGBoot model so that it can learn the complex relationship between the input features and the net energy output. Adjust the hyperparameters of the model through methods such as cross-validation, grid search, or random search to improve its prediction performance.

[0051] Use the test set data to evaluate the prediction accuracy of the trained preset prediction model based on the XGBoot model. Commonly used evaluation metrics include the mean squared error (MSE), the root mean squared error (RMSE), the R² score, etc.

[0052] Through the embodiments of the present invention, the XGBoost model is an efficient gradient boosting decision tree algorithm that constructs a strong learner by integrating multiple weak learners (usually decision trees) and can capture complex non-linear relationships in the data. By training the model with a large amount of sludge sample data, the precise mapping relationship between the input features (physical and chemical property parameters, hydrothermal hydrolysis process parameters, anaerobic digestion parameters) and the output feature (net energy output) can be learned, thereby improving the prediction accuracy and reliability.

[0053] According to a method for dynamically regulating the hydrothermal hydrolysis temperature of sludge based on the XGBoost model provided by the present invention, data preprocessing is performed based on the input features and the output features to obtain a sludge sample data set, including: Remove missing data from the input features and the output feature and perform statistical analysis to obtain statistical data; Normalize the statistical data to obtain a sludge sample dataset. Among them, the normalization process includes: ; Among them, represents the -dimensional data in the sludge sample dataset, represents the -dimensional data of the statistical data, represents the mean of the statistical data, represents the standard deviation of the statistical data.

[0054] In the embodiments of the present invention, missing data in the input features and output features are removed and statistical analysis is performed to obtain statistical data. It should be noted that, preferably, the statistical analysis in the embodiments of the present invention is performed in the form of drawing. For example, the statistical data obtained by performing statistical analysis on the detailed information of multiple groups of data containing multiple input features and 1 output feature.

[0055] In some embodiments, traverse the entire dataset to check whether there are missing values (i.e., null values or NaN values) in each input feature and output feature; for missing values, different strategies can be adopted for processing, such as deleting samples containing missing values (if the missing values are not many and do not affect the data representativeness), filling them using methods such as mean, median, mode, or interpolation method.

[0056] In some embodiments, the statistical analysis includes descriptive statistics, correlation analysis, and data visualization. Among them, descriptive statistics include: calculating basic statistics of each input feature and output feature, such as mean, standard deviation, minimum value, maximum value, quartiles, etc., to understand the distribution and characteristics of the data. Correlation analysis includes: calculating the correlation coefficients between input features (such as Pearson correlation coefficient, Spearman rank correlation coefficient, etc.), and the correlation between input features and output features, to identify which features may have a significant impact on the output. Data visualization includes: using visualization tools such as histograms, box plots, scatter plots, etc. to display the relationship between the distribution of the data and the characteristics, so as to more intuitively understand the data.

[0057] Normalize the statistical data to obtain a sample dataset; among them, the calculation formula for normalizing the statistical data includes: ; Among them, represents the -dimensional data in the sludge sample dataset, represents the -dimensional data of the statistical data, represents the mean of the statistical data, Represents the standard deviation of statistical data.

[0058] Through the embodiments of the present invention, by removing missing data and then performing statistical analysis, the statistical data can be made redundant-free, facilitating better training of the subsequent preset prediction model. Then, through normalization, incomparable data becomes comparable while maintaining the relative relationship between the two compared data, achieving training of more dimensional objectives and greatly improving the generalization ability.

[0059] According to a method for dynamically regulating the sludge hydrothermal hydrolysis temperature based on the XGBoost model provided by the present invention, a preset prediction model based on the XGBoot model is trained based on a sludge sample data set to obtain a trained target prediction model, including: Input the sludge sample data set into the preset prediction model based on the XGBoot model to obtain the sample predicted net energy output output by the preset prediction model; Determine the target loss function of the sample predicted net energy output and the sample true net energy output through the mean square error; Perform supervised training on the preset prediction model based on the target loss function, and obtain the trained target prediction model through iterative optimization.

[0060] In the embodiments of the present invention, the sludge sample data set is input into the preset prediction model based on the XGBoot model, and the target loss function of the model is calculated through the mean square error.

[0061] It should be noted that in the embodiments of the present invention, the XGBoost model is trained using the integrated algorithm structure of gradient boosting decision trees to simulate the non-linear influence of each input feature on the net energy output during the sludge hydrothermal hydrolysis process. The training process includes the following steps: First, construct a gradient boosting model based on trees, where each decision tree fits the residual of the previous model to continuously optimize the overall prediction performance. The number of input features determines the input dimension of the model, and each feature is used to construct the split nodes of the decision tree. The model is initialized with a global prediction value and then gradually improves the prediction accuracy through a series of iterations. In each iteration, the gradient of the objective function (the derivative of the loss function) is calculated based on the input data and the current model prediction value, and these gradient values are used to construct a new regression tree. The new regression tree helps the model correct errors by fitting the current residual. During the tree construction process, parameters such as the number of trees, maximum depth, minimum sum of leaf node weights, learning rate, etc. play a role in controlling the model complexity and improving the generalization ability.

[0062] During the training process, the performance of the model on the validation set is monitored through an early stopping mechanism to avoid overfitting. If the loss function of the validation set does not improve for several rounds, the training is terminated early. After all iterations are completed, the XGBoost model integrates multiple optimized regression trees to form an integrated model with strong non-linear modeling capabilities.

[0063] After training is completed, the machine learning model (the trained target prediction model) can dynamically predict the net energy output according to the new input feature combinations, realizing the intelligent regulation of the sludge hydrothermal hydrolysis temperature. The model is supervised and trained through the target loss function, and the final target prediction model is obtained through iterative optimization.

[0064] The XGBoost model is an efficient, flexible, and scalable gradient boosting algorithm, especially suitable for processing data with highly non-linear and complex relationships. In the embodiments of the present invention, the XGBoost model uses an ensemble tree structure based on the additive model to model multi-dimensional input data, including continuous values and discrete values, thus showing excellent performance in the target prediction model. Compared with traditional machine learning methods, the XGBoost model can capture the complex interaction relationships between input features through the gradient boosting framework and achieve high-precision prediction. The XGBoost model has multiple advantages: First, it has strong robustness to noise and outliers because the model reduces the impact of noise by gradually optimizing the residuals. Second, the XGBoost model can control the model complexity through parameters such as regularization terms (such as L1 and L2 regularization) to avoid overfitting problems. In terms of hyperparameter optimization, the XGBoost model provides rich tuning options, and these hyperparameters can flexibly adapt to different data features, thereby improving the generalization ability of the model.

[0065] Through the embodiments of the present invention, the XGBoost model adopts an efficient splitting algorithm and cache optimization technology during the training process, which can significantly reduce the training time and accelerate the model iteration process. In the prediction of sludge hydrothermal hydrolysis process parameters, the XGBoost model can quickly adapt to the non-linear relationships in the data, and provide a scientific basis for optimizing the process parameters by accurately predicting the net energy output. Combining feature importance analysis and interpretability methods (such as SHAP value analysis), the XGBoost model can also identify the key variables affecting the prediction results and provide decision-making support for actual operations.

[0066] According to a method for dynamically regulating the sludge hydrothermal hydrolysis temperature based on the XGBoost model provided by the present invention, the above method further includes: Adjust the hyperparameters of the target prediction model through the cross-validation error to obtain multiple adjusted hyperparameters; Determine the target hyperparameters based on the multiple adjusted hyperparameters through the genetic algorithm.

[0067] In an embodiment of the present invention, the hyperparameters of the target prediction model for the XGBoost model are adjusted through cross-validation error; the hyperparameters include the type of base learner, the number of trees, the maximum depth of the trees, the subsampling ratio, the learning rate, the regularization method, and the parameters.

[0068] It should be noted that the cross-validation error refers to the prediction error of the model on the validation dataset during the cross-validation process. Cross-validation divides the dataset into several parts, and each time one part is used as the validation dataset, and the remaining parts are used as the training dataset to train the model. Then, the trained model is used to make predictions on the validation dataset, and the prediction error is calculated. This process is repeated multiple times, with different data groupings used each time. Finally, the prediction errors in all repeated experiments are averaged to obtain the cross-validation error. The cross-validation error can be used to evaluate the generalization ability of the model, that is, the performance of the model on unknown data.

[0069] Ten-fold cross-validation means dividing the dataset into ten parts and then performing ten cross-validations. In each cross-validation, one part of the data is selected as the validation set, and the remaining nine parts are used as the training set. This can ensure that each part of the data is used as the validation set once, and the results of each cross-validation are independent. By this method, the sensitivity of the model to data division can be reduced, making the evaluation results of the model more stable and reliable. Especially in the case of a small amount of data, ten-fold cross-validation can effectively improve the generalization ability of the model.

[0070] The adjustment of the hyperparameters is supervised through a genetic algorithm to obtain the optimal hyperparameters (i.e., the target hyperparameters).

[0071] Through the embodiment of the present invention, more stable and reliable model evaluation results are obtained through cross-validation error. Since the results of a single cross-validation may be affected by the data division method, performing multiple cross-validations can reduce this influence, thereby obtaining more accurate model evaluation results, and thus better adjusting the hyperparameters. Combining the optimization algorithm with cross-validation can find the optimal solution more efficiently and improve the efficiency of adjusting the hyperparameters.

[0072] According to a method for dynamically regulating the sludge hydrothermal hydrolysis temperature based on the XGBoost model provided by the present invention, the target loss function includes: ; wherein, represents the target loss function, represents the number of samples in the sludge sample dataset, represents the sample index of the sludge sample dataset, represents the true net energy output of the represents the predicted net energy output of the th sample, represents the regularization coefficient, represents the sum of the absolute values of all weights.

[0073] In the embodiment of the present invention, through supervised training of the target loss function, a target prediction model is obtained, and the result of the target loss function for each iterative training of the target prediction model is calculated; wherein, the calculation formula of the target loss function result includes: ; wherein, represents the target loss function, represents the number of samples in the sludge sample data set, represents the sample index of the sludge sample data set, represents the th sample's true net energy output, represents the th sample's predicted net energy output, represents the regularization coefficient, represents the sum of the absolute values of all weights (i.e., the L1 regularization term), which can control the model complexity and prevent overfitting.

[0074] Refer to Figure 2 , Figure 2 is the experimental and evaluation result data graph of the target prediction model provided by the present invention. In this embodiment, 391 data sets in the literature are trained to obtain the final target model. The evaluation results R2 (Training = 0.958, Test = 0.929), RMSE (Training = 0.261, Test = 0.286) and correlation R (Training = 0.982, Test = 0.906) indicate that the model prediction value can better reflect the real situation and has high reliability and representativeness.

[0075] In the embodiment of the present invention, the convergence of the target loss model is ensured through the loss threshold and the number threshold. If the result of the target loss function is lower than the loss threshold or the number of iterative trainings is higher than the number threshold, the final target prediction model is obtained, and the model result is referred to Figure 2 .

[0076] In some embodiments, after the preset prediction model is supervised and trained based on the target loss function and the trained target prediction model is obtained through iterative optimization, the above method further includes: Dividing the sludge sample data set to obtain a test set, and inputting the test set into the target prediction model to obtain the prediction result output by the target prediction model.

[0077] Calculate the coefficient of determination and root mean square error corresponding to the target prediction model based on the prediction results (i.e., the predicted net energy output of the samples); the calculation formulas for the coefficient of determination and root mean square error include: ; ; where, represents the coefficient of determination, represents the root mean square error, represents the number of samples in the sludge sample dataset, represents the sample index of the sludge sample dataset, represents the true net energy output of the th sample, represents the predicted net energy output of the th sample,

[0078] Here, ranges from 0 to 1. The closer the value is to 1, the better the model fitting effect. The smaller the value of

[0079] , the smaller the prediction error of the model, and the higher the prediction accuracy of the model.

[0080] Calculate the evaluation result of the target prediction model based on the coefficient of determination and root mean square error.

[0081] Next, an example of the dynamic regulation method for sludge thermohydrolysis temperature based on the XGBoost model provided by the present invention in practical applications will be described.

[0082] Refer to Figure 3 , Figure 3 is a structural schematic diagram of the dynamic regulation method for sludge thermohydrolysis temperature based on the XGBoost model provided by the present invention, including: input and output, model prediction and evaluation, and online prediction platform and regulation Among them, the input and output include: inputting sludge physical and chemical characteristics, thermohydrolysis process parameters, and anaerobic digestion process parameters, constructing an original dataset, data preprocessing, and feature partitioning; the model prediction and evaluation include: dividing the data into a training set and a test set, performing cross-validation and optimization algorithms based on the training set to obtain optimal hyperparameters, model training, and performing model evaluation on the trained model through the test set to obtain an optimal model; the online prediction platform and regulation include: an online prediction platform, online rapid prediction, regulating the thermohydrolysis temperature, and accurate verification experiments.

[0083] Reference Figure 3 , in some embodiments of the present invention, a dynamic regulation method for the sludge hydrothermal hydrolysis temperature based on the XGBoost model is provided, including: Consult relevant literature on sludge hydrothermal hydrolysis research, and collect the physical and chemical property parameters of sludge samples, hydrothermal hydrolysis process parameters, anaerobic digestion parameters, and the corresponding net energy output.

[0084] Take the physical and chemical property parameters of sludge, hydrothermal hydrolysis process parameters, and anaerobic digestion parameters as input features, take the net energy output as the output feature, and perform data preprocessing on the input features and output features to obtain a sample data set.

[0085] Divide the sample data set to obtain a training set, and train a target prediction model through the training set.

[0086] Deploy the target prediction model to an online system to obtain an online prediction website.

[0087] Detect the physical and chemical properties of the sludge to be regulated, determine the hydrothermal hydrolysis technical parameters of the sewage treatment plant and the anaerobic digestion conditions of the sludge biogas project, and input the physical and chemical property parameters of the sludge, hydrothermal hydrolysis process parameters, and anaerobic digestion parameters into the online prediction website to obtain the net energy output of the sludge anaerobic digestion output by the online prediction website; Take the net energy output of the sludge anaerobic digestion as the objective function, and calculate the hydrothermal hydrolysis temperature parameter of the sludge to be regulated when the net energy output is maximized.

[0088] This method can first reflect the relationship between the sludge raw material characteristics, hydrothermal hydrolysis-anaerobic digestion process parameters, and net energy output according to the physical and chemical property parameters of sludge, hydrothermal hydrolysis process parameters, anaerobic digestion parameters, and the net energy output of sludge anaerobic digestion collected in the samples; through preprocessing of the physical and chemical property parameters of sludge, hydrothermal hydrolysis process parameters, anaerobic digestion parameters, and the net energy output of sludge anaerobic digestion, a sample data set can be obtained, which can provide a good data basis for the training of the subsequent target prediction model and can accurately train the mapping relationship between the sludge raw material characteristics, hydrothermal hydrolysis-anaerobic digestion process parameters, and net energy output; secondly, the final target prediction model is trained through the training set, and an online prediction website is obtained according to the target prediction model for online system deployment, providing an online fast prediction platform for the net energy output. At the same time, efficient prediction statistics can be carried out through the online prediction website, saving the time cost of subsequent statistics; finally, the hydrothermal hydrolysis temperature of the sludge to be regulated when the net energy output is maximized is output through the online prediction website, realizing precise regulation operations, and also saving a large amount of experimental and time cost consumption.

[0089] In some embodiments of the present invention, the target prediction model is deployed to an online system to obtain an online prediction website.

[0090] Through Hypertext Markup Language, Cascading Style Sheets, Python, and the Flask framework, the prediction function and regulation function of the final target prediction model are formulated; the prediction function is used for the final target prediction model to predict according to the content in the input box of the front-end interface, and the regulation function is used to regulate the hydrothermal hydrolysis temperature when the model prediction target is maximized according to the content in the input box; The front-end interface and the back-end data of the prediction function and regulation function are linked through a visualization platform to obtain an online prediction website.

[0091] Reference Figure 4 , Figure 4 is the interface diagram of the input, output, and regulation of the online prediction website provided by the present invention. Among them, the sludge physical and chemical properties include organic matter content (VS / TS), soluble COD ratio (SCOD / TCOD), carbohydrate, protein, lipid, hydrothermal hydrolysis temperature (Temperature (TH)), hydrothermal hydrolysis time (Time (TH)), fermentation temperature (Temperature (AD)), fermentation concentration (TS (AD)) represents the sum of all solid substances in water, and the organic loading rate (OLR) refers to the ratio of the mass or volume of organic substances in wastewater to the reactor volume or the unit area of the reactor.

[0092] Under the online prediction website constructed in the embodiment of the present invention, users only need to open the online prediction website, input the sludge physical and chemical properties, hydrothermal hydrolysis process parameters, and anaerobic digestion parameters on the website homepage, and then click the "Prediction" button below to quickly predict the net energy output of sludge anaerobic digestion.

[0093] By forming a link through the visualization platform of the online prediction website and the back-end data of the front-end interface and prediction function, the sludge physical and chemical properties, hydrothermal hydrolysis-anaerobic digestion process parameters, and the net energy output of sludge can be made public and a more convenient online prediction method is provided.

[0094] To verify the effectiveness of the embodiment, an embodiment of a specific experiment on a hydrothermal hydrolysis temperature regulation method based on the XGBoost model is provided: Reference Figure 2 In this embodiment, 391 data sets in the literature are trained to obtain the final target model. The evaluation results R2 (Training = 0.958, Test = 0.929), RMSE (Training = 0.261, Test = 0.286), and correlation R (Training = 0.982, Test = 0.906) indicate that the model prediction values can better reflect the real situation and have high reliability and representativeness.

[0095] Reference Figure 5 , Figure 5 is the regulation result data graph of the target prediction model provided by the present invention; using the obtained final target prediction model, taking the annual sludge raw material characteristics of a sewage treatment plant and the operating parameters of the sludge biogas project coupling thermochemical hydrolysis technology of the sewage treatment plant (as shown in Table 1, where the thermochemical hydrolysis temperature is set as unknown) as an example, as input features, and taking the sludge net energy output as the output feature as the objective function, the thermochemical hydrolysis temperature parameter is optimized by genetic algorithm to maximize the net energy output.

[0096] Table 1 Physicochemical characteristics of sludge and operating parameters of thermochemical hydrolysis-anaerobic digestion in a sewage treatment plant

[0097] Reference Figure 5 , since the physicochemical characteristics of sludge vary greatly throughout the year, the optimal temperature for sludge thermochemical hydrolysis also changes accordingly. Among them, it is particularly significant from June to November. During this period, the organic matter content of sludge is low, and the optimal temperature for sludge thermochemical hydrolysis decreases significantly. Thermochemical hydrolysis is not recommended from August to October. Compared with the constant thermochemical hydrolysis temperature of 170 °C throughout the year, the net energy output of sludge can be increased by more than 1000 MJ / t-TS from August to October.

[0098] In the embodiment of the present invention, a thermochemical hydrolysis temperature regulation system based on the XGBoost model is also provided, including a data acquisition module, a data processing and machine learning module, an online prediction website deployment module, an online prediction module, a thermochemical hydrolysis temperature control module, and a feedback regulation module, where: The data acquisition module is used to obtain the physicochemical characteristic parameters of sludge and the process parameters of thermochemical hydrolysis-anaerobic digestion, and transmit the real-time data to the model through sensors or data interfaces; The data processing and machine learning module includes a data preprocessing unit and an XGBoost model modeling unit. The data preprocessing unit performs standardization and normalization processing on the input features and output features, and the XGBoost model modeling unit is used to train the data set to obtain the final target prediction model; The online prediction website deployment module is used to deploy the final target prediction model online to obtain an online prediction website; The online prediction module inputs the input features into the online prediction website to obtain the net energy output of sludge anaerobic digestion output by the online prediction website; The thermochemical hydrolysis temperature control module includes a dynamic temperature controller and a feedback regulation system, and controls the heating equipment according to the thermochemical hydrolysis temperature when the output feature is maximized by the online prediction website, and adjusts the temperature of sludge thermochemical hydrolysis to the optimal value in real time; The feedback control module continuously optimizes the model through a closed-loop feedback mechanism, enabling the system to adaptively adjust according to local sludge characteristics and process parameters.

[0099] The following describes the sludge hydrothermal hydrolysis temperature dynamic regulation device provided by the present invention based on the XGBoost model. The sludge hydrothermal hydrolysis temperature dynamic regulation device described below can be correspondingly referred to the sludge hydrothermal hydrolysis temperature dynamic regulation method described above based on the XGBoost model.

[0100] Reference Figure 6 , Figure 6 is a schematic diagram of the modules of the sludge hydrothermal hydrolysis temperature dynamic regulation system provided by the present invention based on the XGBoost model.

[0101] The acquisition module 601 is used to acquire the detected physical and chemical characteristic parameters of the sludge to be regulated, the preset hydrothermal hydrolysis process parameters, and the preset anaerobic digestion parameters; The prediction module 602 is used to input the detected physical and chemical characteristic parameters, the preset hydrothermal hydrolysis process parameters, and the preset anaerobic digestion parameters into the trained target prediction model based on the XGBoost model to obtain the predicted net energy output output by the target prediction model; The determination module 603 is used to determine the target hydrothermal hydrolysis process parameters based on the predicted net energy output, where the target hydrothermal hydrolysis process parameters are the preset hydrothermal hydrolysis process parameters when the predicted net energy output is the maximum value.

[0102] Specifically, the above-mentioned sludge hydrothermal hydrolysis temperature dynamic regulation system provided by the present invention based on the XGBoost model can implement all the method steps implemented by the above-mentioned sludge hydrothermal hydrolysis temperature dynamic regulation method embodiment based on the XGBoost model, and can achieve the same technical effects. The same parts and beneficial effects as those in the method embodiment are not specifically described herein.

[0103] Figure 7 is a schematic diagram of the physical structure of the electronic device provided by the present invention, as Figure 7As shown in the figure, the electronic device may include: a processor 710, a communications interface 720, a memory 730, and a communication bus 740. Among them, the processor 710, the communications interface 720, and the memory 730 complete communication with each other through the communication bus 740. The processor 710 may call the logical instructions in the memory 730 to execute the method for dynamically regulating the sludge hydrothermal hydrolysis temperature based on the XGBoost model. The method includes: obtaining the detected physical and chemical characteristic parameters of the sludge to be regulated, the preset hydrothermal hydrolysis process parameters, and the preset anaerobic digestion parameters; inputting the detected physical and chemical characteristic parameters, the preset hydrothermal hydrolysis process parameters, and the preset anaerobic digestion parameters into the trained target prediction model based on the XGBoost model to obtain the predicted net energy output output by the target prediction model; determining the target hydrothermal hydrolysis process parameters based on the predicted net energy output, where the target hydrothermal hydrolysis process parameters are the preset hydrothermal hydrolysis process parameters when the predicted net energy output is the maximum value.

[0104] In addition, when the logical instructions in the above-mentioned memory 730 are implemented in the form of software functional units and sold or used as independent products, they may be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, may be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0105] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the dynamic regulation method for sludge hydrothermal hydrolysis temperature based on the XGBoost model provided by the above-mentioned various methods. The method includes: obtaining the detected physical and chemical characteristic parameters of the sludge to be regulated, the preset hydrothermal hydrolysis process parameters, and the preset anaerobic digestion parameters; inputting the detected physical and chemical characteristic parameters, the preset hydrothermal hydrolysis process parameters, and the preset anaerobic digestion parameters into the trained target prediction model based on the XGBoost model to obtain the predicted net energy output output by the target prediction model; determining the target hydrothermal hydrolysis process parameters based on the predicted net energy output, where the target hydrothermal hydrolysis process parameters are the preset hydrothermal hydrolysis process parameters when the predicted net energy output is the maximum value.

[0106] In another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it realizes the dynamic regulation method for sludge hydrothermal hydrolysis temperature based on the XGBoost model provided by the above-mentioned various methods. The method includes: obtaining the detected physical and chemical characteristic parameters of the sludge to be regulated, the preset hydrothermal hydrolysis process parameters, and the preset anaerobic digestion parameters; inputting the detected physical and chemical characteristic parameters, the preset hydrothermal hydrolysis process parameters, and the preset anaerobic digestion parameters into the trained target prediction model based on the XGBoost model to obtain the predicted net energy output output by the target prediction model; determining the target hydrothermal hydrolysis process parameters based on the predicted net energy output, where the target hydrothermal hydrolysis process parameters are the preset hydrothermal hydrolysis process parameters when the predicted net energy output is the maximum value.

[0107] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0108] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0109] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for dynamic temperature control of sludge thermal hydrolysis based on XGBoost model, characterized in that: include: Obtain the physical and chemical property parameters of the sludge to be regulated, preset thermal hydrolysis process parameters and preset anaerobic digestion parameters; Inputting the detected physicochemical characteristic parameters, the preset thermal hydrolysis process parameters, and the preset anaerobic digestion parameters into a trained target prediction model based on an XGBoost model to obtain a predicted net energy output output by the target prediction model; Based on the predicted net energy output, target thermal hydrolysis process parameters are determined, wherein the target thermal hydrolysis process parameters are preset thermal hydrolysis process parameters when the predicted net energy output is a maximum value.

2. The method for dynamic temperature control of sludge thermal hydrolysis based on the XGBoost model according to claim 1, characterized in that: Before inputting the detected physicochemical characteristic parameters, the preset thermal hydrolysis process parameters, and the preset anaerobic digestion parameters into the trained target prediction model based on the XGBoost model, the method further includes: Obtain sample physicochemical property parameters, sample thermal hydrolysis process parameters, sample anaerobic digestion parameters and sample net energy output of sludge samples; Taking the sample physicochemical property parameters, the sample thermal hydrolysis process parameters, and the sample anaerobic digestion parameters as input features, and taking the sample net energy output as output features; Performing data preprocessing based on the input features and the output features to obtain a sludge sample data set; Based on the sludge sample data set, a preset prediction model based on the XGBoot model is trained to obtain a trained target prediction model.

3. The method for dynamic temperature control of sludge thermal hydrolysis based on the XGBoost model according to claim 2, characterized in that: The data preprocessing based on the input features and the output features to obtain a sludge sample data set includes: Performing missing data removal and statistical analysis on the input features and the output features to obtain statistical data; The statistical data is normalized to obtain a sludge sample data set, wherein the normalization process includes: ； in, Represents the first Dimensional data, The statistical data is represented by Dimensional data, represents the mean of the stated statistics, Represents the standard deviation of the stated statistic.

4. The method for dynamic temperature control of sludge thermal hydrolysis based on the XGBoost model according to claim 2, characterized in that: The preset prediction model based on the XGBoot model is trained based on the sludge sample data set to obtain a trained target prediction model, including: Inputting the sludge sample data set into a preset prediction model based on the XGBoot model to obtain the sample predicted net energy output output by the preset prediction model; Determine the target loss function of the sample predicted net energy output and the sample actual net energy output through mean square error; The preset prediction model is supervised and trained based on the target loss function, and the trained target prediction model is obtained through iterative optimization.

5. The method for dynamic temperature control of sludge thermal hydrolysis based on the XGBoost model according to claim 4, characterized in that: The method further comprises: Adjusting the hyperparameters of the target prediction model through cross-validation errors to obtain multiple adjusted hyperparameters; A target hyperparameter is determined based on the multiple adjusted hyperparameters through a genetic algorithm.

6. The method for dynamic temperature control of sludge thermal hydrolysis based on XGBoost model according to claim 4, characterized in that: The objective loss function includes: ； in, represents the objective loss function, represents the number of samples in the sludge sample dataset, represents the sample index of the sludge sample dataset, Indicates The true net energy output of samples is Indicates The sample predicts the net energy output of samples, represents the regularization coefficient, Represents the sum of the absolute values ​​of all weights.

7. A sludge thermal hydrolysis temperature dynamic control system based on XGBoost model, characterized in that: include: An acquisition module is used to obtain the physical and chemical property parameters of the sludge to be regulated, the preset thermal hydrolysis process parameters and the preset anaerobic digestion parameters; A prediction module, used for inputting the detected physicochemical characteristic parameters, the preset thermal hydrolysis process parameters and the preset anaerobic digestion parameters into a trained target prediction model based on an XGBoost model to obtain a predicted net energy output output by the target prediction model; A determination module is used to determine a target thermal hydrolysis process parameter based on the predicted net energy output, wherein the target thermal hydrolysis process parameter is a preset thermal hydrolysis process parameter when the predicted net energy output is a maximum value.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the method for dynamically controlling the sludge thermal hydrolysis temperature based on the XGBoost model as described in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for dynamically controlling the sludge thermal hydrolysis temperature based on the XGBoost model as described in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for dynamically controlling the sludge thermal hydrolysis temperature based on the XGBoost model as described in any one of claims 1 to 6 is implemented.

Images