A sludge drying treatment method and system based on real-time data analysis

Abstract

The present invention discloses a sludge drying treatment method and system based on real-time data analysis, which relates to the technical field of sludge drying treatment, including: pre-processing sludge drying process parameters by using Kalman filtering and linear interpolation methods; analyzing the pre-treated sludge drying process parameters by using edge computing, detecting sludge moisture content, equipment temperature, energy consumption and tail gas composition by using an isolation forest method, and generating working condition status summary information by using a state coding method; combining the working condition status summary information with the pre-treated sludge drying process parameters, and predicting the future drying state trend by using an ARIMA model; adjusting the operating parameters of the drying equipment by using a reinforcement learning method based on the working condition status summary information and the prediction results; optimizing the control strategy by using the reinforcement learning method, significantly reducing energy consumption and reducing tail gas emission concentration, and achieving the goal of improving energy efficiency and environmental protection performance.

Description

Technical Field

[0001] The present invention relates to the technical field of sludge drying treatment, and in particular to a sludge drying treatment method and system based on real-time data analysis. Background Art

[0002] As an indispensable part of the sewage treatment process, sludge drying has received extensive attention in the field of environmental protection in recent years. Traditional sludge drying methods mainly include thermal drying, solar drying and mechanical dehydration. Among them, thermal drying technology is widely used in industrial-scale sludge treatment due to its high efficiency and stability. However, traditional sludge drying systems usually rely on fixed parameters to operate, lack the ability to dynamically analyze real-time data, and are difficult to adapt to complex and changing actual working conditions.

[0003] Although the existing technology has made certain progress in the field of sludge drying, there are still some shortcomings. First, the traditional sludge drying system is relatively simple in data collection and processing, and often relies on only a few sensors to obtain limited parameter information, resulting in incomplete state monitoring of the drying process and inability to accurately reflect the equipment operation status and changes in sludge characteristics. Secondly, in terms of prediction and control, most of the existing drying systems use static models or empirical rules for parameter adjustment, and lack the ability to effectively predict dynamic trends. This not only limits the adaptability of the system, but may also lead to increased energy consumption or unstable drying effects. Summary of the invention

[0004] In view of the above existing problems, the present invention provides a sludge drying treatment method based on real-time data analysis to solve the problems of inaccurate state monitoring and insufficient energy consumption optimization.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] In the first aspect, the present invention provides a sludge drying treatment method based on real-time data analysis, which includes: collecting sludge drying process parameters in real time through sensors; preprocessing the sludge drying process parameters by using Kalman filtering and linear interpolation methods; analyzing the pretreated sludge drying process parameters by edge computing, detecting sludge moisture content, equipment temperature, energy consumption and exhaust gas composition by using the isolation forest method, and generating operating status summary information by using the state coding method; combining the operating status summary information with the pretreated sludge drying process parameters, and using the ARIMA model to predict future drying state trends; adjusting the drying equipment operating parameters by using the reinforcement learning method based on the operating status summary information and the prediction results; converting the drying equipment operating parameters into control signals by digital signal processing, and transmitting the control signals to each execution unit through the data bus to perform sludge drying treatment.

[0007] As a preferred solution of the sludge drying treatment method based on real-time data analysis of the present invention, wherein: the sludge drying process parameters are collected in real time by sensors, and the specific steps are as follows:

[0008] Through the moisture content sensor, temperature sensor, humidity sensor, energy consumption sensor and exhaust gas composition sensor, the moisture content, temperature, humidity, energy consumption and exhaust gas composition parameters in the sludge drying process are collected in real time to obtain the sludge drying process parameters.

[0009] As a preferred solution of the sludge drying treatment method based on real-time data analysis of the present invention, wherein: the Kalman filter and linear interpolation method are used to pre-process the sludge drying process parameters, and the specific steps are as follows:

[0010] The Kalman filter algorithm is used to denoise the real-time collected sludge drying process parameters, reduce random noise interference, and output smooth and stable sludge drying process parameters;

[0011] Linear interpolation method is used to fill the missing data caused by sensor failure and communication interruption;

[0012] The sludge drying process parameters after denoising and filling were standardized by the Z-score normalization method, and the values ​​were mapped to a uniform range by the Min-Max normalization method to generate high-quality and consistent pretreated sludge drying process parameters.

[0013] As a preferred solution of the sludge drying treatment method based on real-time data analysis described in the present invention, the edge computing is used to analyze the parameters of the sludge drying process after pretreatment, the isolation forest method is used to detect the sludge moisture content, equipment temperature, energy consumption and tail gas composition, and the state coding method is used to generate the working condition summary information. The specific steps are as follows:

[0014] Through edge computing, the parameters of the sludge drying process after pretreatment are analyzed in real time. The physical characteristics, time series characteristics and correlation characteristics are extracted using statistical analysis methods. The normal operation mode, abnormal precursor mode and working condition switching mode are identified using lightweight algorithms.

[0015] The extracted features and patterns are used to generate structured data for anomaly detection through data normalization methods;

[0016] Based on anomaly detection structured data, the isolation forest algorithm is used to detect abnormalities such as excessive moisture content, abnormal temperature, and excessive exhaust gas composition during sludge drying;

[0017] Based on equipment design specifications and historical operating data, statistical analysis is performed to set the sludge drying threshold;

[0018] According to the detected abnormal information, the operating status is obtained in combination with the sludge drying threshold, and the operating status is converted into operating condition summary information using the state encoding method.

[0019] As a preferred solution of the sludge drying treatment method based on real-time data analysis of the present invention, the following specific steps are used to combine the working state summary information with the sludge drying process parameters after pretreatment and use the ARIMA model to predict the future drying state trend:

[0020] The historical sludge drying process parameters after pretreatment are combined with the historical operating status summary information to obtain historical comprehensive status data, which are converted into time series data through the time alignment method;

[0021] Through the rolling forecast origin method, the historical comprehensive state data converted into time series data is divided into training set, validation set and test set, and the ARIMA model is trained;

[0022] The comprehensive status data converted into time series data is input into the trained ARIMA model to output the future drying status trend.

[0023] As a preferred solution of the sludge drying treatment method based on real-time data analysis of the present invention, wherein: based on the working condition summary information and the prediction results, the reinforcement learning method is used to adjust the operating parameters of the drying equipment, and the specific steps are as follows:

[0024] The state space is defined by the summary information of the working condition and the future drying state trend, and the action space is defined by the adjustable control variables and value ranges of the drying equipment;

[0025] The energy consumption reduction percentage, drying efficiency improvement value and exhaust emission concentration reduction are converted into numerical rewards, and the reward function is set by combining the linear weighted weight coefficient and balancing the importance of each target setting;

[0026] Define the reinforcement learning environment using parameterized methods by designing a defined state space, action space, and reward function;

[0027] Use the initialization weight method to randomly initialize the parameters in the reinforcement learning agent;

[0028] The defined reinforcement learning environment interacts with the initialized reinforcement learning agent, and the operating parameters of the drying equipment are obtained through the PPO reinforcement learning algorithm.

[0029] As a preferred solution of the sludge drying treatment method based on real-time data analysis of the present invention, wherein: the operation parameters of the drying equipment are converted into control signals through digital signal processing, and the control signals are sent to each execution unit through data bus transmission to perform sludge drying treatment. The specific steps are as follows:

[0030] According to the adjusted operating parameters of the drying equipment, a corresponding PWM signal is generated through digital signal processing;

[0031] The PWM signal is mapped to the data bus through the industrial communication protocol and transmitted to each execution unit;

[0032] The execution unit receives the control signal and performs corresponding operations according to the specified parameters to carry out sludge drying treatment.

[0033] In the second aspect, the present invention provides a sludge drying treatment system based on real-time data analysis, including a data acquisition module, a data preprocessing module, an anomaly detection and state generation module, a trend prediction module, a parameter optimization module, and a parameter execution and feedback module; the data acquisition module is used to collect sludge drying process parameters in real time through sensors; the data preprocessing module is used to preprocess the sludge drying process parameters using Kalman filtering and linear interpolation methods; the anomaly detection and state generation module is used to use edge computing to analyze the pretreated sludge drying process parameters, and use the isolation forest method to detect the sludge moisture content, setting The module is used to combine the operating status summary information with the parameters of the sludge drying process after pretreatment, and use the ARIMA model to predict the future drying status trend; the module is used to adjust the operating parameters of the drying equipment by the reinforcement learning method based on the operating status summary information and the prediction results; the module is used to convert the operating parameters of the drying equipment into control signals through digital signal processing, and transmit the control signals to each execution unit through the data bus to perform sludge drying treatment.

[0034] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the sludge drying treatment method based on real-time data analysis as described in the first aspect of the present invention is implemented.

[0035] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the sludge drying treatment method based on real-time data analysis as described in the first aspect of the present invention is implemented.

[0036] The beneficial effect of the present invention is that by detecting anomalies in real time and generating summary information of the working state, the system reliability is improved, which helps operators quickly locate the source of the problem and avoid production interruptions. The introduction of edge computing reduces the dependence on cloud resources, saves network communication costs, optimizes resource allocation and improves overall efficiency. At the same time, the anomaly detection results provide high-quality data input for subsequent prediction models, enhancing the ability to predict future drying state trends. In addition, the control strategy is optimized by using reinforcement learning methods, which significantly reduces energy consumption and reduces exhaust emission concentrations, achieving the goal of improving energy efficiency and environmental performance. The system can also automatically adjust operating parameters according to real-time working conditions without manual intervention, greatly enhancing the adaptive ability, especially suitable for complex and changeable sludge drying scenarios, and balances multiple objectives such as energy consumption, efficiency and environmental protection based on the reward function design, ensuring that the system maintains optimal performance under different conditions, and comprehensively improving the intelligent level of sludge drying treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0038] Figure 1 This is a flow chart of the sludge drying treatment method based on real-time data analysis in Example 1.

[0039] Figure 2 This is a schematic diagram of the sludge drying treatment system based on real-time data analysis in Example 1. DETAILED DESCRIPTION

[0040] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0041] Example 1

[0042] Reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, provides a sludge drying treatment method based on real-time data analysis, comprising the following steps:

[0043] S1. Collect sludge drying process parameters in real time through sensors.

[0044] S1.1. Through the moisture content sensor, temperature sensor, humidity sensor, energy consumption sensor and exhaust gas composition sensor, the moisture content, temperature, humidity, energy consumption and exhaust gas composition parameters in the sludge drying process are collected in real time to obtain the sludge drying process parameters.

[0045] It should be noted that by deploying high-precision moisture content sensors, the changing trend of moisture content in sludge can be dynamically monitored; temperature sensors are used to accurately measure the temperature distribution inside the drying equipment and on the sludge surface to evaluate the heat transfer efficiency; humidity sensors are responsible for capturing changes in ambient humidity to provide a basis for optimizing drying conditions. At the same time, energy consumption sensors can record the power or heat energy consumption data during the operation of the equipment to help analyze energy utilization efficiency.

[0046] S2. Kalman filtering and linear interpolation methods are used to preprocess the sludge drying process parameters.

[0047] S2.1. The Kalman filter algorithm is used to denoise the real-time collected sludge drying process parameters, reduce random noise interference, and output smooth and stable sludge drying process parameters.

[0048] It should be noted that the Kalman filter algorithm is used to denoise the parameters of the sludge drying process collected in real time, which effectively reduces the interference of random noise and outputs smooth and stable parameter data. The algorithm is based on the dynamic model and observation model of the system state, combined with prior knowledge and current measurement values, to achieve accurate prediction and correction of parameter estimation results. In the sludge drying process, facing the possible noise in the sensor collected data, the Kalman filter finds the best balance between the predicted value and the actual measured value by calculating the optimal weight, eliminating abnormal fluctuations; this recursive processing method not only reduces the computational complexity, but also ensures the stability and accuracy of the output data. In addition, the unique advantage of the Kalman filter is that it can combine historical data and the latest measurement information to provide more accurate data estimation, which is crucial to improving the overall control accuracy of the sludge drying process.

[0049] S2.2. Use linear interpolation method to fill in the missing data caused by sensor failure and communication interruption.

[0050] It should be noted that the linear interpolation method is used to fill the missing data caused by sensor failure and communication interruption, and the linear relationship between the missing data points and their adjacent known data points is constructed, which effectively handles the missing values ​​caused by sensor failure or communication interruption. Specifically, when the data missing at a certain moment is detected, the algorithm will select the two closest valid data points before and after the moment, assume that the change between the two points is a linear relationship, and then calculate the estimated value of the missing point according to the slope formula. This method is particularly suitable for scenarios where the parameter changes are relatively stable during the sludge drying process, such as slowly changing variables such as temperature and humidity. Although linear interpolation cannot completely restore the true value, it can significantly reduce the impact of missing data on the overall analysis and ensure the continuity and integrity of the data series. Combining linear interpolation with Kalman filtering, first using Kalman filtering to remove noise, and then using linear interpolation to fill the possible data gaps, forming a complete set of data preprocessing solutions. This combination is novel and efficient in the field of sludge drying treatment, greatly improving data quality and processing efficiency.

[0051] S2.3. The sludge drying process parameters after denoising and filling are standardized by the Z-score normalization method, and the values ​​are mapped to a uniform range by the Min-Max normalization method to generate high-quality and consistent pre-treated sludge drying process parameters.

[0052] It should be noted that Z-score standardization and Min-Max normalization are two commonly used numerical processing methods to improve data quality and consistency. First, through Z-score standardization, the denoised and filled sludge drying process parameters are converted into a standard normal distribution form. Specifically, the mean of each parameter value is subtracted and divided by the standard deviation so that the data has zero mean and unit variance, thereby eliminating dimensional differences. Next, the Min-Max normalization method is used to linearly map the standardized data to a preset range to ensure that all parameter values ​​are within a uniform interval. The combination of these two methods can not only retain the original distribution characteristics of the data, but also avoid the influence of extreme values, and ultimately generate high-quality and consistent preprocessed data.

[0053] S3. Use edge computing to analyze the parameters of the sludge drying process after pretreatment, use the isolation forest method to detect the sludge moisture content, equipment temperature, energy consumption and exhaust gas composition, and generate operating status summary information through the state coding method.

[0054] S3.1. Use edge computing to analyze the parameters of the sludge drying process after pretreatment in real time, use statistical analysis methods to extract physical property characteristics, time series characteristics and correlation characteristics, and use lightweight algorithms to identify normal operating modes, abnormal precursor modes and operating condition switching modes.

[0055] It should be noted that through edge computing technology, the mean formula is used to extract the mean of physical characteristics including temperature, humidity and other parameters; the time series characteristics capture the trend of parameters over time, such as the slope in the sliding window; the correlation characteristics analyze the relationship between different parameters, such as the correlation between temperature and energy consumption. On this basis, a lightweight algorithm is used to perform pattern recognition on the features. The normal operating mode corresponds to the situation where the parameters are stable and meet expectations; the abnormal precursor mode warns of potential faults by detecting early signals that deviate from the normal range; the operating condition switching mode identifies parameter mutations caused by equipment adjustments or changes in external conditions.

[0056] Extract physical characteristics and use the mean formula, the expression is:

[0057] ;

[0058] in, is the average temperature, is the total number of temperature data points collected in a period of time, It is The temperature value collected.

[0059] Extract time series features and use the sliding window slope formula, the expression is:

[0060] k i = x i+w - x i w ,i ∈ [1,N - w] ;

[0061] in, The temperature is The rate of change within a time period, is the temperature value collected at the end of the sliding window, is the temperature value collected at the beginning of the sliding window, is a sliding window containing 10 data points, is 10.

[0062] Correlation feature extraction uses the Pearson coefficient formula, which is expressed as follows:

[0063] ;

[0064] in, is the correlation between temperature and energy consumption, is the difference between the collected temperature value and the average temperature, yes Temperature value of time is the mean temperature, is the difference between the collected energy consumption value and the average energy consumption, yes The energy consumption value of time, is the mean energy consumption, is the number of temperature values ​​and energy consumption values.

[0065] S3.2. The extracted features and patterns are used to generate structured data for anomaly detection through data normalization methods.

[0066] It should be noted that, first, the extracted features and patterns are quantized to ensure that the data format is uniform and easy to analyze. Then, these feature values ​​are processed using data normalization methods to eliminate the problems of dimensional differences and inconsistent numerical ranges. The standardized data can more accurately reflect the relative relationship between features while reducing the impact of outliers. Finally, the generated structured data contains clearly defined feature vectors for subsequent anomaly detection algorithms.

[0067] Through the data standardization method, anomaly detection structured data is generated, and the expression is:

[0068] ;

[0069] in, is the standardized temperature value, is the original temperature value, is the average temperature, is the standard deviation of the temperature data.

[0070] S3.3. Based on anomaly detection structured data, the isolation forest algorithm is used to detect abnormalities such as excessive moisture content, abnormal temperature, and excessive exhaust gas composition during the sludge drying process.

[0071] It should be noted that the isolation forest algorithm quickly isolates abnormal data points by constructing multiple randomized decision trees, thereby achieving efficient detection of anomalies in the sludge drying process. Based on anomaly detection structured data, the isolation forest uses the distribution characteristics of the data, assuming that normal data points are similar, while abnormal points are more easily isolated. Specifically, the algorithm randomly selects features and divides the data space, recursively generates decision trees until all data points are completely separated. Data points with excessive moisture content, abnormal temperature or excessive exhaust gas composition are sparsely distributed and have a short path length in the tree, so they can be quickly identified as anomalies.

[0072] Using the isolated Senli algorithm, the average path length is calculated, and the expression is:

[0073] ;

[0074] in, It is a data point The expected value of the average path length is is the number of data points contained in a subtree in the current isolation forest, It is a constant related to the distribution of data points and is used to measure the average path length of data points in the tree.

[0075] S3.4. Based on the equipment design specifications and historical operating data, statistical analysis is performed to set the sludge drying threshold.

[0076] It should be noted that, first, the range of key parameters during normal operation of the equipment, such as the theoretical upper and lower limits of temperature, humidity, and energy consumption, are obtained based on the equipment design specifications. Then, combined with historical operation data, statistical analysis methods are used to calculate the mean, standard deviation, and distribution characteristics of each parameter to further verify and adjust the theoretical threshold. For example, by analyzing the 95% confidence interval in historical data, the normal fluctuation range of the parameter is determined; the boundary values ​​with frequent abnormalities are evaluated, and a more reasonable sludge drying threshold is set.

[0077] S3.5. Based on the detected abnormal information and the sludge drying threshold, the operating status is obtained and the operating status is converted into operating condition summary information using the state coding method.

[0078] It should be noted that, first, the detected excessive moisture content, abnormal temperature or excessive exhaust gas composition is compared with the pre-set sludge drying threshold to determine whether each parameter exceeds the normal range, so as to determine whether the current operating state is normal, warning or faulty. Next, the state coding method is used to quantify and simplify the operating state. For example, binary coding or numerical coding is used to represent different states: 001 may represent a normal state, 010 represents a warning state, and 100 represents a fault state. If there are multiple abnormalities, complex working conditions can be reflected through combined coding. Finally, the coding results are integrated into structured working condition summary information.

[0079] S4. Combine the operating status summary information with the sludge drying process parameters after pretreatment, and use the ARIMA model to predict the future drying status trend.

[0080] S4.1. Combine the historical sludge drying process parameters after pretreatment with the historical operating status summary information to obtain historical comprehensive status data, and convert it into time series data through the time alignment method.

[0081] It should be noted that, first, the historical sludge drying process parameters after pretreatment are integrated with the historical operating status summary information to form historical comprehensive status data containing parameter values ​​and corresponding status. Then, the time alignment method is used to ensure that data from different sources are consistent in the time dimension. Specifically, based on a unified timestamp, the parameter data and status information are arranged in chronological order to fill in the missing points caused by different sampling frequencies. Finally, the sorted data is converted into a time series format, and each record contains a timestamp, parameter value, and status code, thereby generating time series data that can be used for analysis and modeling.

[0082] S4.2. The historical comprehensive status data converted into time series data are divided into training set, validation set and test set through the rolling forecast origin method, and the ARIMA model is trained.

[0083] It should be noted that when training the ARIMA model, a sufficiently long period of historical data is first selected as the initial training set. For example, if there are 100 time points in total, the data of the first 70 time points can be used as the training set for preliminary analysis. Based on the selected training set, the key parameters (p, d, q) of the ARIMA model are preliminarily determined by analyzing the autocorrelation function (ACF) and partial autocorrelation function (PACF). These parameters represent the order of the autoregressive term, the difference order, and the order of the moving average term, respectively; these parameters are then used to fit the ARIMA model so that the model can learn the characteristics of trends, seasonality, and random fluctuations in the data. After the initial training is completed, the first time point of the validation set is gradually added to the training set using the rolling forecast origin method. The updated training set is used to recalculate or fine-tune the model parameters, and the model performance is evaluated by calculating the prediction error such as the mean square error. Further adjustments are made based on the evaluation results until satisfactory results are achieved. Finally, after multiple rounds of rolling updates and optimizations, the remaining 15% of the data was used as a test set for the final model performance evaluation. By comparing the predicted values ​​with the actual observed values, the performance of the model in the real application scenario was understood, so as to decide whether to adopt the model for future state trend prediction of the sludge drying process, thereby completing the training of the ARIMA model.

[0084] The analytical autocorrelation function (ACF) and partial autocorrelation function (PACF) are used to measure the correlation of time series data at different lag orders. The expressions are:

[0085] ;

[0086] in, is hysteresis The autocorrelation coefficient of order, is the value of the time series at time point t, is the mean of the time series, is the total length of the time series.

[0087] When using the validation set to evaluate model performance, the mean square error is used as the evaluation indicator, and its expression is:

[0088] ;

[0089] in, is the mean square error, represents the actual observation value t, which represents a specific time point in the time series. represents the predicted value and n represents the number of observations.

[0090] The final performance of the model is evaluated on the test set using the mean absolute error, expressed as:

[0091] ;

[0092] in, It is the average absolute error between the predicted value and the true value, which is used to evaluate the prediction performance of the model. is the total number of data points, It's time. It is The true value at a time point, It is The predicted value at a time point.

[0093] S4.3. The comprehensive state data converted into time series data is input into the trained ARIMA model to output the future drying state trend.

[0094] It should be noted that, first, the time-aligned and integrated comprehensive state data is passed as input to the trained ARIMA model. The ARIMA model performs prediction calculations on the input data based on the time series features learned in the previous training process. Specifically, the model uses autoregressive terms to capture the dependencies of historical data, differential terms to eliminate non-stationarity, and moving average terms to deal with residual effects, thereby inferring the state value at future moments. After calculation, the model outputs the trend prediction results of the future drying state, and the drying state change trends of temperature, humidity and moisture content.

[0095] S5. Based on the operating status summary information and prediction results, the reinforcement learning method is used to adjust the operating parameters of the drying equipment.

[0096] S5.1. Define the state space by the summary information of the operating state and the trend of the future drying state.

[0097] It should be noted that, first, the state space is constructed using the summary information of the working state, such as normal, warning or fault codes, and the future drying state trends, such as the change prediction of temperature and humidity. The state space describes all possible operating states in the sludge drying process, and each state is determined by the current working condition and the predicted trend. Then, the action space is defined by analyzing the adjustable control variables of the drying equipment, such as heating power, wind speed, feed speed, etc., and their allowed value ranges.

[0098] S5.2. The percentage of energy consumption reduction, the improvement of drying efficiency and the reduction of exhaust emission concentration are converted into numerical rewards, and the reward function is set by combining the linear weighted weight coefficient and balancing the importance of each target setting.

[0099] It should be noted that, first of all, the percentage of energy consumption reduction, the improvement value of drying efficiency and the reduction in exhaust emission concentration are converted into specific numerical rewards. For example, define a certain reward value for every 1% reduction in energy consumption, 1% increase in drying efficiency or 1 unit reduction in exhaust emission concentration. Then, according to actual needs and priorities, set the linear weighted weight coefficients of each target, such as a weight of 0.4 for energy consumption, a weight of 0.5 for drying efficiency, and a weight of 0.1 for exhaust emissions. By multiplying the reward value of each target with the corresponding weight and summing them, a comprehensive reward function is constructed. This method can fully consider the relationship between different targets and ultimately achieve a reward function that balances the importance of each target.

[0100] S5.3. Define the reinforcement learning environment using parameterized methods by designing a defined state space, action space, and reward function.

[0101] It should be noted that, first, based on the defined state space and action space, all possible states of the system and the operations that the device can perform are clarified. Then, through the design of the reward function, the effect of each action in a specific state is quantified to provide a learning goal for the agent. Next, the state, action, and reward function are formally represented using a parameterized method, such as using vectors to represent the state and action, and using functions or matrices to describe the reward mechanism. Finally, these parameterized elements are integrated into the reinforcement learning framework to build a complete interactive system, thereby defining the reinforcement learning environment.

[0102] S5.4. Use the initialization weight method to randomly initialize the parameters in the reinforcement learning agent.

[0103] It should be noted that in reinforcement learning, random initialization methods are usually used to ensure that the parameters of the agent's neural network have a good initial state. By selecting a suitable random distribution and assigning initial values ​​to the network weights, the learning symmetry problem caused by the same initial values ​​of the parameters can be avoided. Common methods include Xavier initialization and He initialization, which adjust the weight range according to the number of input and output nodes of the network layer, which helps to stabilize the training process and accelerate convergence. This process effectively prevents the occurrence of gradient vanishing or explosion phenomena, provides a basis for subsequent optimization strategies, and realizes random initialization of parameters in reinforcement learning agents.

[0104] S5.5. The defined reinforcement learning environment and the initialized reinforcement learning agent interact with each other, and the operating parameters of the drying equipment are obtained through the PPO reinforcement learning algorithm.

[0105] It should be noted that, first, the defined reinforcement learning environment is connected to the initialized agent, so that the agent can learn through trial and error in the environment. The agent selects an action based on the current state, and the environment returns the new state and the corresponding reward value after receiving the action. The PPO algorithm is used to maximize the cumulative reward while ensuring the stability of the strategy update. PPO evaluates the pros and cons of actions by calculating the advantage function, and uses the truncated probability ratio to limit the strategy update amplitude to avoid large fluctuations in parameters. After multiple rounds of iterations, the agent continuously optimizes the strategy and eventually converges to a parameter configuration that can make the drying equipment operate with the highest efficiency, the lowest energy consumption and the least emissions, thereby obtaining the optimal drying equipment operating parameters.

[0106] S6. Through digital signal processing, the operating parameters of the drying equipment are converted into control signals, and the control signals are sent to each execution unit through data bus transmission to perform sludge drying treatment.

[0107] S6.1. Generate corresponding PWM signals through digital signal processing according to the adjusted operating parameters of the drying equipment.

[0108] It should be noted that, first, the specific control target values, such as heating power or fan speed, are determined based on the adjusted operating parameters of the drying equipment. Then, these target values ​​are converted into duty cycle parameters of the pulse width modulation signal through digital signal processing technology to ensure that the signal can accurately reflect the required equipment state changes. Using a microcontroller or signal processor, combined with the preset frequency and amplitude requirements, a PWM waveform that meets the requirements of the execution unit is calculated and generated. This process converts the optimized parameters into executable control instructions and finally generates the corresponding PWM signal.

[0109] S6.2, map the PWM signal to the data bus through the industrial communication protocol and transmit it to each execution unit.

[0110] It should be noted that, first, the PWM signal is encapsulated according to the selected industrial communication protocol (such as Modbus or CAN) to ensure that it meets the transmission format requirements of the data bus. Then, the encapsulated signal is mapped to the network through the data bus to achieve efficient transmission from the main controller to the execution unit. In this process, the communication protocol ensures the reliability and real-time performance of the signal and avoids noise interference and data loss. Finally, each execution unit receives and parses the PWM signal transmitted on the bus, thereby completing the signal transmission and transmitting it to each execution unit.

[0111] S6.3, the execution unit receives the control signal, performs corresponding operations according to the specified parameters, and performs sludge drying treatment.

[0112] It should be noted that after receiving the control signal, the execution unit first decodes the signal and converts it into specific control parameters, such as temperature setting value, fan speed or valve opening. Then, based on the analysis results, it drives the internal actuator to perform corresponding operations to ensure that the sludge drying process runs stably according to the preset parameters. By precisely controlling the parameters of each link, it can effectively remove the water in the sludge and complete efficient sludge drying treatment.

[0113] This embodiment also provides a sludge drying treatment system based on real-time data analysis, including: a data acquisition module, a data preprocessing module, an anomaly detection and state generation module, a trend prediction module, a parameter optimization module, and a parameter execution and feedback module;

[0114] Data acquisition module, used to collect sludge drying process parameters in real time through sensors;

[0115] A data preprocessing module is used to preprocess the parameters of the sludge drying process using Kalman filtering and linear interpolation methods;

[0116] The anomaly detection and state generation module is used to use edge computing to analyze the parameters of the sludge drying process after pretreatment, use the isolation forest method to detect the sludge moisture content, equipment temperature, energy consumption and exhaust gas composition, and generate the working condition status summary information through the state coding method;

[0117] The trend prediction module is used to combine the operating status summary information with the sludge drying process parameters after pretreatment, and use the ARIMA model to predict the future drying status trend;

[0118] Parameter optimization module, which is used to adjust the operating parameters of the drying equipment using reinforcement learning method based on the operating status summary information and prediction results;

[0119] The parameter execution and feedback module is used to convert the operating parameters of the drying equipment into control signals through digital signal processing, and transmit the control signals to each execution unit through the data bus to perform sludge drying treatment.

[0120] This embodiment also provides a computer device, which is suitable for the sludge drying treatment method based on real-time data analysis, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the sludge drying treatment method based on real-time data analysis proposed in the above embodiment.

[0121] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0122] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the sludge drying treatment method based on real-time data analysis as proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0123] In summary, the present invention achieves the effect of improving system reliability by: detecting anomalies in real time and generating summary information of operating status, helping operators to quickly locate the source of the problem and avoid production interruptions. The introduction of edge computing reduces dependence on cloud resources, saves network communication costs, optimizes resource allocation and improves overall efficiency. At the same time, the anomaly detection results provide high-quality data input for subsequent prediction models, enhancing the ability to predict future drying state trends. In addition, the control strategy is optimized using reinforcement learning methods, which significantly reduces energy consumption and reduces exhaust emission concentrations, achieving the goal of improving energy efficiency and environmental performance. The system can also automatically adjust operating parameters according to real-time operating conditions without manual intervention, greatly enhancing its adaptive capabilities, and is particularly suitable for complex and changeable sludge drying scenarios. It balances multiple objectives such as energy consumption, efficiency and environmental protection based on the reward function design, ensuring that the system maintains optimal performance under different conditions, and comprehensively improving the level of intelligence in sludge drying treatment.

[0124] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A sludge drying treatment method based on real-time data analysis, characterized in that: include: Collect sludge drying process parameters in real time through sensors; Kalman filtering and linear interpolation methods are used to preprocess the parameters of the sludge drying process. The Kalman filter and linear interpolation method are used to pre-process the sludge drying process parameters. The specific steps are as follows: The Kalman filter algorithm is used to denoise the real-time collected sludge drying process parameters, reduce random noise interference, and output smooth and stable sludge drying process parameters; Linear interpolation method is used to fill the missing data caused by sensor failure and communication interruption; The sludge drying process parameters after denoising and filling were standardized by the Z-score normalization method, and the values ​​were mapped to a uniform range by the Min-Max normalization method to generate high-quality and consistent pre-treated sludge drying process parameters; Edge computing is used to analyze the parameters of the sludge drying process after pretreatment, and the isolation forest method is used to detect the sludge moisture content, equipment temperature, energy consumption and exhaust gas composition. The state coding method is used to generate the working condition summary information; The method uses edge computing to analyze the parameters of the sludge drying process after pretreatment, uses the isolation forest method to detect the sludge moisture content, equipment temperature, energy consumption and tail gas composition, and generates the working condition summary information through the state coding method. The specific steps are as follows: Through edge computing, the parameters of the sludge drying process after pretreatment are analyzed in real time. The physical characteristics, time series characteristics and correlation characteristics are extracted using statistical analysis methods. The normal operation mode, abnormal precursor mode and working condition switching mode are identified using lightweight algorithms. The extracted features and patterns are used to generate structured data for anomaly detection through data normalization methods; Based on anomaly detection structured data, the isolation forest algorithm is used to detect abnormalities such as excessive moisture content, abnormal temperature, and excessive exhaust gas composition during sludge drying; Based on equipment design specifications and historical operating data, statistical analysis is performed to set the sludge drying threshold; According to the detected abnormal information, the operating status is obtained in combination with the sludge drying threshold, and the operating status is converted into operating condition summary information using the state coding method; Combine the operating status summary information with the sludge drying process parameters after pretreatment, and use the ARIMA model to predict the future drying status trend; The process of combining the working condition summary information with the sludge drying process parameters after pretreatment and using the ARIMA model to predict the future drying state trend is as follows: The historical sludge drying process parameters after pretreatment are combined with the historical operating status summary information to obtain historical comprehensive status data, which are converted into time series data through the time alignment method; Through the rolling forecast origin method, the historical comprehensive state data converted into time series data is divided into training set, validation set and test set, and the ARIMA model is trained; The comprehensive state data converted into time series data is input into the trained ARIMA model to output the future drying state trend; Based on the summary information of the working status and the prediction results, the reinforcement learning method is used to adjust the operating parameters of the drying equipment; Based on the working condition summary information and prediction results, the reinforcement learning method is used to adjust the operating parameters of the drying equipment. The specific steps are as follows: Define the state space through the summary information of the working state and the trend of the future drying state; The energy consumption reduction percentage, drying efficiency improvement value and exhaust emission concentration reduction are converted into numerical rewards, and the reward function is set by combining the linear weighted weight coefficient and balancing the importance of each target setting; Define the reinforcement learning environment using parameterized methods by designing a defined state space, action space, and reward function; Use the initialization weight method to randomly initialize the parameters in the reinforcement learning agent; The defined reinforcement learning environment interacts with the initialized reinforcement learning agent, and the operating parameters of the drying equipment are obtained through the PPO reinforcement learning algorithm; Through digital signal processing, the operating parameters of the drying equipment are converted into control signals, and the control signals are sent to each execution unit through data bus transmission to carry out sludge drying treatment.

2. The sludge drying treatment method based on real-time data analysis according to claim 1, characterized in that: The sludge drying process parameters are collected in real time by sensors, and the specific steps are as follows: Through the moisture content sensor, temperature sensor, humidity sensor, energy consumption sensor and exhaust gas composition sensor, the moisture content, temperature, humidity, energy consumption and exhaust gas composition parameters in the sludge drying process are collected in real time to obtain the sludge drying process parameters.

3. The sludge drying treatment method based on real-time data analysis according to claim 2, characterized in that: The operation parameters of the drying equipment are converted into control signals through digital signal processing, and the control signals are sent to each execution unit through data bus transmission to perform sludge drying treatment. The specific steps are as follows: According to the adjusted operating parameters of the drying equipment, a corresponding PWM signal is generated through digital signal processing; The PWM signal is mapped to the data bus through the industrial communication protocol and transmitted to each execution unit; The execution unit receives the control signal and performs corresponding operations according to the specified parameters to carry out sludge drying treatment.

4. A sludge drying treatment system based on real-time data analysis, based on the sludge drying treatment method based on real-time data analysis according to any one of claims 1 to 3, characterized in that: Including data acquisition module, data preprocessing module, anomaly detection and state generation module, trend prediction module, parameter optimization module, parameter execution and feedback module; Data acquisition module, used to collect sludge drying process parameters in real time through sensors; A data preprocessing module is used to preprocess the parameters of the sludge drying process using Kalman filtering and linear interpolation methods; The anomaly detection and state generation module is used to use edge computing to analyze the parameters of the sludge drying process after pretreatment, use the isolation forest method to detect the sludge moisture content, equipment temperature, energy consumption and exhaust gas composition, and generate the working condition status summary information through the state coding method; The trend prediction module is used to combine the operating status summary information with the sludge drying process parameters after pretreatment, and use the ARIMA model to predict the future drying status trend; Parameter optimization module, which is used to adjust the operating parameters of the drying equipment using reinforcement learning method based on the operating status summary information and prediction results; The parameter execution and feedback module is used to convert the operating parameters of the drying equipment into control signals through digital signal processing, and transmit the control signals to each execution unit through the data bus to perform sludge drying treatment.

5. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the sludge drying treatment method based on real-time data analysis described in any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the sludge drying treatment method based on real-time data analysis described in any one of claims 1 to 3 are implemented.

Images