Sludge drying treatment optimization control method and system

Abstract

The invention relates to the technical field of sludge treatment and intelligent control, in particular to a sludge drying treatment optimization control method and system. Through multi-source data acquisition and standardization processing, real-time equipment operation core parameters are compared with a dynamically updated preset safety threshold, abnormal working conditions are judged, a preset safety strategy is triggered immediately, and complex working conditions are classified by extracting feature vectors and calculating the Euclidean distance between the feature vectors and a typical working condition clustering center. Automatically loading a hierarchical control strategy matched with the working condition, outputting a single-equipment optimal operation parameter adapted to the current specific working condition by adjusting the priority and parameter constraint of a control target, and constructing a multi-target global optimization function fusing three core targets, namely total energy consumption, total treatment capacity and average moisture content standard-reaching rate; and a lightweight particle swarm optimization algorithm is adopted to cooperatively optimize feed quantity distribution and operation parameter fine tuning of each device, and finally a global optimal coordination instruction is output.

Description

Technical Field

[0001] This invention relates to the field of sludge treatment and intelligent control technology, specifically to an optimal control method and system for sludge drying treatment. Background Technology

[0002] Sludge drying is a crucial step in sludge reduction, stabilization, and resource recovery. It involves supplying heat energy to wet sludge to evaporate moisture, thereby significantly reducing the sludge volume. Sludge drying is an industrial process involving complex heat transfer, mass transfer, momentum transfer, and physicochemical changes. Its core process parameters are interdependent, exhibiting complex dynamic characteristics and placing extremely high demands on automated control.

[0003] The sludge drying process is characterized by significant complexity and dynamic fluctuations. Different drying conditions require different control strategies. Fixed parameter control cannot adapt to the dynamic changes in sludge characteristics, which can easily lead to excessive drying effect or energy waste. Single-objective feedback regulation is difficult to balance energy consumption and processing capacity. In the case of multiple devices in parallel, load optimization cannot be achieved. If differentiated control strategies cannot be matched, it will be difficult to achieve stable, efficient and low-consumption operation of sludge drying. Therefore, we propose an optimal control method and system for sludge drying treatment. Summary of the Invention

[0004] The purpose of this invention is to provide an optimal control method and system for sludge drying treatment, which can solve any of the technical problems raised in the background art.

[0005] To achieve the above objectives, the present invention provides an optimal control method for sludge drying treatment, comprising the following steps: S10. Real-time collection of core equipment operation parameters, core sludge characteristic parameters, and energy consumption and effect feedback parameters during the sludge drying process, and preprocessing of the collected data to form a standardized dataset. S20. Compare the core operating parameters of the device in the standardized dataset with the preset threshold. If the parameters exceed the preset threshold range, the device is determined to be in an abnormal operating condition, and the safety policy is triggered immediately. S30. For operating conditions that do not exceed the preset threshold range, extract the core parameters of sludge characteristics and energy consumption and effect feedback parameters, construct feature vectors, calculate the Euclidean distance between the feature vectors and the corresponding cluster centers of normal stable operating conditions and sludge characteristic fluctuation operating conditions, determine the type of operating condition, load the hierarchical control strategy for the corresponding operating condition, and perform inference calculation on the standardized dataset through a lightweight hybrid optimization model to output the optimal operating parameters for a single device. S40. Combine the optimal operating parameters of a single device with the drying effect data and the device load status to construct a multi-objective global optimization function with the core objectives of minimizing total energy consumption, maximizing total processing capacity, and maximizing the average moisture content compliance rate. Use a lightweight particle swarm optimization algorithm to iteratively solve for the minimum value of this function and output the optimal feed rate allocation instructions for each device and the operation parameter coordination and fine-tuning instructions.

[0006] As a further improvement to this technical solution, S20 determines the theoretical safety boundary of the core operating parameters of the equipment based on the equipment manual and sludge drying process specifications, and uses it as the initial threshold value. The confidence interval of the parameter is calculated using the quartile method. The initial threshold value is compared with the confidence interval. If the initial threshold value exceeds the confidence interval, the boundary of the confidence interval is used as the final threshold value, and the preset threshold value is periodically updated.

[0007] As a further improvement to this technical solution, the specific steps of S30 in determining the operating condition type, adopting a hierarchical control strategy, and outputting the optimal operating parameters for a single device are as follows: S31. Extract core parameters from the preprocessed standardized dataset. The core parameters of sludge characteristics include initial moisture content and viscosity. The energy consumption and effect feedback parameters include unit energy consumption and final moisture content. The parameters are mapped to the [0,1] interval using a normalization method to form a feature vector. S32. Calculate the Euclidean distance between the real-time feature vector and the two cluster centers. Use the minimum distance judgment rule. If the distance between the real-time feature vector and the cluster center of the normal stable operating condition is smaller, it is judged as the normal stable operating condition; if the distance between the real-time feature vector and the cluster center of the sludge characteristic fluctuation condition is smaller, it is judged as the sludge characteristic fluctuation condition. S33. Automatically load the hierarchical control strategy corresponding to the working condition, and adjust the weight of the objective function by setting the priority of differentiated control targets and parameter constraint ranges; S34. Based on the historical mapping relationship between parameters and effects, predict the final moisture content and unit energy consumption under different drying temperature and wind speed combinations according to the real-time feature vector, output candidate parameter combinations, and score and filter the candidate combinations according to the weight strategy of the current working condition, and select the combination with the lowest comprehensive score as the optimal operating parameters.

[0008] As a further improvement to this technical solution, when switching operating conditions, S33 pre-sets the maximum adjustment range and adjustment time interval for a single operation, defines the parameter operating range and equipment safety boundary for different operating conditions, calculates the total adjustment amount of the current parameter and the target parameter of the new operating condition, and then determines the gradient adjustment step number by rounding up according to the maximum adjustment range for a single operation, and adjusts the parameter step by step according to the gradient.

[0009] As a further improvement to this technical solution, S33 constructs a linear gradient function and adopts a weighted linear gradient transition method, which is started synchronously with the gradient adjustment of the running parameters. Each time a parameter fine-tuning is completed, the weight coefficient is updated synchronously.

[0010] As a further improvement to this technical solution, in step S34, when the lowest comprehensive scores of multiple sets of parameters are the same, the maximum feasible processing capacity corresponding to each combination is extracted, and sorted from high to low according to the maximum feasible processing capacity. If the processing capacity of a certain combination is significantly higher than that of other combinations, the combination is directly selected as the optimal operating parameter. If the processing capacity of multiple combinations is similar, the energy consumption per unit processing capacity is further compared, and the combination with lower energy consumption per unit processing capacity is selected first.

[0011] As a further improvement to this technical solution, the specific steps of S40 in constructing a multi-objective global optimization function and outputting the optimal feed rate allocation instruction and operating parameter coordination fine-tuning instruction for each device are as follows: S41. Combine the optimal operating parameters of a single device with the drying effect data and the device load status, normalize the data, map it to a unified range, set global hard constraints, and construct a multi-objective global optimization function. S42. The feed rate allocation ratio and the fine-tuning amount of the operating parameters of each device are used as optimization variables. Each particle corresponds to a set of global optimization schemes. The optimization function value corresponding to each particle is calculated as the fitness. Then, the particle position is iteratively updated according to the rules, gradually moving towards the individual optimal and global optimal schemes. S43. The converged global optimal solution is parsed into two types of instructions: the optimal feed rate allocation instruction and the running parameter coordination and fine-tuning instruction.

[0012] As a further improvement to this technical solution, S41 introduces three major objectives: balancing weight coefficients. By establishing a mapping rule between production scenarios and weight coefficients, a weight allocation scheme for typical scenarios is preset.

[0013] As a further improvement to this technical solution, in each iteration of S42, each particle updates its own motion speed and position based on its own historical best solution and the population's global best solution, moving closer to a better solution. After the position is updated, a constraint check needs to be performed again. If it exceeds the constraint range, it is pulled back to the boundary.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. The sludge drying treatment optimization control method and system integrates and preprocesses the core parameters of equipment operation, core parameters of sludge characteristics, and energy consumption effect feedback parameters through multi-source data acquisition and standardized processing. This effectively eliminates the dimensional differences and random interference of the original data, forming a high-quality standardized dataset that can be directly used for subsequent intelligent analysis. 2. By comparing the real-time core operating parameters of the equipment with dynamically updated preset safety thresholds, abnormal operating conditions of the equipment can be identified in real time. Once an abnormality is determined, the preset safety strategy is triggered immediately, and emergency control commands are output to ensure that the system quickly returns to the safe zone. This achieves millisecond-level response and automated handling of major operational risks, upgrading the traditional passive safety interlock to active intelligent safety protection, fundamentally ensuring the safety of equipment and personnel, and providing a solid foundation for subsequent optimized operation. 3. By extracting feature vectors and calculating the Euclidean distance to the cluster centers of typical working conditions, the system can quickly and accurately classify the current complex working conditions, automatically load hierarchical control strategies that match the working conditions, adjust the priority of control targets and parameter constraints, and use a lightweight hybrid optimization model for reasoning to output the optimal operating parameters of a single device that are suitable for the current specific working conditions. Based on the characteristics of the feed sludge and changes in the operating status, the system can automatically switch to the optimal control mode, which significantly improves the system's adaptability to major disturbances and its local operating efficiency under various working conditions. 4. Based on the optimization of single equipment, a multi-objective global optimization function was constructed that integrates three core objectives: total energy consumption, total processing capacity, and average moisture content compliance rate. A lightweight particle swarm optimization algorithm was adopted to collaboratively optimize the feed rate allocation and fine-tune the operating parameters of each equipment under the premise of meeting safety and operating condition constraints. Finally, the globally optimal coordination command was output. Through dynamic load allocation and parameter coordination, the overall energy-saving potential and processing capacity limit of the system were explored while ensuring product quality, and the comprehensive optimization of safety, quality, energy efficiency, and production capacity was achieved.

[0015] In addition to the objectives, features, and advantages described above, the present invention has other objectives, features, and advantages. The invention will now be described in further detail with reference to the figures. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall method flow of the present invention; Figure 2 This is a schematic diagram of the overall system flow of the present invention.

[0017] The meanings of the labels in the diagram are as follows: 100. Data acquisition module; 200. Anomaly detection module; 300. Working condition classification module; 400. Parameter optimization module. Detailed Implementation

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Currently, the sludge drying process is characterized by significant complexity and dynamic fluctuations. Different drying conditions require different control strategies. Fixed parameter control cannot adapt to the dynamic changes in sludge characteristics, which can easily lead to excessive drying effect or energy waste. Single-objective feedback regulation is difficult to balance energy consumption and processing capacity. In the case of multiple devices in parallel, load optimization cannot be achieved. If differentiated control strategies cannot be matched, it will be difficult to achieve stable, efficient, and low-consumption operation of sludge drying.

[0020] Therefore, this embodiment of the invention proposes to provide a unified and standardized dataset for the entire process by collecting sludge drying process parameters in real time. Based on this dataset, the highest priority safety diagnosis is performed to define insurmountable hard boundaries for subsequent steps. Within the safety boundaries, specific operating conditions are identified and corresponding optimization strategies are loaded. This provides an initial solution and weight guidance adapted to the current operating state for global optimization. Based on the constraints and guidance of the preceding steps, global dynamic optimization is performed through multi-device collaboration, and the optimal feed rate allocation instructions and operating parameter coordination and fine-tuning instructions for each device are output.

[0021] Specifically as follows: Please see Figure 1 As shown in the figure, this embodiment of the invention provides an optimization control method for sludge drying treatment, including the following steps: First, the S10 collects three core parameters: equipment operation, sludge characteristics, and energy efficiency and quality. It then performs preprocessing such as noise filtering, anomaly removal, and standardization to form a high-quality standardized dataset. This solves the problem that the original data is multi-source and heterogeneous, contains noise and has different dimensions, and cannot be directly used for accurate analysis and decision-making. It builds a unified, reliable, and high-quality operational condition perception data foundation, providing accurate and consistent data input for all subsequent advanced analyses. Secondly, the S20 compares the core operating parameters of the equipment with the dynamically preset safety thresholds in real time. Once the threshold is exceeded, the preset emergency control strategy is immediately triggered and a safety boundary signal is generated. This solves the problems of the traditional control system's delayed response to safety risks and the independent and difficult coordination between safety interlocking and process optimization. It achieves millisecond-level active protection against major safety risks and establishes the highest priority and insurmountable hard boundary for safe operation of the entire system, ensuring the safety of equipment and personnel. Then, based on the distance calculation between the feature vector and the cluster center, the S30 quickly identifies the current operating condition type, such as normal stable operating condition or sludge characteristic fluctuation operating condition, automatically loads the matching hierarchical control strategy, and uses a lightweight hybrid optimization model to calculate the optimal single equipment operating parameters for the current operating condition. This solves the problem that the single fixed control strategy has poor adaptability due to the variable sludge characteristics and process nonlinearity, and cannot maintain efficient and stable operation under various operating conditions. It automatically adjusts the control target and strategy according to the changes in operating conditions, which significantly improves the adaptability and robustness to major disturbances, and achieves local optimal operation under different operating conditions. Finally, under the constraints of safety boundaries and operating conditions, S40 constructs a multi-objective global optimization function that integrates total energy consumption, total processing capacity, and quality compliance rate. It then employs a distributed, collaborative, lightweight particle swarm optimization algorithm to solve for the globally optimal feeding allocation and parameter coordination instructions for each device. This solves the problem of multiple devices operating independently in parallel, making it difficult to maximize the overall system energy efficiency and processing capacity under multiple objectives and constraints. It achieves a leap from local optimization of a single device to collaborative optimization of a multi-device system. Under the premise of ensuring safety and quality, it dynamically balances multiple objectives such as energy efficiency and production capacity, taps into the overall energy-saving potential and processing capacity limit of the system, and achieves optimal overall global benefits.

[0022] Among them, S10 collects the core operating parameters of the equipment, the core parameters of sludge characteristics, and the energy consumption and effect feedback parameters in real time during the sludge drying process, and preprocesses the collected data to form a standardized dataset. The core operating parameters of the equipment include drying temperature and hot air velocity. The drying temperature is collected by a thermocouple or platinum resistance temperature sensor installed in the material contact area of ​​the dryer cavity, and the hot air velocity is collected by a thermal wind speed sensor at the hot air inlet of the dryer. Both acquire data at a high frequency of 100ms / time to ensure that the equipment operating status is reflected in real time. The core parameters of sludge characteristics include initial moisture content and viscosity. Initial moisture content is detected non-contactly using an online microwave moisture content meter on the feed conveyor belt (before the dryer inlet), with a sampling frequency of 100ms / time. The equipment is calibrated in advance to ensure data consistency. Viscosity is collected every 5 minutes by an online rotary viscometer installed in the mixing section of the feed pipeline. Energy consumption and performance feedback parameters include unit energy consumption and final moisture content. Unit energy consumption needs to be calculated in real time using the formula (total power consumption + total heat consumption) / total processing volume, based on energy consumption data collected from the electricity meter and heat meter, as well as processing volume data collected from the belt scale. The calculation frequency is 1 second / time. Final moisture content is collected by an online near-infrared moisture content detector at the dryer outlet. Both are collected at a frequency of 100 ms / time, and equipment calibration is completed in advance to ensure data consistency. The 3σ principle is adopted to remove data that exceeds the range of ±3 standard deviations of the parameter mean. For short-term missing data (≤5 collection cycles), linear interpolation is used to ensure data continuity. For long-term missing data (>5 collection cycles), the mean of historical data is used to fill in the missing data. The missing period is marked for manual verification of sensor faults. For high-frequency fluctuating parameters such as temperature, wind speed, and moisture content, a moving average filter with a window size of 5-10 collection cycles is used to eliminate errors caused by sensor noise and airflow disturbances and retain the true trend of data change. The above data is standardized to unify the units. The Min-Max normalization method is used on the preprocessed effective data to map all parameters to the [0,1] interval. The normalization formula is: standardized value = (original data - minimum parameter value) / (maximum parameter value - minimum parameter value). The maximum and minimum values ​​of the parameters are preferably taken from historical extreme values. If there is no historical data, the parameter boundary values ​​in the equipment manual or process specification are referred to. Organize the data into a standardized dataset and conduct quality monitoring. Organize the standardized data into a structured dataset according to parameter categories and collection timestamp formats. High-frequency online acquisition and low-latency transmission ensured the real-time nature of the data. At the same time, preprocessing techniques such as outlier removal, missing value completion, and smoothing denoising eliminated data distortion caused by sensor noise and transmission interference, providing a reliable input basis for subsequent anomaly detection, operating condition identification, and parameter optimization. In addition, standardization processing uniformly mapped parameters of different dimensions and orders of magnitude to the [0,1] interval, effectively eliminating dimensional differences and improving the computational efficiency and convergence accuracy of the lightweight hybrid optimization model and particle swarm optimization algorithm, avoiding optimization result deviations caused by data scale issues.

[0023] In addition, S20 compares the core operating parameters of the device in the standardized dataset with preset thresholds. If the parameters exceed the preset threshold range, it is determined that the device is in an abnormal operating condition and the safety policy is triggered immediately. The standardized values ​​of the core operating parameters of the current device are synchronously extracted from the standardized dataset according to the collection timestamp. Each parameter value is compared with the preset threshold range. If the parameter value is within the range, it is considered normal and no action is triggered. If the parameter value exceeds the range, it is considered a single over-limit and the continuous monitoring mechanism is immediately started. Subsequently, an abnormal operating condition classification judgment is performed: if a parameter exceeds the limit for two consecutive collection cycles, it is judged as abnormal parameter fluctuation and a level one warning is triggered; if a parameter exceeds the limit for five consecutive collection cycles, or multiple parameters exceed the limit simultaneously in a single instance, it is judged as an abnormal operating condition of the equipment and a level two fault is triggered. Standardized data eliminates the interference of dimensions and noise, which can effectively avoid misjudgment and missed judgment. Subsequently, differentiated safety strategies are triggered for different levels of anomalies: In the case of a Level 1 warning, a parameter fine-tuning instruction is automatically issued to redirect the out-of-limit parameter back to the center of the threshold range, and a warning message is pushed to the operation and maintenance terminal; In the case of a Level 2 fault, safety actions are executed according to priority, first cutting off the control loop corresponding to the out-of-limit parameter, then starting the equipment load reduction operation mode, and if the parameter continues to exceed the limit, an emergency shutdown protection is triggered and the control interface is locked, fault location information is pushed, and the operation of the working condition division and parameter optimization module is suspended before the anomaly is cleared to avoid abnormal data interfering with the subsequent control process; Finally, when the standardized values ​​of the equipment operating parameters return to the threshold range for 10 consecutive collection cycles, the system automatically determines that the abnormal operating condition has been resolved. Following the process of parameter reset and module restart, the equipment parameters are reverted to the historical optimal baseline value, the operating condition division and parameter optimization module is restarted, and intelligent control is restored.

[0024] In order to better determine the preset threshold, S20 determines the theoretical safety boundary of the core operating parameters of the equipment based on the equipment manual and sludge drying process specifications, which is used as the initial value of the threshold. The confidence interval of the parameter is calculated using the quartile method. The initial value of the threshold is compared with the confidence interval. If the initial value of the threshold exceeds the confidence interval, the boundary of the confidence interval is used as the final threshold, and the preset threshold is updated periodically. Consult the drying equipment manual to obtain the rated operating range and absolute safety boundary of the core operating parameters of the equipment. At the same time, combine the sludge drying process specification to clarify the process constraint range of the parameters. Combine the two to determine the initial value range of the physical threshold of the parameters and convert it into the [0,1] interval threshold initial value that matches the standardized dataset. Obtain preprocessed valid sample data. Parameter samples under normal operating conditions with no equipment failure and achieving the required drying effect are selected from the historical operation database. After removing outliers, an ordered sample set X={x1,x2,…,xn} is obtained. The sample data is then rearranged in ascending order. The quartiles and interquartile ranges (IQRs) are calculated. Based on the sorted sample set, the lower quartile Q1 and upper quartile Q3 are determined: Q1 is the value at the 25th percentile, representing 25% of the sample data being less than this value; Q3 is the value at the 75th percentile, representing 75% of the sample data being less than this value. The IQR is then calculated using the formula IQR=Q3−Q1. This indicator reflects the dispersion of the middle 50% of the sample data. Based on the quartiles and IQRs, the confidence interval is calculated using the formula [Q1−1.5IQR,Q3+1.5IQR]. Compare the initial threshold value with the confidence interval and calibrate the final threshold. Compare the initial threshold value interval with the confidence interval. If the initial threshold value completely covers the confidence interval without significant deviation, the initial threshold value is retained; if the initial threshold value exceeds the confidence interval, the boundary of the confidence interval is used as the final threshold. Based on standardized unified data scales, the system can quickly and accurately identify the out-of-limit status of key parameters such as drying temperature and hot air velocity, avoiding abnormal misjudgments and omissions caused by differences or distortions in data dimensions. At the same time, by triggering safety strategies in real time, measures such as emergency parameter reversion, reduced equipment load operation, or even shutdown protection can be taken immediately to prevent problems such as serious over-limit drying effect and equipment component damage caused by continuous parameter exceedances. This ensures the safety and stability of the sludge drying process, reduces the cost of fault handling, and builds a solid safety line for subsequent recovery of operating conditions and normal optimization control.

[0025] In addition, for operating conditions that do not exceed the preset threshold range, S30 extracts the core parameters of sludge characteristics and energy consumption and effect feedback parameters, constructs feature vectors, calculates the Euclidean distance between the feature vectors and the cluster centers corresponding to normal stable operating conditions and sludge characteristic fluctuation operating conditions, determines the type of operating condition, loads the hierarchical control strategy corresponding to the operating condition, and performs inference calculations on the standardized dataset through a lightweight hybrid optimization model to output the optimal operating parameters for a single device. By relying on the precise clustering of feature vectors, the system can quickly identify and distinguish operating conditions, avoiding control strategy mismatch caused by sludge characteristic fluctuations. At the same time, the combination of hierarchical control strategy and lightweight optimization model not only ensures the pertinence and smoothness of parameter adjustment, but also improves the efficiency and real-time performance of optimization calculation. It can quickly output the optimal operating parameters of a single device that are adapted to the current operating conditions, laying a reliable local parameter foundation for subsequent global multi-objective optimization. Ultimately, it ensures that the sludge drying process can achieve efficient, stable and energy-saving operation under different operating conditions.

[0026] The specific steps by which S30 determines the operating condition type, adopts a hierarchical control strategy, and outputs the optimal operating parameters for a single device are as follows: S31. Extract core parameters from the preprocessed standardized dataset. The core parameters of sludge characteristics include initial moisture content and viscosity. The energy consumption and effect feedback parameters include unit energy consumption and final moisture content. The parameters are mapped to the [0,1] interval using a normalization method to form a feature vector. Four core parameters are precisely extracted according to parameter categories: initial moisture content and viscosity for sludge characteristics, and unit energy consumption and final moisture content for energy consumption and effect feedback. During extraction, the data must be aligned according to the collection timestamp to ensure that the four parameters at the same time correspond one-to-one, forming a set of parameter data tuples to be processed. For each core parameter to be normalized, determine its historical extreme values ​​or process boundary values, i.e., the minimum and maximum values ​​of the parameter. Historical extreme values ​​are preferentially extracted from pre-treated historical operating data; if no historical data is available, the parameter boundary values ​​set in the sludge drying process specifications and equipment manual are referenced. Next, for the original values ​​of each pre-treated parameter set, substitute them into the Min-Max normalization formula for calculation. The formula is: ; in, These are the normalized parameters. These are the original parameters. For the minimum value of the parameter, This represents the maximum value of the parameter.

[0027] This formula eliminates the differences in dimensions and orders of magnitude between different parameters. Finally, the normalized result is checked for interval. If the value exceeds the interval [0,1], it is corrected to the interval boundary value (0 for less than 0, 1 for greater than 1) to ensure that the normalized value of all core parameters is within the interval [0,1], providing parameter data of a uniform scale for subsequent feature vector construction. Following a fixed parameter order, the four normalized parameter values ​​corresponding to a single timestamp are arranged sequentially to form a feature vector of dimension 4. The vector format is uniformly as follows: ; in, Normalized value of initial moisture content Normalized viscosity value Normalized value of unit energy consumption This is the normalized value of the final moisture content.

[0028] S32 calculates the Euclidean distance between the real-time feature vector and the two cluster centers. The minimum distance judgment rule is adopted. If the distance between the real-time feature vector and the cluster center of the normal and stable operating condition is smaller, it is judged as the normal and stable operating condition; if the distance between the real-time feature vector and the cluster center of the sludge characteristic fluctuation condition is smaller, it is judged as the sludge characteristic fluctuation condition. Two types of operating conditions are pre-defined cluster center vectors. Effective feature vector sample sets for normal and stable operating conditions and sludge characteristic fluctuation conditions are selected from historical operating data. The normal and stable operating condition samples are selected from data with stable equipment operating parameters, small sludge characteristic fluctuations, and satisfactory drying effect. The sludge characteristic fluctuation condition samples are selected from data with sudden changes in initial sludge moisture content or viscosity and small fluctuations in drying effect but without triggering anomalies. The mean vectors of the two sample sets are calculated to obtain the cluster centers for normal and stable operating conditions and the cluster centers for sludge characteristic fluctuation conditions. Both cluster centers are 4-dimensional vectors with the same dimension as the real-time feature vectors. Extract the 4-dimensional real-time feature vector at the current moment, and substitute it into the Euclidean distance calculation formula to calculate the distance between the real-time feature vector and the two cluster centers. The formula is as follows: ; in, The distance between the feature vector and the cluster center. For operating condition type, The first real-time feature vector Dimensional parameter values, For the corresponding cluster center Dimensional parameter values.

[0029] The distance between the real-time feature vector and the cluster center under normal and stable operating conditions is calculated. Distance from the cluster center of sludge characteristic fluctuation conditions The minimum distance criterion is used to determine the working condition type, and the comparison is performed. and Size: If Less than If so, the current condition is determined to be a normal and stable operating condition; if Greater than If the values ​​are equal, the current condition is determined to be a fluctuating sludge characteristic condition; if the two values ​​are equal, the condition is determined to be a normal and stable condition by default, and it is marked as a critical state and continuous monitoring is started to avoid misjudgment of the condition due to instantaneous data fluctuations.

[0030] S33. Automatically load the hierarchical control strategy corresponding to the working condition, and adjust the weight of the objective function by setting the priority of differentiated control targets and parameter constraint ranges; Differentiated stratified control strategies were formulated for normal and stable operating conditions and sludge characteristic fluctuation conditions. For normal and stable operating conditions, the priority of control objectives was "lowest total energy consumption > maximum treatment capacity > highest moisture content compliance rate"; the parameter constraint range was taken as the middle 80% of the equipment's rated operating range to ensure stable operation of the equipment under low load; the objective function weight benchmark values ​​were set as energy consumption weight 0.5, treatment capacity weight 0.3, and moisture content weight 0.2. Sludge characteristic fluctuation operating condition strategy: The control target priority is adjusted to "highest moisture content compliance rate > maximum treatment capacity > lowest total energy consumption"; the parameter constraint range is widened to the full range of rated equipment operation, and parameter adjustment space is reserved to adapt to sludge characteristic fluctuations; the objective function weight benchmark values ​​are set to moisture content weight 0.5, treatment capacity weight 0.3, and energy consumption weight 0.2. Based on the pre-set weight benchmark values ​​of the strategy, and combined with the parameter fluctuations under real-time operating conditions, the weight coefficients of each indicator in the objective function are dynamically adjusted. Based on the priority of the control objectives, upper and lower limits and basic adjustment step sizes are set for the weights of each objective to avoid excessive weight adjustment that could lead to imbalance in the objective function. Highest priority target: weight cap set at 0.7 to 0.8, lower limit set at 80% of the baseline value, and basic adjustment step size at 0.1; Medium priority target: The upper limit of the weight is set to 120% of the baseline value, the lower limit is set to 50% of the baseline value, and the basic adjustment step size is 0.05; Lowest priority target: The upper limit of the weight is set to 100% of the baseline value, the lower limit is set to 30% of the baseline value, and the basic adjustment step size is 0.05; For each control objective's corresponding core parameter, the real-time deviation is calculated, serving as the core basis for weight adjustment. The formula is as follows: ; in, For real-time deviation, This is the real-time value of the parameter. The target value for the parameter.

[0031] The adjustment is determined based on the principle of prioritizing high-priority targets and passively adapting to medium and low-priority targets, combined with the degree of deviation: the adjustment amount is 0 when the highest priority target deviates slightly, the basic step size is used for medium deviation, and twice the basic step size is used for severe deviation, not exceeding the upper limit; medium and low priority targets are adjusted proportionally to the adjustment amount of the highest priority target, with their own baseline weight being reduced in proportion to the sum of the baseline weights of medium and low priority targets. The adjustment amount is verified to ensure that the weight of a single target is within the preset boundary and the sum of all target weights is 1. If there is a deviation, the boundary value is taken or the weight is finely adjusted according to the weight ratio. When the parameter deviation returns to a slight level or meets the standard, the weight is gradually pulled back to the benchmark value by the reverse step size to avoid system oscillation caused by sudden weight changes.

[0032] Considering that parameter mutations may cause equipment shocks and process fluctuations during operating condition switching, in order to achieve a smooth transition of operating conditions and ensure the safety, stability and controllability of the sludge drying process, S33 pre-sets the maximum single adjustment range and adjustment time interval when switching operating conditions. At the same time, it defines the parameter operating range and equipment safety boundary for different operating conditions, calculates the total adjustment amount of the current parameters and the target parameters of the new operating conditions, and then determines the gradient adjustment steps by rounding up according to the maximum single adjustment range, and adjusts the parameters step by step according to the gradient. Based on the mechanical characteristics of the equipment and the lag in process response, a maximum single adjustment threshold is set for core operating parameters such as drying temperature and hot air velocity to avoid sudden parameter changes impacting the equipment or affecting the drying effect. Combined with the process stabilization cycle after parameter adjustment, an adjustment time interval is set to ensure that the system has sufficient time to reach a steady state after each adjustment. At the same time, based on the hierarchical control strategy for new operating conditions, the target operating range and absolute safety boundary of the parameters are defined. The target range must be completely within the safety boundary, and the safety boundary must not be exceeded during the adjustment process. After the operating condition switch is triggered, the real-time values ​​of the current operating parameters of the equipment are extracted, the target parameter value of the new operating condition is determined (the median value of the target operating range is taken), and the total adjustment amount of the two is calculated by the absolute value formula; if the target parameter exceeds the safety boundary, it is corrected to the safety boundary value and the total adjustment amount is recalculated. Using the maximum adjustment range in a single step as a constraint, the ratio of the total adjustment amount to the maximum adjustment range in a single step is rounded up to obtain the number of gradient adjustment steps. This ensures that the remaining adjustment amount that is insufficient for a single step can be completed in the last step, guaranteeing that the parameters can eventually reach the target value. The total adjustment amount is allocated to each step. The first N-1 steps are executed according to the maximum adjustment range in a single step, and the remaining adjustment amount is executed in the last step. Adjustment instructions are issued strictly according to the preset time interval. After each adjustment, the steady-state value of the parameter is checked to see if it is within the safety boundary. If it meets the requirements, wait for the interval to end and execute the next step. After all steps are completed, check whether the parameter falls into the target operating range of the new operating condition. If it does, the switch is judged to be completed and the new operating condition optimization control is started. If it does not, an additional adjustment is made to ensure that the target is met. By using a closed-loop logic of preset constraints, quantified adjustment amounts, and gradient step execution, a smooth transition of equipment operating parameters is achieved during the switching of operating conditions. This avoids the load impact and component wear caused by sudden parameter changes, and also prevents large fluctuations in the drying effect. Precise step-by-step adjustments ensure that parameters meet the standards, laying a reliable foundation for subsequent optimization calculations.

[0033] Considering that when the weights jump directly from the old value to the new value during the switching of operating conditions, the objective function will change abruptly, causing the running parameters output by the optimization model to fluctuate drastically, which contradicts the principle of smoothness of parameter gradient adjustment. S33 constructs a linear gradient function and adopts a transition method of linear gradient of weights, which starts synchronously with the gradient adjustment of running parameters. Each time the parameter fine-tuning is completed, the weight coefficients are updated synchronously. Using the number of gradient adjustment steps during the switching of operating conditions as the independent variable, and taking the initial weight value of the old operating condition and the target weight value of the new operating condition as the boundary, a linear gradient function is constructed, the formula of which is: ; in, for The weight value at that time, These are the initial values ​​for the weights. The target value for the weight. Adjust the number of steps for the current parameters. Adjust the number of steps for the gradient.

[0034] Clearly define the weight calculation rules for each step of parameter adjustment, and ensure that the sum of the weights of each target is always 1 during the weight transition process; When the parameter gradient adjustment command triggers the switching of operating conditions, the linear gradual change process of weights is started synchronously. It is set that after each parameter fine-tuning is completed and the steady state is verified, a weight update is triggered to ensure that the timing of the two is completely synchronized. The weight coefficients are updated step by step according to the parameter adjustment steps: After each parameter adjustment step is completed, the current step number is substituted into the linear gradual function to calculate the weight coefficient of the corresponding step and update the target function weight set; this process is repeated until the last parameter adjustment step is completed, and the weights are synchronously updated to the target value of the new working condition, so as to achieve a smooth transition of the weights of the old and new working conditions. By relying on a linear gradient function that matches the number of parameter adjustment steps, a gradual transition of weights between old and new operating conditions is achieved, avoiding system oscillations caused by sudden weight changes. At the same time, by synchronizing the timing of parameter fine-tuning and weight updates, each step of parameter adjustment corresponds to the weight coefficients of the transition phase, ensuring that the parameters output by the optimization model always fit the control objectives during the operating condition switching process.

[0035] Based on the historical mapping relationship between parameters and effects, S34 predicts the final moisture content and unit energy consumption under different drying temperature and wind speed combinations according to real-time feature vectors, outputs candidate parameter combinations, and scores and filters the candidate combinations according to the weight strategy of the current working condition, selecting the combination with the lowest comprehensive score as the optimal operating parameters. The input consists of 4-dimensional parameters (initial sludge moisture content, viscosity, drying temperature, and hot air velocity) standardized to [0,1], and the output consists of 2-dimensional drying effect parameters (final moisture content and unit energy consumption). A multi-output architecture is used to simultaneously fit the two objectives. Samples covering normal and stable conditions as well as fluctuating sludge characteristics are extracted from the historical operation library to ensure that the input and output of each set of data are complete and consistent in time. The training set and the test set are randomly divided in a 7:3 ratio to ensure that both datasets cover the full range of operating conditions. Lightweight frameworks such as LightGBM or XGBoost are used to initialize the model. Initial hyperparameters are set to 100 to 200 decision trees, 5 to 8 tree depths, and a learning rate of 0.05 to 0.1. Based on the training set, the first decision tree is constructed to fit the initial correlation between the input and output and the residual is calculated. Each subsequent tree is optimized iteratively by using gradient descent to fit the residual of the previous model. A multi-objective joint training mode is adopted to simultaneously reduce the prediction error of the final water content and unit energy consumption. The hyperparameters are adjusted by grid search, and the optimal parameter combination is determined with the goal of minimizing the root mean square error (RMSE) of the test set. The optimized model is then encapsulated into a standardized prediction interface and embedded into the system parameter prediction module. By inputting real-time sludge characteristic parameters and candidate temperature and wind speed combinations, the predicted values ​​of final moisture content and unit energy consumption can be directly output, providing support for the scoring and selection of subsequent candidate parameter combinations. Considering that when multiple sets of parameters have the same lowest comprehensive score, it is impossible to select the optimal parameter combination, S34 extracts the maximum feasible processing capacity corresponding to each combination when multiple sets of parameters have the same lowest comprehensive score, sorts them from high to low according to the maximum feasible processing capacity. If the processing capacity of a certain combination is significantly higher than that of other combinations, the combination is directly selected as the optimal operating parameter. If the processing capacity of multiple combinations is similar, the energy consumption per unit processing capacity is further compared, and the combination with lower energy consumption per unit processing capacity is selected first. Extract all candidate parameter combinations with the lowest overall scores, query and determine the maximum feasible processing capacity for each combination (the maximum feasible processing capacity must be determined based on the current operating condition parameter constraint range, the rated processing capacity of the equipment, and the characteristics of the sludge, and must not exceed the safe operating boundary of the equipment). These candidate combinations are sorted from highest to lowest maximum feasible processing capacity, and the differences in processing capacity among the groups are compared: if the maximum feasible processing capacity of a certain combination is higher than that of other combinations, then the combination is directly selected as the optimal operating parameter to prioritize processing efficiency. If the maximum feasible processing capacity of multiple combinations after sorting is similar (the difference is within a preset threshold), then the energy consumption per unit processing capacity of each combination is further calculated (energy consumption per unit processing capacity = energy consumption per unit / maximum feasible processing capacity). The energy consumption per unit processing capacity of each group is compared, and the combination with the lower energy consumption per unit processing capacity is selected as the optimal operating parameter to balance processing efficiency and energy saving goals. While ensuring the original weighted strategy control objectives, further optimization of both treatment efficiency and energy consumption economy is achieved: firstly, by prioritizing the comparison of the maximum feasible treatment volume, the sludge treatment capacity of the equipment is maximized, which meets the core need for efficiency improvement in industrial production; secondly, when the treatment volume is similar, by comparing the energy consumption per unit treatment volume, a more energy-efficient parameter combination is selected, taking into account the control of operating costs.

[0036] In addition, S40 combines the optimal operating parameters of a single device with the drying effect data and the device load status to construct a multi-objective global optimization function with the core objectives of minimizing total energy consumption, maximizing total processing capacity, and maximizing the average moisture content compliance rate. The minimum value of this function is solved iteratively using a lightweight particle swarm optimization algorithm, and the optimal feed rate allocation instructions and operating parameter coordination fine-tuning instructions for each device are output. By integrating the operating parameters of individual devices, drying effects, and load status, a global optimization objective function is constructed that takes into account total energy consumption, total processing capacity, and moisture content compliance rate. Relying on the efficient iterative solution capability of the lightweight particle swarm optimization algorithm, the optimal feed allocation and parameter coordination fine-tuning instructions for each device can be quickly output, realizing the collaborative linkage and optimal resource allocation among multiple devices, and ultimately achieving a globally optimal operating state with lower overall system energy consumption, higher processing efficiency, and more stable drying effect.

[0037] The specific steps of S40 in constructing a multi-objective global optimization function and outputting the optimal feed rate allocation instructions and operating parameter coordination fine-tuning instructions for each device are as follows: S41. Combine the optimal operating parameters of a single device with the drying effect data and the device load status, normalize the data, map it to a unified range, set global hard constraints, and construct a multi-objective global optimization function. The decision variables are the feed rate and fine-tuning of operating parameters of each device, and the target variables are the total energy consumption, total processing capacity and average moisture content compliance rate. Since the optimization directions of the three target variables are different (energy consumption needs to be minimized, and processing capacity and compliance rate need to be maximized) and the dimensions are different, they need to be mapped to the [0,1] interval and uniformly adjusted to the optimization direction of smaller values ​​being better. The formula for the comprehensive optimization function is: ; in, For the comprehensive target value, The weight of the total energy consumption normalization value, This is the normalized value of total energy consumption. The weights for the normalized values ​​of the total processing volume. This is the normalized value of the total processing volume. The weight of the normalized value of the average moisture content compliance rate. This is the normalized value of the average moisture content compliance rate.

[0038] The inviolable boundaries are defined from three dimensions: safe equipment operation, achievement of process objectives, and overall system balance. Specifically, these include: Equipment operation constraints: The drying temperature of each piece of equipment must be within the upper and lower limits allowed by the process. The equipment load rate must not be higher than the full load threshold or lower than the low load idling threshold. The equipment health must be maintained above the minimum standard for stable operation to avoid equipment overload or inefficient operation. Process performance constraints: The final moisture content of the dried product of a single device must be lower than the preset threshold, and the average moisture content compliance rate of the system must not be lower than the minimum standard required by the process, so as to ensure that the drying effect meets the downstream use or disposal specifications. System global constraints: The total energy consumption of the system must not exceed the preset energy consumption limit, and the total processing capacity must reach the minimum target value set in the production plan to ensure the economic efficiency of system operation and the achievement of production capacity targets; Based on the core production goals of the enterprise, the weight ratio is adjusted to match the needs of cost pressure, urgent tasks, and product standards; based on process constraints, the weight of secondary goals is ensured not to be too low, so as to avoid exceeding the thresholds of processing volume, energy consumption, and compliance rate, which would lead to system instability; based on historical data, the comprehensive target value is calculated by substituting historical data of similar working conditions, the weight combination that best fits the expectations is selected, and the weight scheme is iteratively optimized.

[0039] In order to better balance the three objectives of minimizing total energy consumption, maximizing total processing capacity, and achieving the highest average moisture content compliance rate, S41 introduces weight coefficients to balance the three objectives. By establishing a mapping rule between production scenarios and weight coefficients, a weight allocation scheme for typical scenarios is preset. By matching the priority requirements of different production scenarios with differentiated weight proportions, the optimization results are ensured to align with actual production orientation. A mapping rule between production scenarios and weight coefficients is established: using core production needs and process constraints as dual-dimensional judgment criteria, production scenarios are first divided according to production tasks (e.g., routine production, emergency response, energy saving and consumption reduction), cost pressures (high energy consumption costs, routine costs), and product standards (strict handling standards, general handling standards). Then, the priority ranking of the three major objectives under each scenario is clarified, mapping the corresponding weight allocation logic, for example: In a typical production scenario (balancing efficiency and cost): the priority is that the throughput meets the target = energy consumption is controllable > the effect is satisfactory, and the weight allocation is (total energy consumption 0.35, total throughput 0.35, moisture content compliance rate 0.3). Emergency response scenario (efficiency priority): The priority is: maximum processing volume > satisfactory effect > controllable energy consumption, with the weight allocation as follows (total energy consumption 0.2, total processing volume 0.5, moisture content compliance rate 0.3). Energy saving and consumption reduction scenario (energy saving priority): the priority is the lowest energy consumption > the processing volume meets the standard > the effect is qualified, and the weight allocation is (total energy consumption 0.5, total processing volume 0.3, moisture content compliance rate 0.2). Strict treatment scenario (efficiency first): The priority is highest moisture content compliance rate > treatment volume compliance > controllable energy consumption, with the weight allocation as follows (total energy consumption 0.2, total treatment volume 0.3, moisture content compliance rate 0.5). By precisely balancing the three objectives of total energy consumption, total processing capacity, and moisture content compliance rate through differentiated weight allocation, the system imbalance caused by single-objective optimization is avoided. At the same time, relying on the mapping rules between production scenarios and weight coefficients, the weight scheme can be standardized and quickly invoked, eliminating the need to repeatedly debug parameters in different scenarios and greatly improving the efficiency and adaptability of multi-objective global optimization.

[0040] S42 uses the feed rate allocation ratio and the fine-tuning amount of the operating parameters of each device as optimization variables. Each particle corresponds to a set of global optimization schemes. The optimization function value corresponding to each particle is calculated as the fitness. Then, the particle position is iteratively updated according to the rules, gradually moving towards the individual optimal and global optimal schemes. The feed rate allocation ratio of each device and the fine-tuning of drying temperature and wind speed are taken as core optimization variables. Each particle corresponds to a global optimization scheme containing all optimization variables. Then, the particle population is initialized by randomly generating a number of particles to form an initial population, ensuring that the parameter values ​​of each particle are within the range of global hard constraints. At the same time, the initial position of each particle is set as the individual optimal position. The fitness value of all particles is calculated (i.e., the comprehensive objective value corresponding to the multi-objective global optimization function; the smaller the value, the higher the fitness). The position of the particle with the best fitness is selected as the initial global optimal position. In order to better iterate and update the particle swarm, S42 updates the speed and position of each particle in each iteration based on its own historical best solution and the global best solution of the swarm, moving closer to the better solution. After the position is updated, the constraint is checked again. If it exceeds the constraint range, it is pulled back to the boundary. According to the particle swarm optimization (PSO) velocity update rule, and combining inertia weight, learning factor, and random number, the particle's velocity in the next round is calculated using the following formula: ; in, For the speed of movement in the next round, Inertial weights control the degree to which particles inherit historical velocities. The particle's current velocity. This serves as a learning factor, guiding particles to learn in the direction of their individual optimality. Use random numbers in the interval [0,1] to enhance population diversity. This represents the optimal position for an individual particle. This is the particle's current position. This serves as a learning factor, guiding particles to learn towards the global optimum. Use random numbers in the interval [0,1] to enhance population diversity. This is the globally optimal position for the population.

[0041] The particle position is updated based on the updated velocity. The new position corresponds to a new set of equipment feed rate allocation ratios and operating parameter fine-tuning schemes, achieving an approximation of a better solution. Constraint verification is immediately performed after the position is updated, checking item by item whether the feed rate allocation ratio, operating parameter fine-tuning amount, etc., exceed the global hard constraints. If there are out-of-bounds parameters, the boundary truncation method is used to pull them back into the constraint boundary. The finally confirmed compliant position is used as the final position of this iteration for subsequent fitness calculations and the next iteration. Relying on the dual guidance mechanism of individual optimality and global optimality, particles can quickly move towards a better solution, avoid getting trapped in local optima, and improve the efficiency and accuracy of multi-objective global optimization. The constraint verification and boundary pull-back operation performed after each iteration can effectively prevent optimization variables from breaking through the hard constraints of equipment operation, process effect and system global, and ensure that the output feed allocation ratio and parameter fine-tuning amount meet the safety and compliance requirements of actual production.

[0042] S43 parses the converged global optimal solution into two types of instructions: the optimal feed rate allocation instruction and the operating parameter coordination and fine-tuning instruction. Extract the core parameters corresponding to the global optimal solution, separate the optimal feed rate allocation ratio of each device, and the optimal fine-tuning amount of temperature and wind speed. Multiply the feed rate allocation ratio of each device with the target value of the total feed rate of the system to calculate the actual optimal feed rate of a single device. Organize the data by device number, clarify the feed rate of each device, the total feed rate, the execution priority and the effective time, and form a standardized feed allocation instruction. By combining the baseline operating parameters of each device, the temperature and wind speed fine-tuning amounts are converted into specific target operating parameters. At the same time, multi-device coordination rules are embedded (such as maintaining temperature difference and wind speed matching between devices). The target parameters, fine-tuning range, coordination requirements and effective time of each device are specified according to the device number, forming standardized parameter fine-tuning instructions. Finally, compliance checks are performed on the two types of instructions to confirm that the feed rate matches the overall target, the operating parameters meet the constraints, and the coordination logic is reasonable. After the checks are passed, the instructions are converted into a format that the control system can recognize and then issued for execution.

[0043] like Figure 2 As shown, a sludge drying treatment optimization control system is used for employing the sludge drying treatment optimization control method according to any one of claims 1 to 9, comprising: The data acquisition module 100 collects the core operating parameters of the equipment, the core parameters of sludge characteristics, and the energy consumption and effect feedback parameters in real time during the sludge drying process, and preprocesses the collected data to form a standardized dataset. The anomaly detection module 200 compares the core operating parameters of the device in the standardized dataset with preset thresholds. If the parameters exceed the preset threshold range, the device is identified as being in an abnormal operating condition, and a safety policy is immediately triggered. The working condition classification module 300 extracts the core parameters of sludge characteristics and energy consumption and effect feedback parameters for working conditions that do not exceed the preset threshold range, constructs feature vectors, calculates the Euclidean distance between the feature vectors and the cluster centers corresponding to normal and stable working conditions and sludge characteristic fluctuation working conditions, determines the working condition type, loads the hierarchical control strategy for the corresponding working condition, and performs inference calculations on the standardized dataset through a lightweight hybrid optimization model to output the optimal operating parameters for a single device. The parameter optimization module 400 combines the optimal operating parameters of a single device with the drying effect data and the device load status to construct a multi-objective global optimization function with the core objectives of minimizing total energy consumption, maximizing total processing capacity, and maximizing the average moisture content compliance rate. The module uses a lightweight particle swarm optimization algorithm to iteratively solve for the minimum value of the function and outputs the optimal feed rate allocation instructions for each device and the operation parameter coordination and fine-tuning instructions.

[0044] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for optimizing and controlling sludge drying treatment, characterized in that, Includes the following steps: S10. Real-time collection of core equipment operation parameters, core sludge characteristic parameters, and energy consumption and effect feedback parameters during the sludge drying process, and preprocessing of the collected data to form a standardized dataset. S20. Compare the core operating parameters of the device in the standardized dataset with the preset threshold. If the parameters exceed the preset threshold range, the device is determined to be in an abnormal operating condition, and the safety policy is triggered immediately. S30. For operating conditions that do not exceed the preset threshold range, extract the core parameters of sludge characteristics and energy consumption and effect feedback parameters, construct feature vectors, calculate the Euclidean distance between the feature vectors and the corresponding cluster centers of normal stable operating conditions and sludge characteristic fluctuation operating conditions, determine the type of operating condition, load the hierarchical control strategy for the corresponding operating condition, and perform inference calculation on the standardized dataset through a lightweight hybrid optimization model to output the optimal operating parameters for a single device. S40. Combine the optimal operating parameters of a single device with the drying effect data and the device load status to construct a multi-objective global optimization function with the core objectives of minimizing total energy consumption, maximizing total processing capacity, and maximizing the average moisture content compliance rate. Use a lightweight particle swarm optimization algorithm to iteratively solve for the minimum value of this function and output the optimal feed rate allocation instructions for each device and the operation parameter coordination and fine-tuning instructions.

2. The sludge drying treatment optimization control method according to claim 1, characterized in that: S20 determines the theoretical safety boundary of the core operating parameters of the equipment based on the equipment manual and sludge drying process specifications, and uses it as the initial threshold value. The confidence interval of the parameter is calculated using the quartile method. The initial threshold value is compared with the confidence interval. If the initial threshold value exceeds the confidence interval, the boundary of the confidence interval is used as the final threshold value, and the preset threshold value is updated periodically.

3. The sludge drying treatment optimization control method according to claim 2, characterized in that: The specific steps of S30 in determining the operating condition type, adopting a hierarchical control strategy, and outputting the optimal operating parameters for a single device are as follows: S31. Extract core parameters from the preprocessed standardized dataset. The core parameters of sludge characteristics include initial moisture content and viscosity. The energy consumption and effect feedback parameters include unit energy consumption and final moisture content. The parameters are mapped to the [0,1] interval using a normalization method to form a feature vector. S32. Calculate the Euclidean distance between the real-time feature vector and the two cluster centers. Use the minimum distance judgment rule. If the distance between the real-time feature vector and the cluster center of the normal stable operating condition is smaller, it is judged as the normal stable operating condition; if the distance between the real-time feature vector and the cluster center of the sludge characteristic fluctuation condition is smaller, it is judged as the sludge characteristic fluctuation condition. S33. Automatically load the hierarchical control strategy corresponding to the working condition, and adjust the weight of the objective function by setting the priority of differentiated control targets and parameter constraint ranges; S34. Based on the historical mapping relationship between parameters and effects, predict the final moisture content and unit energy consumption under different drying temperature and wind speed combinations according to the real-time feature vector, output candidate parameter combinations, and score and filter the candidate combinations according to the weight strategy of the current working condition, and select the combination with the lowest comprehensive score as the optimal operating parameters.

4. The sludge drying treatment optimization control method according to claim 3, characterized in that: When switching operating conditions, S33 pre-sets the maximum adjustment range and adjustment time interval for a single operation, defines the parameter operating range and equipment safety boundary for different operating conditions, calculates the total adjustment amount of the current parameter and the target parameter of the new operating condition, and then determines the gradient adjustment step number by rounding up according to the maximum adjustment range for a single operation, and adjusts the parameter step by step according to the gradient.

5. The sludge drying treatment optimization control method according to claim 4, characterized in that: S33 constructs a linear gradient function and adopts a weighted linear gradient transition method, which is started synchronously with the gradient adjustment of the running parameters. Each time the parameter fine-tuning is completed, the weight coefficients are updated synchronously.

6. The sludge drying treatment optimization control method according to claim 3, characterized in that: When multiple sets of parameters have the same lowest comprehensive score, S34 extracts the maximum feasible processing capacity corresponding to each combination and sorts them from high to low according to the maximum feasible processing capacity. If the processing capacity of a certain combination is significantly higher than that of other combinations, the combination is directly selected as the optimal operating parameter. If the processing capacity of multiple combinations is similar, the energy consumption per unit processing capacity is further compared, and the combination with lower energy consumption per unit processing capacity is selected first.

7. The sludge drying treatment optimization control method according to claim 1, characterized in that: The specific steps of S40 in constructing a multi-objective global optimization function and outputting the optimal feed rate allocation instruction and operating parameter coordination fine-tuning instruction for each device are as follows: S41. Combine the optimal operating parameters of a single device with the drying effect data and the device load status, normalize the data, map it to a unified range, set global hard constraints, and construct a multi-objective global optimization function. S42. The feed rate allocation ratio and the fine-tuning amount of the operating parameters of each device are used as optimization variables. Each particle corresponds to a set of global optimization schemes. The optimization function value corresponding to each particle is calculated as the fitness. Then, the particle position is iteratively updated according to the rules, gradually moving towards the individual optimal and global optimal schemes. S43. The converged global optimal solution is parsed into two types of instructions: the optimal feed rate allocation instruction and the running parameter coordination and fine-tuning instruction.

8. The sludge drying treatment optimization control method according to claim 7, characterized in that: The S41 introduces three major objectives: balancing weight coefficients. It establishes a mapping rule between production scenarios and weight coefficients and presets a weight allocation scheme for typical scenarios.

9. The sludge drying treatment optimization control method according to claim 7, characterized in that: In each iteration of S42, each particle updates its speed and position based on its own historical best solution and the population's global best solution, moving closer to a better solution. After the position is updated, a constraint check is performed again. If the position exceeds the constraint range, it is pulled back to the boundary.

10. A sludge drying treatment optimization control system, used in accordance with the sludge drying treatment optimization control method according to any one of claims 1 to 9, characterized in that, include: The data acquisition module (100) collects the core operating parameters of the equipment, the core parameters of sludge characteristics, and the energy consumption and effect feedback parameters in real time during the sludge drying process, and preprocesses the collected data to form a standardized dataset. The anomaly detection module (200) compares the core operating parameters of the device in the standardized dataset with preset thresholds. If the parameters exceed the preset threshold range, the device is determined to be in an abnormal operating condition, and a safety policy is immediately triggered. The working condition classification module (300) extracts the core parameters of sludge characteristics and energy consumption and effect feedback parameters for working conditions that do not exceed the preset threshold range, constructs feature vectors, calculates the Euclidean distance between the feature vectors and the corresponding cluster centers of normal stable working conditions and sludge characteristic fluctuation working conditions, determines the working condition type, loads the hierarchical control strategy for the corresponding working condition, and performs inference calculations on the standardized dataset through a lightweight hybrid optimization model to output the optimal operating parameters for a single device. The parameter optimization module (400) combines the optimal operating parameters of a single device with the drying effect data and the device load status to construct a multi-objective global optimization function with the core objectives of minimizing total energy consumption, maximizing total processing capacity, and maximizing the average moisture content compliance rate. The module uses a lightweight particle swarm optimization algorithm to iteratively solve for the minimum value of the function and outputs the optimal feed rate allocation instructions and operating parameter coordination fine-tuning instructions for each device.

Images