Autonomous optimization control method and system for intelligent sludge treatment equipment

Abstract

The invention provides an autonomous optimization control method and system for intelligent sludge treatment equipment, and the method comprises the steps: deploying an intelligent agent system based on a neural network in a sludge treatment unit, receiving the water content and concentration of sludge as input, and outputting a stirring speed and a pressure control value; the method comprises the following steps: establishing a star network system supporting multi-agent parallel learning, wherein the star network system comprises a central control node and a plurality of sludge treatment unit nodes; and collaborative learning and knowledge migration among the neural networks of the plurality of processing units are realized by using a distributed algorithm. The technical problems that a traditional sludge treatment device is fixed in control strategy and cannot be adaptively adjusted according to different working conditions are solved. Through multi-agent collaborative learning and a knowledge migration mechanism, autonomous optimization of a sludge treatment unit control strategy is realized, the equipment treatment efficiency is improved, the energy consumption is reduced, and meanwhile, the collaborative operation effect among a plurality of treatment units is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of sludge treatment equipment control, and in particular to an autonomous optimization control method and system for intelligent sludge treatment equipment. Background Art

[0002] With the rapid development of industry and urbanization, sludge treatment has become an important topic in the field of environmental protection. The amount of sludge generated in the process of industrial wastewater and urban sewage treatment continues to increase, and the requirements for sludge treatment equipment are also constantly increasing. Existing sludge treatment equipment mainly achieves sludge reduction treatment by adjusting parameters such as stirring speed and pressure. During the treatment process, the stirring speed directly affects the shear force and floc crushing effect of the sludge, while the pressure parameter determines the degree of sludge dehydration. The regulation of these two key parameters has a decisive influence on the sludge treatment effect. At present, most sludge treatment equipment adopts a fixed control strategy or a simple PID control method, which is controlled according to preset parameters, and the control logic is relatively simple.

[0003] However, the prior art has the following significant deficiencies in practical applications: The properties of sludge are complex and changeable. Sludge from different sources and different processes has great differences in moisture content, organic matter content, particle size distribution and other characteristics. Traditional fixed control strategies are difficult to adapt to these differences and often use unified control parameters, which cannot be accurately adjusted for different sludge characteristics, resulting in low treatment efficiency; During the sludge treatment process, key parameters such as sludge moisture content and concentration will change dynamically over time. A single control mode cannot adaptively adjust the control strategy according to the real-time changes of these parameters, which can easily lead to problems of insufficient or excessive treatment. Especially when the treatment conditions suddenly change, the adaptability of existing control methods is obviously insufficient; Large-scale sewage treatment plants are usually equipped with multiple sludge treatment units running in parallel, but there is a lack of effective coordination mechanism between these units. Each treatment unit operates independently, and it is impossible to share experience and transfer optimization strategies, resulting in repeated trial and error under similar working conditions, which not only reduces the overall efficiency of the system, but also causes waste of resources; Energy consumption has always been an important issue in the sludge treatment process. Existing control methods rely too much on empirical parameters, making it difficult to achieve energy conservation and consumption reduction while ensuring treatment effects. Especially when facing different types of sludge, due to the lack of intelligent optimization mechanisms, conservative control strategies are often adopted, resulting in unnecessary energy waste; Traditional control systems lack learning and optimization capabilities and are unable to summarize experience from historical operating data and improve control strategies. This results in low efficiency of equipment operation over the long term, and makes it difficult to cope with new sludge treatment requirements, and the equipment has poor adaptability and scalability.

[0004] These technical problems restrict the treatment efficiency and operation effect of sludge treatment equipment, which not only affects the quality and stability of sludge treatment, but also causes waste of energy resources and increases operating costs. Therefore, an intelligent control method is urgently needed to improve the performance of sludge treatment equipment and realize precise control and continuous optimization of the sludge treatment process. Summary of the invention

[0005] The purpose of the present invention is to propose an autonomous optimization control method and system for intelligent sludge treatment equipment to solve the technical problem that the control strategy of existing sludge treatment equipment is fixed and cannot be adjusted adaptively, thereby improving sludge treatment efficiency and reducing energy consumption. To achieve the above purpose, the technical solution adopted by the present invention is as follows: An autonomous optimization control method for intelligent sludge treatment equipment comprises the following steps: Deploy a neural network-based intelligent agent system in each sludge treatment unit, each of the neural networks receives the sludge moisture content and sludge concentration of the unit as input, and outputs control values ​​of the stirring speed and pressure of the unit; Establishing a star network system supporting multi-agent parallel learning, including a central control node and a plurality of sludge treatment unit nodes; Based on the star network system, the coordinated learning and knowledge transfer between the neural networks of the plurality of sludge treatment units are realized through a distributed algorithm to optimize the neural network of each unit; Each of the sludge treatment units utilizes the optimized neural network to generate and execute control instructions for the unit based on the real-time status data of the unit.

[0006] A further technical solution is that the step of deploying a neural network-based intelligent system in each sludge treatment unit includes: Constructing a three-layer fully connected neural network in each of the sludge treatment units as the strategy network of the unit, wherein the input layer of the strategy network receives the state parameters of the sludge treatment unit, the strategy network comprises at least one hidden layer for feature extraction and nonlinear mapping, and the output layer of the strategy network generates the control parameters of the sludge treatment unit; The number of nodes in the input layer of the strategy network is the same as the dimension of the unit state vector, the state vector includes the sludge moisture content and sludge concentration of the unit, the hidden layer of the strategy network includes a plurality of neurons, the number of nodes in the output layer of the strategy network is the same as the dimension of the unit action vector, and the action vector includes the stirring speed and pressure of the unit; A reward function based on sludge treatment efficiency is constructed for each sludge treatment unit, and the reward function is calculated by the ratio of the sludge reduction value of the unit to the energy consumption value per unit processing volume of the unit, wherein the sludge reduction value is the ratio of the difference in mass of the inlet and outlet sludge of the unit per unit time to the processing time, and the energy consumption value per unit processing volume is the ratio of the total energy consumption of the unit to the total amount of sludge treated.

[0007] A further technical solution is that the steps of establishing a star network system supporting multi-agent parallel learning include: The star network system is established by using a network communication protocol, with the central control node as the central node and the sludge treatment unit as the branch node; A local experience pool is set in each of the sludge treatment units, and the local experience pool adopts a circular queue structure to store experience data including the unit's identification, the unit's current state vector, the unit's action vector, the unit's reward value, and the unit's next state vector; A global experience pool is set at the central control node for storing and distributing high-quality experience data, wherein the high-quality experience data is screened based on the relative size of the reward value calculated by the reward function of each of the sludge treatment units; A communication retransmission and heartbeat detection mechanism is set up to ensure the reliability of the star network system.

[0008] A further technical solution is that the star network system comprises: A central control node to coordinate the overall optimization process; The plurality of sludge treatment unit nodes respectively establish direct two-way communication connections with the central control node; An independent communication channel is established between the central control node and each of the sludge treatment unit nodes; A further technical solution is that the step of implementing collaborative learning and knowledge transfer between the neural networks of the plurality of sludge treatment units through a distributed algorithm comprises: A distributed gradient descent algorithm is used to simultaneously train the respective strategy networks of the plurality of sludge treatment units to achieve parallel optimization of the sludge treatment units; Based on the cosine similarity, the similarity between the current state of the target sludge treatment unit and the historical state in the global experience pool is calculated, and when the cosine similarity is greater than a preset similarity threshold, the corresponding experience data is selected, and the experience data is input into the strategy network of the target sludge treatment unit for training, so as to optimize the control strategy of the unit for similar working conditions; The efficiency score is the average value of the reward values ​​obtained by the sludge treatment unit within a preset time period; The central control node periodically performs strategy parameter migration. When the efficiency score difference between the two sludge treatment units is greater than a preset efficiency threshold, the central control node obtains the strategy network parameters from the sludge treatment unit with a higher efficiency score, and updates the strategy network parameters to the strategy network of the sludge treatment unit with a lower efficiency score; A further technical solution is that the step of generating and executing control instructions based on the real-time status data of the sludge treatment unit comprises: Each of the sludge treatment units collects the sludge moisture content and sludge concentration data of the unit in real time, performs outlier detection, missing value processing and numerical normalization on the data, and constructs a state vector of the unit; The stirring speed value and the pressure value of the unit are calculated by the strategy network of the unit, and validity verification is performed, wherein the validity verification includes determining whether the stirring speed value and the pressure value are within the operating range allowed by the equipment, and determining whether the changes of the stirring speed value and the pressure value conform to the dynamic response characteristics of the equipment; When the variation range of the stirring speed value and the pressure value of the unit exceeds the control parameter smoothing threshold, a linear interpolation method is used to generate a transition control sequence of the unit; The control parameters verified by the unit are encoded into standard control instructions and executed.

[0009] The present invention also provides an autonomous optimization control system for intelligent sludge treatment equipment, comprising: a plurality of sludge treatment units and a central control node, wherein: The sludge treatment unit is a sludge reduction treatment functional unit with a stirring device and a pressurizing device, comprising: A feeding system for conveying sludge to be treated; A processing chamber is provided with the stirring device and the pressurizing device; Discharging system, used to discharge the treated sludge; Data collection module, used to collect sludge moisture content, sludge concentration data, inlet and outlet sludge quality data, processing time data and energy consumption data; A strategy network module, including a three-layer fully connected neural network structure, is used to output a stirring speed value and a pressure value according to the sludge moisture content and the sludge concentration data; A control execution module, used for executing control instructions corresponding to the stirring speed value and the pressure value; The central control node is connected to the plurality of sludge treatment units via a network system, and is used to coordinate collaborative learning and knowledge transfer of the plurality of sludge treatment units.

[0010] A further technical solution is that: the sludge treatment unit further comprises a local experience pool, the local experience pool being used to store experience data including a unit identification, a current state vector, an executed action vector, an obtained reward value and a next state vector; The central control node includes a global experience pool and a distributed training module. The global experience pool is used to store and distribute high-quality experience data. The distributed training module is used to execute a distributed gradient descent algorithm to train the policy network and perform policy parameter migration based on efficiency scores.

[0011] The beneficial effects of the present invention are as follows: (1) Significant intelligent adaptive control effect: The present invention adopts an intelligent system based on a neural network, which receives sludge moisture content and concentration data in real time and dynamically outputs the optimal stirring speed and pressure control value. Compared with the traditional fixed control strategy, the present invention can automatically adjust the processing parameters according to different sludge characteristics, making the control strategy more accurate and flexible.

[0012] (2) The advantages of multi-unit collaborative learning are outstanding: The present invention innovatively establishes a star network system and a distributed algorithm framework to achieve collaborative learning and knowledge transfer among multiple processing units. When a unit obtains high-quality processing experience, it can be quickly shared with other units to avoid repeated trial and error.

[0013] (3) The experience pool is designed scientifically and efficiently: The present invention sets up a local experience pool in each processing unit and a global experience pool in the central control node, and adopts a circular queue structure to efficiently store operating experience. Through cosine similarity calculation, the system can quickly retrieve and utilize historical high-quality experience to provide optimization reference for new working conditions.

[0014] (4) Strong system reliability and scalability: The present invention adopts a distributed gradient descent algorithm for parallel training and sets up a communication retransmission and heartbeat detection mechanism to ensure the stable and reliable operation of the system. At the same time, the star network architecture facilitates the rapid access of new processing units and the migration of optimization strategies. The system has strong scalability and good adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of the autonomous optimization control method for intelligent sludge treatment equipment.

[0016] Figure 2 Deployment flow chart for neural network-based agent systems.

[0017] Figure 3 This is a star network system architecture diagram.

[0018] Figure 4 Flowchart of collaborative learning and knowledge transfer of neural network in sludge treatment unit.

[0019] Figure 5 This is the architecture diagram of the autonomous optimization control system for intelligent sludge treatment equipment. DETAILED DESCRIPTION

[0020] The following is a clear and complete description of the technical solutions in the embodiments of the invention in conjunction with the drawings in the embodiments of the invention. Obviously, the described embodiments are only part of the embodiments of the invention, not all of the embodiments. Based on the embodiments of the invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the invention.

[0021] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0022] Reference Figure 1 The main steps of the method provided by the present invention are as follows: (1) A neural network-based intelligent agent system is deployed in each sludge treatment unit. Each neural network receives the sludge moisture content and sludge concentration of the unit as input and outputs the control values ​​of the stirring speed and pressure of the unit.

[0023] (2) A star network system that supports multi-agent parallel learning is established, including a central control node and multiple sludge treatment unit nodes. Each sludge treatment unit establishes a two-way communication connection with the central control node to realize the storage and sharing of experience data.

[0024] (3) Based on the star network system, the coordinated learning and knowledge transfer between the neural networks of multiple sludge treatment units are realized through a distributed algorithm, including distributed gradient descent training, similar experience reuse and strategy parameter migration, to optimize the neural network of each unit.

[0025] (4) Each sludge treatment unit uses the optimized neural network to generate and execute control instructions based on real-time status data, including data preprocessing, control parameter calculation, validity verification, and instruction execution.

[0026] As a preferred embodiment of the present invention, Figure 2 As shown, in step (1), the deployment process of the neural network-based agent system is as follows: (1-1) For each sludge treatment unit, a three-layer fully connected neural network is constructed as the policy network of the unit. The policy network includes an input layer, a hidden layer and an output layer. The input layer is used to receive the state parameters of the sludge treatment unit; the hidden layer includes multiple neuron nodes for feature extraction and nonlinear mapping of the input data; the output layer is used to generate the control parameters of the sludge treatment unit. Specifically, the number of nodes in the input layer of the policy network is the same as the dimension of the unit state vector, and the state vector includes the sludge moisture content and sludge concentration data collected by the unit in real time; preferably, the hidden layer of the policy network adopts the ReLU activation function, preferably, includes 30-50 neuron nodes; the number of nodes in the output layer of the policy network is the same as the dimension of the unit action vector, and the action vector includes the stirring speed and pressure control value of the unit.

[0027] (1-2) Construct a reward function based on sludge treatment efficiency for each sludge treatment unit. The reward function is calculated by the ratio of the unit's sludge reduction value to the unit's unit processing energy consumption value, and is used to evaluate the pros and cons of the control strategy. The sludge reduction value calculation formula is: the ratio of the unit's inlet and outlet sludge quality difference per unit time to the processing time; the unit processing energy consumption value calculation formula is: the ratio of the unit's total energy consumption to the total amount of sludge treated. Preferably, this calculation method can simultaneously consider the sludge treatment reduction effect and energy consumption, and achieve a balance between treatment efficiency and energy consumption.

[0028] (1-3) Preferably, data preprocessing is performed in each sludge treatment unit, and the collected sensor data is standardized. Preferably, it specifically includes: performing outlier detection on the sludge moisture content and sludge concentration data, and eliminating data that obviously deviates from the normal range; using interpolation methods to repair missing data; normalizing the data to the [-1, 1] interval to improve the training effect and prediction accuracy of the neural network. Preferably, a data caching mechanism is set to regularly save the preprocessed historical data for offline training and online optimization of the neural network.

[0029] As a preferred embodiment of the present invention, refer to Figure 3 In step (2), the star network system includes a central control node and multiple sludge treatment unit nodes. The central control node serves as a central node, the sludge treatment unit serves as a branch node, and each sludge treatment unit node establishes a direct two-way communication connection with the central control node. The network structure can support parallel learning of multiple sludge treatment units to achieve collaborative optimization of control strategies.

[0030] Step (2-1) The communication module design process of the star network system includes: using a network communication protocol to establish a star network system. Preferably, the TCP / IP protocol is used to achieve reliable communication between nodes to ensure the integrity and real-time performance of data transmission. Preferably, the design of the communication protocol includes message type definition, data format specification and communication timing requirements, and supports multiple business requirements such as empirical data transmission and parameter synchronization.

[0031] Step (2-2) The design process of the experience pool module of the star network system includes: setting a local experience pool in each sludge treatment unit, wherein the local experience pool adopts a circular queue structure and is used to store experience data including the unit's identification, the unit's current state vector, the unit's action vector, the unit's reward value, and the unit's next state vector. Preferably, the capacity of the local experience pool is 1000-2000 experience data, and when the storage space is insufficient, the new experience data will replace the earliest stored data.

[0032] Step (2-3) The central management module design process of the star network system includes: Setting a global experience pool at the central control node for storing and distributing high-quality experience data, wherein the high-quality experience data is screened based on the relative size of the reward value calculated by the reward function of each sludge treatment unit. Preferably, a reward value screening threshold is set to save only the experience data with a reward value in the top 30%. Preferably, the global experience pool is updated regularly to keep the data up to date.

[0033] Step (2-4) The reliability assurance process of the star network system includes: Setting up a communication retransmission and heartbeat detection mechanism to ensure the reliability of the star network system. Specifically, when a communication failure is detected, the system automatically retransmits the data packet; the online status of each node is detected through regular heartbeat messages, and when a node is found to be offline, fault handling is performed in a timely manner. Preferably, the upper limit of the number of retransmissions is set to 3 times, and the heartbeat detection period is set to 5 seconds. Preferably, the system also includes an encryption and authentication mechanism for communication data to ensure the security of data transmission.

[0034] As a preferred embodiment of the present invention, refer to Figure 4 In step (3), the collaborative learning and knowledge transfer between the neural networks of multiple sludge treatment units are realized by a distributed algorithm, including the following steps: Step (3-1) A distributed gradient descent algorithm is used to simultaneously train the respective policy networks of multiple sludge treatment units to achieve parallel optimization of the sludge treatment units. Preferably, the training process is as follows: each sludge treatment unit calculates the parameter gradient of the policy network based on local empirical data; the locally calculated gradient information is sent to the central control node; the central control node aggregates the collected gradient information to generate a global gradient update value; the global gradient update value is distributed to each sludge treatment unit for updating the respective policy network parameters. Preferably, the Adam optimizer is used for parameter update, the initial value of the learning rate is set to 0.001, and it is dynamically adjusted during the training process.

[0035] Step (3-2) calculates the similarity between the current state of the target sludge treatment unit and the historical state in the global experience pool based on cosine similarity. When the cosine similarity is greater than the preset similarity threshold, the corresponding experience data is selected, and the experience data is input into the policy network of the target sludge treatment unit for training to optimize the control strategy of the unit for similar working conditions. Preferably, the specific implementation process includes: calculating the cosine similarity between the current state vector of the target sludge treatment unit and the historical state vector in the global experience pool; preferably, setting the similarity threshold to 0.85; screening experience data with a similarity greater than the threshold from the global experience pool, and adding these experience data to the local experience pool of the target unit. Preferably, the amount of experience data added each time does not exceed 20% of the capacity of the local experience pool. In supplementary training, a small batch stochastic gradient descent method is used. Preferably, the batch size is set to 32, the number of training rounds is set to 10 rounds, and the learning rate is set to 0.0001.

[0036] Step (3-3) sets up an efficiency evaluation mechanism and performs policy parameter migration. The efficiency score is the average value of the reward value obtained by the sludge treatment unit within a preset time period. The central control node periodically performs policy parameter migration. When the difference in efficiency scores between the two sludge treatment units is greater than a preset efficiency threshold, the central control node obtains its policy network parameters from the sludge treatment unit with a high efficiency score, and updates the policy network parameters to the policy network of the sludge treatment unit with a low efficiency score. Preferably, the specific implementation process is as follows: calculate the efficiency score of each sludge treatment unit within a preset time period (preferably set to 24 hours); preferably, set the efficiency threshold to 20%; sort the sludge treatment units according to the efficiency score; when a unit pair whose efficiency score difference exceeds the threshold is detected, trigger the policy parameter migration; preferably, during the policy parameter migration process, the network parameters of the high-efficiency unit are directly copied to replace the network parameters of the low-efficiency unit to achieve rapid knowledge migration.

[0037] Through the above-mentioned collaborative learning and knowledge transfer mechanism, the operating experience of each sludge treatment unit can be fully utilized, the optimization process of the strategy network can be accelerated, and the control effect of the entire system can be improved. Among them, distributed training ensures learning efficiency, similarity calculation ensures the effectiveness of experience reuse, and strategy migration promotes knowledge sharing between different units.

[0038] As a preferred embodiment of the present invention, in step (4), each sludge treatment unit generates and executes control instructions based on real-time status data, including the following steps: Step (4-1) Each sludge treatment unit collects the sludge moisture content and sludge concentration data of the unit in real time, performs outlier detection, missing value processing and numerical normalization on the data, and constructs the state vector of the unit. Preferably, a sensor is used to collect real-time data, and the sampling period is set to 1 second. The collected data is firstly subjected to outlier detection, and preferably, the 3σ criterion is used to identify abnormal data points. The detected outliers and missing values ​​are repaired, and preferably, the nearest neighbor interpolation method is used for data repair. The processed data is then normalized and mapped to the [-1, 1] interval, and finally a standardized state vector is constructed.

[0039] Step (4-2) The stirring speed value and pressure value of the unit are calculated by the strategy network of the unit, and the validity is verified. The validity verification includes judging whether the stirring speed value and the pressure value are within the operating range allowed by the equipment, and judging whether the changes in the stirring speed value and the pressure value conform to the dynamic response characteristics of the equipment. Specifically, firstly, the standardized state vector is input into the strategy network to obtain the initial control value; then check whether the control value is within the operating range of the equipment, preferably, the stirring speed range is 0-1000rpm, and the pressure range is 0-2MPa. At the same time, the rate of change of the control value is calculated. Preferably, the maximum rate of change of the stirring speed is set to 100rpm / s, and the maximum rate of change of the pressure is set to 0.2MPa / s. When the control value exceeds the limit range, truncation processing is performed to ensure the validity of the control parameters.

[0040] Step (4-3) When the variation range of the stirring speed value and the pressure value of the unit exceeds the control parameter smoothing threshold, a linear interpolation method is used to generate a transition control sequence of the unit, and the verified control parameters are encoded as standard control instructions and executed. Specifically, a control parameter smoothing threshold is first set. Preferably, the stirring speed variation threshold is 50rpm, and the pressure variation threshold is 0.1MPa. When the parameter variation exceeds the threshold, preferably, a smooth transition sequence is generated by linear interpolation. The control parameters are then encoded according to the device communication protocol, a control instruction in a standard format is generated, and the instruction is sent to the device execution layer through the control execution module. Preferably, an instruction execution feedback mechanism is set to monitor the execution status of the control instruction to ensure the reliability of the control process.

[0041] Through the above control instruction generation and execution mechanism, the smooth operation of the sludge treatment equipment can be ensured, and the impact of drastic changes in control parameters on the equipment can be avoided. At the same time, the perfect validity verification and smooth processing mechanism can improve the reliability and stability of the control system.

[0042] Reference Figure 5 The autonomous optimization control system of the intelligent sludge treatment equipment provided by the present invention includes multiple sludge treatment units and a central control node. The system adopts a distributed architecture to realize intelligent control and optimization of sludge treatment.

[0043] As a specific embodiment of the present invention, the sludge treatment unit is a sludge reduction treatment functional unit having a stirring device and a pressurizing device, and its specific structure includes: The feeding system is used to transport the sludge to be treated. Specifically, it includes a feeding pump, a piping system and a flow metering device. Preferably, the feeding pump adopts a variable frequency speed regulating screw pump, and the feeding rate can be adjusted according to the process requirements; preferably, the flow metering device adopts an electromagnetic flowmeter to monitor the feeding amount in real time.

[0044] The treatment chamber is provided with a stirring device and a pressurizing device. Specifically, the stirring device includes a variable frequency motor and a stirring paddle, which are used to achieve full mixing of the sludge; the pressurizing device includes a pressure pump and a pressure sensor, which are used to adjust and monitor the chamber pressure. Preferably, the stirring paddle adopts a double propeller structure to improve the mixing effect; preferably, the pressure sensor has a range of 0-5MPa and an accuracy level of 0.5.

[0045] The discharge system is used to discharge the treated sludge. Specifically, it includes a discharge pump, a discharge pipeline and a sludge metering device. Preferably, the discharge system is equipped with an automatic regulating valve to control the discharge rate according to process requirements.

[0046] The data acquisition module is used to collect sludge moisture content, sludge concentration data, inlet and outlet sludge quality data, processing time data and energy consumption data. Specifically, it includes various sensors, data acquisition cards and signal conditioning circuits. Preferably, the moisture content is measured by a microwave moisture meter, the concentration is monitored by an online concentration meter, and the energy consumption data is collected by an electric energy meter.

[0047] The strategy network module includes a three-layer fully connected neural network structure, which is used to output the stirring speed value and the pressure value according to the sludge moisture content and sludge concentration data. Specifically, the module is deployed on an industrial computer and uses a real-time operating system to run the neural network algorithm. Preferably, the industrial computer is equipped with an industrial-grade CPU and memory to ensure computing performance and stability.

[0048] The control execution module is used to execute the control instructions corresponding to the stirring speed value and the pressure value. Specifically, it includes a PLC controller, a frequency converter and an actuator. Preferably, a PLC system with a redundant function is used to improve control reliability.

[0049] The local experience pool is used to store experience data including unit identification, current state vector, executed action vector, reward value obtained and next state vector. Specifically, an industrial-grade database is used to implement data storage and management. Preferably, a regular data backup mechanism is set to prevent data loss.

[0050] As a specific implementation of the present invention, the central control node includes: The global experience pool is used to store and distribute high-quality experience data. Specifically, a distributed database system is used to implement data storage and management, supporting multi-node concurrent access. Preferably, a data sharding storage mechanism is set to improve data access efficiency.

[0051] The distributed training module is used to execute the distributed gradient descent algorithm to train the policy network and perform policy parameter migration based on the efficiency score. Specifically, it includes a training scheduler, a parameter synchronizer, and a performance evaluator. Preferably, a GPU accelerator card is used to increase the training speed, and a breakpoint resume function is configured to avoid training interruptions.

[0052] As a specific implementation of the present invention, the components of the system are connected via industrial Ethernet. Specifically, a ring network topology is used to improve communication reliability, and a network redundancy protection function is configured. Preferably, network security protection measures are set, including firewalls, access control and other mechanisms.

[0053] Through the above system implementation plan, intelligent control and optimization of sludge treatment equipment can be achieved. Among them, the various functional modules of the sludge treatment unit work together to ensure the stable operation of a single treatment unit; the central control node realizes the coordinated optimization of multiple units through the experience pool and distributed training; the reliable system integration solution ensures the stability and scalability of the entire system.

[0054] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be within the protection scope of the present invention.

Claims

1. An autonomous optimization control method for intelligent sludge treatment equipment, characterized in that: The following steps are involved: Deploy a neural network-based intelligent agent system in each sludge treatment unit, each of the neural networks receives the sludge moisture content and sludge concentration of the unit as input, and outputs control values ​​of the stirring speed and pressure of the unit; Establishing a star network system supporting multi-agent parallel learning, including a central control node and a plurality of sludge treatment unit nodes; Based on the star network system, the coordinated learning and knowledge transfer between the neural networks of the plurality of sludge treatment units are realized through a distributed algorithm to optimize the neural network of each unit; Each of the sludge treatment units utilizes the optimized neural network to generate and execute control instructions for the unit based on the real-time status data of the unit.

2. The autonomous optimization control method of an intelligent sludge treatment equipment according to claim 1 is characterized in that: The steps of deploying a neural network-based intelligent agent system in each sludge treatment unit include: Constructing a three-layer fully connected neural network in each of the sludge treatment units as the strategy network of the unit, wherein the input layer of the strategy network receives the state parameters of the sludge treatment unit, the strategy network comprises at least one hidden layer for feature extraction and nonlinear mapping, and the output layer of the strategy network generates the control parameters of the sludge treatment unit; The number of nodes in the input layer of the strategy network is the same as the dimension of the unit state vector, the state vector includes the sludge moisture content and sludge concentration of the unit, the hidden layer of the strategy network includes a plurality of neurons, the number of nodes in the output layer of the strategy network is the same as the dimension of the unit action vector, and the action vector includes the stirring speed and pressure of the unit; A reward function based on sludge treatment efficiency is constructed for each sludge treatment unit, and the reward function is calculated by the ratio of the sludge reduction value of the unit to the energy consumption value per unit processing volume of the unit, wherein the sludge reduction value is the ratio of the difference in mass of the inlet and outlet sludge of the unit per unit time to the processing time, and the energy consumption value per unit processing volume is the ratio of the total energy consumption of the unit to the total amount of sludge treated.

3. The autonomous optimization control method of an intelligent sludge treatment equipment according to claim 1 is characterized in that: The steps of establishing a star network system supporting multi-agent parallel learning include: The star network system is established by using a network communication protocol, with the central control node as the central node and the sludge treatment unit as the branch node; A local experience pool is set in each of the sludge treatment units, and the local experience pool adopts a circular queue structure to store experience data including the unit's identification, the unit's current state vector, the unit's action vector, the unit's reward value, and the unit's next state vector; A global experience pool is set at the central control node for storing and distributing high-quality experience data, wherein the high-quality experience data is screened based on the relative size of the reward value calculated by the reward function of each of the sludge treatment units; A communication retransmission and heartbeat detection mechanism is set up to ensure the reliability of the star network system.

4. The autonomous optimization control method of an intelligent sludge treatment equipment according to claim 3 is characterized in that: The star network system comprises: A central control node to coordinate the overall optimization process; The plurality of sludge treatment unit nodes respectively establish direct two-way communication connections with the central control node; An independent communication channel is established between the central control node and each of the sludge treatment unit nodes.

5. The autonomous optimization control method of an intelligent sludge treatment equipment according to claim 1 is characterized in that: The step of implementing collaborative learning and knowledge transfer between the neural networks of the plurality of sludge treatment units through a distributed algorithm comprises: A distributed gradient descent algorithm is used to simultaneously train the respective strategy networks of the plurality of sludge treatment units to achieve parallel optimization of the sludge treatment units; Based on the cosine similarity, the similarity between the current state of the target sludge treatment unit and the historical state in the global experience pool is calculated, and when the cosine similarity is greater than a preset similarity threshold, the corresponding experience data is selected, and the experience data is input into the strategy network of the target sludge treatment unit for training, so as to optimize the control strategy of the unit for similar working conditions; The efficiency score is the average value of the reward values ​​obtained by the sludge treatment unit within a preset time period; The central control node periodically performs strategy parameter migration. When the efficiency score difference between the two sludge treatment units is greater than a preset efficiency threshold, the central control node obtains the strategy network parameters from the sludge treatment unit with a higher efficiency score, and updates the strategy network parameters to the strategy network of the sludge treatment unit with a lower efficiency score.

6. The autonomous optimization control method of an intelligent sludge treatment equipment according to claim 1 is characterized in that: The step of generating and executing control instructions based on the real-time status data of the sludge treatment unit comprises: Each of the sludge treatment units collects the sludge moisture content and sludge concentration data of the unit in real time, performs outlier detection, missing value processing and numerical normalization on the data, and constructs a state vector of the unit; The stirring speed value and the pressure value of the unit are calculated by the strategy network of the unit, and validity verification is performed, wherein the validity verification includes determining whether the stirring speed value and the pressure value are within the operating range allowed by the equipment, and determining whether the changes of the stirring speed value and the pressure value conform to the dynamic response characteristics of the equipment; When the variation range of the stirring speed value and the pressure value of the unit exceeds the control parameter smoothing threshold, a linear interpolation method is used to generate a transition control sequence of the unit; The control parameters verified by the unit are encoded into standard control instructions and executed.

7. An autonomous optimization control system for intelligent sludge treatment equipment, characterized in that: It includes multiple sludge treatment units and central control nodes, including: The sludge treatment unit is a sludge reduction treatment functional unit with a stirring device and a pressurizing device, comprising: A feeding system for conveying sludge to be treated; A processing chamber is provided with the stirring device and the pressurizing device; Discharging system, used to discharge the treated sludge; Data collection module, used to collect sludge moisture content, sludge concentration data, inlet and outlet sludge quality data, processing time data and energy consumption data; A strategy network module, including a three-layer fully connected neural network structure, is used to output a stirring speed value and a pressure value according to the sludge moisture content and the sludge concentration data; A control execution module, used for executing control instructions corresponding to the stirring speed value and the pressure value; The central control node is connected to the plurality of sludge treatment units via a network system, and is used to coordinate collaborative learning and knowledge transfer of the plurality of sludge treatment units.

8. The autonomous optimization control system of an intelligent sludge treatment equipment according to claim 7 is characterized in that: The sludge treatment unit also includes a local experience pool, which is used to store experience data including a unit identifier, a current state vector, an executed action vector, an obtained reward value, and a next state vector; The central control node includes a global experience pool and a distributed training module. The global experience pool is used to store and distribute high-quality experience data. The distributed training module is used to execute a distributed gradient descent algorithm to train the policy network and perform policy parameter migration based on efficiency scores.

Images