Solenoid valve control method and system of jacquard machine

Abstract

The invention relates to the technical field of jacquard control, in particular to an electromagnetic valve control method and system for a jacquard, and the method comprises the steps: collecting electromagnetic valve state information and fabric weaving state information, and obtaining a comprehensive feature vector; based on the comprehensive feature vector, utilizing an experience recovery mechanism and a target network technology to train an intelligent agent; pre-training the CNN model and migrating to a target model for fine tuning, and establishing a control parameter adjustment rule; an alliance chain network is constructed, and cooperative control among nodes is realized in combination with edge calculation; a high-speed switching device mathematical model is established, a PID control algorithm is operated, and a driving signal is adjusted in real time according to the deviation between a feedback signal and a set value; integrating a load sensor to construct a load model, and designing an energy recovery and energy saving related mechanism; and establishing an energy-saving effect evaluation index system, and optimizing a scheme according to data and adapting to different types of jacquard machines. The system operation efficiency and collaboration are improved by designing the reward function and other operations, the energy consumption is reduced, and the energy utilization rate is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of jacquard machine control, and in particular to a solenoid valve control method and system for the jacquard machine. Background Art

[0002] Traditional jacquard machines struggle to accurately and in real time adjust the solenoid valve drive signals based on the weaving task and machine status. The solenoid valves exhibit significant timing errors, leading to inaccurate fabric patterns. This is especially true when weaving complex patterns, where coordination between the solenoid valves cannot be guaranteed, severely impacting fabric quality. Furthermore, the solenoid valves consume a lot of energy, resulting in low energy efficiency and increased production costs. Furthermore, traditional control methods lack effective adaptability to the diverse jacquard machine models and complex and changing weaving processes and operating conditions, making it difficult to quickly adjust control strategies to meet production needs.

[0003] Jacquard looms generate multi-source, heterogeneous data during operation, such as operational data from solenoid valves and related components, and information on fabric weaving status. Traditional systems lack the ability to effectively collect, fuse, and analyze this multimodal data. The inability to comprehensively and accurately collect all types of data leads to an incomplete understanding of the equipment's operating status. Without efficient data fusion algorithms, it is difficult to extract comprehensive features from multimodal data that reflect the jacquard loom's operating status, thus failing to provide sufficient data support for intelligent control. Long-term, continuous operation of jacquard looms poses a significant energy consumption problem. Traditional jacquard looms lack systematic energy-saving control measures and are unable to adjust energy consumption in real time based on load changes, environmental parameters, and operating modes. Furthermore, there is no effective mechanism for recycling the back-electromotive force generated by the solenoid valves, resulting in significant energy waste. When multiple jacquard looms operate collaboratively, traditional systems lack efficient distributed collaborative control mechanisms. Data transmission between nodes is insecure and susceptible to tampering and forgery. Furthermore, a lack of standardized data sharing rules and collaborative operation processes leads to low production efficiency and prevents optimal resource allocation. Moreover, traditional jacquard machines cannot be optimized specifically when faced with the special needs of different types of jacquard machines (such as high-speed jacquard machines, large jacquard machines, jacquard machines for special fabric weaving, etc.), and it is difficult to meet diverse production scenarios. Summary of the Invention

[0004] The present invention can achieve precise and intelligent control of the solenoid valve, improve the system operation efficiency and coordination, adapt to different types of jacquard machines through evaluation and optimization, and reduce energy consumption while ensuring weaving accuracy and stability.

[0005] The technical solution proposed by the present invention is: a solenoid valve control method for a jacquard machine, the method comprising: The solenoid valve status information and fabric weaving status information are collected, and after preprocessing, a comprehensive feature vector is obtained through weighted summation and fusion; Based on the comprehensive feature vector, the reward function is designed by combining the solenoid valve action time error, energy consumption and fabric defect rate. The experience recycling mechanism and target network technology are used to train the intelligent agent. Collect data from multiple jacquard machine models to build a dataset, pre-train the CNN model and migrate it to the target machine model for fine-tuning, and establish control parameter adjustment rules based on working condition identification; Build a consortium chain network, use the PBFT algorithm to reach transaction consensus, and combine edge computing to achieve collaborative control between nodes; Establish a mathematical model for high-speed switching devices, design a priority scheduling algorithm to optimize the transmission sequence and time interval of the drive signal, and run a PID control algorithm to adjust the drive signal in real time based on the deviation between the feedback signal and the set value; Integrate load sensors to build load models, adjust PWM signals based on environmental parameters, and design energy recovery and energy-saving mechanisms; Establish an energy-saving effect evaluation index system, monitor data and conduct in-depth analysis, optimize solutions based on the data and adapt to different types of jacquard machines.

[0006] Preferably, the process of obtaining the comprehensive feature vector includes the following steps: The operating data of the solenoid valve and related components and the fabric weaving status information are collected; the collected multimodal data are cleaned, filtered and normalized; the image data is feature extracted through CNN; the time series features are extracted through LSTM; and the comprehensive feature vector is obtained by feature fusion through weighted summation.

[0007] Preferably, the agent training process is as follows: Based on the comprehensive feature vector, the reward function is designed in combination with the solenoid valve action time error, energy consumption, and fabric defect rate. The experience replay mechanism and target network technology are used, the mean square error is used as the DQN loss function, and the gradient descent method is used to update the main network parameters. The reward function formula is as follows: ; in: is the reward function; 、 、 is the weighting coefficient, and ; is the reward term related to the action time error; It is a reward item related to energy consumption; It is a bonus item related to the fabric defect rate.

[0008] Preferably, the migration process includes the following: The solenoid valve control data of multiple models of jacquard machines under different weaving processes and working conditions were collected to construct a data set. CNN was used as a pre-training model. After sufficient training with historical data, the model parameters were transferred to the target machine model. The stochastic gradient descent algorithm was used for fine-tuning based on the actual operating data of the target machine model. The weaving speed, fabric material, ambient temperature and humidity working condition information were monitored in real time through sensors, and the working conditions were analyzed in real time using Support Vector Machine (SVM). According to different working condition categories, corresponding control parameter adjustment rules were established to achieve adaptive optimization of the solenoid valve control of the target machine model.

[0009] Preferably, the specific content of the collaborative control is as follows: Build a consortium chain network consisting of multiple jacquard machine nodes, adopt asymmetric encryption technology, establish a node access mechanism, and ensure the safe and stable operation of the network; use the PBFT algorithm to elect the master node through view change to complete the transaction consensus process; write smart contracts to clarify the data sharing rules, collaborative operation procedures and fault handling mechanisms between nodes; deploy edge computing devices at each jacquard machine node to pre-process the data collected by local sensors and execute control algorithms; the edge computing devices interact with the processed data with the blockchain network to achieve collaborative control between nodes and improve the overall operation efficiency and coordination of the jacquard machine system.

[0010] Preferably, the optimization process of the driving signal is as follows: Establish a mathematical model of high-speed switching devices; design a priority scheduling algorithm based on the weaving requirements of the jacquard machine and the working status of the solenoid valve to optimize the transmission order and time interval of the drive signal; study the working principles and performance characteristics of different circuit topologies, use software algorithms to simulate their functions, and dynamically adjust the circuit topology according to the operating status and load changes of the jacquard machine; establish a solenoid valve voltage-current-action characteristic model, adopt a fuzzy control algorithm, and adjust the drive signal according to the difference between the actual current and the expected current and the error change rate; collect the solenoid valve action feedback signal, filter and amplify it, and then use the PID control algorithm to adjust the drive signal in real time according to the deviation between the feedback signal and the set value.

[0011] Preferably, the design process of the energy recovery and energy saving related mechanisms is as follows: Integrated load sensors collect current, pressure, and flow data, and use neural networks to build a load model; based on the load power and weaving requirements predicted by the load model, the PWM signal parameters are adjusted in real time and optimized using genetic algorithms; based on changes in ambient temperature and humidity parameters, compensation algorithms are designed to achieve energy-saving control; high-precision voltage sensors are used to monitor the back electromotive force generated by the solenoid valve in real time, energy recovery circuits and algorithms are designed, an energy management system is established, and a collaborative power supply mode and intelligent monitoring optimization system are designed; based on the working scenarios and task requirements of the jacquard machine, working modes are divided, and cluster analysis algorithms are used to identify working modes in real time, energy-saving strategies are formulated for different modes, a working mode switching mechanism is designed, and user-defined energy-saving options are provided.

[0012] Preferably, the construction process of the energy-saving effect evaluation index system is as follows: Through sensors installed in various key locations, energy consumption data, energy recovery data, and working mode switching data are monitored continuously over a long period of time. Data analysis tools are used to conduct in-depth analysis of the monitoring data. Based on the results of long-term monitoring and data analysis, innovative solutions for energy-saving control enhancement are continuously optimized and iterated. Targeted adjustments and adaptations are made to different types of jacquard machines based on their characteristics to continuously improve the advancement and adaptability of the solutions.

[0013] The present invention also provides a solenoid valve control system for a jacquard machine, wherein the system is used to execute the solenoid valve control method for a jacquard machine.

[0014] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the solenoid valve control method of a jacquard machine.

[0015] Beneficial effects of the present invention: By using a variety of high-precision sensors to comprehensively collect data, preprocessing and integrating it to generate a comprehensive feature vector, the intelligent agent is provided with rich information. Based on this design, the reward function comprehensively considers multiple key factors, allowing the trained intelligent agent to precisely adjust the solenoid valve drive signal. This enables the jacquard control system to fully perceive the operating status, breaking away from the traditional reliance on a single data type. This effectively improves weaving quality, reduces energy consumption, and enhances the precision of solenoid valve movement, achieving preliminary intelligent control and laying the foundation for more complex control strategies.

[0016] The consortium chain network is built, employing asymmetric encryption and strict node access mechanisms for security. Transaction consensus is achieved through the PBFT algorithm, smart contracts standardize data sharing and collaborative operations, and edge computing devices locally process and interact with the blockchain. This enables efficient collaboration between jacquard machine nodes, ensuring data security and traceability, improving overall system efficiency, reducing data transmission pressure, and enabling nodes to work collaboratively based on real-time data, adapting to the complex operational requirements of large-scale, multi-node jacquard machine systems.

[0017] The integrated load sensor builds a load model, adjusts PWM signal parameters in real time based on load and environmental parameters, designs energy recovery circuits and algorithms, establishes an energy management system, divides operating modes, formulates energy-saving strategies, and provides user-defined energy-saving options. This reduces jacquard machine energy consumption in multiple ways, improves energy efficiency, lowers production costs, and minimizes energy waste, meeting environmental protection and sustainable development requirements. Energy-saving strategies can also be flexibly adjusted to suit different operating scenarios and user needs, enhancing system adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a flow chart of a solenoid valve control method for a jacquard machine according to the present invention; Figure 2 The present invention is a flow chart of the control process of a solenoid valve control method for a jacquard machine. DETAILED DESCRIPTION

[0019] The following description is intended to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are for illustrative purposes only, and those skilled in the art will readily appreciate other obvious variations. The basic principles of the present invention defined in the following description may be applied to other embodiments, variations, improvements, equivalents, and other technical solutions that do not depart from the spirit and scope of the present invention.

[0020] It is to be understood that the term "one" should be understood as "at least one" or "one or more", that is, in one embodiment, the number of an element may be one, while in another embodiment, the number of the elements may be multiple, and the term "one" should not be understood as a limitation on the quantity.

[0021] like Figure 1 and Figure 2 As shown, high-precision, high-response current sensors, voltage sensors, temperature sensors, pressure sensors, flow sensors, and image sensors are selected. Based on the structure and operating principle of the jacquard loom, the sensors are rationally arranged to ensure comprehensive and accurate collection of operating data from the solenoid valve and related components, as well as information on the fabric weaving status.

[0022] The collected multimodal data is cleaned, filtered, normalized and other pre-processing operations are performed. Taking normalization as an example, for a data sequence collected by a sensor , using the minimum-maximum normalization formula: ; in: and The data series are The maximum and minimum values ​​in .

[0023] For example, the temperature sensor collects a set of temperature data [20, 22, 25, 23, 21]. =20, =25, then =22After normalization, .

[0024] Convolutional neural network (CNN) is used to extract features from image data, and recurrent neural network (RNN) or long short-term memory network (LSTM) is used to process time series data. Assume that the image data It is a 224×224 RGB image, and the feature vector is obtained after CNN processing ; Use long short-term memory network (LSTM) to process time series data, such as current, voltage and other data sequences over a period of time , after LSTM processing, the feature vector is obtained The feature vectors of different types of data are input into the fusion layer and fused by weighted summation to form a comprehensive feature vector : ; in: and is the weighting coefficient, and .

[0025] For example, when weaving complex floral patterns, the image captured by the image sensor is processed by CNN to extract the texture, color and other features of the pattern to form a The current, voltage and other time series data are processed by LSTM to obtain the working stability of the solenoid valve. If you set , , then the fused comprehensive feature vector Provides a more complete description of the current state.

[0026] Comprehensive consideration of the action time error of the solenoid valve (Unit: milliseconds), energy consumption (Unit: Joule), fabric defect rate (percentage) and other factors to design the reward function : ; in: 、 、 is the weighting coefficient, and , which are used to measure the relative importance of each factor in the reward function. is the reward term related to the action time error, which can be expressed as: ; in: 、 is the reward coefficient, It is the set action time error threshold (in milliseconds); when the action time error is within the threshold range, a reward is given , indicating that the solenoid valve acts in time, which plays a positive role in the knitting process; when the action time error exceeds the threshold, a reward is given ,and , used to penalize action delays.

[0027] is the reward item related to energy consumption, which can be expressed as: ; in: is the maximum energy consumption allowed for a specific knitting task (unit: joules), is the reward coefficient. This formula reflects the proportional relationship between actual energy consumption and maximum allowable energy consumption. The lower the energy consumption, the higher the reward, which incentivizes the system to reduce energy consumption.

[0028] is a bonus item related to the fabric defect rate and can be expressed as: ; in: is the set fabric defect rate threshold (percentage), and Is the reward coefficient. When the fabric defect rate is within the threshold range, a reward is given , indicating good weaving quality; when the fabric defect rate exceeds the threshold, a reward is given ,and , used to penalize poor weaving quality.

[0029] For example, in a knitting task, set , =0.3, =0.3, =5 ms, =1%, =10, =-5, =8, =12, =-8, =100 joules. If during a certain weaving process, = 3 ms, = 60 joules, =0.8%, then =10, , =12, reward function =0.4×10+0.3×3.2+0.3×12=4+0.96+3.6=8.56.

[0030] Use the experience replay mechanism and target network technology to improve the training efficiency and stability of the agent. Stored in experience pool During training, Randomly sample a batch of experiences for learning. The target network Parameters Regularly from the main network The loss function of DQN is the mean square error (MSE): ; in: is a discount factor, which usually ranges from [0,1] and is used to balance the importance of current rewards and future rewards. For example, when =0.9, indicating that the agent pays more attention to the long-term cumulative rewards rather than just the immediate rewards of the current step.

[0031] Hypothetical experience pool 1000 experiences are stored in the dataset, and 32 experiences are randomly sampled from them during each training. During the training process, the loss function is calculated based on the sampled experiences. , and then use gradient descent to update the main network Parameters For example, initially is a set of random values, calculated by back propagation (Loss function with respect to parameters The gradient of ), the learning rate is set to 0.001, and the parameter update formula is , after multiple iterative training, the agent learns a better strategy.

[0032] In actual training, as the number of training sessions increases, the agent's ability to select the optimal action under different conditions becomes increasingly accurate. For example, in a task involving weaving complex patterns on a jacquard loom, after thousands of training sessions, the agent can more precisely adjust the solenoid valve drive signal based on different conditions, such as fabric material and weaving speed, resulting in higher-quality patterns while reducing energy consumption.

[0033] The specific content of the adaptive control strategy based on transfer learning is as follows: Solenoid valve control data for various jacquard loom models (such as Model A and Model B) under different weaving processes (plain, twill, and satin), including normal operation data, fault data, and data under different operating conditions (such as ambient temperature, humidity, and weaving speed), is collected to construct a jacquard loom control dataset. For example, data on the solenoid valve opening and closing time, current, and voltage are collected for Model A when weaving plain fabric at an ambient temperature of 25°C and a humidity of 50%. Data is also collected for Model B when weaving twill fabric at a weaving speed of 30 meters per minute.

[0034] Convolutional neural network (CNN) is selected as the pre-training model. Assuming that the input data x is pre-processed image data (size is (128×128) pixels), the convolution operation after the convolution layer can be expressed as: ; in: It is The output feature map of the convolutional layer is at position ( ), It is The layer input feature map is at position ( ), is the weight of the convolution kernel, and its size is , is the bias. Assume that the convolution kernel size is 3×3 (i.e. , =3), with a stride of 1, in the first convolutional layer, =1, input feature map The size is 128×128. After the convolution operation, the output feature map The size of is (128-3+1)×(128-3+1)=126×126. In the pre-training stage, a large amount of relevant data of different models of jacquard machines (including images, operating parameters and other data related to solenoid valve control) is used to train the CNN model to learn common features and patterns.

[0035] The pre-trained model parameters are transferred to the control model of the target machine model (e.g., Model C jacquard loom). Based on a small amount of actual operating data of Model C jacquard loom, the stochastic gradient descent (SGD) algorithm is used for fine-tuning. The parameter update formula is: ; in: It is The parameters of the iterations, Is the learning rate, which is used to control the step size of each parameter update. For example, it is set to 0.001. is the loss function About parameters The gradient of the loss function It can be the mean square error between the model prediction result and the actual label (such as the actual control state label of the solenoid valve). In one iteration, the loss function is calculated with respect to the parameter Gradient , and then update the parameters according to the formula. For example, when fine-tuning a model C jacquard loom, after 100 iterations, the model's prediction accuracy of the solenoid valve control state on the C jacquard loom increased from an initial 70% to 85%.

[0036] Real-time monitoring of weaving speed through sensors (Unit: m / min), fabric material (such as cotton, linen, silk, etc.), ambient temperature (unit: degrees Celsius) and humidity (Unit:%) and other working condition information. Support vector machine (SVM) is used to identify and classify working conditions in real time. The decision function of SVM is: ; in: is the Lagrange multiplier, is the sample label (corresponding to different working condition category labels), It is the kernel function, here we choose the radial basis kernel function , is the kernel function parameter, for example, set to 0.1, is the bias. Assume there is a set of working condition data ,After the SVM model calculation, it is judged that the working condition belongs to the ,category of “medium speed weaving linen fabric and slightly high ,ambient temperature”.

[0037] According to different working conditions, corresponding control parameter adjustment rules are established. For example, when the weaving speed When accelerating, the driving voltage amplitude of the solenoid valve and pulse width The adjustment formula is: ; ; in: and is the initial driving voltage amplitude and pulse width, is the initial weaving speed, and is the adjustment factor.

[0038] For example, the initial driving voltage amplitude =10V, initial pulse width =4 milliseconds, initial weaving speed =20 m / min, adjustment factor =0.3, =0.2. When the knitting speed is increased to =30 m / min, the driving voltage amplitude is adjusted to =10+0.3×(30-20)=13V, the pulse width is adjusted to =4+0.2×(30-20)=6 milliseconds to ensure that the solenoid valve can adapt to the new knitting speed and ensure the knitting quality.

[0039] Utilizing distributed collaborative control that integrates blockchain and edge computing, the blockchain network construction process is as follows: A consortium chain network consisting of multiple jacquard machine nodes is constructed. Each jacquard machine participates in this network as an independent node, connected by a high-speed communication network to ensure fast and stable data transmission. To ensure network security, asymmetric encryption technology is used to generate a unique public-private key pair for each node. The public key is used to encrypt data transmission between nodes, while the private key is securely stored by each node and used to digitally sign transmitted data to prevent tampering and forgery. A strict node access mechanism is established, ensuring that only authorized jacquard machine nodes can join the consortium chain network. The authorization process is overseen by the consortium chain's management node, which verifies the node's identity, device legitimacy, and other factors before deciding whether to allow it to join. Furthermore, joined nodes are monitored in real time. If a node exhibits abnormal behavior (such as frequently sending erroneous data or engaging in network attacks), the management node has the authority to remove it from the network to maintain the stability and security of the entire consortium chain network.

[0040] In the PBFT algorithm, the operation of the entire system is divided into multiple views. Each view has a master node responsible for coordinating the processing of transactions. When the system starts or the current master node fails, the view change process will be triggered. During the view change process, each node will broadcast the latest view information it believes to other nodes. By comparing the view number and node priority (node ​​priority can be pre-set based on factors such as device performance and historical transaction processing records), a new master node is elected. For example, in a view change, nodes A, B, C, D, and E participate in the election. Node B has the highest view number and a higher priority. In the end, node B is elected as the new master node. The newly elected master node is in the view Next, responsible for generating pre-prepared messages .in, The message sequence number is used to sort transaction messages and ensure sequential processing. The message digest is generated by hashing the transaction content (such as the control instructions for a solenoid valve) and is used to verify message integrity. This is the actual transaction content. The master node broadcasts the preliminary message to other nodes.

[0041] After receiving the pre-prepare message, other nodes will first verify the legitimacy of the message, including whether the view number is correct, whether the message sequence number is reasonable, whether the message summary matches the transaction content, etc. If the verification is successful, the node will send the pre-prepare message to other nodes. ,in Before sending a prepare message, the node will store the prepare message and its own signature locally for subsequent verification.

[0042] When a node receives ( is the number of tolerable Byzantine nodes. Assume there are 5 nodes in the network. ,but ) matching prepare messages, indicating that the transaction has reached a preliminary consensus among the majority of nodes. At this point, the node will send a commit message to other nodes When a node receives After a matching submission message is received, the transaction finally reaches a consensus, and the node will execute the corresponding transaction operation, such as adjusting the working state of the solenoid valve according to the solenoid valve control instruction.

[0043] The specific definition of smart sharing rules involves writing smart contracts to define data sharing rules between nodes. For example, after completing a weaving task, each node is required to upload the solenoid valve control data related to that task (including control instructions, operating status, energy consumption data, etc.) to the blockchain network and set data access permissions. Some data (such as equipment operating status) is publicly accessible to all nodes for collaborative decision-making, while some sensitive data (such as equipment fault diagnosis information) is only visible to specific management nodes or authorized maintenance personnel. The smart contract specifies the collaborative operation process between nodes. When a jacquard machine node receives a new weaving task, the smart contract automatically checks whether the required resources (such as raw materials and energy supply) are sufficient. If resources are sufficient, the smart contract coordinates task allocation and resource scheduling among relevant nodes. For example, it can assign some weaving tasks to adjacent idle jacquard machine nodes to balance network load. The smart contract also monitors task execution progress in real time to ensure that each node completes the task according to the predetermined process and timeframe. The smart contract also incorporates a comprehensive fault handling mechanism for potential jacquard machine failures. When a jacquard machine detects a solenoid valve failure, it immediately triggers a fault alarm function in the smart contract. This function records the fault information (including fault type, occurrence time, and location) on the blockchain, ensuring immutability and traceability. The smart contract also notifies other nodes, which then decide whether to assist in resolving the fault based on their own operational status. For example, if there are idle nodes, they can proactively request to take over some of the faulty machine's tasks, minimizing production interruption.

[0044] The collaborative process of edge computing devices is as follows: The edge computing devices deployed at each jacquard machine node possess powerful computing and data processing capabilities. They first perform real-time preprocessing on the large amount of data collected by local sensors, including data cleaning, filtering, and denoising, to improve data quality. They then execute complex control algorithms, such as deep reinforcement learning algorithms, to analyze the solenoid valve's operating data and predict its failure risk. For example, by monitoring and analyzing solenoid valve current, voltage, and temperature data in real time, a deep reinforcement learning model is used to predict the probability of solenoid valve failure within a certain period of time. If the predicted probability exceeds a set threshold, a timely warning signal is issued. The edge computing devices then upload the locally processed key control data (such as optimized solenoid valve control parameters) and decision results (such as fault diagnosis results) to the blockchain network. Before uploading the data, the edge computing devices encrypt the data and attach a digital signature to ensure its security and integrity. Edge computing devices at other nodes then retrieve this data from the blockchain network and perform collaborative control based on their own actual conditions. For example, when a jacquard machine's weaving task changes, its edge computing device uploads the new control parameters to the blockchain. The edge computing devices of other jacquard machines then retrieve these parameters and adjust their weaving strategies to ensure consistency throughout the production process. Simultaneously, edge computing devices can also obtain information such as the operating status and energy consumption data of other nodes from the blockchain network, analyze and compare this data, and further optimize their own control strategies.

[0045] Through experiments and data analysis, a mathematical model of high-speed switching devices is established to describe its on- and off-characteristics, delay time, power consumption and other parameters. Assuming that the on-resistance of the switching device is , the off resistance is , the turn-on delay time is , the turn-off delay time is , then in the on state, the current through the switching device and the voltage across The relationship is: ; In shutdown state: For example, if the on-resistance of the switching device is Ohm, when current flows through Ampere-hour, the voltage across the two ends in the on-state Volts; if the resistor is turned off Ohms, leakage current The voltage across the two terminals in the off state is in ampere hours. Volt. According to the weaving requirements of the jacquard machine and the working status of the solenoid valve, a scheduling algorithm is designed to optimize the transmission order and time interval of the drive signal. A priority scheduling algorithm is used to assign a priority to the drive signal of each solenoid valve. Priority is determined based on factors such as the complexity of the weaving pattern and the urgency of the solenoid valve's action. For example, a higher priority is assigned to the solenoid valve drive signal for key locations in a complex weaving pattern. Drive signals are sorted and dispatched based on priority, with high-priority drive signals being sent first to ensure the solenoid valve can respond to critical actions promptly. When weaving a complex floral pattern, the solenoid valve drive signal that controls the flower's outline has a higher priority and is dispatched first, ensuring the pattern's precision.

[0046] Research the operating principles and performance characteristics of different circuit topologies (such as H-bridge and half-bridge circuits) and simulate their functions using software algorithms. Taking the H-bridge circuit as an example, controlling the on and off states of four switches can achieve forward and reverse control of a solenoid valve. In the software simulation, logic code is written to simulate the state changes of the switches. For example, when the solenoid valve needs to be energized in the forward direction, two switches are turned on and the other two are turned off. When the solenoid valve needs to be energized in the reverse direction, the combination of on and off switches is switched. The simulated circuit topology is dynamically adjusted based on the jacquard machine's operating status and load changes. Under high load conditions, the simulated circuit topology is switched from a simple half-bridge circuit to a more efficient full-bridge circuit to improve the solenoid valve's driving capability. For example, when weaving heavy fabrics, the load increases, and the software algorithm adjusts the circuit topology to a full-bridge circuit, increasing the driving voltage and current to ensure the solenoid valve operates properly and maintain weaving quality.

[0047] Establish the voltage-current-action characteristic model of the solenoid valve, assuming that the current of the solenoid valve and driving voltage and movement displacement The relationship between can be expressed by the following nonlinear model: By monitoring the current and voltage of the solenoid valve in real time, the model parameters are estimated using the least square method and other methods to predict its action state and performance. For example, a series of experiments are conducted to set the drive voltage 8V, 10V, 12V, etc., respectively, to measure the displacement corresponding to different actions Current at , substitute these data into the least squares formula for calculation, and then obtain the estimated values ​​of each parameter in the model, so as to predict the action state and performance of the solenoid valve under different voltage inputs.

[0048] According to the prediction results and weaving requirements, the fuzzy control algorithm is used to accurately adjust the amplitude, frequency and pulse width of the driving voltage and current. The input of the fuzzy control is the error (the difference between the actual current and the expected current) and the error change rate, and the output is the control quantity (the parameter used to adjust the driving voltage or current). When weaving complex patterns, the expected current is , and the actual measured current is , then the error Calculate the error rate of change , assuming that after a period of time, the current at the previous moment is ,but According to the pre-set fuzzy rule table, if Belongs to fuzzy sets (Indicates that the error is small), Belongs to fuzzy sets (Indicates that the error change rate is moderate), and the control quantity is obtained through fuzzy reasoning Belongs to fuzzy sets , and then through the defuzzification operation, the fuzzy control quantity is converted into the actual control parameter, such as increasing the driving voltage by 0.3V to adjust the current to make it close to the expected current, ensuring the accuracy and stability of the solenoid valve action during the weaving process.

[0049] Collect the action feedback signal of the solenoid valve, such as displacement ,speed ,pressure Etc., filtering, amplification and other processing are performed to improve the accuracy of the feedback signal. Taking a first-order low-pass filter as an example, its transfer function is: ,in, Is the time constant. Assuming that the collected displacement feedback signal contains high-frequency noise, by setting a suitable time constant , the displacement signal is input into the first-order low-pass filter for processing. In the actual circuit, the resistor and capacitors The circuit composed of realizes the filter function, according to the formula , select appropriate resistor and capacitor values. For example, when Ohm, we can calculate After filtering, the displacement signal becomes smoother and can more accurately reflect the actual operation of the solenoid valve.

[0050] The proportional-integral-differential (PID) control algorithm is used to adjust the feedback signal according to the deviation between the feedback signal and the set value. , adjust the drive signal in real time. The output of the PID controller for: ,in, 、 and are proportional, integral and differential coefficients respectively. During the jacquard weaving process, the target displacement of the solenoid valve is set to Millimeters, the actual displacement of the solenoid valve is collected in real time through the displacement sensor , then the deviation . Proportional coefficient Determines the controller's quick response to the deviation, the integral coefficient Used to eliminate the steady-state error of the system, differential coefficient The trend of deviation change can be predicted and adjustments can be made in advance. mm, if , then the proportional term ; Integral term By integrating the deviation over a period of time, assuming that the integration time is 0.1 seconds and the deviation is relatively stable during this period, the integral value is 0.04. , then the integral term If the deviation change rate mm / s, , then the differential term Then, the output of the PID controller is , and adjust the driving signal according to the output, such as increasing the amplitude of the driving voltage or changing the pulse width, so that the actual displacement of the solenoid valve approaches the target displacement, ensuring the weaving accuracy.

[0051] The integrated load sensor collects data and uses neural networks to build a load model. The PWM signal parameters are adjusted in real time according to the model and weaving requirements, and optimized using genetic algorithms. At the same time, a compensation algorithm is designed based on changes in environmental parameters to achieve energy-saving control. In the jacquard system, a high-precision current sensor is tightly connected in series in the power supply circuit of the solenoid valve to accurately collect working current data; the pressure sensor is installed in the air circuit close to the solenoid valve to obtain pressure parameters in real time; the flow sensor is installed on the fluid transmission pipeline to accurately measure the flow value. These sensors are connected to the data acquisition module by wire or wireless means, and the collected load parameters, including current ,pressure ,flow The data is transmitted to the subsequent data processing unit at a specified frequency (e.g., 10 times per second). For example, at a certain moment, the current value collected by the current sensor is 2.5A, the pressure sensor measures the pressure as 0.3MPa, and the flow sensor records the flow as 5L / min.

[0052] The collected load data is analyzed and modeled using a neural network. Assume that the neural network has an input layer, two hidden layers, and an output layer. The input layer receives the load parameter vector , each hidden layer contains several neurons. Tier For example, a neuron has an input for: in, Is connected Tier neurons and Tier The weights of neurons, is the bias, It is The number of neurons in the layer. The output of the neurons Through the activation function Calculated, such as the commonly used ReLU activation function After calculations at each layer, the final output layer outputs load power In actual training, the neural network is trained using a large amount of sample data with known load power, and the weights and biases are continuously adjusted to enable the model to accurately predict load power based on the input load parameters. For example, when a set of load parameters (2.5A, 0.3MPa, 5L / min) is input, the trained neural network predicts an output load power value of 100W.

[0053] Load power according to load forecast and weaving requirements, adjusting the frequency of the pulse width modulation (PWM) signal in real time and duty cycle In the case of light load, in order to reduce the energy consumption of the solenoid valve and reduce unnecessary energy consumption, the frequency and duty cycle of the PWM signal should be appropriately reduced. For example, when the load power Below the set light load threshold When the load is high, the PWM signal frequency is reduced from the default 1000Hz to 500Hz, and the duty cycle is reduced from 50% to 30%. Under heavy load conditions, to ensure that the solenoid valve can work normally and meet the requirements of the weaving process, the frequency and duty cycle of the PWM signal are increased. Assuming the load power Above the set heavy load threshold When the frequency is increased to 2000Hz, the duty cycle is increased to 70%. Assuming the load power The relationship between the frequency and duty cycle of the PWM signal is obtained through experimental fitting as follows: The genetic algorithm is used to optimize the parameters of the PWM signal. In the genetic algorithm, the individual Represents the frequency and duty cycle combination of the PWM signal. Fitness function Defined as: .in, is an individual The corresponding load power. In the iterative process of the genetic algorithm, the individuals in the population are first selected. According to the size of the fitness value, individuals with high fitness (i.e., low load power) are selected to increase their probability of appearing in the next generation. For example, using the roulette wheel selection method, the probability of an individual being selected is proportional to its fitness value. Then a crossover operation is performed to randomly select two individuals and exchange some of their genes (i.e., the values ​​of frequency and duty cycle) to generate new individuals. Assume that the two individuals are and , new individuals may be generated after crossover and Finally, a mutation operation is performed to randomly change the genes of the individual with a certain probability (such as 0.01) to prevent the algorithm from falling into a local optimum. After multiple generations of iteration, the PWM signal parameter combination that minimizes the load power is found.

[0054] In the working environment of the jacquard machine, a temperature sensor and a humidity sensor are installed. The temperature sensor is a thermistor type sensor, which is installed near the solenoid valve and can monitor the temperature of the working environment of the jacquard machine in real time. , the accuracy can reach ±0.5℃. The humidity sensor uses a capacitive humidity sensor, which is installed in the working area to obtain the ambient humidity in real time , with an accuracy of ±2%. These sensors transmit the collected temperature and humidity data to the control system via a data transmission line. For example, at a certain moment, the temperature sensor measures 28°C and the humidity sensor measures 60%.

[0055] The study found that the ambient temperature has an impact on the resistance of the solenoid valve. There is an influence, and the relationship can be expressed as: ; in: is the initial resistance, is the temperature coefficient (for example, for a certain solenoid valve, ℃), is the initial temperature (set to 25℃). When the ambient temperature changes, the resistance change is calculated according to the above formula, and the parameters of the PWM signal are adjusted accordingly. For example, when the temperature =28℃, assuming =10Ω, then the resistance becomes =10 × (1 + 0.004 × (28 - 25)) = 10.12Ω. Since resistance changes affect the solenoid valve's power consumption and operating performance, to compensate for this effect, the frequency and duty cycle of the PWM signal are appropriately adjusted based on a pre-established table of resistance and PWM parameter relationships to ensure energy-saving control under varying environmental conditions.

[0056] Real-time monitoring of the back electromotive force generated by the solenoid valve through a high-precision voltage sensor , the collected back electromotive force signal is transmitted to the signal analysis module. Assume that the amplitude of the back electromotive force is , the frequency is , Fast Fourier Transform (FFT) is used to perform spectrum analysis on the back EMF signal. In actual operation, the back EMF signal collected within a period of time (such as 1s) is transformed by FFT, and the time domain signal is converted into a frequency domain signal to obtain the spectrum of the back EMF, so as to clearly analyze its amplitude and frequency components. For example, after FFT analysis, the amplitude of the back EMF is obtained. =50V, frequency =100Hz.

[0057] Design a special energy recovery circuit, using bidirectional DC-DC converter and super capacitor and other energy storage components. Bidirectional DC-DC converter can realize bidirectional energy transmission according to the size of back electromotive force and the state of energy storage components. Its efficiency It can be expressed as: ; in, is the input power, is the output power. In the design, select a high-efficiency bidirectional DC-DC converter, such as a model with an efficiency of up to 95%. Supercapacitors, as energy storage components, offer advantages such as fast charge and discharge speeds and a long lifespan. Based on the solenoid valve's operating characteristics and the magnitude of its back EMF, select a supercapacitor of appropriate capacity, such as a 10F supercapacitor. By properly designing circuit parameters, the energy recovery circuit can efficiently recover and store the back EMF energy in the supercapacitor.

[0058] Develop an energy recovery algorithm to adjust the working mode and parameters of the energy recovery circuit in real time according to the characteristics of the back electromotive force and the state of the energy storage element. Greater than the set threshold =40V, the energy recovery circuit is activated, controlling the bidirectional DC-DC converter to convert the back EMF energy into the appropriate voltage and current, storing it in the supercapacitor. Simultaneously, the energy recovery rate is adjusted based on the supercapacitor's charge level to avoid overcharging or over-discharging. When the supercapacitor's charge reaches 80%, the energy recovery rate is appropriately reduced to ensure the stability and efficiency of the energy recovery process.

[0059] Establish an energy management system to monitor the energy requirements of each component in real time through sensors , such as the power requirements of components such as solenoid valves, sensors, and controllers, as well as the remaining power of energy storage elements (supercapacitors) Assuming a total of Parts, The energy requirement of each component is , then the energy distribution formula is: ; For example, the energy requirements of the five components are =20W, =15W, =10W, =25W, W, remaining capacity of supercapacitor =1000 , then the energy allocated to the first component is: .

[0060] A collaborative power supply mode is designed to prioritize recycled energy for powering other low-power components, such as sensors and controllers. When recycled energy becomes insufficient, the system automatically switches to an external power source to ensure the proper operation of the jacquard. For example, at a certain point, the recycled energy can meet the energy needs of the sensors and controller, which are then powered by the supercapacitor. When the jacquard enters a high-load operating state and the recycled energy is insufficient to supply all components, the system automatically switches to an external power source while continuing to recycle back EMF energy for subsequent use.

[0061] Design an intelligent monitoring system for the energy recovery process. Use sensors to monitor the working status of the energy recovery circuit (such as the working current and voltage of the bidirectional DC-DC converter), the power of the energy storage element, the energy recovery efficiency and other parameters in real time. These parameters are transmitted to the monitoring terminal through the data transmission network and displayed intuitively in the form of charts on the monitoring interface, allowing operators to understand the energy recovery status in real time. Use fuzzy control algorithms to optimize the energy recovery process in real time based on the monitoring data. The input of fuzzy control is the energy recovery efficiency. and the charge of the energy storage element , the output is the control parameter of the energy recovery circuit , such as the operating frequency and duty cycle of a bidirectional DC-DC converter. Pre-defined fuzzy rules allow, for example, to increase the energy recovery rate when the energy recovery efficiency is high and the energy storage element's charge is low, and to decrease it when the energy recovery efficiency is low and the energy storage element's charge is high. Through fuzzy reasoning and defuzzification, specific control parameter values ​​are derived to optimize the energy recovery process.

[0062] According to the working scenarios and task requirements of the jacquard machine, the working modes are divided into normal working mode, standby mode, energy-saving mode, etc. In normal working mode, the jacquard machine operates according to the conventional weaving process and speed, and the solenoid valve moves frequently according to the pattern requirements; in standby mode, the jacquard machine temporarily stops weaving, but keeps some systems powered on, waiting for new task instructions; in energy-saving mode, the jacquard machine takes a series of energy-saving measures to reduce energy consumption while meeting basic operating requirements. Using cluster analysis algorithms (such as K-means clustering), the current working mode of the jacquard machine is identified in real time based on the operating data and working condition information collected by the sensor. The goal of K-means clustering is to minimize the sum of the distances from each sample to the center of the cluster to which it belongs: ; in: is the number of clusters (for working pattern recognition, =3, corresponding to normal working mode, standby mode, and energy-saving mode respectively). It is clusters, It is The center of each cluster is determined. In practical applications, a large amount of operating data (such as current, voltage, and weaving speed) under different operating modes is collected as samples. Through multiple iterative calculations, the center of each cluster is determined. For example, after cluster analysis, when the characteristic data combination of current between 1-3A and weaving speed between 20-30m / min is determined to be normal operating mode, while when the characteristic data combination of current between 0.1-0.3A and weaving speed of 0m / min is determined to be standby mode.

[0063] Develop energy-saving strategies tailored to different operating modes. In standby mode, reduce the solenoid valve's drive voltage and frequency to minimize energy consumption. For example, reduce the drive voltage from 12V to 5V during normal operation, and the frequency from 1000Hz to 100Hz. In energy-saving mode, further optimize PWM signal parameters to improve energy recovery efficiency. For example, adjust the PWM signal's duty cycle to a more appropriate value, ensuring the solenoid valve meets operating requirements while consuming less energy.

[0064] A working mode switching mechanism has been designed to quickly and smoothly switch to the appropriate working mode when the jacquard machine's operating state changes, automatically adjusting control parameters to ensure energy savings and weaving quality. For example, if the jacquard machine has not received any new weaving task instructions for a period of time, it automatically switches from normal working mode to standby mode and adjusts relevant control parameters. If it detects that the jacquard machine is entering a low-load operating scenario with low weaving speed requirements, it automatically switches to energy-saving mode and optimizes control parameters to achieve energy-saving operation.

[0065] A user interface has been designed to provide energy-saving option settings, allowing users to customize energy-saving parameters and operating modes based on their needs. The interface uses a graphical design to intuitively display various energy-saving options, such as setting the drive voltage and frequency adjustment range in standby mode and the parameter adjustment range of the PWM signal in energy-saving mode. Users can set parameters using interactive elements such as sliders and input boxes.

[0066] The user's customized settings are stored in the system's configuration file. During the operation of the jacquard machine, the control system reads the configuration file and automatically adjusts the control strategy based on the user's settings to meet the user's personalized energy-saving needs. For example, if the user sets the drive voltage to 4V and the frequency to 80Hz in standby mode, the system will automatically adjust the drive voltage and frequency to the user-set values ​​when it detects that the machine has entered standby mode.

[0067] Establish a comprehensive energy-saving effect evaluation indicator system, covering indicators such as energy consumption reduction rate, energy recovery efficiency improvement value, and energy consumption comparison under different working modes.

[0068] The formula for calculating the energy consumption reduction rate is: (energy consumption per unit time before optimization - energy consumption per unit time after optimization) ÷ energy consumption per unit time before optimization × 100%.

[0069] For example, the energy consumption per unit time of the machine before optimization is 500W, and after optimization it is 400W. The energy consumption reduction rate is (500-400) ÷ 500 × 100% = 20%.

[0070] The improvement in energy recovery efficiency is obtained by comparing the recovery and conversion efficiency of the back electromotive force energy of the energy recovery circuit before and after optimization.

[0071] Compare energy consumption under different working modes, and collect energy consumption data of normal working mode, standby mode, and energy-saving mode within the same operating time for horizontal comparative analysis.

[0072] During the actual operation of the jacquard machine, sensors installed in key locations continuously monitor energy consumption, energy recovery, and operating mode switching data. Data collection is set to once per minute to ensure that even subtle changes in energy consumption are captured. The collected data is stored in a dedicated database and analyzed in depth using data analysis tools (such as the Python data analysis library Pandas and the data visualization library Matplotlib). For example, by plotting energy consumption over time, fluctuations in energy consumption can be observed across different time periods and operating modes. The trend of energy recovery efficiency as it changes with the knitting task can be analyzed to identify the key factors affecting energy recovery efficiency.

[0073] Based on the results of long-term monitoring and data analysis, the innovative scheme for energy-saving control enhancement is continuously optimized and iterated. If it is found that the energy consumption of the energy-saving mode is still high under a specific weaving process, the working characteristics of the solenoid valve under this process will be further studied, the PWM signal parameter optimization algorithm will be adjusted, and the neural network load model will be retrained to improve the energy-saving effect. In view of the problem of low energy recovery efficiency under certain working conditions during the energy recovery process, the parameter design of the energy recovery circuit will be improved, and the threshold setting and control logic in the energy recovery algorithm will be optimized. The interface layout and function settings of user-defined energy-saving options will be updated regularly, and new energy-saving parameter settings will be added based on user feedback to enhance the user experience. At the same time, with the development of jacquard machine technology and the application of new materials and components, these new elements will be promptly incorporated into the energy-saving control scheme to maintain the advancement and adaptability of the scheme.

[0074] Targeted adjustments and adaptations are made to innovative solutions for different types of jacquard looms (such as high-speed jacquard looms, large jacquard looms, and jacquard looms for specialty fabric weaving). For high-speed jacquard looms, the focus is on testing and optimizing the real-time performance and response speed of the control algorithm to ensure precise control and energy savings even under high-frequency solenoid valve operation. For example, by increasing the sensor data acquisition frequency to 50 times per second and optimizing the computing resource allocation of the deep reinforcement learning algorithm, it can quickly process rapidly changing operating data.

[0075] For large jacquard looms, given their complex mechanical structures and higher energy consumption requirements, the energy recovery and redistribution system was expanded and optimized. The supercapacitor capacity was increased to store more recovered energy, and the energy distribution strategy was optimized to ensure that all components received the appropriate energy supply under different operating conditions. On jacquard looms used for specialty fabric weaving, such as those with special textures or functions, the adaptability of innovative solutions to different weaving process requirements was tested. Fuzzy control rules and PID parameters were adjusted to meet the special requirements for solenoid valve operation accuracy and stability during the specialty fabric weaving process.

[0076] The embodiments disclosed in the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. The embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part, and / or installed from a removable medium. When the computer program is executed by the central processing unit (CPU), the above-mentioned functions defined in the method of the present application are executed. It should be noted that the computer-readable medium mentioned above in the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, but not limited to, a system, device or component of an electrical, magnetic, optical, electromagnetic, infrared segment, or semiconductor, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more wire segments, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take a variety of forms, including, but not limited to, an electromagnetic signal, an optical signal, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transfer a program for use by or in conjunction with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, electrical wire, optical fiber cable, RF, etc., or any suitable combination thereof.

[0077] The flow charts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, program segment, or a part of code, and the module, program segment, or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0078] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The objectives of the present invention have been fully and effectively achieved. The functional and structural principles of the present invention have been demonstrated and explained in the embodiments. Without departing from the principles, the implementation methods of the present invention may be subject to any deformation or modification.

Claims

1. A solenoid valve control method for a jacquard machine, characterized in that: The method comprises: The solenoid valve status information and fabric weaving status information are collected, and after preprocessing, a comprehensive feature vector is obtained through weighted summation and fusion; Based on the comprehensive feature vector, the reward function is designed by combining the solenoid valve action time error, energy consumption and fabric defect rate. The experience recycling mechanism and target network technology are used to train the intelligent agent. Collect data from multiple jacquard machine models to build a dataset, pre-train the CNN model and migrate it to the target machine model for fine-tuning, and establish control parameter adjustment rules based on working condition identification; Build a consortium chain network, use the PBFT algorithm to reach transaction consensus, and combine edge computing to achieve collaborative control between nodes; Establish a mathematical model for high-speed switching devices, design a priority scheduling algorithm to optimize the transmission sequence and time interval of the drive signal, and run a PID control algorithm to adjust the drive signal in real time based on the deviation between the feedback signal and the set value; Integrate load sensors to build load models, adjust PWM signals based on environmental parameters, and design energy recovery and energy-saving mechanisms; Establish an energy-saving effect evaluation index system, monitor data and conduct in-depth analysis, optimize solutions based on the data and adapt to different types of jacquard machines.

2. The solenoid valve control method of a jacquard machine according to claim 1, characterized in that: The process of obtaining the comprehensive feature vector includes the following steps: The operating data of the solenoid valve and related components and the fabric weaving status information are collected; the collected multimodal data are cleaned, filtered and normalized; the image data is feature extracted through CNN; the time series features are extracted through LSTM; and the comprehensive feature vector is obtained by feature fusion through weighted summation.

3. The solenoid valve control method of a jacquard machine according to claim 2, characterized in that: The agent training process is as follows: Based on the comprehensive feature vector, the reward function is designed in combination with the solenoid valve action time error, energy consumption, and fabric defect rate. The experience replay mechanism and target network technology are used, the mean square error is used as the DQN loss function, and the gradient descent method is used to update the main network parameters. The reward function formula is as follows: ； in: is the reward function; 、 、 is the weighting coefficient, and ; is the reward term related to the action time error; It is a reward item related to energy consumption; It is a bonus item related to the fabric defect rate.

4. The solenoid valve control method of a jacquard machine according to claim 1, characterized in that: The migration process includes the following: Collect solenoid valve control data of multiple models of jacquard machines under different weaving processes and working conditions to build a data set; Using CNN as the pre-training model and after sufficient training with historical data, the model parameters are transferred to the target model; based on the actual operating data of the target model, the stochastic gradient descent algorithm is used for fine-tuning; the weaving speed, fabric material, ambient temperature and humidity working condition information are monitored in real time through sensors, and the support vector machine (SVM) is used to perform real-time analysis of the working conditions; according to different working condition categories, corresponding control parameter adjustment rules are established to achieve adaptive optimization of the solenoid valve control of the target model.

5. The solenoid valve control method of a jacquard machine according to claim 4, characterized in that: The specific content of the collaborative control is as follows: Build a consortium chain network consisting of multiple jacquard machine nodes, use asymmetric encryption technology, establish a node access mechanism, and ensure network and stable operation. Use the PBFT algorithm to elect the master node through view changes to complete the transaction consensus process. Write smart contracts to clarify data sharing rules, collaborative operation procedures, and fault handling mechanisms between nodes. Edge computing devices are deployed at each jacquard machine node to pre-process the data collected by local sensors and execute control algorithms. The edge computing devices interact with the processed data with the blockchain network to achieve collaborative control between nodes and improve the overall operating efficiency and coordination of the jacquard machine system.

6. The solenoid valve control method of a jacquard machine according to claim 1, characterized in that: The optimization process of the driving signal is as follows: Establish a mathematical model of high-speed switching devices; design a priority scheduling algorithm based on the weaving requirements of the jacquard machine and the working status of the solenoid valve to optimize the transmission order and time interval of the drive signal; study the working principles and performance characteristics of different circuit topologies, use software algorithms to simulate their functions, and dynamically adjust the circuit topology according to the operating status and load changes of the jacquard machine; establish a solenoid valve voltage-current-action characteristic model, adopt a fuzzy control algorithm, and adjust the drive signal according to the difference between the actual current and the expected current and the error change rate; collect the solenoid valve action feedback signal, filter and amplify it, and then use the PID control algorithm to adjust the drive signal in real time according to the deviation between the feedback signal and the set value.

7. The solenoid valve control method of a jacquard machine according to claim 6, characterized in that: The design process of the energy recovery and energy saving related mechanisms is as follows: Integrated load sensors collect current, pressure, and flow data, and use neural networks to build a load model; based on the load power and weaving requirements predicted by the load model, the PWM signal parameters are adjusted in real time and optimized using genetic algorithms; based on changes in ambient temperature and humidity parameters, compensation algorithms are designed to achieve energy-saving control; high-precision voltage sensors are used to monitor the back electromotive force generated by the solenoid valve in real time, energy recovery circuits and algorithms are designed, an energy management system is established, and a collaborative power supply mode and intelligent monitoring optimization system are designed; based on the working scenarios and task requirements of the jacquard machine, working modes are divided, and cluster analysis algorithms are used to identify working modes in real time, energy-saving strategies are formulated for different modes, a working mode switching mechanism is designed, and user-defined energy-saving options are provided.

8. The solenoid valve control method of a jacquard machine according to claim 7, characterized in that: The construction process of the energy-saving effect evaluation index system is as follows: Through sensors installed in various key locations, energy consumption data, energy recovery data, and working mode switching data are monitored continuously over a long period of time. Data analysis tools are used to conduct in-depth analysis of the monitoring data. Based on the results of long-term monitoring and data analysis, innovative solutions for energy-saving control enhancement are continuously optimized and iterated. Targeted adjustments and adaptations are made to different types of jacquard machines based on their characteristics to continuously improve the advancement and adaptability of the solutions.

9. A solenoid valve control system for a jacquard machine, characterized in that: The system is used to execute the solenoid valve control method of a jacquard machine according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the solenoid valve control method for a jacquard machine according to any one of claims 1 to 8.

Images