Temperature regulation and control system

Abstract

The invention, which relates to the technical field of temperature control, discloses a temperature regulation and control system comprising a target making module, a comprehensive monitoring module, a central decision module, an optimization control module and a temperature control execution module. The target formulating module generates a real-time target temperature table by considering the environmental parameters based on the environmental correction strategy; the comprehensive monitoring module obtains and feeds back a real-time monitoring parameter table; the central decision-making module considers the environmental parameter change and the real-time temperature change of each key component based on a progressive change perception strategy to decide whether to update a target thermometer or carry out temperature regulation and control; the optimization control module adopts a grey wolf optimization fuzzy PID algorithm to optimize a control factor, and obtains a real-time optimal temperature control power adjustment scale; and the temperature control execution module starts temperature control equipment through a conventional temperature control power meter, receives an adjustment execution instruction and updates real-time temperature control power, so that multi-component cooperative temperature regulation and control which dynamically cope with the environment and adjust the total power to be minimum is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of temperature control, and particularly to a temperature regulation system. Background Art

[0002] In industrial production, temperature regulation systems are widely used in various production processes, such as metallurgy, chemical industry, electric power, food processing, pharmaceutical and electronics industries, etc. In these processes, temperature changes will affect the operation effect and service life of production equipment. Therefore, precise temperature control is crucial for ensuring product quality and improving production efficiency.

[0003] A patent application with the publication number CN119087788A proposes a furnace tube Heater heating control system based on the PID control algorithm, including a temperature sensor for real-time collecting the temperature inside the furnace tube; a PID controller for receiving the temperature data of the temperature sensor and calculating according to the deviation between the set temperature and the measured temperature, and outputting a control instruction; an adjustment device for receiving the control instruction of the PID controller and adjusting the power of the heating element to achieve precise control of the heating temperature of the furnace tube Heater and reduce temperature fluctuations.

[0004] However, the existing temperature control systems have the following defects:

[0005] 1. Due to the different structures and functions of equipment, the corresponding temperature control requirements are also different. For example, a semi-solid die casting machine needs to control the temperature of three components, namely the alloy melting furnace, the pressure chamber and the mold, simultaneously, while a motor usually only needs overall temperature control. The existing temperature control systems generally conduct targeted temperature control on the whole or specific components of the controlled equipment, with low versatility, and do not consider the association between the whole and the local parts of the equipment for coordinated temperature control;

[0006] 2. The existing temperature control technologies do not consider the influence of external environmental factors on the heat exchange of equipment, and simply conduct temperature control through real-time monitoring combined with prediction algorithms or control algorithms, which may lead to frequent temperature control adjustments of the equipment. For example, when the external temperature is low and the humidity is high, the heat exchange between some metal equipment and the outside increases, resulting in faster heat loss, and it is necessary to frequently adjust the heating power to keep the equipment at a stable temperature;

[0007] 3. When conducting temperature control, the existing temperature control technologies rarely consider the cost and / or efficiency issues, and only obtain the operating power of the temperature control equipment through the control algorithm to achieve the constant temperature of the production equipment, but ignore whether there is room for optimization in the temperature control power and / or temperature control efficiency. Summary of the Invention

[0008] In view of the deficiencies of the existing technology, the present invention proposes a temperature regulation system to provide a coordinated temperature regulation system that can dynamically respond to external environmental changes and consider the temperature requirements of different key components of the controlled equipment.

[0009] The technical solution adopted to achieve the object of the present invention is as follows:

[0010] A temperature control system includes a target setting module, a comprehensive monitoring module, a central decision-making module, an optimization control module, and a temperature control execution module;

[0011] The target setting module stores the optimal temperature table and the conventional parameter table , receives the adjustment instruction at , generates the target temperature table at based on the real-time environmental parameters considering the environmental correction strategy and feeds it back to the central decision-making module;

[0012] The comprehensive monitoring module receives the monitoring instruction at , obtains the monitoring parameter table at and feeds it back to the central decision-making module;

[0013] The central decision-making module sets the start time and sets subsequent times according to the monitoring interval , sends the monitoring instruction to the comprehensive monitoring module at each time, receives and stores the monitoring parameter table at , decides whether to send the instruction at considering the environmental parameter change and the real-time temperature change of several key components based on the progressive change perception strategy;

[0014] The optimization control module receives the control instruction at , optimizes the control factor by using the grey wolf optimization fuzzy PID algorithm, and obtains the optimal temperature control power adjustment scale table at , packs it with the adjustment execution flag to generate the adjustment execution instruction and sends it to the temperature control execution module;

[0015] The temperature control execution module receives the execution instruction and operates the temperature control equipment with the conventional temperature control power table , receives the adjustment execution instruction at , updates the temperature control power of each temperature control equipment at to the temperature control power at plus the optimal temperature control power adjustment amount at at

[0016] Further, the optimal temperature table ,in, The first The optimal temperature of each key component, general parameter table ,in, and are normal ambient temperature and normal ambient humidity, respectively. For the The temperature control equipment corresponding to each key component is at normal ambient temperature and normal ambient humidity Next Key components reach optimal temperature The required conventional temperature control power, , The total number of key components of the controlled equipment.

[0017] Furthermore, the target setting module receives Time adjustment instructions , based on the environment correction strategy generation Target temperature table at each moment , including the following specific steps:

[0018] from Time adjustment instructions Get real-time ambient temperature changes And real-time environmental humidity changes And synchronize the input environment correction network group, output Real-time fitting temperature change table at each moment ;

[0019] From the optimal temperature table Extract the Optimal temperature for key components , set the The key components in Target temperature at the moment For the best temperature minus Real-time temperature change at any time ;

[0020] statistics The key components in The target temperature at the moment, get Target temperature table at each moment .

[0021] Furthermore, the environmental correction network group includes The environment correction network with the same structure and requiring pre-training, among which An environmental correction network fits the temperature change of the th key component with respect to the environmental temperature change and the environmental humidity change nonlinear mapping function. , is the total number of environmental correction networks, which is equal in quantity to

[0022] Furthermore, the pre-training of the th environmental correction network includes the following specific steps:

[0023] Fix the conventional temperature control power of the temperature control device corresponding to the th key component , and measure and obtain the different temperatures of the th key component at different environmental temperatures and environmental humidities ;

[0024] Calculate the temperature change corresponding to each pair of environmental temperature changes and environmental humidity changes ;

[0025] Use a number of environmental temperature changes , a number of environmental humidity changes and the corresponding number of temperature changes to construct a data set and divide it into a training set and a test set;

[0026] Select an environmental correction network, define the environmental temperature change and the environmental humidity change as the input, and the temperature change as the output;

[0027] Define the loss function as the mean square error, and use the Adam optimizer to perform multiple rounds of training to adjust the hyperparameters of the th environmental correction network;

[0028] After the Adam optimizer training is completed, fix the hyperparameters of the th environmental correction network, and input the environmental temperature change and the environmental humidity change in the test set into the th environmental correction network to obtain the mean square error corresponding to the test set and determine whether it is less than or equal to the error threshold;

[0029] If it is less than or equal to the error threshold, the pre-training of the th environmental correction network is completed. If it is greater than the error threshold, the Adam optimizer is adjusted to re-train.

[0030] Furthermore, the central decision-making module receives the monitoring parameter table at and decides whether to send the instruction at

[0031] based on the progressive change perception strategy, including the following specific steps: Obtain the monitoring parameter table at and extract the real-time environmental temperature and the real-time environmental humidity , call the general parameter table and extract the general environmental temperature and the general environmental humidity ;

[0032] Based on whether it is the start time at decide to compare the environmental temperature and compare the environmental humidity select the general environmental temperature and the general environmental humidity or the real-time environmental temperature at and the real-time environmental humidity ,

[0033] Calculate the environmental temperature change gradient at and the environmental humidity change gradient and judge whether both are less than or equal to the change gradient threshold;

[0034] If the environmental temperature change gradient at and the environmental humidity change gradient are both less than or equal to the change gradient threshold, judge whether it is the start time at

[0035] If it is the start time, use the optimal temperature table as the target temperature table at store, generate an execution mark , extract from the general parameter table the general temperature control power corresponding to key components to form a general temperature control power table , further package the execution mark And the conventional temperature control power meter Generate The execution instruction at the moment And send it to the temperature control execution module;

[0036] If it is not the startup moment, then The target temperature table at the moment As The target temperature table at the moment Store, from The monitoring parameter table at the moment Extract the real-time temperature table ;

[0037] Calculate respectively The temperature change gradients of the key components at The moment , and judge whether they are all less than or equal to the change gradient threshold;

[0038] If The temperature change gradients of the key components at The moment Are all less than or equal to the change gradient threshold, no operation is required;

[0039] If there is a single component whose temperature change gradient at The moment is greater than the change gradient threshold, generate a control flag , and with The target temperature table at the moment Pack and generate a control instruction And send it to the optimization control module;

[0040] If The ambient temperature change gradient at the moment And / or the ambient humidity change gradient Is greater than the change gradient threshold, calculate The real-time ambient temperature change amount at the moment And the real-time ambient humidity change amount , and with the adjustment flag Pack and generate an adjustment instruction And send it to the target formulation module, receive and store The target temperature table at the moment , and with the control flag Pack and generate The control instruction at the moment And send it to the optimization control module.

[0041] Furthermore, the optimization control module uses the grey wolf optimization fuzzy PID algorithm to obtain The optimal temperature control power adjustment scale table at the moment , including the following specific steps:

[0042] Define the solution space And initialize The position of the head gray wolf and set the maximum number of optimization rounds , define the fitness function ;

[0043] Execute a single-round optimization. Substitute the position of the Head gray wolf into the fuzzy PID algorithm and calculate The fitness function at the moment , select the Fitness function at the moment in the first-round optimization The three gray wolves with the smallest fitness are respectively used as Wolf, Wolf and Wolf, Wolf, Wolf and The positions of the wolves are respectively denoted as , And ;

[0044] Randomly generate the first random number and the second random number for the first-round optimization within . Calculate the first coefficient in the first-round optimization based on the number of optimization rounds and the first random number, and calculate the second coefficient in the first-round optimization based on the second random number;

[0045] Based on the second coefficient in the first-round optimization and the Wolf position , Wolf position And Wolf position Calculate the distances of the remaining Head gray wolves with respect to Wolf, Wolf and Wolf;

[0046] Further, based on the first coefficient in the first-round optimization and the distance of each gray wolf with respect to Wolf, Wolf and Wolf, obtain the updated position of each gray wolf;

[0047] Repeat the single-round optimization. In each round of optimization, re-select Wolf, Wolf and Wolf and make the remaining gray wolves move closer to Wolf, Wolf and Wolf until the number of optimization rounds is equal to the maximum number of optimization rounds , substitute the position of the wolf in the nth round of optimization into the optimal temperature control power adjustment scale at the moment output by the fuzzy PID algorithm. .

[0048] Furthermore, substitute the position of a single grey wolf into the fuzzy PID algorithm and calculate the fitness function, including the following specific steps:

[0049] Extract the target temperature table from the control instruction at the moment, call the monitoring parameter table at the moment and the and monitoring parameter table at the moment, and respectively extract the real-time temperature table at the and real-time temperature table at the ;

[0050] Subtract the target temperature table from the real-time temperature table to obtain the real-time target temperature difference table at the moment. Subtract the real-time temperature table at the from the real-time temperature table at the moment and divide by the monitoring interval to obtain the real-time temperature change rate table at the moment;

[0051] Multiply the real-time target temperature difference and real-time temperature change rate of the key components at the moment by the corresponding first quantization factor and second quantization factor respectively as the input of the fuzzy controller. The output of the fuzzy controller is the proportional factor table , integral factor table and differential factor table ;

[0052] Select the Gaussian membership function as the membership function of the fuzzy controller, define fuzzy subsets for the input and output of the fuzzy controller and establish a fuzzy control rule table;

[0053] Select the centroid method as the defuzzification method to obtain the proportional factor table , integral factor table ​And differential factor table , obtain the temperature control power adjustment amount and control rounds corresponding to each key component at moment through the PID algorithm;

[0054] Sum respectively the temperature control power adjustment amounts and control rounds of key components to obtain the total temperature control power adjustment amount at moment and the total control rounds , and further accumulate and sum to obtain the fitness function at .

[0055] Compared with the prior art, the significant advantages of the present invention are as follows:

[0056] 1. Based on the synergistic effect of the environmental correction strategy and the progressive change perception strategy, considering the change of real-time environmental parameters relative to the pre-stored conventional environmental parameters, dynamically generate a real-time target temperature table through the environmental correction network group. The target temperature table is always the working temperature of the device or key component that is most adaptable to the real-time external environment; and at the system startup moment and non-startup moment, judge in real time whether the environmental parameters change and whether the real-time temperatures of all key components change. When the environmental parameters change, update the real-time target temperature table and regulate the temperature control device. When the environmental parameters do not change but the temperatures of the key components change, perform temperature regulation to ensure that the device or key component always maintains the target temperature, and can achieve high-efficiency temperature regulation adapting to the change of external environmental factors;

[0057] 2. Design the grey wolf optimization fuzzy PID algorithm, introduce grey wolf optimization on the basis of the fuzzy PID algorithm. The grey wolf optimization searches for the optimal control factor based on the control rounds and the total temperature control power adjustment amount fed back by the fuzzy PID algorithm, and obtains the optimal temperature control power adjustment table with the highest real-time efficiency and the smallest total temperature control power adjustment amount of each key component, realizing the coordinated temperature control of the whole and local parts of the controlled device, and reducing the temperature control cost and improving the temperature control efficiency. Description of the Drawings

[0058] Figure 1 is a schematic diagram of a temperature regulation system of the present invention;

[0059] Figure 2 is a flow chart of the progressive change perception strategy in the present invention;

[0060] Figure 3 is a flow chart of the grey wolf optimization fuzzy PID algorithm in the present invention;

[0061] Figure 4 is the fuzzy control rule table proposed in the present invention. Detailed Embodiment

[0062] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments.

[0063] Embodiment 1

[0064] As Figure 1 shown, a specific embodiment of the present invention discloses a temperature control system, including a target setting module, a comprehensive monitoring module, a central decision-making module, an optimization control module, and a temperature control execution module;

[0065] The target setting module stores the optimal temperature table and the conventional parameter table , receives the adjustment instruction at , generates the target temperature table at based on the real-time environmental parameters at considering the environmental correction strategy and feeds it back to the central decision-making module;

[0066] The comprehensive monitoring module receives the monitoring instruction at , obtains the monitoring parameter table at in real time through multiple sensors and feeds it back to the central decision-making module, where and are respectively the real-time environmental temperature and real-time environmental humidity at , are respectively the real-time temperatures of key components 1 to of the controlled device, is the total number of key components of the controlled device, and the number of sensors is equal to the total number of key components plus 2. The key components specifically refer to the components whose real-time temperatures can have a greater impact on the operation of the controlled device, which are determined by those skilled in the art;

[0067] The central decision-making module sets the start time at system startup and periodically sets subsequent times according to the monitoring interval . At each time, it sends the monitoring instruction to the comprehensive monitoring module, receives and stores the monitoring parameter table at , and decides whether to send the instruction at based on the progressive change perception strategy by comprehensively considering the changes in environmental parameters and the real-time temperature changes of key components of the controlled device. The instruction includes the adjustment instruction at , the execution instruction and the control instruction ;

[0068] The optimization control module receives the control instruction at moment, and uses the grey wolf optimization fuzzy PID algorithm to consider the control round and the total amount of temperature control power adjustment at moment to optimize the control factors of the fuzzy PID algorithm, and obtain the control round at moment and the total amount of temperature control power adjustment, and generate an optimal temperature control power adjustment scale table with both being the smallest , generate an adjustment execution mark for uniquely identifying the adjustment execution instruction and package it with the optimal temperature control power adjustment scale table to obtain the adjustment execution instruction and send it to the temperature control execution module. Among them, is the optimal temperature control power adjustment amount of the temperature control device corresponding to the th key component at moment, , is the total number of key components of the controlled device;

[0069] The temperature control execution module includes multiple temperature control devices, and the total number of temperature control devices is the same as the total number of key components , and respectively performs temperature regulation on the key components of the controlled device, receives the execution instruction at the start moment and operates the temperature control device with the conventional temperature control power table , receives the adjustment execution instruction at and respectively extracts from it the optimal temperature control power adjustment amount of the temperature control device corresponding to the key components at moment, so that the temperature control power of each temperature control device at moment is updated to the temperature control power at moment plus the optimal temperature control power adjustment amount at

[0070] Furthermore, the best temperature table , where is the best temperature of the th key component of the controlled device, and the total number of key components is determined based on the controlled device applied in the temperature regulation system. The best temperature table records the best working temperatures of each key component of the controlled device. Taking the semi-solid die casting machine as an example, the semi-solid die casting machine includes a total of 3 key components, that is, the total number of key components , the three key components are an alloy furnace, a shot chamber, and a mold, respectively. Then, the optimal temperature table records the optimal furnace temperature, shot chamber temperature, and mold temperature when the semi-solid die casting machine processes specific raw materials. For the semi-solid die casting machine, the optimal temperature table for the three key components will also change with the change of raw materials.

[0071] Furthermore, the conventional parameter table , where and are the conventional ambient temperature and conventional ambient humidity, respectively, specifically referring to the ambient temperature and ambient humidity corresponding to the working environment of the controlled equipment under normal conditions. Then, is the conventional temperature control power required for the temperature control equipment corresponding to the th key component to make the th key component of the controlled equipment reach the optimal temperature under the conditions of conventional ambient temperature and conventional ambient humidity. The conventional temperature control power includes conventional heating power and conventional cooling power, which is determined by the key components of the controlled equipment. Taking the semi-solid die casting machine as an example, the three key components of the semi-solid die casting machine all need to be heated. The alloy furnace melts the raw materials into soup through heating, the shot chamber ensures that the soup can maintain a semi-solid state through heating, and the mold prevents the soup from solidifying prematurely and affecting the final quality of the casting. Therefore, the three conventional temperature control powers corresponding to the three key components of the semi-solid die casting machine are all conventional heating powers. Furthermore, the target setting module receives

[0072] the adjustment instruction at time, and generates the target temperature table at time based on the environmental correction strategy, including the following specific steps:

[0073] Obtain from the adjustment instruction at time the real-time ambient temperature change amount and the real-time ambient humidity change amount . The real-time ambient temperature change amount is the real-time ambient temperature at time minus the conventional ambient temperature . The real-time ambient humidity change amount is the real-time ambient humidity at time minus the conventional ambient humidity . is the adjustment flag, which uniquely identifies the adjustment instruction;

[0074] The real-time environmental temperature change at a moment and the real-time environmental humidity change are synchronously input into the environmental correction network group and output the real-time fitting temperature change table at a moment , where is the real-time fitting temperature change of the th key component;

[0075] Extract the optimal temperature of the th key component from the optimal temperature table . To make up for the real-time fitting temperature change of the th key component , the target temperature of the th key component at the moment should be the optimal temperature minus the real-time temperature change at the moment , that is ;

[0076] Statistically analyze the target temperatures of the key components of the controlled device at the moment, and obtain the target temperature table at the moment.

[0077] Furthermore, the environmental correction network group includes environmental correction networks with the same structure. The environmental correction network can be any feedforward neural network with non-linear fitting ability and needs to be pre-trained. In this embodiment, a BP neural network is selected. The th environmental correction network fits the temperature change of the th key component under the conventional temperature control power with respect to the environmental temperature change and the environmental humidity change non-linear mapping function, , is the total number of environmental correction networks, is equal to in quantity. Taking the shot sleeve in a semi-solid die casting machine as an example, under the conventional environmental temperature and the conventional environmental humidity The pressure chamber is heated with a conventional temperature control power, which can maintain the optimal temperature of the pressure chamber. When the ambient temperature and humidity change, although the conventional temperature control power of the pressure chamber remains unchanged, the heat exchange between the pressure chamber and the outside world changes, causing the temperature of the pressure chamber to deviate from the optimal temperature. The ambient correction network is used to fit the curve of the pressure chamber temperature changing with the ambient temperature and humidity.

[0078] Furthermore, the pre-training of the th ambient correction network includes the following specific steps:

[0079] Determine the conventional temperature control power required for the temperature control device corresponding to the th key component to reach the optimal temperature under the conventional ambient temperature and conventional ambient humidity ;

[0080] Fix the conventional temperature control power , and measure and obtain the different temperatures of the th key component at different ambient temperatures and ambient humidities ;

[0081] Calculate the temperature change corresponding to each pair of ambient temperature change , ambient humidity change of the th key component

[0082] ; Use several ambient temperature changes , several ambient humidity changes and the corresponding several temperature changes of the th key component to construct a data set, and divide the training set and test set according to the

[0083] Select an ambient correction network, define the ambient temperature change and ambient humidity change as the input of the ambient correction network, and the temperature change as the output of the ambient correction network; define the loss function as the mean square error, that is, the error between the fitted temperature change output by the th ambient correction network and the temperature change in the training set. Use the Adam optimizer for multiple rounds of training to adjust the Hyperparameters of an environmental correction network, where the learning rate and number of training epochs of the Adam optimizer are initially set to 0.01 and 100;

[0084] After the Adam optimizer finishes training, fix the hyperparameters of the th environmental correction network, and use the environmental temperature change and environmental humidity change in the test set as the input to the th environmental correction network to obtain the mean squared error corresponding to the test set;

[0085] If the mean squared error corresponding to the test set is less than or equal to the error threshold, the pre-training of the th environmental correction network is completed. If the mean squared error corresponding to the test set is greater than the error threshold, adjust the learning rate and number of training epochs of the Adam optimizer and retrain.

[0086] As Figure 2 shown, further, the central decision-making module receives the monitoring parameter table at time and decides whether to send an instruction at time based on the progressive change perception strategy, including the following specific steps:

[0087] Obtain the monitoring parameter table at time and extract the real-time environmental temperature and real-time environmental humidity at time. Call the general parameter table stored in the target formulation module and extract the general environmental temperature and general environmental humidity from it; ;

[0088] Judge whether time is the start time. If time is the start time, use the general environmental temperature and general environmental humidity as the comparison environmental temperature and comparison environmental humidity respectively. If time is not the start time, use the real-time environmental temperature and real-time environmental humidity at time as the comparison environmental temperature and comparison environmental humidity respectively, is the monitoring interval;

[0089] Calculate the environmental temperature change gradient at time​ and the environmental humidity change gradient , where the environmental temperature change gradient indicates the temperature fluctuation range from the comparison environmental temperature to the real-time environmental temperature , and the environmental humidity change gradient indicates the humidity fluctuation range from the comparison environmental humidity to the real-time environmental humidity ;

[0090] Judge whether the environmental temperature change gradient and the environmental humidity change gradient at a certain moment are both less than or equal to the change gradient threshold;

[0091] If the environmental temperature change gradient and the environmental humidity change gradient at a certain moment are both less than or equal to the change gradient threshold, it indicates that the temperature from the comparison environmental temperature to the real-time environmental temperature has almost no fluctuation and the humidity from the comparison environmental humidity to the real-time environmental humidity has almost no fluctuation, that is, the environmental parameters have not changed significantly. Based on whether the moment is the start moment, determine the target temperature table at a certain moment and decide whether to generate an execution instruction and send it to the temperature control execution module, where is the target temperature of the th key component of the controlled device at a certain moment;

[0092] When a certain moment is the start moment, since the environmental parameters at a certain moment have no significant change compared with the conventional environmental parameters, directly call the best temperature table of the target formulation module and the conventional parameter table , and extract from the conventional parameter table the conventional temperature control power corresponding to key components to generate a conventional temperature control power table , where is the temperature control device corresponding to the th key component to make the th key component of the controlled device reach the best temperature under the conditions of conventional environmental temperature and conventional environmental humidity The required conventional temperature control power, and the optimal temperature table as the target temperature table at the moment and store it, generating an execution mark for uniquely identifying the execution instruction , and package it with the conventional temperature control power table to generate the execution instruction at the moment and send it to the temperature control execution module;

[0093] When the moment is a non-starting moment, since the environmental parameters at the moment have no significant change compared with the environmental parameters at the moment the target temperature table at the moment is directly used as the target temperature table at the moment and store it;

[0094] Extract from the monitoring parameter table at the moment the real-time temperature table at the moment , where are respectively the real-time temperatures of the 1st to the th key components of the controlled device;

[0095] Calculate respectively the temperature change gradients of the key components at the moment , where the temperature change gradient of the th key component of the controlled device at the moment is , is the real-time temperature of the th key component of the controlled device, and further judge whether there is a temperature change gradient of a single component greater than the change gradient threshold at the moment ;

[0096] If the temperature change gradients of the key components at the moment are all less than or equal to the change gradient threshold, that is, the temperatures of the key components of the controlled device from the moment to the moment practically have no change, then no operation needs to be performed, and the temperature control execution module continues to run with the real-time temperature control power table at the moment;

[0097] If Among the key components, if there is a single component whose temperature change gradient at is greater than the change gradient threshold, a control mark is generated to uniquely identify the control instruction , and it is packaged with the target temperature table at to generate the control instruction at and sent to the optimization control module;

[0098] If the environmental temperature change gradient at and / or the environmental humidity change gradient is greater than the change gradient threshold, it indicates that there are significant fluctuations in the temperature from the compared environmental temperature to the real-time environmental temperature and / or in the humidity from the compared environmental humidity to the real-time environmental humidity , that is, the environmental parameters have changed significantly. The target temperature table at needs to be readjusted and generated. Calculate the real-time environmental temperature change amount at and the real-time environmental humidity change amount , generate an adjustment mark , and it is packaged with the real-time environmental temperature change amount at and the real-time environmental humidity change amount to generate the adjustment instruction at and sent to the target formulation module. Receive and store the target temperature table at sent back by the target formulation module, generate a control mark , and it is packaged with the target temperature table at to generate the control instruction at and sent to the optimization control module.

[0099] Specifically, taking a semi-solid die casting machine as an example, the central decision-making module determines whether the moment is the starting moment when the temperature control system acts on the semi-solid die casting machine;

[0100] If it is the starting moment, based on the real-time environmental temperature at , the real-time environmental humidity and the conventional environmental temperature , the conventional environmental humidity to determine whether the environmental parameters have changed significantly;

[0101] If the environmental parameters do not change significantly, the conventional temperature control power meter is used to heat the alloy furnace, the pressure chamber and the mold respectively, so that the best furnace temperature, pressure chamber temperature and mold temperature can be achieved at all times;

[0102] If the environmental parameters change significantly, the target temperatures of the furnace, the pressure chamber and the mold at the moment need to be re-determined by the target setting module, and the re-determined target temperature table at the moment is informed to the optimization control module through the control instruction to adjust the heating power of the furnace, the pressure chamber and the mold at the moment;

[0103] If it is not the starting moment, based on the real-time environmental temperature at the moment and the real-time environmental humidity and the real-time environmental temperature at the moment and the real-time environmental humidity to determine whether the environmental parameters change significantly, which is the monitoring interval;

[0104] If the environmental parameters do not change, and the temperatures of the furnace, the pressure chamber and the mold at the moment must be maintained as the target temperature table , and further judge whether the furnace temperature, the pressure chamber temperature and the mold temperature change significantly based on the temperature change gradients of the furnace, the pressure chamber and the mold at the moment;

[0105] If the furnace temperature, the pressure chamber temperature and the mold temperature change significantly at the moment, the target temperature table at the moment is used as the target temperature table at the moment and is informed to the optimization control module through the control instruction to adjust the heating power of the furnace, the pressure chamber and the mold at the moment;

[0106] If the furnace temperature, the pressure chamber temperature and the mold temperature do not change significantly at the moment, it indicates that from the moment to the moment, the furnace, the pressure chamber and the mold are still operating stably at their respective target temperatures and no adjustment is required;

[0107] If the environmental parameters change, then the furnace, the pressure chamber and the mold at The target temperature at a moment needs to be re-determined by the target formulation module, and the new target temperature table is passed through the control instruction at the moment to inform the optimization control module to adjust the heating power of the furnace, the injection chamber, and the mold at the moment.

[0108] As Figure 3 shown, further, the optimization control module uses the grey wolf optimization fuzzy PID algorithm to obtain the optimal temperature control power adjustment scale table at the moment , including the following specific steps:

[0109] Define the solution space and initialize the positions of the grey wolf group in the solution space . The grey wolf group includes n heads of grey wolves. The position of each grey wolf represents a set of control factors of the fuzzy PID algorithm in the solution space . Among them, is the position of the th head of grey wolf in the grey wolf group, , n is the total number of grey wolves. The dimension of the solution space is equal to the total number of control factors of the fuzzy PID algorithm. The control factors of the fuzzy PID algorithm include the first quantization factors corresponding to m key components and the second quantization factors . The control factors determine the output of the fuzzy PID algorithm;

[0110] Set the maximum number of optimization rounds T, define the fitness function at the moment t, where t is the total number of control rounds at the moment, Δu is the total amount of temperature control power adjustment at the moment;

[0111] Execute a single-round optimization. At this time, the number of optimization rounds is 1. Substitute the positions of n heads of grey wolves into the fuzzy PID algorithm respectively and calculate the fitness function at the moment t, and select the three grey wolves with the smallest fitness function in the first-round optimization as α wolf, β wolf, and γ wolf in turn. The positions of the α wolf, β wolf, and γ wolf are respectively recorded as , and , at this time, the wolf, the wolf and the wolf can be regarded as the three gray wolves with the best positions in the first-round optimization;

[0112] Calculate the distances of the positions of the remaining gray wolves with respect to the wolf, the wolf and the wolf respectively, and make the remaining gray wolves move closer to the wolf, the wolf and the wolf to obtain the updated positions of the remaining gray wolves. Assuming that the th gray wolf is not the wolf, the wolf and the wolf, taking the position of the th gray wolf as an example, the updated position of the th gray wolf after the first-round optimization is calculated as follows:

[0113] ;

[0114] Among them, is the first coefficient in the first-round optimization, and the first coefficient in the th round of optimization , is the th random number generated within in the th round of optimization, , and are the distances between the positions of the th gray wolf before the first-round optimization and the positions of the wolf, the wolf, the wolf and the wolf respectively. The specific calculation formulas are as follows:

[0115] ;

[0116] ;

[0117] ;

[0118] Among them, is the second coefficient in the first-round optimization, and the The second coefficient in the round optimization , is the second random number generated within in the round optimization;

[0119] Repeatedly execute single-round optimization, and reselect wolves, wolves and wolves in each round of optimization, and order the remaining grey wolves to move closer to wolves, wolves and wolves until the number of optimization rounds is equal to the maximum number of optimization rounds . Then, take the position of the wolves in the round optimization as the optimal control factor of the fuzzy PID algorithm, and substitute it into the optimal temperature control power adjustment table at the moment. Among them, is the optimal temperature control power adjustment amount of the temperature control device corresponding to the th key component at the th moment.

[0120] Furthermore, substitute the position of a single grey wolf into the fuzzy PID algorithm and calculate the fitness function , including the following specific steps:

[0121] Extract the target temperature table at the moment from the received control instruction at the moment, call the monitoring parameter table at the moment and the monitoring parameter table at the moment stored in the central decision-making module, and extract the real-time temperature table at the moment and the real-time temperature table at the moment respectively. Subtract the target temperature table from the real-time temperature table to obtain the real-time target temperature difference table at the moment. Subtract the real-time temperature table at the moment from the real-time temperature table ​to obtain real-time temperature change rate table at a moment , where and are respectively the real-time target temperature difference and the real-time target temperature change rate of the th key component at a moment;

[0122] Multiply the real-time target temperature difference and the real-time temperature change rate of the key components at a moment by the corresponding first quantization factor and second quantization factor respectively and input them into a fuzzy controller. The fuzzy controller is the first half of the fuzzy PID algorithm, and the output of the fuzzy controller is the proportional factor table , integral factor table and differential factor table , where , and are respectively the proportional factor, integral factor and differential factor of the th key component at a moment. The PID algorithm is the second half of the fuzzy PID algorithm;

[0123] Select the membership function of the fuzzy controller as the Gaussian membership function and define fuzzy subsets for the input and output of the fuzzy controller , where , , , , , and represent larger negative number, moderate negative number, smaller negative number, 0, smaller positive number, moderate positive number and larger positive number respectively;

[0124] Establish a fuzzy control rule table based on the inference rules between the input and output of the fuzzy controller. The fuzzy control rule table is specifically as shown in Figure 4 . Taking the th key component as an example, assume that the real-time target temperature difference and the real-time target temperature change rate of the th key component at a moment both belong to . According to the fuzzy rule table, it can be known that the proportional factor , integral factor and differential factor of the th key component respectively belong to , and ;

[0125] The centroid method is selected as the defuzzification method to obtain the proportional factor table of the PID algorithm , integral factor table and differential factor table , and further obtain the temperature control power adjustment amount and control rounds corresponding to each key component at moment through the PID algorithm. Sum the temperature control power adjustment amounts and control rounds of key components respectively to obtain the total temperature control power adjustment amount at moment and the total control rounds , so as to obtain the fitness function at moment. Among them, the centroid method and the PID algorithm are existing methods and will not be elaborated too much.

[0126] The present invention discloses a temperature control system, including a target setting module, a comprehensive monitoring module, a central decision-making module, an optimization control module and a temperature control execution module; the target setting module generates a real-time target temperature table based on the environmental correction strategy considering the change of environmental parameters; the comprehensive monitoring module obtains and feeds back the real-time monitoring parameter table; the central decision-making module decides whether to update the target temperature table or perform temperature control based on the progressive change perception strategy considering the change of environmental parameters and the real-time temperature change of each key component; the optimization control module optimizes the control factor by using the grey wolf optimization fuzzy PID algorithm to obtain the real-time optimal temperature control power adjustment amount table; the temperature control execution module starts the temperature control device with the conventional temperature control power table, receives the adjustment execution instruction and updates the real-time temperature control power, realizing the multi-component collaborative temperature control that dynamically responds to the environment and has the minimum total adjustment power.

[0127] The above is only the preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions within the idea of the present invention belong to the protection scope of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A temperature control system, characterized in that, It includes a target setting module, a comprehensive monitoring module, a central decision-making module, and an optimization control module; The target setting module receives real-time adjustment instructions, generates a real-time target temperature table based on the environmental correction strategy and real-time environmental parameters, and feeds it back to the central decision-making module; The comprehensive monitoring module receives real-time monitoring instructions and feeds back a real-time monitoring parameter table to the central decision-making module; The central decision-making module periodically sends real-time monitoring instructions, stores the real-time monitoring parameter table, and decides whether to send real-time instructions based on the progressive change perception strategy considering environmental parameter changes and real-time temperature changes of key components; The optimization control module receives real-time control instructions, optimizes control factors using the grey wolf optimization fuzzy PID algorithm, obtains a real-time optimal temperature control power adjustment scale table, and generates an adjustment execution instruction.

2. The temperature control system according to claim 1, characterized in that, The obtaining of the real-time optimal temperature control power adjustment scale table includes the following specific steps: Define the solution space and initialize the positions of grey wolves, set the maximum number of optimization rounds, and define the fitness function; Execute single-round optimization, substitute the positions of each grey wolf into the fuzzy PID algorithm and calculate the fitness function, select the three grey wolves with the smallest fitness function and record their positions respectively; Randomly generate the first random number and the second random number for single-round optimization, calculate the first coefficient based on the number of optimization rounds and the first random number, and calculate the second coefficient based on the second random number; Calculate the distances of the remaining grey wolves from the three grey wolves with the smallest fitness function based on the second coefficient; Obtain the updated position of each grey wolf by combining the first coefficient and the distances of each grey wolf from the three grey wolves with the smallest fitness function; Repeat the execution of single-round optimization, re-select the three grey wolves with the smallest fitness function in each round of optimization and obtain the updated positions of the grey wolves until the maximum number of optimization rounds is reached, substitute the position of the grey wolf with the smallest fitness function into the fuzzy PID algorithm to output the real-time optimal temperature control power adjustment scale table.

3. A temperature control system according to claim 2, characterized in that, The substituting of the position of each grey wolf into the fuzzy PID algorithm and calculating the fitness function includes the following specific steps: Extract the target temperature table at the current moment, obtain the real-time temperature table at the previous moment and the real-time temperature table at the current moment; Subtract the target temperature table at the current moment from the real-time temperature table at the current moment to obtain the real-time target temperature difference table at the current moment, subtract the real-time temperature table at the previous moment from the real-time temperature table at the current moment and divide by the monitoring interval to obtain the real-time temperature change rate table at the current moment; Multiply the real-time target temperature difference and real-time temperature change rate of all key components at the current moment by the corresponding first quantization factor and second quantization factor respectively as the input of the fuzzy controller, and define the output of the fuzzy controller as the proportional factor table, integral factor table, and differential factor table; Select the Gaussian membership function, define the fuzzy subsets and establish the fuzzy control rule table; Select the centroid method as the defuzzification method, obtain the proportional factor table, integral factor table, and differential factor table, and obtain the temperature control power adjustment amount and control rounds corresponding to each key component at the current moment through the PID algorithm; Sum up the temperature control power adjustment amounts and control rounds of all key components to obtain the total temperature control power adjustment amount and total control rounds at the current moment, and accumulate to obtain the fitness function at the current moment.

4. A temperature control system according to claim 1, characterized in that, The decision on whether to send a real-time instruction includes the following specific steps: Determine that the current moment is the start moment, obtain the real-time ambient temperature and real-time ambient humidity at the current moment, and use the normal ambient temperature and normal ambient humidity as the comparison ambient temperature and comparison ambient humidity respectively; Calculate the ambient temperature change gradient and ambient humidity change gradient at the current moment and determine whether both are less than or equal to the change gradient threshold; If both are less than or equal to the change gradient threshold, use the optimal temperature table as the target temperature table at the current moment and store it, and package the execution flag and the normal temperature control power of all key components to generate the execution instruction at the current moment; If the ambient temperature change gradient and / or ambient humidity change gradient at the current moment is greater than the change gradient threshold, calculate the real-time ambient temperature change amount and real-time ambient humidity change amount at the current moment, package them with the adjustment flag to generate an adjustment instruction and send it to the target formulation module, receive and store the target temperature table at the current moment, and package it with the control flag to generate the control instruction at the current moment and send it to the optimization control module.

5. The temperature control system according to claim 1, wherein, The decision on whether to send a real-time instruction further includes the following specific steps: Determine that the current moment is not the start moment, obtain the real-time ambient temperature and real-time ambient humidity at the current moment, and use the real-time ambient temperature and real-time ambient humidity at the previous moment as the comparison ambient temperature and comparison ambient humidity respectively; Calculate the ambient temperature change gradient and ambient humidity change gradient at the current moment and determine whether both are less than or equal to the change gradient threshold; If both are less than or equal to the change gradient threshold, use the target temperature table at the previous moment as the target temperature table at the current moment and store it, and obtain the real-time temperature table at the current moment; Calculate the temperature change gradient of all key components at the current moment respectively and determine whether both are less than or equal to the change gradient threshold. If both are less than or equal to the change gradient threshold, no operation needs to be performed. If there is a single component whose temperature change gradient at the current moment is greater than the change gradient threshold, generate a control flag, package it with the target temperature table at the current moment to generate a control instruction and send it to the optimization control module; If the ambient temperature change gradient and / or ambient humidity change gradient at the current moment is greater than the change gradient threshold, calculate the real-time ambient temperature change amount and real-time ambient humidity change amount at the current moment, package them with the adjustment flag to generate an adjustment instruction and send it to the target formulation module, receive and store the target temperature table at the current moment, and package it with the control flag to generate the control instruction at the current moment and send it to the optimization control module.

6. The temperature control system according to claim 5, wherein The target formulation module receives the real-time adjustment instruction and generates a real-time target temperature table based on the environment correction strategy combined with the real-time environment parameters, including the following specific steps: Input the real-time ambient temperature change amount and real-time ambient humidity change amount into the environment correction network group to output the real-time fitting temperature change table; Obtain the optimal temperature of a single key component and subtract the real-time temperature change amount to obtain the real-time target temperature of a single key component; Statistically calculate the real-time target temperatures of all key components to obtain the real-time target temperature table.

7. A temperature control system according to claim 6, characterized in that, The environmental correction network group includes environmental correction networks whose quantity is equal to the total number of key components and which need to be pre-trained. Among them, a single environmental correction network fits a non-linear mapping function of the temperature change amount of a corresponding single key component under the conventional temperature control power with respect to the environmental temperature change amount and the environmental humidity change amount.

8. A temperature control system according to claim 7, wherein The pre-training of the single environmental correction network includes the following specific steps: Fix the conventional temperature control power of a single temperature control device, and measure and obtain different temperatures of a single key component at different environmental temperatures and environmental humidities respectively; Calculate the temperature change amount corresponding to each pair of environmental temperature change amount and environmental humidity change amount; Construct a data set using the environmental temperature change amount, the environmental humidity change amount, and the corresponding temperature change amount, and divide it into a training set and a test set; Select an environmental correction network, define the input as the environmental temperature change amount and the environmental humidity change amount, the output as the temperature change amount, define the loss function as the mean square error, and use the Adam optimizer for training; After the training is completed, fix the hyperparameters of the single environmental correction network, and use the test set for testing to obtain the mean square error corresponding to the test set and determine whether it is less than or equal to the error threshold; If it is less than or equal to the error threshold, the pre-training of the single environmental correction network is completed. If it is greater than the error threshold, adjust the Adam optimizer and re-train.

9. The temperature control system according to claim 1, characterized in that, The target formulation module stores an optimal temperature table and a conventional parameter table. The optimal temperature table includes the optimal temperatures of all key components, and the conventional parameter table includes the conventional environmental temperature, the conventional environmental humidity, and the conventional temperature control power of the temperature control devices corresponding to all key components.

10. A temperature control system according to claim 1, characterized in that, The temperature control system further includes a temperature control execution module. The temperature control execution module receives an execution instruction and operates the temperature control device according to the conventional temperature control power table, and receives a real-time adjustment execution instruction and updates the real-time temperature control power of each temperature control device.

Images