Injection mold temperature control system based on PID closed loop algorithm

Abstract

The application belongs to the technical field of intelligent control of injection molding, and particularly discloses an injection mold temperature control system based on a PID closed-loop algorithm. The system comprises a multi-partition temperature control execution unit, a temperature sensing network, a central collaborative optimization controller, a thermodynamic state evaluation module and a dynamic set value generation unit. The central collaborative optimization controller models each temperature control partition as a non-cooperative game participant, solves a Nash equilibrium solution that takes into account product quality stability and minimum energy consumption under the constraint of thermodynamic entropy increase, and updates the PID target value of each partition in real time by the dynamic set value generation unit. The application suppresses the heat fighting phenomenon between adjacent partitions, reduces the invalid start and stop of the heating and cooling device, reduces energy consumption, and improves the temperature control robustness and the stability of the molding quality.

Description

Technical Field

[0001] This invention belongs to the field of intelligent control technology for injection molding, specifically relating to a temperature control system for injection molds based on a PID closed-loop algorithm. Background Technology

[0002] With the continuous evolution of precision injection molding technology, precise temperature control of injection molds has become a crucial factor determining the molding quality and production stability of plastic products. Traditional industrial temperature control systems mainly achieve closed-loop regulation through proportional-integral-differential algorithms to maintain a constant thermal environment inside the mold. In highly automated production processes, the system needs to precisely schedule heating elements and cooling circuits to ensure that the material is in its optimal thermodynamic state during the filling, holding, and cooling stages. This is of great significance for improving the dimensional consistency and mechanical strength of the product.

[0003] Multi-zone independent temperature control technology has been widely used in the manufacturing of complex mold structures. By arranging independent sensing and actuation units in different parts of the mold, it aims to achieve more refined management of local temperature field distribution. The basic goal of this technology is to utilize feedback control strategies to respond to fluctuations in heat load in real time and dynamically compensate the output power of each zone according to preset process parameters, thereby meeting the stringent requirements of precision molds for thermal balance.

[0004] However, existing temperature control solutions typically employ regulation logic based on fixed setpoints. This leads to frequent heater start-ups and shutdowns in an attempt to maintain an absolutely constant temperature, resulting in redundant energy consumption. Furthermore, due to a lack of in-depth characterization of the thermal coupling patterns across multiple zones, cross-conduction of heat between adjacent temperature control zones easily causes control gain conflicts, leading to severe temperature field oscillations or thermal interference during dynamic adjustment. In addition, traditional control strategies struggle to balance molding quality boundary conditions with system operating efficiency, failing to establish a nonlinear balance between ensuring product process stability and reducing macroscopic energy consumption. This results in limited overall robustness and resource allocation efficiency in the injection molding process. Therefore, a temperature control system for injection molds based on a PID closed-loop algorithm is needed. Summary of the Invention

[0005] The purpose of this invention is to provide a temperature control system for injection molds based on a PID closed-loop algorithm, which can solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A temperature control system for injection molds based on a PID closed-loop algorithm includes a multi-zone temperature control execution unit, a temperature sensing network, a central collaborative optimization controller, a thermodynamic state assessment module, and a dynamic setpoint generation unit.

[0008] The multi-zone temperature control execution unit is configured in different functional areas of the injection mold and is used to independently adjust the heating and cooling power output of each zone according to the received control commands.

[0009] The temperature sensing network is deployed at key temperature measurement points inside the mold to collect real-time temperature data of each zone and transmit it to the central collaborative optimization controller.

[0010] The central collaborative optimization controller receives real-time temperature data from the temperature sensing network and, in conjunction with preset product quality boundary conditions and energy consumption constraints, coordinates the control strategies between different zones.

[0011] The thermodynamic state assessment module is used to quantitatively assess the current thermal energy distribution state of the system based on the principle of thermodynamic entropy increase, and inputs the assessment results as constraints to the central collaborative optimization controller.

[0012] The dynamic setpoint generation unit dynamically generates the target setpoint of the PID closed-loop control loop of each zone based on the game equilibrium solution output by the central collaborative optimization controller, and sends it to the corresponding multi-zone temperature control execution unit.

[0013] Preferably, the central collaborative optimization controller models each temperature control zone of the mold as an independent participant in a non-cooperative game, uses product quality stability and system comprehensive energy consumption minimization as the payoff function, and solves the Nash equilibrium solution under the premise of satisfying the thermodynamic entropy increase constraint.

[0014] Furthermore, the product quality stability is characterized by whether the temperature fluctuation range is within a preset threshold range, and the system's overall energy consumption is minimized by measuring the start-stop frequency and total power output of the heating and cooling devices.

[0015] Furthermore, the thermodynamic state assessment module calculates the entropy growth rate of the system in the current operating stage based on the overall heat flux density distribution and local heat conduction rate of the mold, and determines whether there is a risk of heat energy waste or thermal field imbalance.

[0016] Preferably, the dynamic setpoint generation unit updates the PID target setpoint once in each control cycle, so that the actual temperature of each zone is dynamically adjusted within the allowable process fluctuation range, avoiding frequent power switching caused by forced constant temperature.

[0017] Furthermore, the multi-zone temperature control execution unit includes independently controllable electric heating elements and fluid cooling channels, and its power output response speed and adjustment accuracy meet the requirements of precision injection molding for dynamic stability of the thermal field.

[0018] Furthermore, the temperature sensing network employs high-precision thermocouples or infrared temperature measurement arrays, whose sampling frequency and spatial resolution are sufficient to capture the transient thermal behavior characteristics of the mold during the filling, holding, and cooling stages.

[0019] Preferably, the central collaborative optimization controller has a built-in multi-objective collaborative optimization engine. This multi-objective collaborative optimization engine integrates game theory models from microeconomics with the basic laws of thermodynamics, and can automatically find the temperature control path with the optimal energy consumption while ensuring the quality of product molding.

[0020] Furthermore, the solution process of the Nash equilibrium takes into account the thermal coupling interference effect between adjacent temperature control zones. By introducing a cross-thermal conduction coefficient matrix, the strategy space of each participant is modified to suppress the occurrence of thermal conflict.

[0021] Furthermore, the dynamic setpoint generation unit and the local PID controller of each partition adopt a closed-loop feedback mechanism to ensure that the actual temperature response after the setpoint adjustment can quickly converge to a new equilibrium point.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] The injection mold temperature control system based on PID closed-loop algorithm provided by this invention successfully constructs a multi-objective collaborative optimization mechanism that takes into account both product quality and energy efficiency by treating each temperature control zone as a game participant and introducing the principle of thermodynamic entropy increase as a constraint.

[0024] The system no longer mechanically maintains a fixed temperature setpoint, but dynamically seeks the lowest energy consumption operating state within the allowable fluctuation range of product quality, reducing the number of ineffective start-ups and shutdowns of heating and cooling devices and significantly saving energy consumption in the production process.

[0025] This invention proactively coordinates multi-zone thermal coupling interference through a non-cooperative game theory model, suppressing control conflicts and thermal field oscillations between adjacent regions, thereby improving the robustness and stability of the entire temperature control system. This system deeply integrates microeconomics, thermodynamics, and automatic control theory, breaking through the limitations of traditional single-objective control logic and providing a new paradigm for intelligent, energy-efficient, and adaptive temperature control in the precision injection molding manufacturing field. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the overall technical solution architecture of the present invention;

[0027] Figure 2 This is a schematic diagram of the core principle framework of the central collaborative optimization that integrates the non-cooperative game model and the thermodynamic entropy increase constraint in this invention.

[0028] Figure 3 This is a flowchart illustrating the logical flow of this invention, from real-time temperature acquisition and thermodynamic state assessment to Nash equilibrium solution.

[0029] Figure 4 This is a schematic diagram of the multi-level interaction relationship and data flow between the multi-zone temperature control execution unit and the central collaborative optimization controller in this invention;

[0030] Figure 5 This is a schematic diagram of the principle framework of the core principle of this invention, which is to suppress partitioned thermal coupling interference by modifying the strategy space based on the cross-thermal conduction coefficient matrix. Detailed Implementation

[0031] Example 1: Please refer to the appendix Figure 1 To be continued Figure 5 To make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to specific embodiments.

[0032] A temperature control system for injection molds based on a PID closed-loop algorithm includes a multi-zone temperature control execution unit, a temperature sensing network, a central collaborative optimization controller, a thermodynamic state assessment module, and a dynamic setpoint generation unit.

[0033] The temperature sensing network is configured to be distributed across key geometric topological points of the injection mold, enabling real-time, high-frequency capture of changes in the microscopic thermal field within the mold. It converts the acquired multidimensional raw temperature signals into a standardized digital thermal information stream. The network integrates a high-precision thermocouple array and a non-contact infrared sensing matrix. The high-precision thermocouple array is embedded in sensing apertures near the mold cavity surface to acquire the transient conduction temperature of the metal substrate. The non-contact infrared sensing matrix is ​​positioned at key cross-sections of the mold parting surface and material flow channels to monitor the characterizing thermal radiation data of the melt during the filling process. The network also includes a signal conditioning unit configured to amplify the weak voltage signal output from the sensing terminals at a high rate with low noise, compensate for zero-point drift, and filter out random interference based on a digital filtering algorithm. This ensures that the temperature data transmitted to the central collaborative optimization controller has extremely high confidence and time resolution.

[0034] The multi-zone temperature control actuator is physically coupled to different functional areas of the injection mold, including the fixed mold cooling zone, the moving mold heating zone, the slider independent temperature control zone, and the gate hot runner adjustment zone. The multi-zone temperature control actuator independently and concurrently drives the heating power devices and cooling medium regulating valves of each zone according to received control commands. The multi-zone temperature control actuator includes an electric heating control subunit and a fluid circulation control subunit; the electric heating control subunit adopts a solid-state relay drive scheme based on pulse width modulation technology, achieving stepless fine adjustment of the output power of the heating rod or heating coil by adjusting the duty cycle; the fluid circulation control subunit includes multi-way proportional solenoid valves and a variable frequency circulating pump, used to dynamically adjust the flow rate and pressure of cooling water or guide oil according to heat load requirements, achieving rapid removal of local heat. Although the zones are physically independent, they have complex conduction and radiation coupling relationships at the thermophysical level.

[0035] The central collaborative optimization controller, serving as the decision-making core of the entire system, is based on a high-performance multi-core industrial processing platform, incorporating a large-capacity synchronous dynamic random access memory and high-speed non-volatile storage media. The central collaborative optimization controller is configured to receive real-time temperature data streams from the temperature sensor network and, in conjunction with product quality boundary conditions, energy efficiency constraints, and material thermophysical property parameters stored in a preset database, to globally coordinate the temperature control strategy of the entire system. The core logic of the central collaborative optimization controller lies in abstracting and modeling each temperature control zone of the mold as an independent participant in a non-cooperative game, with each participant possessing a specific strategy space, i.e., the adjustment range of the PID setpoint. The central collaborative optimization controller utilizes a built-in concurrent computing engine to solve the payoff function of each participant in real time. The payoff function is defined as the weighted difference between the product quality stability score and the overall system energy consumption cost. By executing a non-cooperative game algorithm, the central collaborative optimization controller automatically finds the Nash equilibrium solution among the zones, while satisfying the thermodynamic boundary constraints.

[0036] The thermodynamic state assessment module, which engages in deep data interaction with the central collaborative optimization controller, has a core function of performing a deep quantitative assessment of the system's current thermal energy distribution state based on the principle of thermodynamic entropy increase. The thermodynamic state assessment module includes a heat flux density calculation unit, an entropy generation rate analysis unit, and a thermal field balance evaluation unit. The heat flux density calculation unit analyzes the temperature gradient between adjacent temperature measurement points and, combined with the thermal conductivity matrix of the mold material, calculates the cross-interface heat flux vector field in real time. The entropy generation rate analysis unit, based on the second law of thermodynamics, calculates the rate of increase of the system's total entropy caused by irreversible heat conduction, fluid friction, and electrothermal conversion processes. The thermodynamic state assessment module uses the calculated entropy growth rate as a feedback constraint signal and transmits it to the central collaborative optimization controller in real time. If the current entropy growth rate exceeds a preset efficiency threshold, it determines that the system has a serious risk of energy waste or thermal interference, triggering the optimization controller to adjust the boundary of the game strategy space and force the system to converge towards a metastable state with low entropy production and high energy efficiency.

[0037] The dynamic setpoint generation unit, located at the output of the central collaborative optimization controller, is used to convert the game equilibrium solution into a physical quantity recognizable by the local PID control loops of each partition. The dynamic setpoint generation unit is configured to receive the optimal strategy vector from the central collaborative optimization controller at the beginning of each control cycle and, combined with the current actual temperature value of each partition, dynamically generate a new target temperature setpoint. This dynamic setpoint generation unit breaks away from the rigid strategy of traditional constant temperature control, allowing the target setpoint to move slightly within the fluctuation range allowed by the product process. For example, when the system detects that slightly reducing the temperature setpoint of a non-critical area can significantly reduce cooling medium consumption without affecting molding quality, the dynamic setpoint generation unit issues a corrected target value. Furthermore, the dynamic setpoint generation unit also integrates a setpoint smoothing filter to ensure that the target value issued to each partition PID controller remains continuous in the time dimension, preventing severe oscillations in the actuators caused by step changes in the target value.

[0038] When constructing the non-cooperative game model, the central collaborative optimization controller defines the control loop of each temperature control zone as a decision-making entity. The strategy set of the decision-making entity is the allowable temperature setpoint deviation for that zone under the current process stage. The design depth of the payoff function considers the physical properties of injection molding production: the deviation between the actual temperature and the theoretical process value is used to characterize the product quality stability cost, which increases exponentially with the increase of the deviation; the energy consumption cost is characterized by calculating the product of the real-time power output value of each zone's actuator and the rated energy efficiency ratio, and adding the start-up and shutdown losses caused by the actuator switching action. The central collaborative optimization controller is configured to search for the Nash equilibrium point in the multidimensional strategy space through an iterative optimization algorithm, which makes the payoff functions of all decision-making entities reach optimal balance. At the Nash equilibrium point, any single zone attempting to further reduce energy consumption by changing its own setpoint will lead to a disproportionate increase in the product quality cost, achieving a non-linear balance between quality and energy consumption.

[0039] In its specific hardware implementation, the multi-zone temperature control execution unit also includes a cross-coupling compensation circuit, which is used to suppress electromagnetic interference between adjacent heating elements at the physical layer. A high-sensitivity turbine flow meter and pressure sensor are installed inside the fluid cooling channel. The fluid state parameters collected by these sensors are fed back in real time to the local drive controller of the multi-zone temperature control execution unit for second-order closed-loop correction of the proportional solenoid valve opening. Each sensor node in the temperature sensing network is connected to a data concentrator via an industrial fieldbus protocol. The data concentrator employs a redundant link design to ensure the integrity of signal transmission in complex electromagnetic environments.

[0040] Furthermore, the thermodynamic state assessment module employs a simplified multidimensional thermal field assessment model based on finite element analysis during quantitative assessment. The module stores a three-dimensional topology mapping table of the injection mold and reconstructs the continuous temperature distribution field inside the mold using a Kriging interpolation algorithm based on discrete point temperature values ​​fed back from the temperature sensing network. The entropy generation rate analysis unit is configured to perform gradient calculations on the reconstructed continuous temperature field, calculating the difference between the product of the heat transfer amount and the reciprocal of the temperature for each volume element per unit time. By spatially integrating the entire mold volume, an overall entropy increase index reflecting the macroscopic energy efficiency level of the system is obtained. The thermodynamic state assessment module also compares this overall entropy increase index with a preset golden energy efficiency curve in real time, generating an early warning factor to characterize the risk of thermal field imbalance.

[0041] The central collaborative optimization controller also includes a cross-thermal conductivity coefficient matrix correction unit. Given the non-negligible metal thermal conduction between different zones within the mold, heating one zone will inevitably cause a temperature increase in adjacent zones. The cross-thermal conductivity coefficient matrix correction unit establishes a dynamic matrix describing the degree of interference between zones based on pre-set thermal response experimental data. During the game-theoretic optimization process, the central collaborative optimization controller uses this matrix to compensate for and correct the predicted temperature rise of each zone, achieving proactive prediction and elimination of thermal interference at the algorithm level.

[0042] Before issuing instructions, the dynamic setpoint generation unit also executes a safety verification logic. This safety verification logic is configured to retrieve the safety process package stored in the central collaborative optimization controller and check whether the dynamically generated setpoint exceeds the limit temperature range allowed by the mold structure strength or material degradation point. If the verification fails, the dynamic setpoint generation unit will automatically clamp the setpoint to the safety boundary and issue a policy anomaly alarm to the operator terminal.

[0043] The dynamic setpoint generation unit also integrates a heartbeat monitoring module, which includes a heartbeat frame transceiver unit, a timeout determination unit, and an emergency execution unit. During system operation, the heartbeat frame transceiver unit periodically sends bidirectional heartbeat frames to the local controllers of each zone in the multi-zone temperature control execution unit according to a preset 10ms heartbeat cycle. Each heartbeat frame contains a frame sequence number, a timestamp, and a link checksum. After receiving the heartbeat frame, each zone's local controller must send back a feedback response frame carrying the corresponding frame sequence number within a single heartbeat cycle. The heartbeat frame transceiver unit then transmits the received response frame to the timeout determination unit in real time.

[0044] The timeout determination unit has a preset timeout determination threshold of 3 consecutive heartbeat cycles. When the local controller of the corresponding partition fails to send back a valid feedback response frame within 3 consecutive heartbeat cycles, the timeout determination unit determines that the partition has a communication feedback loss fault and immediately transmits the fault signal to the emergency execution unit.

[0045] Upon receiving a fault signal, the emergency execution unit immediately triggers the emergency thermal balance procedure: First, it immediately stops sending dynamic target setpoints to the faulty zone, blocking the abnormal control link; Second, it sends a preset safe temperature value stored in the safety process package to the local actuator of the faulty zone, switching to open-loop protection mode. This preset safe temperature value is the constant safe temperature allowed by the injection molding process in the corresponding zone, ensuring that the mold thermal field will not fluctuate significantly and avoiding the risk of temperature runaway due to communication interruption; Third, it simultaneously sends a communication abnormality alarm signal to the central collaborative optimization controller and the operator terminal. After receiving the alarm signal, the central collaborative optimization controller synchronously adjusts the temperature control strategy of adjacent normal zones, maintaining the overall thermal field balance of the mold through cross-heat conduction compensation.

[0046] Example 2: Based on Example 1, this example provides an injection mold temperature control system based on edge computing and cloud collaborative architecture, which aims to solve the problem of centralized management and global energy efficiency optimization of multiple mold temperature controllers in a large-scale production workshop.

[0047] A temperature control system for injection molds based on a PID closed-loop algorithm includes an edge control cluster distributed across various production nodes, a global thermodynamic strategy server, a high-bandwidth industrial IoT access gateway, and a multi-sensor execution terminal deployed on the mold.

[0048] The edge control cluster consists of multiple independently deployed edge computing nodes next to the injection molding machine. Each node directly connects to a set of multi-zone temperature control execution units. Each edge computing node has a built-in lightweight Nash equilibrium solver for real-time processing of non-cooperative game logic between different zones within the mold. Unlike Embodiment 1, this edge control cluster is configured with local autonomous decision-making capabilities, maintaining efficient operation based on locally stored thermodynamic constraint models even when disconnected from the cloud. The edge computing nodes employ a real-time industrial operating system to ensure the control cycle of the temperature control loop remains stable within 10 milliseconds.

[0049] The global thermodynamic strategy server, deployed in the workshop control center or cloud platform, collects operational data from all edge control clusters. It contains a massively parallel processing matrix for performing in-depth thermodynamic entropy-increasing data analysis. This server is configured to analyze historical energy consumption characteristics under different molds, materials, and ambient temperatures, extracting the optimal payoff function weighting coefficients through machine learning algorithms. The server periodically pushes updated heat transfer coefficient matrices and game-theoretic optimization models to each edge computing node, achieving a leap from local optimization to global collaboration.

[0050] The high-bandwidth industrial IoT access gateway, serving as a data highway between the edge and the cloud, employs a hardware architecture supporting latency-sensitive network protocols. This gateway preprocesses and compresses the massive amounts of thermal field data from the edge control cluster, uploading only key feature vectors reflecting thermodynamic state changes, entropy increase trends, and game convergence to the global thermodynamic policy server, thus alleviating bandwidth pressure on the backbone network. Simultaneously, the access gateway integrates a physical layer encryption module to ensure the security of production process parameters during transmission.

[0051] The multi-sensor execution terminal incorporates a more integrated sensing and actuation component. Within each temperature control zone, sensors and actuators align data via a local wireless synchronization protocol. The multi-sensor execution terminal integrates a microelectromechanical system (MEMS) accelerometer and a pressure sensing unit, which, in addition to monitoring temperature, also monitor the vibration characteristics of the mold and changes in mold closing pressure during the injection molding cycle. This multi-dimensional sensing data is transformed into auxiliary constraint variables in a game theory model through an edge control cluster. For example, when the pressure sensing unit detects fluctuations in mold closing force, the edge computing node automatically determines that the current mold contact thermal resistance has changed and corrects the cross-thermal conductivity matrix in real time, improving the predictability of temperature regulation.

[0052] The global thermodynamic strategy server incorporates group game theory when performing optimization tasks. It considers not only the conflicting interests within each partition of a single mold, but also the allocation of cooling water resources and power load across the entire injection molding workshop as resource items in the game. The server is configured to dynamically adjust the temperature control weights of each mold by solving for a generalized Nash equilibrium at the workshop scale, avoiding instantaneous energy spikes caused by multiple devices simultaneously starting high-power heating or high-flow cooling. This scheduling strategy based on macroscopic thermodynamic entropy reduction can further improve the resource allocation efficiency of the manufacturing process from a systems engineering perspective.

[0053] In this embodiment, the dynamic setpoint generation unit of each edge computing node in the edge control cluster is equipped with a heartbeat monitoring module consistent with that in Embodiment 1. Its heartbeat frame transceiver unit establishes a low-latency bidirectional communication link with the local controller of each partition of the corresponding mold. The timeout determination unit can flexibly adjust the heartbeat cycle and timeout determination threshold according to the electromagnetic environment of the workshop. When a feedback loss fault is detected, in addition to executing the local emergency thermal balance program, the fault information will also be synchronously uploaded to the global thermodynamic strategy server through the high-bandwidth industrial IoT access gateway to realize global fault registration and cross-device collaborative scheduling.

[0054] The dynamic setpoint generation unit is further divided into a local fast-loop unit and a cloud-based long-cycle unit. The local fast-loop unit is responsible for fine-tuning the PID setpoint at the microsecond level based on the output of the edge computing node to cope with transient thermal shocks in the injection molding cycle. The cloud-based long-cycle unit, based on the analysis results of the global strategy server, adjusts the baseline operating point of the temperature control system on an hourly or daily timescale to adapt to fluctuations in the environmental thermal baseline caused by seasonal changes. This two-level linkage mechanism ensures that the system can maintain its thermodynamically optimal operating range at different time scales.

[0055] Furthermore, the thermodynamic state assessment module in Example 2 adds an online energy efficiency diagnostic function. By performing cross-correlation analysis on the energy consumption data and temperature response curves uploaded by each edge node, the module can identify physical defects such as aging of heating pipes, scaling of cooling pipes, or failure of the insulation layer. When an abnormal increase in the entropy generation rate and a slowdown in the game convergence speed of a certain partition are detected, the system automatically generates maintenance suggestions and guides the central collaborative optimization controller to temporarily reduce the performance weight of that partition, allowing adjacent partitions to maintain the stability of the overall thermal field of the mold through heat conduction compensation. This self-healing control logic significantly improves the long-term robustness of the injection molding temperature control system.

[0056] Example 3: This example describes an enhanced temperature control system implementation scheme for ultra-precision, micro-nano-scale injection molding, which focuses on strengthening the ability to sense and synergistically suppress extremely small disturbances in the thermal field.

[0057] A temperature control system for injection molds based on a PID closed-loop algorithm includes a high-frequency transient sensor array, a microfluidic array execution unit, a hyperparallel game calculation engine, and a thermodynamic entropy flow monitoring module.

[0058] The high-frequency transient sensing array employs sensing nodes based on thin-film thermocouple technology, which are directly sputtered and deposited onto the nanostructured surface of the mold cavity. The sampling frequency of the sensing array is increased to the kilohertz level, enabling it to capture extremely subtle temperature jumps at the moment of melt contact. The high-frequency transient sensing array transmits signals to a hyperparallel game theory computing engine via a dedicated serial bus, providing ultra-high-dimensional real-time state space input for subsequent Nash equilibrium solutions.

[0059] The microfluidic array execution unit consists of a micro-semiconductor cooling chip and a microelectromechanical pulse jet cooling system integrated inside the mold. Unlike traditional macro-cooling circuits, the microfluidic array execution unit can achieve precise removal of local hot spots at the micrometer level. Each micro-execution node is configured as a game participant, and its payoff function incorporates mass weights based on surface roughness and molecular chain orientation consistency. By frequently switching between heating and cooling states, the execution unit constructs a dynamically stable thermal environment at the microscopic level, supporting the perfect replication of micro- and nano-structures.

[0060] The hyperparallel game computing engine employs a customized field-programmable gate array (FPGA) architecture, specifically designed to handle non-cooperative games between tens of thousands of micro-partitions. Internally, the engine implements a hardware-based matrix operation accelerator, capable of solving ultra-large-scale linear equation systems in sub-milliseconds and calculating the Nash equilibrium solution for each micro-partition. This extremely rapid decision-making capability allows the system to counteract thermal disturbances through localized, predictive adjustments before thermal diffusion causes overall oscillations.

[0061] The thermodynamic entropy flow monitoring module utilizes laser interferometry to non-contactly monitor the entropy flow direction and entropy generation distribution on the mold surface. This module is configured to convert real-time acquired thermal field distribution images into entropy density cloud maps, which are then input as visual constraints to the game theory calculation engine. By actively guiding the entropy flow direction, the system can achieve directional heat transfer, i.e., using game theory strategies to actively direct heat from mass-sensitive areas to less sensitive areas with higher heat capacity. This significantly optimizes the thermal stability of key parts of the mold without increasing total energy consumption.

[0062] In this embodiment, the dynamic setpoint generation unit employs a policy mapping network pre-trained based on deep reinforcement learning. This policy mapping network, based on game equilibrium solutions, learns how to generate setpoint combinations for complex microstructures in a very short time through large-scale offline simulation training. The setpoint generation process also considers the coupling effects of rheological parameters, ensuring that each dynamic adjustment of the temperature setpoint is highly synchronized with the melt flow state.

[0063] The system also features an active thermal inertia compensation mechanism. Just before the injection molding cycle switches to the holding pressure stage, the central collaborative optimization controller predicts the upcoming thermal load impact based on a game theory model. Through a dynamic setpoint generation unit, it issues preheating or precooling commands to specific zones in advance. This time-lead-based control strategy utilizes the delayed characteristics of heat conduction, ensuring that the mold temperature is precisely at the set process equilibrium point when the actual impact arrives, eliminating the lag effect commonly found in traditional PID control.

[0064] Furthermore, in Embodiment 3, the multi-zone temperature control actuator integrates surface acoustic wave (SAW) technology to improve the wettability of the melt and mold wall through micro-vibration while achieving high-precision temperature control. This cross-domain actuator integration transforms the temperature control system from a passive environmental maintenance system into a comprehensive control platform that actively participates in the physical process of product molding.

[0065] In summary, this invention transforms the traditional industrial automation problem of injection mold temperature control into an interdisciplinary issue of resource optimization and physical system energy efficiency improvement by introducing non-cooperative game theory and the principle of thermodynamic entropy increase. Through dynamic game theory among multiple participants in each zone, the system finds a Nash equilibrium point that dynamically balances product molding quality and operational energy efficiency. During actual operation, the temperature sensing network continuously provides multi-dimensional visual and physical feedback to the system; the thermodynamic state assessment module monitors the system's disorder and energy loss in real time; the central collaborative optimization controller acts as the brain, coordinating the strategy space of hundreds or thousands of zone actuators; and the dynamic setpoint generation unit translates the optimal decision into every minute action of the actuators. Compared with existing technologies, this invention not only solves the market pain points caused by multi-zone thermal coupling interference but also establishes a new paradigm for the deep integration of green manufacturing and intelligent manufacturing in the field of precision injection molding.

[0066] In the system description of this invention, the logical relationships such as greater than, less than, and equal to refer to the textual comparison of physical parameters on the numerical axis. For example, when the actual temperature is greater than the upper limit of the set value, the system is configured to increase the circulation flow rate of the cooling medium; when the calculated entropy growth rate is less than the preset energy efficiency benchmark value, the system is configured to prevent a step contraction of the current game strategy space. All processes involving numerical calculations are described as logical processing behaviors within each functional unit. The arithmetic sum, arithmetic difference, product, quotient, as well as integral and differential operations are all considered as algorithmic logic configured within the module and are not expressed through any algebraic symbols.

[0067] The technical features described in each embodiment can be combined in various ways without conflict. For example, the hardware connection method in Embodiment 1 can be applied to the edge computing architecture of Embodiment 2; the high-frequency sensing technology in Embodiment 3 can also be a preferred configuration to improve the sensing accuracy in the system of Embodiment 1. This modular and scalable design ensures that the injection mold temperature control system based on the PID closed-loop algorithm proposed in this invention can adapt to various levels of production needs, from basic industrial parts to high-end precision optical components. In future engineering implementation, the system can be further integrated with artificial intelligence prediction models to achieve more precise predictive temperature control and energy efficiency optimization through data mining of all elements of the production environment.

[0068] All functional units mentioned in this invention are not limited to the FPGA, ARM, or industrial server described herein; any computing device capable of performing the aforementioned logic operations and signal processing is within the scope of protection of this invention. Furthermore, the physical forms of the heating elements, cooling valves, and other actuators described herein can be equivalently replaced according to the specific structure of the mold. For example, electromagnetic induction heating can be used instead of electric heating rods, or supercritical carbon dioxide fluid can be used instead of water cooling. These variations in implementation fall within the substantial protection scope of the system architecture of this invention.

[0069] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A temperature control system for injection molds based on a PID closed-loop algorithm, characterized in that, include: Multi-zone temperature control actuators are configured in different functional areas of the injection mold to independently adjust the heating and cooling power output of each zone according to the received control commands. A temperature sensing network is deployed at various temperature measurement points inside the mold to collect real-time actual temperature data of each zone. The central collaborative optimization controller is connected to the temperature sensing network to receive the actual temperature data and coordinate the control strategies between the various zones by combining preset product quality boundary conditions and energy consumption constraint information. The thermodynamic state assessment module interacts with the central collaborative optimization controller to quantitatively assess the current thermal energy distribution state of the system based on the principle of thermodynamic entropy increase, and inputs the assessment results as constraints to the central collaborative optimization controller. The dynamic setpoint generation unit is connected to the output of the central collaborative optimization controller. It is used to dynamically generate the target setpoint of the PID closed-loop control loop of each zone according to the game equilibrium solution output by the central collaborative optimization controller, and send the target setpoint to the multi-zone temperature control execution unit. The central collaborative optimization controller is configured to model each temperature control zone of the injection mold as an independent player in a non-cooperative game, and execute the following logic using a built-in concurrent computing engine: Construct a strategy space for each participant, whereby the strategy space is defined as the range of allowable offsets of the target setting value under the current process stage. Define a payment function for each participant, which is a weighted difference between the product quality stability score and the overall energy consumption cost of the system. The product quality stability score is determined based on the deviation between the actual temperature data and the theoretical process mean, and the weighted difference increases exponentially with the increase of the deviation. Under the premise of satisfying the constraints output by the thermodynamic state assessment module, an iterative optimization algorithm is used to search for the Nash equilibrium solution in the multidimensional policy space that makes the payoff functions of all participants reach optimal balance. The thermodynamic state assessment module includes: The heat flux density calculation unit is used to analyze the temperature gradient between adjacent temperature measurement points and, in conjunction with the thermal conductivity matrix of the mold material, to calculate the cross-interface heat flux vector field in real time. The entropy generation rate analysis unit is used to calculate the rate of increase of the total entropy of the system caused by irreversible heat conduction, fluid friction, and electrothermal conversion processes, based on the second law of thermodynamics. The thermal field balance evaluation unit is used to reconstruct the discrete point temperature values ​​into a continuous temperature distribution field inside the injection mold using the Kriging interpolation algorithm, and to perform gradient calculation on the continuous temperature distribution field to calculate the difference between the product of the heat transfer amount of each volume element and the reciprocal of the temperature per unit time. The overall entropy increase index is obtained by spatial integration over the entire mold volume. The thermodynamic state assessment module compares the overall entropy increase index with the preset golden energy efficiency curve in real time to generate an early warning factor for characterizing the risk of thermal imbalance.

2. The injection mold temperature control system based on PID closed-loop algorithm according to claim 1, characterized in that, The temperature sensing network includes: A high-precision thermocouple array is embedded in a sensing aperture close to the surface of the injection mold cavity to obtain the transient conduction temperature of the metal substrate; A non-contact infrared sensing matrix is ​​arranged on the parting surface and key sections of the material flow channel of the injection mold to monitor the characterizing thermal radiation data of the melt during the mold filling process. The signal conditioning unit is electrically connected to the high-precision thermocouple array and the non-contact infrared sensing matrix, respectively. The signal conditioning unit is configured to amplify the weak voltage signal output from the sensing end with high-rate low noise, compensate for zero-point drift, and filter out random interference based on digital filtering algorithm, converting the multidimensional original temperature signal into a standardized digital thermal information stream and transmitting it to the central collaborative optimization controller.

3. The injection mold temperature control system based on PID closed-loop algorithm according to claim 1, characterized in that, The multi-zone temperature control actuator is physically coupled to the fixed mold cooling zone, moving mold heating zone, slider independent temperature control zone, and gate hot runner adjustment zone of the injection mold. The multi-zone temperature control actuator includes: The electric heating control subunit adopts a solid-state relay drive scheme based on pulse width modulation technology, which realizes stepless fine adjustment of the output power of the electric heating device by adjusting the duty cycle; The fluid circulation control subunit includes a multi-way proportional solenoid valve, a variable frequency circulating pump, a turbine flow meter, and a pressure sensor. It is used to dynamically adjust the flow rate and pressure of the cooling medium according to the heat load requirements, and to perform second-order closed-loop correction on the opening of the proportional solenoid valve using the collected fluid state parameters. Cross-coupling compensation circuitry is used to suppress electromagnetic interference between adjacent heating elements at the physical layer.

4. The injection mold temperature control system based on PID closed-loop algorithm according to claim 1, characterized in that, The dynamic setpoint generation unit includes: A setpoint smoothing filter is used to ensure that the target setpoint sent to each partition PID control loop remains continuous in the time dimension, preventing actuator oscillations caused by target value steps. The safety verification logic unit is configured to retrieve a preset safety process package, check whether the dynamically generated target setting value exceeds the limit temperature range allowed by the mold structure strength or material degradation point, and clamp the target setting value to the safety boundary and issue a strategy abnormality alarm when the verification fails. The heartbeat monitoring module is used to maintain real-time communication and monitoring with the local controller of each zone, and to start an emergency thermal balance program when feedback loss is detected, switching to an open-loop protection mode based on preset safety values.

5. The injection mold temperature control system based on PID closed-loop algorithm according to claim 1, characterized in that, The central collaborative optimization controller also includes a cross-thermal conduction coefficient matrix correction unit. The cross-thermal conduction coefficient matrix correction unit establishes a dynamic matrix describing the degree of mutual interference between each temperature control zone through preset thermal response experimental data, and uses the dynamic matrix to compensate and correct the predicted temperature rise of each zone during the game optimization process, thereby actively eliminating thermal coupling interference between adjacent zones at the algorithm level.

6. The injection mold temperature control system based on PID closed-loop algorithm according to claim 5, characterized in that, The system also includes a high-bandwidth industrial IoT access gateway and a global thermodynamic strategy server: the high-bandwidth industrial IoT access gateway adopts a hardware architecture that supports latency-sensitive network protocols and is used to preprocess and compress key feature vectors of the actual temperature data, thermodynamic state and game convergence. The global thermodynamic strategy server receives the compressed feature vector through the high-bandwidth industrial IoT access gateway, performs workshop-level thermodynamic entropy increase data analysis using a large-scale parallel processing matrix, extracts the optimal payoff function weighting coefficients through machine learning algorithms, and periodically pushes the updated cross-thermal conduction coefficient matrix correction unit to the central collaborative optimization controller.

7. The injection mold temperature control system based on PID closed-loop algorithm according to claim 1, characterized in that, The multi-zone temperature control execution unit also integrates a surface acoustic wave actuator, which is used to improve the wettability between the melt and the mold wall by generating micro-vibrations while performing temperature control. The temperature sensing network also includes a microelectromechanical system accelerometer and a pressure sensing unit integrated inside the mold, used to monitor vibration characteristics and mold closing pressure changes during the injection molding cycle, and input the vibration characteristics and mold closing pressure changes as auxiliary constraint variables to the central collaborative optimization controller.

8. The injection mold temperature control system based on PID closed-loop algorithm according to claim 1, characterized in that, The system also has an active thermal inertia compensation mechanism. The central collaborative optimization controller is configured to predict the upcoming thermal load impact based on a non-cooperative game model within a preset lead time before the injection cycle switches to the holding pressure stage, and issue preheating or precooling commands to specific temperature control zones in advance through the dynamic setpoint generation unit, using the delayed characteristics of heat conduction to offset the thermal field fluctuations of the injection mold during the holding pressure stage.

Images