Soft-start control method for intelligent temperature control of injection molds

Abstract

This invention discloses a soft-start control method for intelligent temperature control of injection molds, specifically relating to the field of injection mold technology. The method involves: collecting the initial temperatures of multiple heating zones of the mold to construct a temperature distribution model; calculating thermal inertia scores based on the thermal response characteristics and heat capacity parameters of each zone; predicting the temperature rise rate of different zones using the thermal inertia scores, thermal coupling anomalies, and historical control deviation frequencies, and setting differentiated soft-start limiting parameters; collecting temperature feedback in real time during the heating process to construct a temperature rise rate error sequence, and using a dynamic Bayesian network to predict future deviation risks; and dynamically correcting the PWM duty cycle or conduction angle of the heating zone when the risk value exceeds the limit to achieve adaptive power output adjustment. This method possesses advantages such as strong predictability, accurate response, and intelligent control, making it suitable for high-precision heating control applications in complex injection molds.

Description

Technical Field

[0001] This invention relates to the field of injection molding technology, and more specifically to a soft-start control method for intelligent temperature control of injection molding molds. Background Technology

[0002] The soft-start control of intelligent temperature control in injection molding molds refers to the use of an intelligent temperature control system to regulate the temperature of the mold's heating elements during the injection molding process. In the initial heating stage, a "soft start" method is employed to gradually increase the temperature, thereby avoiding sudden current surges that could impact the equipment, extending the lifespan of the heating elements, and ensuring a uniform temperature rise in the mold, thus improving molding quality and production stability. This control method combines temperature feedback with program settings to achieve precise and safe heating management.

[0003] The existing technology has the following shortcomings:

[0004] In large or multi-cavity complex molds, due to the lag in heat conduction and significant differences in thermal inertia, existing temperature control systems often rely on single-point temperature sensors. This may cause the controller to misjudge the temperature rise process as too slow and prematurely release the soft-start limit, thereby causing a sudden increase in local temperature that exceeds the material's tolerance limit. This can lead to stress concentration, microcracks, and even damage to the sealing structure inside the mold, ultimately affecting the dimensional accuracy and appearance quality of the product. The risk is particularly prominent in high-cavity pressure molds such as automotive lamp housings or optical lenses. Summary of the Invention

[0005] The purpose of this invention is to provide a soft-start control method for intelligent temperature control of injection molds, so as to overcome the shortcomings of the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a soft-start control method for intelligent temperature control of injection molds, comprising:

[0007] Initial temperature data for multiple regions of the mold, including the mold cavity, nozzle, and material channel, are acquired, and a temperature distribution model is constructed.

[0008] Based on the temperature distribution model and the preset thermal inertia score, the temperature rise rate of different regions is predicted, and regions that may have thermal hysteresis are identified.

[0009] During the soft start phase, differentiated current or power limiting parameters are set for each heating zone to form a dynamic soft start strategy at the zone level.

[0010] Real-time temperature feedback data from various regions is collected, and power output is dynamically adjusted by combining the predicted temperature rise trajectory with the actual temperature rise rate.

[0011] Once the target temperature threshold is reached, the limiting is lifted, and the system switches to PID closed-loop control to achieve uniform and stable temperature control of the mold.

[0012] Preferably, the calculation of the thermal inertia score includes:

[0013] Collect historical temperature rise curves for each heating zone of the mold;

[0014] Calculate the temperature rise response time per unit power in each region as a thermal response delay index;

[0015] Calculate the thermal inertia score based on the thermal response delay index and the regional heat capacity parameters;

[0016] If the thermal inertia score of the heating area is higher than the preset threshold, it is marked as a thermal hysteresis area and a limiting strategy is adapted in the soft start control.

[0017] Preferably, predicting the temperature rise rate in different regions includes: inputting the current initial temperature and the target set temperature into the prediction model; the prediction model adopts a multi-factor weighted method, wherein the weighting factors include thermal inertia score, regional thermal coupling anomaly value and historical control deviation frequency, and calculates the expected temperature rise time and rate.

[0018] Preferably, the method for obtaining the regional thermal coupling degree anomaly value is as follows: temperature change data are collected synchronously in multiple preset heating areas to construct a temperature time series matrix for multiple time periods; the temperature covariance between any two areas is calculated based on the temperature series, and the total thermal coupling degree of each area is solved accordingly; the global average value of the thermal coupling degree of all areas is calculated, and the thermal coupling degree deviation value of each area is calculated as the regional thermal coupling degree anomaly value.

[0019] Preferably, the method for obtaining the historical control deviation frequency is as follows: record the instantaneous deviation value between the actual temperature and the target temperature of each heating zone in multiple historical soft-start cycles; set a deviation tolerance threshold, and statistically analyze the proportion of time points in each soft-start cycle where the deviation exceeds the threshold, which is defined as the regional deviation frequency; summarize the deviation frequencies of multiple cycles and take their average to form the historical control deviation frequency.

[0020] Preferably, the dynamically corrected power output includes:

[0021] Real-time temperature feedback values ​​of each heating zone are collected, and the rate of temperature change per unit time is calculated as the actual temperature rise rate.

[0022] The temperature rise trajectory output by the prediction model is called, and the error between the actual temperature rise rate and the prediction model is calculated at each time step to construct the temperature rise rate error sequence.

[0023] Based on the temperature rise rate error sequence, a probabilistic model is performed on the difference between the predicted and actual temperatures to estimate the risk value of future temperature control deviations.

[0024] If the temperature control deviation risk value exceeds the risk threshold, the PWM duty cycle or conduction angle of the heating area will be dynamically corrected to achieve real-time optimization of power output.

[0025] Preferably, the estimated future temperature control deviation risk values ​​include:

[0026] The predicted temperature rise rate and the actual temperature rise rate of each heating zone are collected at multiple time steps, and the difference between the two is calculated to form a temperature rise rate error sequence.

[0027] Using the error value as an observable variable, a dynamic Bayesian network model is constructed. The hidden state at each time step is defined as the control deviation state, which depends only on the hidden state at the previous time step. A state transition probability relationship is established.

[0028] Based on historical control data or real-time observation sequences, the state transition probability and observation probability distribution are estimated using the expectation-maximization algorithm or Bayesian filtering method.

[0029] At any given moment, based on the known error sequence, infer the probability that the system will be in a high deviation state in the future, and use this as the future temperature control deviation risk value.

[0030] Preferably, the dynamic correction of the PWM duty cycle or conduction angle of the heating area includes:

[0031] The adjustment priority of each heating zone is determined based on the predicted deviation risk value, with priority given to adjusting zones with a risk probability higher than the threshold.

[0032] If the heating area uses PWM control, the adjustment increment is calculated based on the current error direction and amplitude, and the PWM duty cycle is increased or decreased to refine the power output. The change in duty cycle is limited by the preset maximum adjustment step size.

[0033] If the heating zone uses a silicon controlled rectifier or thyristor control method, the conduction angle can be adjusted by phase control to shorten or extend the conduction time, thereby changing the average power in a single cycle.

[0034] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0035] 1. This invention introduces multi-dimensional analysis parameters such as thermal inertia scoring, thermal coupling degree anomaly identification, and historical control deviation frequency to realize differentiated amplitude limiting control and predictive temperature rise strategy in the heating area. It effectively overcomes the problems of thermal hysteresis misjudgment, local overheating, and large temperature fluctuation in traditional mold temperature control systems, and significantly improves the response accuracy and system safety of the heating process.

[0036] 2. This invention integrates a dynamic Bayesian network algorithm to predict the risk of future temperature control deviations, and combines real-time feedback to dynamically correct the PWM duty cycle or conduction angle. After reaching the target temperature, it automatically switches to PID closed-loop control, thus constructing an integrated intelligent temperature control system with prediction, decision-making, and adaptive adjustment capabilities. This not only improves the uniformity of mold heating and the consistency of molding quality, but also has good scalability and industrial applicability. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0038] Figure 1 This is a mind map of the method of the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] For examples, please refer to Figure 1 As shown in this embodiment, the soft-start control method for intelligent temperature control of injection molds includes:

[0041] Initial temperature data for multiple regions of the mold, including the mold cavity, nozzle, and material channel, are acquired, and a temperature distribution model is constructed.

[0042] Based on the temperature distribution model and the preset thermal inertia score, the temperature rise rate of different regions is predicted, and regions that may have thermal hysteresis are identified.

[0043] During the soft start phase, differentiated current or power limiting parameters are set for each heating zone to form a dynamic soft start strategy at the zone level.

[0044] Real-time temperature feedback data from various regions is collected, and power output is dynamically adjusted by combining the predicted temperature rise trajectory with the actual temperature rise rate.

[0045] Once the target temperature threshold is reached, the limiting is lifted, and the system switches to PID closed-loop control to achieve uniform and stable temperature control of the mold.

[0046] The process of acquiring initial temperature data for multiple regions of the mold, including the mold cavity, nozzle, and material channel, and constructing a temperature distribution model, can be broken down into the following technical steps and key points:

[0047] Area division and sensor deployment:

[0048] Mold cavity: Install thermocouples or PT100 temperature sensors near the molding cavity, cooling channel or stress concentration area;

[0049] Nozzle area: A sensor is installed at the nozzle root or flow channel interface to reflect the temperature at the initial injection point of the material;

[0050] Material channel area: Sensors are installed near the main runner, branch runners and end gate to cover the heat conduction path;

[0051] The goal is to achieve multi-point coverage of the main thermal control areas within the mold, ensuring that the temperature control strategy takes into account local differences.

[0052] Initial temperature data acquisition:

[0053] After the control system is powered on, it collects real-time temperature data from all sensors.

[0054] If it is a cold start (long-term shutdown), the temperature of each area will differ after natural cooling.

[0055] Data records are linked by timestamps to form an initial temperature state matrix;

[0056] Construct a temperature distribution model:

[0057] Based on the collected data, a thermal distribution map of the mold is constructed through interpolation or thermal field simulation.

[0058] Finite element modeling, spatial thermal network method, and other methods can be used to quantify the thermal inertia and heat capacity of each region;

[0059] The system establishes a "mold region-temperature response relationship" for subsequent temperature rise prediction and differentiated current limiting control;

[0060] Integrating historical operational data or model training (such as machine learning) can further enhance model accuracy and responsiveness.

[0061] To achieve uniform temperature control and equipment protection during the initial heating phase of the injection mold, this invention proposes a soft-start control method for intelligent temperature control of the injection mold, specifically designed with a modeling and analysis mechanism for the thermal response characteristics of each heating zone. This method optimizes the power distribution strategy during the soft-start process by collecting historical temperature data, establishing a thermal response model, and assessing thermal coupling characteristics and control deviation risks, thereby improving temperature control accuracy and system stability.

[0062] In heating control systems, the temperature rise response of different heating zones varies significantly due to differences in structure, materials, and heat dissipation environment. To model and quantify this difference, this invention proposes the concept of a "thermal inertia score".

[0063] First, during multiple typical heating cycles, the system collects the temperature rise curves of each heating zone in the mold, i.e., the trajectory of temperature change in each zone per unit time. This data is used to analyze the time required for each zone to rise from the initial temperature to the target temperature.

[0064] Subsequently, given the known heating power input, the temperature rise response time per unit power is calculated. Even with a heating power of 1 watt, the time required for each heating zone to rise from the initial temperature to the set temperature is used as the "thermal response delay index" for that zone.

[0065] Next, we introduce the regional heat capacity parameter. This parameter can be calculated from the volume of the heated region, the specific heat capacity of the material, and the density, representing the region's heat storage capacity. By combining the thermal response delay index with the heat capacity parameter, we can construct a thermal inertia score. The larger the score, the slower the response and the stronger the inertia of the region.

[0066] To further identify potential control risks, this method sets a thermal inertia scoring threshold. If the score of a certain region exceeds this threshold, it is considered to have a thermal hysteresis risk and is marked as a "thermal hysteresis region." In subsequent soft-start control, a more conservative power limiting parameter is set for this region to prevent it from being prematurely heated to an excessively high temperature due to hysteresis, thus avoiding local overheating or control errors.

[0067] To improve the predictive capability of temperature control strategies, this invention proposes to input the current initial temperature and the target set temperature into the prediction model, and to predict the temperature rise trend in different regions through weighted analysis.

[0068] The prediction model is a multi-factor weighted model, and its weighting factors include:

[0069] Thermal inertia score: used to reflect the rate of thermal response in each region;

[0070] Regional thermal coupling anomaly: used to indicate the intensity of thermal interference between this region and other regions;

[0071] Historical control deviation frequency: used to measure the stability of deviation in the region during past control.

[0072] In practice, the model assigns weights to these factors and extrapolates the evolution of the temperature difference between the initial and target temperatures, predicting the time required for heating and the rate of temperature rise. This predicted value is used to generate a temperature rise trend curve, which is compared with the target temperature rise curve. If the deviation exceeds the allowable range, the system proactively adjusts the soft-start current ratio or the slope of the heating curve to correct the temperature rise trend.

[0073] There are often heat conduction paths between different heating zones in a mold, meaning that heating up one zone can affect temperature fluctuations in adjacent zones. This invention designs a calculation method to identify such "thermal coupling" relationships and uses their fluctuation anomalies to identify potential system interference or structural coupling risks.

[0074] The specific method is as follows:

[0075] First, temperature sensors are installed in multiple preset heating zones of the mold to collect real-time temperature change data during the heating process, forming a temperature time series for multiple time periods. For example, the temperature value of each zone is recorded every second, and continuous sampling is performed for several minutes to obtain a detailed temperature change curve.

[0076] Then, select any two regions and analyze whether there is synchronicity in the changing trends of their temperature sequences. The covariance method is used, which involves observing the degree to which the temperatures of the two regions rise or fall simultaneously over a certain period of time and calculating their covariance value as a measure of thermal coupling strength.

[0077] For each region, the sum of the absolute values ​​of its covariance with all other regions is calculated to obtain its "total thermal coupling degree" index. Then, the average value of the total thermal coupling degree of all regions is calculated, and the deviation between the thermal coupling degree of each region and the average value is analyzed. If the deviation exceeds a set threshold, the region can be determined to have thermal coupling anomalies and marked as a "thermal coupling degree anomaly region".

[0078] This judgment method allows the system to identify sensitive areas that may be significantly affected by neighboring areas, and then provide them with key protection in soft-start control.

[0079] To further improve the reliability of the control strategy, this invention introduces a historical control deviation frequency parameter, which is used to evaluate the probability of a significant deviation occurring in a certain region in multiple past control operations.

[0080] The evaluation steps are as follows:

[0081] In multiple historical soft-start cycles, the difference between the actual temperature and the target temperature in each region was recorded;

[0082] Set a temperature deviation tolerance threshold, such as ±2 degrees Celsius;

[0083] If the difference between the actual temperature and the target temperature exceeds this threshold, it is considered that a "deviation event" has occurred.

[0084] The proportion of time points in each cycle that are deviation events is statistically analyzed and defined as the "regional deviation frequency" for that cycle.

[0085] The "historical control deviation frequency" is obtained by averaging the deviation frequencies of multiple historical periods.

[0086] The higher the frequency, the greater the error in the temperature control model for that region, or the stronger the influence of disturbances. In subsequent prediction models, this region can be weighted or the feedback sensitivity can be increased to intervene and control earlier, thereby improving the stability of temperature control.

[0087] This invention constructs a soft-start control method with learning capabilities and adaptive adjustment functions through the above four core data modeling and fusion judgment mechanisms. It is particularly suitable for injection molding systems with high requirements for temperature control stability and complex mold structures. Compared with existing control methods that rely on single-point temperature feedback, this method can effectively identify and suppress thermal hysteresis regions and thermal coupling errors, thereby improving the consistency of mold heating and the quality of product molding.

[0088] Traditional soft-start control for mold heating typically employs a uniform limiting strategy, setting the same current limit or power slope for all heating zones. However, in actual mold applications, different zones vary in structure, materials, heat capacity, and heat dissipation environment, resulting in inconsistent thermal response characteristics.

[0089] For example, the mold cavity area typically dissipates heat quickly but responds slowly, while the nozzle or sprue area, due to its small size and rapid material heat transfer, heats up significantly faster than other areas. If uniform limiting parameters are used, it will inevitably cause uneven heating, local overheating, or insufficient heating, which may lead to mold damage or fluctuations in molding quality in severe cases.

[0090] Therefore, it is necessary to set differentiated soft-start parameters based on the thermal characteristics of each heating zone to achieve regional-level heating strategy optimization.

[0091] The specific implementation steps of the differentiated limiting strategy include:

[0092] Obtain the thermal inertia score (representing the speed of thermal response) for each heating zone;

[0093] Obtain regional thermal coupling anomalies (representing the strength of interference from neighboring regions);

[0094] Obtain the historical control deviation frequency (representing the temperature rise stability and model fit).

[0095] For regions with slower response and higher inertia scores (such as mold cavities), set higher initial power or current limits to compensate for thermal response hysteresis;

[0096] For areas with fast response or significant thermal coupling effects (such as nozzles), set a lower limit to prevent the temperature from rising rapidly;

[0097] For regions with high deviation frequencies, set a gradual rise curve or multi-level limiting strategy to improve the controllability of temperature rise;

[0098] For example, in a system with a total rated power of 100%, the power limit during the soft start phase is allocated as follows: 60% for the cavity area, 30% for the material channel area, and 10% for the nozzle area.

[0099] The limiting parameters are not statically set, but dynamically adjusted based on real-time temperature feedback and prediction deviations;

[0100] The controller monitors the difference between the current rate of temperature change in each region and the output of the prediction model. If a region heats up too quickly, its limit is tightened appropriately.

[0101] If the temperature rise in a certain area is significantly lagging and there is no risk of thermal coupling, the limit can be temporarily relaxed to increase the temperature rise rate.

[0102] Using a PID controller in conjunction with an adjustable phase controller (such as SCR or Triac), the heating power can be precisely controlled by dynamically adjusting the conduction angle;

[0103] Alternatively, the heating circuit on-time can be controlled via PWM to achieve high-resolution power limiting regulation;

[0104] All control parameters are controlled by the soft-start control algorithm module deployed in the temperature control motherboard or industrial-grade PLC.

[0105] For example, in an automotive headlight injection molding system, the cavity area has a large mass and slow thermal response, while the nozzle area is a small, metal cone. The system employs the differentiated limiting strategy of this invention, allowing the cavity area to gradually increase its power from 50% to 90% from the start, while the nozzle area remains at no more than 30%. Furthermore, the heating rate is corrected in real-time by a predictive model, ultimately achieving overall mold temperature difference control within ±1.5℃, significantly superior to traditional uniform heating methods.

[0106] During the heating process, this method collects the temperature values ​​of each heating zone in real time and calculates the rate of temperature change per unit time as the actual temperature rise rate. Simultaneously, the system calls upon a pre-built prediction model to generate a predicted temperature rise trajectory for each zone based on input parameters such as initial temperature, thermal inertia score, and historical deviation.

[0107] Subsequently, at each time step, the actual temperature rise rate is compared with the predicted temperature rise rate, and the difference between them is calculated, i.e., the temperature rise rate error. This error value not only reflects the prediction accuracy but also serves as an immediate feedback basis for the system's state stability. The set of error data formed within consecutive time steps constitutes the "temperature rise rate error sequence," which is used for subsequent modeling and control decisions.

[0108] This method innovatively employs a dynamic Bayesian network (DBN) algorithm to probabilistically model the temperature rise rate error sequence in order to determine whether there is a high risk of deviation in the future, thereby providing prior decision-making for control strategies.

[0109] The temperature rise rate error at each time point is denoted as Et, which is the difference between the current predicted value and the actual temperature rise rate.

[0110] At the same time, the "hidden state" St at each moment is defined to represent the deviation level of the current temperature control of the system, which is divided into several levels such as "normal", "slight deviation" and "serious deviation".

[0111] Set St only to the previous moment It is relevant and satisfies the Markov assumption. That is:

[0112] The transition probability is It can be learned from historical training data.

[0113] The observation model is used to describe the observation error Et that the system may produce in a certain hidden state St. It is usually modeled as a Gaussian distribution; that is, a certain state corresponds to a range of mean and standard deviation of error.

[0114] Using the prior probability of the state at the previous time step The state transition probability and observation probability are calculated using Bayesian inference or particle filtering algorithms after observing the error Et at the current time t. This probability is denoted as the temperature control deviation risk value Rt+k. If this value is greater than 0.7 (i.e., a 70% risk probability), the system is considered to have a significant deviation trend.

[0115] Once the prediction module determines that there is a risk of a deviation in the heating trend in a certain area, this method achieves dynamic optimization of power output by finely adjusting the heating control parameters.

[0116] The system first determines the adjustment priority of each region based on the magnitude of the deviation risk value. Regions with higher risk values ​​are adjusted first, so as to concentrate resources on controlling key areas of temperature rise.

[0117] If the area uses PWM (Pulse Width Modulation) to control the electric heating element, the system calculates the PWM duty cycle adjustment increment based on the direction and magnitude of the current error: if the actual temperature rise is slower than predicted, the duty cycle is increased; if the actual temperature rise is too fast, the duty cycle is decreased; the adjustment step size is limited by the set maximum variation range (such as maximum ±10% / cycle) to prevent temperature control oscillation or heating instability.

[0118] If the region employs silicon controlled rectifier (SCR) or thyristor phase control technology, power regulation is achieved by changing the conduction angle. A larger conduction angle results in a longer heating time and higher average power; conversely, a smaller conduction angle results in a shorter heating time and higher average power. Based on the predicted risk, the controller appropriately shortens or lengthens the conduction angle to bring the actual heating rate closer to the predicted target.

[0119] All PWM duty cycle or conduction angle adjustments are completed within the system control cycle, forming a closed-loop control circuit. This circuit is linked with the DBN prediction module, enabling the entire temperature control system to achieve intelligent control logic of "real-time monitoring → dynamic prediction → rapid response".

[0120] During the soft-start phase, due to issues such as uneven thermal inertia and structural thermal coupling in the initial heating stage, the system typically uses power limiting (limiting the PWM duty cycle or conduction angle) to gradually increase the temperature and prevent local overheating or system current surges. However, if power limiting control is maintained after the mold temperature approaches the set target value, it may lead to insufficient heating and slow temperature control response, thereby affecting production efficiency and mold temperature uniformity.

[0121] Therefore, when the temperature reaches a specific "target threshold", the system needs to remove the amplitude limit and instead rely on PID closed-loop control to stably maintain the set temperature, thereby achieving higher precision temperature control.

[0122] The system typically sets a "target temperature judgment threshold," which can be any of the following forms:

[0123] Absolute temperature difference determination: The actual temperature reaches a certain percentage of the target set value (such as 95%, 98%, 100%).

[0124] Temperature rise rate determination: When the temperature rise rate decreases significantly and approaches zero, it indicates that the system is close to steady state;

[0125] Temperature stability determination: The temperature fluctuation range is less than the set range (e.g., ±0.5°C) within a certain period of time (e.g., 30 seconds).

[0126] Comprehensive condition judgment: Combining the above conditions to form a stability judgment model, further avoiding misjudgment.

[0127] Once the above conditions are met, the controller will trigger the "release limit" flag.

[0128] Removing the limit includes: canceling the soft-start power limit parameter; allowing the PWM duty cycle or conduction angle to return to the normal adjustable range (e.g., 0–100%); and the system entering a dynamic output response state with full power regulation capability.

[0129] The PID controller is activated, using the set temperature as the target value and the current temperature as the feedback value, to calculate the error. Based on the error, the proportional (P), integral (I), and derivative (D) outputs are calculated in real time. The output control signal drives the heating power adjustment, so that the actual temperature accurately tracks the set value.

[0130] The system continuously operates PID control to dynamically respond to external disturbances. If the temperature in a certain area deviates from the target, the PID controller responds immediately and adjusts the power output to restore temperature balance. The control objective is to maintain temperature stability with minimal energy consumption, and the temperature fluctuation range is controlled within ±0.3°C.

[0131] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0132] It should be understood that the term "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone, where A and B can be singular or plural. Additionally, the character " / " in this document generally indicates an "or" relationship between the preceding and following related objects, but it may also indicate an "and / or" relationship; please refer to the context for specific understanding. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0133] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A soft-start control method for intelligent temperature control of injection molds, characterized in that: include: Initial temperature data for multiple regions of the mold, including the mold cavity, nozzle, and material channel, are acquired, and a temperature distribution model is constructed. Based on the temperature distribution model and the preset thermal inertia score, the temperature rise rate of different regions is predicted, and regions that may have thermal hysteresis are identified. The calculation of the thermal inertia score includes: collecting historical temperature rise curves of each heating area of ​​the mold; calculating the temperature rise response time per unit power for each area as a thermal response delay index; Based on the thermal response delay index and the regional heat capacity parameters, the thermal inertia score is calculated; if the thermal inertia score of the heating area is higher than the preset threshold, it is marked as a thermal hysteresis area and a limiting strategy is adapted in the soft start control. Predicting the temperature rise rate in different regions includes: inputting the current initial temperature and the target set temperature into the prediction model; the prediction model adopts a multi-factor weighting method, and the weighting factors include thermal inertia score, regional thermal coupling anomaly value and historical control deviation frequency, to calculate the expected temperature rise time and rate; The method for obtaining the regional thermal coupling degree anomaly value is as follows: temperature change data are collected synchronously in multiple preset heating areas to construct a temperature time series matrix for multiple time periods; the temperature covariance between any two areas is calculated based on the temperature series, and the total thermal coupling degree of each area is solved accordingly; the global average value of the thermal coupling degree of all areas is calculated, and the thermal coupling degree deviation value of each area is calculated as the regional thermal coupling degree anomaly value. During the soft start phase, differentiated current or power limiting parameters are set for each heating zone to form a dynamic soft start strategy at the zone level. Real-time temperature feedback data from various regions is collected, and power output is dynamically adjusted by combining the predicted temperature rise trajectory with the actual temperature rise rate. The dynamic power output correction includes: real-time acquisition of temperature feedback values ​​from each heating zone, calculation of its temperature change rate per unit time as the actual temperature rise rate; calling the temperature rise trajectory output by the prediction model, and calculating the error between the prediction and the actual temperature rise rate at each time step to construct a temperature rise rate error sequence; based on the temperature rise rate error sequence, performing probabilistic modeling on the difference between the predicted and actual temperatures to estimate the future temperature control deviation risk value; if the temperature control deviation risk value exceeds the risk threshold, dynamically correcting the PWM duty cycle or conduction angle of the heating zone to achieve real-time optimization of power output; Once the target temperature threshold is reached, the limiting is lifted, and the system switches to PID closed-loop control to achieve uniform and stable temperature control of the mold.

2. The soft-start control method for intelligent temperature control of injection molds according to claim 1, characterized in that: The method for obtaining the historical control deviation frequency is as follows: record the instantaneous deviation between the actual temperature and the target temperature of each heating zone in multiple historical soft-start cycles; Set a deviation tolerance threshold, and statistically analyze the proportion of time points in each soft-start cycle where the deviation exceeds the threshold. Define this as the regional deviation frequency. Summarize the deviation frequencies of multiple cycles and take their average to form the historical control deviation frequency.

3. The soft-start control method for intelligent temperature control of injection molds according to claim 1, characterized in that: The estimated future temperature control deviation risk values ​​include: The predicted temperature rise rate and the actual temperature rise rate of each heating zone are collected at multiple time steps, and the difference between the two is calculated to form a temperature rise rate error sequence. Using the error value as an observable variable, a dynamic Bayesian network model is constructed. The hidden state at each time step is defined as the control deviation state, which depends only on the hidden state at the previous time step, and a state transition probability relationship is established. Based on historical control data or real-time observation sequences, the state transition probability and observation probability distribution are estimated using the expectation-maximization algorithm or Bayesian filtering method. At any given moment, based on the known error sequence, infer the probability that the system will be in a high deviation state in the future, and use this as the future temperature control deviation risk value.

4. The soft-start control method for intelligent temperature control of injection molds according to claim 3, characterized in that: The dynamic correction of the PWM duty cycle or conduction angle of the heating area includes: The adjustment priority of each heating zone is determined based on the predicted deviation risk value, with priority given to adjusting zones with a risk probability higher than the threshold. If the heating area uses PWM control, the adjustment increment is calculated based on the current error direction and amplitude, and the PWM duty cycle is increased or decreased to refine the power output. The change in duty cycle is limited by the preset maximum adjustment step size. If the heating zone uses a silicon controlled rectifier or thyristor control method, the conduction angle can be adjusted by phase control to shorten or extend the conduction time, thereby changing the average power in a single cycle.

Images