Cooling system control method and device for precision injection mold

Abstract

The application relates to the technical field of injection molding process control, and discloses a cooling system control method and device for a precision injection mold. The method comprises the following steps: acquiring multi-source interference data in a workshop environment and injection equipment, and constructing a preliminary interference data set; performing weighted fusion processing on the data set to generate a weighted interference information matrix; analyzing the correlation of interference factors based on the matrix and determining a key interference influence vector; taking the key interference influence vector as an optimization input, cooperatively optimizing cooling parameters under the cooling balance constraint condition among multiple molds, and generating a cooling parameter candidate set; performing adaptability screening and filtering on the candidate set to obtain a refined parameter set; implementing cooling control based on the refined parameter set and evaluating the cooling effect, and generating a final adjustment instruction when a stable threshold is not reached; updating the cooling parameter configuration according to the adjustment instruction to form a global multi-mold cooperative cooling parameter regulation scheme. The application improves the balance and stability of the multi-mold cooling process.

Description

Technical Field

[0001] This application relates to the field of injection molding process control technology, and in particular to a cooling system control method and device for precision injection molds. Background Technology

[0002] Injection molding is widely used in the mass production of precision plastic products. The mold cooling system, as a crucial component of the injection molding cycle, directly impacts the product's molding quality, dimensional accuracy, and production efficiency. Especially in precision injection molding, the uniformity of temperature distribution during the mold cooling stage determines the product's shrinkage behavior and residual stress level. Improper cooling control can easily lead to problems such as warping, shrinkage marks, and decreased dimensional consistency.

[0003] As injection molding manufacturing moves towards higher precision, higher efficiency, and parallel production of multiple molds, existing mold cooling control technology has evolved from traditional experience-based adjustment and independent temperature control for a single mold to closed-loop control based on sensor monitoring. Existing technologies typically control mold temperature by adjusting coolant flow rate, cooling water temperature, or cooling time, but these are mostly focused on localized adjustments under single-mold operating conditions.

[0004] In a multi-mold collaborative production environment, multiple sources of interference, such as fluctuations in workshop ambient temperature, changes in cooling channel thermal resistance due to equipment aging, differences in mold surface temperature gradients, and batch-specific characteristics of raw materials, can dynamically affect the cooling process and lead to inconsistent cooling states among different molds. Traditional control methods struggle to effectively model and analyze these multi-source interferences and lack collaborative optimization and adaptive adjustment mechanisms for balancing cooling among multiple molds, easily resulting in insufficient matching of cooling parameters and a decline in overall cooling stability.

[0005] Therefore, how to achieve coordinated control and adaptive parameter regulation of the cooling process of precision injection molds under multi-source interference conditions has become a key technical problem that urgently needs to be solved in this field. Summary of the Invention

[0006] To address the problems of poor cooling uniformity, difficulty in quantifying the influence of multiple sources of interference, lack of collaborative optimization of cooling parameters, and insufficient dynamic adjustment capability in existing precision injection mold cooling control methods under multi-mold parallel production conditions, this application provides a cooling system control method and device for precision injection molds to achieve collaborative regulation and stable control of the cooling process of multiple molds, thereby improving the uniformity and adaptability of the cooling process.

[0007] In a first aspect, this application provides a cooling system control method for precision injection molds, the method comprising:

[0008] S1. Obtain multi-source interference data affecting the mold cooling process in real time from the workshop environment and injection equipment of precision injection molding production, and construct a preliminary interference dataset;

[0009] S2. Perform weighted fusion processing on the preliminary interference dataset to obtain a weighted interference information matrix;

[0010] S3. Based on the weighted interference information matrix, analyze the correlation between various interference factors and their impact on the cooling effect to determine the key interference influence vector;

[0011] S4. Using the key interference influence vector as the optimization input and achieving cooling balance among multiple molds as the constraint, the cooling parameters of the mold cooling system are collaboratively optimized to generate an optimized set of cooling parameter candidates.

[0012] S5. Perform adaptability screening and filtering on the candidate set of cooling parameters, eliminate parameter combinations that do not meet the current production conditions, and obtain the set of refining parameters.

[0013] S6. Implement cooling control based on the refining parameter set, evaluate the thermal balance state and cooling effect of the mold cooling process, and generate the final adjustment command when the evaluation result does not reach the preset stability threshold.

[0014] S7. Update the cooling parameter configuration according to the final adjustment instruction, and generate a cooling parameter control scheme for global multi-mold collaborative cooling control.

[0015] Secondly, this application provides a cooling system control device for precision injection molds, the device comprising:

[0016] The data acquisition module is used to acquire multi-source interference data affecting the mold cooling process in real time from the workshop environment and injection equipment of precision injection molding production, and to build a preliminary interference dataset.

[0017] The data fusion module is used to perform weighted fusion processing on the initial interference dataset to obtain a weighted interference information matrix;

[0018] The interference analysis module is used to analyze the correlation between various interference factors and their impact on the cooling effect based on the weighted interference information matrix, and to determine the key interference impact vector.

[0019] The collaborative optimization module is used to collaboratively optimize the cooling parameters of the mold cooling system with the key interference influence vector as the optimization input and the constraint of achieving cooling balance among multiple molds, and generate an optimized candidate set of cooling parameters.

[0020] The parameter filtering module is used to perform adaptability screening and filtering on the candidate set of cooling parameters, eliminate parameter combinations that do not meet the current production conditions, and obtain the refined parameter set.

[0021] The effect evaluation module is used to implement cooling control based on the refining parameter set, evaluate the thermal balance state and cooling effect of the mold cooling process, and generate the final adjustment instruction when the evaluation result does not reach the preset stability threshold.

[0022] The global control module is used to update the cooling parameter configuration according to the final adjustment command and generate a cooling parameter control scheme for global multi-mold collaborative cooling control.

[0023] Compared with the prior art, the beneficial effects of the technical solution of this application are at least as follows:

[0024] 1. By collecting multi-source interference data such as workshop environment, equipment aging status, mold surface temperature gradient, and batch characteristic changes of raw materials, and constructing a weighted interference information matrix, unified modeling and credibility fusion of interference factors in the cooling process were achieved, improving the accuracy of interference analysis.

[0025] 2. Based on the correlation analysis of interference factors, key interference influence vectors are extracted, and the cooling parameters are collaboratively optimized under the constraint of multi-mold cooling balance. This can effectively improve the consistency and global balance of cooling parameter configuration under the condition of parallel production of multiple molds.

[0026] 3. By adaptively screening and filtering the candidate optimization parameters, and combining the cooling effect evaluation and stability threshold judgment to generate the final adjustment command, the closed-loop adaptive update of the cooling control parameters is realized, which enhances the system's response to dynamic operating condition changes and long-term operational stability.

[0027] 4. The final global multi-mold collaborative cooling parameter control scheme can reduce defects such as warping and dimensional deviations in precision injection molded products caused by uneven cooling, thereby improving product quality and production efficiency. Attached Figure Description

[0028] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a flowchart of the cooling system control method for precision injection molds according to this application;

[0030] Figure 2 This is a schematic diagram of multi-mold thermal flux coupling in an embodiment of this application;

[0031] Figure 3 This is a schematic diagram of the particle swarm optimization convergence curve in an embodiment of this application;

[0032] Figure 4 This is a schematic diagram comparing the heat flux density before and after optimization in an embodiment of this application.

[0033] Figure 5 This is a schematic diagram comparing the effect of the proposed solution with that of a traditional fixed-base cooling solution;

[0034] Figure 6 This is a schematic diagram of the cooling system control device for precision injection molds according to this application. Detailed Implementation

[0035] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms “comprising” or “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0036] For ease of understanding, the specific process of the embodiments of this application is described below. Figure 1 The diagram shows a flowchart of a cooling system control method for precision injection molds provided by the present invention. The flowchart specifically includes the following steps:

[0037] S1. Obtain multi-source interference data affecting the mold cooling process in real time from the workshop environment and injection equipment of precision injection molding production, and construct a preliminary interference dataset.

[0038] In one specific embodiment, the process of performing step S1 may specifically include the following steps:

[0039] Sensors deployed in the workshop and equipment are used to collect real-time data on ambient temperature fluctuations, increased thermal resistance of cooling channels due to equipment aging, and temperature gradient values ​​on the mold surface.

[0040] Simultaneously monitor batch-to-batch heat capacity differences and melt flow index changes of raw materials used in injection molding, and determine the set of dynamic interference factors caused by changes in material properties.

[0041] By aligning and integrating ambient temperature fluctuation data, thermal resistance increase data, temperature gradient values, and dynamic interference factors over time, a preliminary interference dataset is generated.

[0042] Specifically, in a multi-mold parallel workshop environment for precision injection molding production, multiple sensors are deployed at key locations in the workshop and on injection equipment to collect multi-source interference data affecting the mold cooling process in real time. These sensors include temperature sensors distributed in different areas of the workshop, flow and pressure sensors installed at the inlet and outlet of the cooling channel, thermocouple arrays embedded in the mold cavity surface or near the cooling pipe, and infrared thermal imagers, thereby acquiring data on ambient temperature fluctuations, data on the increase in thermal resistance of the cooling channel caused by equipment aging, and temperature gradient values ​​on the mold surface.

[0043] The temperature sensor records the changes in the workshop ambient temperature once per second, and the recorded values ​​are expressed in degrees Celsius and are accompanied by a timestamp.

[0044] The flow sensor and pressure sensor simultaneously measure the inlet flow rate Q of the coolant in each mold cooling channel. in and export flow Q out (Primarily used to verify flow consistency and monitor potential leaks), and inlet temperature T in and outlet temperature T out Meanwhile, by embedding high-precision thermocouples or thin-film temperature sensors near the wall of the mold cooling channel or on the surface of the cavity, the mold wall temperature T is measured in real time. wall (Typically, 3 to 5 measuring points are arranged at multiple representative locations in each mold, and the average value is taken as T.) wall ), and then through the thermal resistance calculation formula: R th = (T wall - (T in + T out ) / 2) / [ρ × Q in × c× (T out - T in The real-time thermal resistance of the cooling channel is calculated, where ρ is the coolant density, c is the coolant specific heat capacity, and (T) in + T out ) / 2 represents the average temperature of the coolant in the channel (when the coolant maintains single-phase flow and the temperature rise is within the allowable range of the process, the arithmetic mean is used to approximate the average coolant temperature), (T out - T in Used to calculate the heat flow rate Q absorbed by the coolant. dot = ρ × Q in × c × (T out - T in Thermal resistance R thThis characterizes the overall thermal resistance from the mold wall to the coolant, reflecting the overall heat transfer efficiency of the cooling channel. The aforementioned thermal resistance is an equivalent convective thermal resistance index between the mold wall and the coolant, used to characterize the relative change in the heat transfer state of the cooling channel, rather than a complete thermal resistance model of the system. When equipment aging leads to scaling or deformation of the channel, the heat flow rate Q under the same heat load... dot Reduce or drive temperature difference (T) wall - (T in + T out Increasing ) / 2) makes the actual thermal resistance R th Gradually deviating from the initial calibration value R th0 (R) th0 Under standard test conditions—such as constant coolant inlet temperature, flow rate, and mold thermal load—calibrated by averaging multiple sets of data during the initial operation phase), the increase in relative thermal resistance ΔR = (R th - R th0 ) / R th0 Interference caused by aging of quantification equipment.

[0045] The temperature gradient value on the mold surface is calculated by collecting temperature data from multiple measuring points on the mold cavity surface using a thermocouple array. For example, nine measuring points are arranged on the mold surface to form a 3×3 grid, and the temperature difference ΔT between adjacent measuring points is calculated. ij = T i - T j The maximum value is taken as the gradient index, or the gradient vector is extracted after capturing the temperature distribution image of the entire mold surface using an infrared thermal imager. T, this gradient value directly reflects the difference in thermal stress distribution caused by uneven local cooling of the mold. These acquisition processes are performed simultaneously to ensure that all data carry a unified timestamp, which is based on the synchronous clock signal provided by the workshop central controller and marked with millisecond-level precision.

[0046] Simultaneously, batch-to-batch heat capacity differences and melt flow index (MFI) variations of raw materials used in injection molding are monitored by integrating online detection equipment into the raw material feeding system and the injection molding machine barrel. For example, a near-infrared spectrometer installed at the hopper outlet scans the spectral characteristics of raw material particles in real time and compares them with a pre-stored batch database. The database stores the standard ranges of heat capacity values ​​(Cp) and melt flow index (MFI) for different batches of raw materials. After the spectrometer outputs the absorption peak intensity, it is converted into the heat capacity deviation ΔCp = Cp of the current batch through a preset calibration model. current - Cp standard And the melt flow index deviation ΔMFI = MFI current - MFI standardThe calibration model is based on multiple linear regression. The inputs are the absorbance values ​​A1, A2, ..., An at multiple wavelengths, and the output is Cp or MFI. The regression coefficients are obtained through laboratory training on calibration data of known batches of samples. The training process includes collecting measured values ​​of heat capacity and MFI from more than 50 different batches of samples and their corresponding spectral data. The coefficient matrix is ​​solved using the least squares method, ensuring a prediction error of less than 2%. Furthermore, during the screw advance process of the injection molding machine, the MFI deviation is indirectly verified by monitoring the melt viscosity change using a torque sensor. A decrease in the melt flow index leads to an increase in screw torque, thus forming a dynamic set of interference factors. This set includes ΔCp, ΔMFI, and the viscosity interference factor η = ΔTorque / Torque, calculated from the torque deviation ΔTorque. nominal These raw material-related data are also timestamped to keep them synchronized with the sensor data.

[0047] During the initial interference dataset generation process, time alignment is performed by aligning all data sources within the same time window, using the central controller's timestamp as a reference. For example, a 10-second acquisition cycle is used to aggregate the latest data points from all sensors within that cycle. If a sensor's data is missing, it is filled with the value from the previous cycle or marked as invalid. The integration process then organizes this data into a structured matrix, where rows correspond to time series points and columns correspond to different interference types. For example, the first column represents the ambient temperature fluctuation value T. env The second column shows the relative increase in thermal resistance ΔR, and the third column shows the maximum temperature gradient. T max The fourth column is the heat capacity deviation ΔCp, the fifth column is the melt flow index deviation ΔMFI, and the sixth column is the viscosity interference factor η, thus forming the preliminary interference dataset Dpreliminary = [Tpreliminary]. env , ΔR, T max [, ΔCp, ΔMFI, η;…]. This dataset is stored in the industrial controller's memory as a time-series table or a multidimensional array, facilitating direct access in subsequent steps.

[0048] This solution combines the synchronous deployment of multiple sensors with online detection equipment to achieve comprehensive coverage and real-time performance of interference data. This ensures that the initial interference dataset reflects all major interference sources in the actual production process, thus providing a reliable data foundation for subsequent interference correlation analysis and collaborative optimization. It avoids lag in cooling parameter adjustment due to missing or delayed data, improves the accuracy and completeness of interference perception, and enables early detection of potential thermal imbalance risks in multi-mold parallel precision injection molding scenarios.

[0049] S2. Perform weighted fusion processing on the preliminary interference dataset to obtain the weighted interference information matrix.

[0050] In one specific embodiment, the process of performing step S2 may specifically include the following steps:

[0051] Based on the freshness of the acquisition timestamp of each interference data in the preliminary interference dataset and the calibration reliability of the corresponding sensor, a corresponding weight coefficient is assigned to each interference data.

[0052] The weighted interference information matrix is ​​generated by weighting the various interference data in the preliminary interference dataset based on the weighting coefficients.

[0053] Specifically, the initial interference dataset contains interference data from multiple sources, which differ in collection frequency, accuracy, and timeliness. Direct use of these data can lead to subsequent analysis being sensitive to noise or over-reliance on outdated data. Therefore, weighted fusion is used to improve the overall credibility and consistency of the dataset, thereby obtaining a weighted interference information matrix.

[0054] The freshness of the acquisition timestamp for each piece of interfering data is quantified by the difference between the current processing time and the data acquisition timestamp. The time delay between the current central controller time and the acquisition timestamp is calculated using the current central controller time as the base time. For example, the freshness sub-weights are calculated using an exponential decay method, with the formula: W f,k = exp(-α × δt_k), where α is the attenuation coefficient (e.g., 0.05), δt_k is the time delay of the k-th data point, and W f,k Let be the freshness sub-weight for the k-th data point. Based on this, newer data is given higher weights, while data older than a certain time period has its weights rapidly reduced, thus avoiding interference from outdated data in the fusion results.

[0055] The calibration reliability of the corresponding sensor is determined by pre-assigning a fixed reliability sub-weight W based on the sensor's factory calibration records and periodic calibration history. r For example, the reliability sub-weight W of the workshop ambient temperature sensor. r,T The value is set to 0.98 because these sensors are calibrated with high-precision platinum resistance thermometers and are less susceptible to environmental interference; the reliability sub-weight W for the cooling channel flow sensor and pressure sensor is also considered. r,Q The value is set to 0.92 because slight drift may occur during long-term operation; the reliability sub-weight W of the thermocouple array on the mold surface. r,G The reliability sub-weight W of the near-infrared spectrometer is set to 0.95. r,C The value is set to 0.90 because spectral scanning is slightly affected by particle uniformity. These reliability sub-weights are stored in the sensor configuration table of the industrial controller, loaded during system initialization, and updated after each sensor calibration.

[0056] The overall weighting coefficient is obtained by multiplying the freshness sub-weight and the reliability sub-weight, i.e., W.k = W f,k × W r W k Let W be the comprehensive weight coefficient for the k-th data point. This takes into account both the timeliness of the data and the inherent accuracy of the sensor itself. For each row of time series points in the initial interference dataset, the comprehensive weight of each column of data is calculated to form a weight vector W. row .

[0057] When performing weighted calculations on the various interference data in the initial interference dataset based on the comprehensive weighting coefficient, the interference data in each column is first weighted and averaged within a sliding time window. For example, the time window length is fixed at 60 seconds to cover sufficient production dynamics. All valid data points within the window are normalized according to the comprehensive weight and then summed. For instance, for the environmental temperature fluctuation column, there are n data points in the current window, and their weighted fusion value T... fused = (∑(W T,i2 ×T env,i2 )) / (∑W T,i2 ), where T env,i2 W represents the ambient temperature fluctuation value for the i2th data point. T,i2 This represents the comprehensive temperature weighting coefficient corresponding to the i2th data point. This operation merges multiple original data points within each time window into a representative data point for that window. After traversing all interference parameter columns and performing the same window-weighted averaging operation, a set containing fused values ​​of all interference types is obtained for each time window. Subsequently, these fused values ​​are normalized to eliminate differences in the dimensions of different interferences. For example, the min-max normalization method is used to map each fused value to the interval [0,1], thereby obtaining the normalized interference vector V_norm = [T_norm, ΔR_norm, ... T_norm, ΔCp_norm, ΔMFI_norm, η_norm].

[0058] The weighted interference information matrix M_weighted is constructed in the form of a time series, with the rows being the fused time points and the columns being the normalized interference components, i.e., M_weighted = [V_norm_1; V_norm_2; … ; V_norm_m], where m is the length of the fused time series. This matrix is ​​directly stored in the buffer of the industrial controller for subsequent interference analysis steps.

[0059] By combining the product of exponentially decaying freshness weights and fixed reliability weights, along with sliding window weighted averaging and normalization, the credibility and dimensionality of interference data are improved. This ensures that the weighted interference information matrix accurately reflects the true interference level under current production conditions, providing high-quality input for subsequent correlation analysis and key interference vector extraction, and avoiding misjudgments or optimization biases caused by low-quality data. This feature enhances the system's robustness to dynamic interference, and in multi-mold parallel precision injection molding scenarios, it can suppress the impact of sensor drift or temporary noise on cooling control.

[0060] S3. Based on the weighted interference information matrix, analyze the correlation between various interference factors and their impact on the cooling effect to determine the key interference influence vector.

[0061] In one specific embodiment, the process of performing step S3 may specifically include the following steps:

[0062] Extract data sequences of at least two interference factors from the weighted interference information matrix, and calculate the correlation strength between the interference factors through correlation analysis;

[0063] Based on the temperature gradient data sequence in the weighted interference information matrix, the comprehensive influence of each interference factor on the cooling state of multiple molds is calculated through a preset evaluation model.

[0064] Based on the correlation strength and the degree of comprehensive influence, the core interference factors that play a dominant role in the cooling effect are identified, and key interference influence vectors are generated.

[0065] Specifically, data sequences of at least two interfering factors are extracted from the weighted interference information matrix. These data sequences are time series vectors in the corresponding columns of the matrix. For example, the normalized sequence of ambient temperature fluctuation T_norm_seq = [T_norm_1, T_norm_2, ..., T_norm_m] and the normalized sequence of thermal resistance increase ΔR_norm_seq = [ΔR_norm_1, ΔR_norm_2, ..., ΔR_norm_m] are extracted, and then the correlation strength between the interfering factors is calculated through correlation analysis. For example, the correlation analysis uses the Pearson correlation coefficient to quantify the correlation strength between the interfering factors. The Pearson correlation coefficient ranges from [-1, 1], with positive values ​​indicating positive correlation and negative values ​​indicating negative correlation. The larger the absolute value, the stronger the correlation. For example, for the ambient temperature fluctuation and thermal resistance increase sequences, ρ is calculated. TR If |ρ TR A correlation of |>0.7 is considered strong because increased temperature in a workshop environment often accelerates scaling in channels, leading to increased thermal resistance. All interfering factors are combined pairwise to form the correlation strength matrix C_corr, where the element C_corr(j2,k) = |ρj2k | Store absolute values ​​to uniformly measure intensity. This matrix is ​​an N1×N1 symmetric matrix with its diagonals set to 0. N1 represents the total number of interfering factors.

[0066] The pre-defined evaluation model is used to objectively quantify the net impact (i.e., the overall impact) of each interfering factor on the temperature gradient, while controlling for the mutual influence of other interfering factors. Standardized coefficients based on multiple linear regression can be used as the impact score; that is, a multiple linear regression model is constructed with the temperature gradient sequence as the dependent variable and all other interfering factor sequences as independent variables. The model is solved using the least squares method or regularization methods (such as ridge regression to handle possible multicollinearity among factors). The absolute value of the standardized regression coefficient of each independent variable is defined as the overall impact score of that interfering factor. Standardized regression coefficients eliminate the influence of dimensions; their absolute values ​​directly reflect the relative contribution of that factor to the temperature gradient change when other interfering factors in the model remain constant, i.e., its independent overall impact.

[0067] In precision injection molding multi-mold cooling control, the temperature gradient sequence directly corresponds to the degree of uneven cooling across the molds and the risk of product defects. Simply comparing the correlation strength between disturbances only reveals synchronous changes and cannot distinguish their actual driving contribution to the cooling effect. Therefore, by calculating the comprehensive influence degree using the aforementioned model, an objective quantification of the influence score is achieved: after controlling for other factors, disturbances with a greater impact on the temperature gradient have higher priority. This helps avoid subjective misjudgments and improves the targeting of subsequent optimization steps.

[0068] Based on the correlation strength and overall impact, screening rules are set. For example, if the absolute value of the correlation coefficient between a certain interfering factor and the temperature gradient is greater than the threshold θ1 (e.g., 0.6), and its overall impact score is greater than the impact threshold (e.g., 1.2 times the average impact of all interfering factors), then it is identified as a core interfering factor. After traversing all interfering factors, a key interfering impact vector V_key is generated, with the vector dimension being the same as the total number of interfering factors. For components identified as core interfering factors, their values ​​are taken as the normalized value of the factor at the current processing time (from the latest row vector corresponding to the component in the weighted interfering information matrix). For non-core interfering factors, their components are set to 0. The length of this vector is fixed at N1 dimensions to match the number of columns in the matrix, facilitating direct input to subsequent optimization algorithms. V_key is stored in the controller buffer with a timestamp appended.

[0069] This solution quantifies the correlation strength using a correlation coefficient matrix and uses a temperature gradient sequence as a proxy for calculating the overall impact of cooling effects. A dual-threshold screening process generates key interference influence vectors, enabling the objective extraction and vectorization of dominant interference factors. This ensures that the key interference influence vectors centrally reflect the most significant interference sources affecting uneven cooling across multiple molds, providing refined input for subsequent particle swarm optimization and avoiding the computational burden and secondary noise interference associated with full-factor optimization. This feature improves the targeting and efficiency of interference analysis. In multi-mold parallel precision injection molding scenarios, it can quickly identify dominant events such as sudden changes in ambient temperature or a sharp increase in the thermal resistance of a specific mold. This allows the system to prioritize adjusting the zoned flow distribution, significantly reducing residual stress unevenness and warpage defects caused by localized overcooling or overheating, and improving thermal balance consistency and production pass rate.

[0070] S4. Using the key interference influence vector as the optimization input and achieving cooling balance among multiple molds as the constraint, the cooling parameters of the mold cooling system are collaboratively optimized to generate an optimized candidate set of cooling parameters.

[0071] In one specific embodiment, the process of performing step S4 may specifically include the following steps:

[0072] Using the key interference influence vector as the optimization input, and under the preset thermal balance constraint between multiple molds, the particle swarm optimization method is used to iteratively optimize the cooling parameters of the mold cooling system. The cooling parameters include at least the total flow rate of the coolant and the partition flow distribution parameters of the cooling channels of each mold.

[0073] In the particle swarm optimization process, a multi-mold thermal flux coupling model is established to simulate the dynamic adjustment process of heat flux density equalization adjustment and coolant flow independent zone control between molds, and to generate dynamic adjustment paths.

[0074] The cooling parameter combinations that satisfy the thermal balance constraints between multiple molds are extracted from the dynamically adjusted path to generate an optimized set of cooling parameter candidates.

[0075] Specifically, the key interference influence vector centrally reflects the normalized components and influence weights of the current dominant interference factors, which is used to indicate the direction and intensity of interference that has a significant impact on the cooling process of multiple molds under the current production conditions, thereby providing a decision-making basis for targeted adjustment of cooling parameters.

[0076] The optimization process first defines a cooling parameter vector X as the decision variable for particle swarm optimization. This vector X includes at least the total coolant flow rate Q_total and the partitioned flow distribution parameters of each mold cooling channel. For example, for a 4-mold parallel system, X = [Q_total, Q1, Q2, Q3, Q4], satisfying Q1+Q2+Q3+Q4= Q_total to ensure flow conservation. The value range of each variable is limited according to the rated capacity of the cooling equipment, pipeline characteristics, and safe operation requirements.

[0077] Subsequently, particle swarm optimization (PSO) is used for iterative optimization. For example, the particle swarm size is set to 50, the maximum number of iterations is 100, the inertia weight ω decreases linearly from 0.9 to 0.4, the cognitive coefficient c1 and social coefficient c2 are both set to 2.0, and each particle position P_n represents a set of candidate cooling parameter vectors. The particle velocity vector is initialized within a preset range and updated iteratively. The particle update process conforms to the conventional iterative rules of PSO, which will not be elaborated here.

[0078] In each iteration, an equivalent heat flow evaluation model is established to assess the thermal influence relationship between multiple molds under different combinations of cooling parameters, and the fitness value corresponding to the particle position is calculated. The equivalent heat flow evaluation model is not intended to precisely describe the three-dimensional transient heat transfer process inside the mold, but rather to quickly evaluate the impact of cooling parameter changes on the equilibrium trend of heat flow distribution across multiple molds during optimization. The model input includes at least the current cooling parameter vector X, the key disturbance influence vector, and the real-time measured mold wall temperature T. wall and the inlet and outlet temperatures of the coolant. For example, for the i1th mold, its equivalent heat flux density evaluation value q i1 It can be represented as: q i1 = h i1 × A i1 × (T wall,i1 - T avg,i1 ), where h i1 A is the equivalent convective heat transfer coefficient. i1 To achieve an effective heat exchange area, T avg,i1 = (T in,i1 + T out,i1 The equivalent convective heat transfer coefficient can be estimated using empirical correlations, experimentally calibrated models, or historical operating data. For example, the Dittus–Boelter correlation is used as an example, h i1 = Nu i1 × λ / D h Nu i1 = 0.023 × Re i1 0.8 ×Pr 0.3Re i1 = ρ × u i1 × D_h / μ,u i1 = Q i1 / (π × (D h / 2) 2 ) represents the flow velocity within the channel, Pr is the Prandtl number, λ is the thermal conductivity of the coolant, and D h denoted as the hydraulic diameter of the channel, and μ as the dynamic viscosity of the coolant.

[0079] To reflect the impact of key disturbances on the cooling adjustment strategy during the optimization process, a disturbance feedforward compensation mechanism is introduced. This mechanism maps the influence vector of key disturbances to an adjustment factor on the target heat flux evaluation value or control setpoint. This mapping relationship guides the optimization algorithm to adjust its search direction when significant disturbances exist. Essentially, it is a feedforward compensation strategy at the control layer, rather than a direct correction to the actual physical heat flux. For example, an equivalent heat flux evaluation target value q is defined for the i1th mold. i1,adjusted It can be represented as: q i1,adjusted =f(q i1 ,V_key_i1), where the function f(·) represents the interference mapping relationship, which can be in the form of proportional adjustment, bias adjustment or a combination thereof. In this embodiment, proportional adjustment is used as an example: q i1,adjusted = q i1× (1 + β × V_key_i1), where β is the disturbance sensitivity coefficient, used to adjust the response strength of the control strategy to changes in disturbance. Its value can be obtained through historical operating data calibration, experimental calibration, or online adaptive adjustment. V_key_i1 is a scalar of the key disturbance effect for the i1th mold, thereby simulating the disturbance effect of key disturbance on heat flux density. V_key_i1 can be synthesized in the following ways: global disturbance components that have approximately consistent effects on all molds (such as changes in ambient temperature, changes in the heat capacity or flow characteristics of raw materials) are directly involved in the synthesis; local disturbance components with different values ​​for different molds (such as changes in the thermal resistance of cooling channels, temperature gradients on the surface of molds) are synthesized using the normalized monitoring values ​​corresponding to the molds; each component is weighted and summed to form V_key_i1, where the weight of local disturbance components can be higher than that of global disturbance components to reflect the differences between molds. For example, weights are determined through offline calibration or online adaptive methods. Offline calibration, based on historical data or simulation models, analyzes the quantitative relationship between various disturbances and mold temperature gradients. Based on the significance of their impact on the differentiation of cooling states between molds, higher weights are assigned to local disturbances and lower weights to global disturbances. Online adaptive methods dynamically adjust weights according to real-time performance. For instance, initial weights can be set empirically and continuously monitored during operation. If a certain type of disturbance consistently becomes the main cause of thermal balance deviation, its weight is automatically increased slightly; conversely, it is decreased, allowing the weight allocation to dynamically match the main contradictions in actual production. Simultaneously, to reflect the thermal state coupling trend under multi-mold parallel production conditions during optimization, a thermal influence term q between molds is introduced. couple,i1 This thermal effect term characterizes the mutual influence of thermal states between molds due to factors such as spatial proximity, shared environment, and radiation. It is a linearized approximate coupling term, used only for evaluating the thermal balance trend during optimization, not for accurately describing the actual transient heat conduction process between molds. For example, this effect term can take the following approximate form: κ is the thermal influence coefficient, which is used to characterize the relative intensity of the mutual influence of thermal states between molds. Its value can be set by regression of historical operating data, offline thermal simulation calibration or based on experience of mold layout structure.

[0080] To simulate the effect of parameter adjustment during optimization iterations, the required mold wall temperature T for fitness calculation is determined. wall,i1 Instead of directly using real-time measurements, a simplified prediction model is employed. This model is based on the cooling parameter X represented by the current particle position and the steady-state temperature from the previous control cycle. It performs a first-order linear extrapolation or rapidly calculates the predicted temperature T using a simplified thermal resistance-thermal capacity network. wall,i1,est and use T wall,i1,est Alternative real-time measurement T wall,i1 To calculate qi1 and q couple,i1 This allows for self-consistent simulation evaluation within the optimization loop.

[0081] Based on the above equivalent heat flux evaluation value and the thermal influence term between molds, a fitness function for particle swarm optimization is constructed to measure the degree of uniformity of heat flux distribution among multiple molds. The formula is: F(X) = ∑|q i1,adjusted + q couple,i1 - q avg | , where q avg This represents the average of the equivalent heat flux evaluation values ​​for all molds under the current parameter combination. The fitness function aims to minimize the heat flux deviation among multiple molds. The key disturbance influence vector is not treated as an independent minimization object, but rather indirectly affects the optimization direction through the aforementioned disturbance mapping relationship and model parameter adjustment, thereby avoiding the optimization process from deviating from the actual disturbance compensation target.

[0082] During the particle swarm optimization iteration process, the dynamic adjustment process under independent control of coolant flow rate zones is simulated in real time using the aforementioned thermal-fluid coupling model. After each generation of particles is updated, the trajectory of the global optimal position G_best is recorded to form a dynamic adjustment path. This path is a sequence of changes from the initial random parameters to the current optimal parameters. Each node in the path corresponds to a set of cooling parameter combinations and the simulated heat flux density distribution of each mold. After the iteration ends or the early stopping condition is reached (no improvement in G_best for 20 consecutive generations), the cooling parameter combination that satisfies the thermal balance constraint between multiple molds is extracted from the dynamic adjustment path. The thermal balance constraint between multiple molds is defined as max(|q i1,adjusted - q avg |) < ε_q and max(|T) wall,i1 - T wall,avg |) <ε_T, where q avg and T wall,avg Let ε_q be the average value of all molds, and exemplarily, let q be the value of all molds. target 5% of the value, ε_T is taken as 2℃. The extraction method is to traverse all nodes of the path. If a node satisfies the constraints, it is included in the candidate set. At the same time, the top K1=10 nodes with the lowest fitness that satisfy the constraints are retained to increase diversity, thereby generating an optimized cooling parameter candidate set S_candidate. This candidate set is a set containing 10-20 sets of parameter vectors, each set with an additional corresponding simulated thermal balance index.

[0083] By employing the above method, particle swarm optimization dynamically searches the cooling parameter space under multi-mold thermal equilibrium constraints. It guides parameter adjustment by using key disturbance influence vectors, enabling the generated candidate cooling parameters to achieve overall equilibrium and stability of the multi-mold heat flux distribution under current disturbance conditions. This scheme avoids local optima and constraint violations, improving the synergy of multi-mold cooling control and its adaptability to changing operating conditions.

[0084] The multi-mold collaborative optimization process of the present invention is as follows: Figures 2 to 4 As shown, the process and effects of collaborative optimization are presented from different perspectives.

[0085] Figure 2 This diagram illustrates the heat flux coupling of multiple molds, showing the heat flux distribution when four molds are operating in parallel before optimization. As can be seen, there are significant differences in heat flux density among the molds in the initial state: molds A and C are in the high heat flux region (red arrows, heat flux values ​​of 120 W / m² and 150 W / m², respectively), mold B is in the medium heat flux region (orange arrow, 80 W / m²), and mold D is in the low heat flux region (green arrow, 60 W / m²). This uneven heat flux distribution is a direct reflection of the uneven cooling problem among multiple molds.

[0086] Figure 3 The convergence curve for particle swarm optimization demonstrates the iterative optimization process of the algorithm. The horizontal axis represents the number of iterations (0-100), and the vertical axis represents the fitness value (reflecting the heat flow deviation of multiple molds). The curve shows that as iterations proceed, the fitness value rapidly decreases from an initial value of approximately 95, converging to a stable value around the 60th iteration, after which it enters the stable convergence region. This proves that the particle swarm optimization algorithm used in this invention can converge effectively, providing a reliable optimization method for achieving balanced cooling of multiple molds.

[0087] Figure 4 The graph shows a comparison of heat flux densities before and after optimization, visually demonstrating the final effect of the synergistic optimization. Before optimization (red bars), the heat flux values ​​of each mold differed significantly, with a maximum difference of 90 W / m². After optimization (green bars), the heat flux values ​​of each mold tended to be consistent, with the maximum difference reduced to only 10 W / m². The graph also indicates the uniformity of each mold (Mold A: 99.5%, Mold B: 95.5%, Mold C: 94.5%, Mold D: 98.5%), with the overall uniformity improved to 97%. The orange dashed line in the graph represents the target uniformity value (100.5 W / m²), showing that the heat flux values ​​of each mold are close to this target value after optimization.

[0088] The above-described results demonstrate that, by using the key disturbance influence vector as the optimization input and employing the particle swarm optimization method to collaboratively optimize cooling parameters under the constraint of balanced cooling among multiple molds, this invention can effectively solve the problem of uneven heat flux distribution among multiple molds, achieve balanced control of heat flux density in each mold, and lay a solid foundation for subsequent refining parameter selection and closed-loop adaptive control.

[0089] S5. Perform adaptability screening and filtering on the candidate set of cooling parameters, eliminate parameter combinations that do not meet the current production conditions, and obtain the set of refining parameters.

[0090] In one specific embodiment, the process of performing step S5 may specifically include the following steps:

[0091] For each parameter combination in the candidate set of cooling parameters, determine whether it matches the batch heat capacity difference and melt flow index change of the current injection molding raw material, and eliminate mismatched parameter combinations.

[0092] For the matched parameter combinations, similarity comparison and filtering are performed based on the geometric feature parameters of each mold cavity, and parameter combinations with geometric similarity lower than the preset similarity threshold are eliminated;

[0093] For parameter combinations selected through geometric features, the consistency of cavity pressure distribution among multiple molds is verified, and parameter combinations that cause pressure distribution inconsistency to exceed the preset allowable range are eliminated.

[0094] Integrate the verified parameter combinations to generate a refined parameter set.

[0095] Specifically, S5 further constrains and filters the candidate set of cooling parameters obtained from the previous co-optimization step to ensure their engineering feasibility. This is to prevent parameter combinations that are only feasible in the sense of mathematical optimization but are difficult to operate stably or are prone to quality fluctuations in actual injection molding production conditions from being directly used for control execution. This achieves convergence from the feasible solution set to the stable application solution set.

[0096] Each parameter combination in the candidate set of cooling parameters includes at least the total coolant flow rate parameter and the flow distribution parameters for each mold or cooling zone, and may further include cooling loop switching parameters or temperature compensation parameters. For this input set, the screening process unfolds sequentially according to the logical order of material adaptation—structural adaptation—molding state adaptation, with each screening sub-step forming a sequential relationship, and the output of the previous sub-step serving as the input for the next sub-step.

[0097] During the material matching and screening stage, for each parameter combination in the candidate set of cooling parameters, it is determined whether the parameter combination matches the batch thermophysical characteristics of the current injection molding raw material. Specifically, the specific heat capacity deviation value ΔCp and melt flow index deviation value ΔMFI corresponding to the current raw material batch are obtained through the production management system or online detection device. For candidate parameter combinations, based on the specific heat capacity parameter of the current batch of raw materials, the estimated cooling heat load Qc within a unit molding cycle is calculated. For example, it can be calculated using the formula Qc = mp × Cp × ΔTp, where mp represents the mass of molding material per unit cycle in a single mold, Cp represents the specific heat capacity of the current batch of material, and ΔTp represents the temperature drop of the material from the molding temperature to the demolding temperature. Based on the specific heat capacity deviation value ΔCp, the adaptable heat load variation range corresponding to the candidate parameter combination is determined (this range is preset according to the rated heat exchange capacity and safety margin of the cooling system). When ΔCp exceeds the allowable heat capacity variation adaptation range of the parameter combination under the current coolant flow distribution conditions, or when the cooling heat load change caused by ΔCp causes Qc to exceed the heat range that the parameter combination can stably handle, the parameter combination is determined to be mismatched with the heat capacity difference of the current raw material batch, and the parameter combination is eliminated. It should be noted that the cooling heat load estimation formula in the material adaptation screening stage is used to quickly screen the adaptability of cooling parameter combinations under different raw material batch conditions, and its calculation result is used to determine whether the parameter combination is within the stable operating range. At the same time, when ΔMFI exceeds the preset allowable fluctuation range, and the corresponding parameter combination cannot compensate for the cavity filling and cooling unevenness caused by the change in melt fluidity through flow rate or cooling time adjustment, the parameter combination is also eliminated. For example, the specific process for determining whether there is a corresponding parameter combination that cannot compensate for the above problem is as follows: First, calculate the theoretical adjustment increment required to maintain the thermal balance of the cavity based on the melt flow index fluctuation ΔMFI; then, convert this increment into the corresponding cooling flow rate and cooling time parameters, and verify whether they meet the preset equipment constraints, including the upper limit of the rated flow rate of the chiller unit, the pressure resistance limit of the mold runner, and the production cycle limit. If the mapped parameters exceed the above constraint boundaries, or the residual thermal deviation after simulation prediction and adjustment is still higher than the preset precision machining threshold, it is determined that the quality fluctuation caused by the change in melt flowability cannot be offset by adjustment, and then the parameter combination rejection instruction is executed. Optionally, the determination process can also be implemented in other equivalent ways, such as calling the pre-stored material process window database for comparison; or performing rapid simulation prediction through a simplified filling-cooling coupling model; or performing logical judgment based on the historical rule base. Through the above processing, a subset of parameters that matches the current production conditions at the level of material thermal properties is obtained, and this subset serves as the input for subsequent geometric adaptation screening.

[0098] In the geometry matching screening stage, based on the parameter combinations selected through material matching, a similarity comparison is further performed using the geometric feature parameters of each mold cavity in the multi-mold system. The geometric feature parameters include at least the cavity volume, cavity surface area, average wall thickness, maximum wall thickness difference, and the average distance from the cooling channel to the cavity surface. The geometric feature parameters of each mold are then used to construct a feature vector G. i1 = [V i1 , S i1 , T avg,i1 , ΔT max,i1 D c,i1 The subscript i1 represents the mold number. After normalizing the feature vectors of multiple molds, the geometric similarity value between any two molds is calculated. For example, it can be expressed in the form of weighted Euclidean distance: G i1,k2 w represents the normalized eigenvalue of the i1th mold in the k2th geometric dimension. k2 This represents the weighting coefficient for the corresponding geometric dimension, which is preset based on the degree of influence of the geometric factor on cooling uniformity. Further, the average geometric similarity index Skeo is calculated for the entire mold group. When Skeo is lower than a preset geometric similarity threshold, it is determined that the current mold system has significant geometric differences, and the candidate parameter combination cannot meet the cooling requirements of molds with different geometric features, thus eliminating the parameter combination. When Skeo meets the threshold requirement, the parameter combination is retained and output to the next screening stage. Through this geometric adaptation screening process, it is possible to avoid incorrectly applying cooling parameter combinations that are only applicable to molds with consistent geometric heights to multi-mold systems with significant geometric differences, thereby solving the problem of uneven cooling among multiple molds.

[0099] In the pressure distribution consistency verification stage, the parameter combinations selected through geometric adaptation are used as input, and the cavity pressure distribution during the injection molding process is considered to verify the feasibility of the candidate parameter combinations at the molding process level. Specifically, in an offline simulation environment or online prediction model, the cavity pressure change process during the injection filling, holding, and cooling stages is simulated based on the parameter combinations, and the average pressure value of each mold cavity is obtained at multiple preset key time points. The pressure values ​​of each mold at the same time point are used to construct a pressure vector P. t = [p 1,t , p 2,t , …, p n,t And calculate the dispersion index of the pressure vector, for example, using the standard deviation. ,in, This represents the average pressure in each mold cavity at the same time point, where n represents the total number of molds participating in the coordinated cooling control. Furthermore, for multiple key time points, σ... tStatistical analysis is performed, and when the maximum or average value exceeds the preset allowable pressure dispersion threshold, it is determined that the parameter combination will lead to inconsistent pressure distribution between molds during the actual molding process, easily causing defects such as flash, shrinkage, or internal stress differences, and thus the parameter combination is eliminated. When the pressure dispersion index is always within the allowable range, the parameter combination is retained. Through this verification process, it is ensured that the cooling parameters are feasible at the thermal control level without causing unacceptable disturbances to the injection molding pressure field, thereby solving the coupling problem between cooling regulation and molding stability.

[0100] Preferably, to avoid the enormous computational burden of online simulation, the pressure distribution consistency verification stage can employ a rapid assessment rule base for pressure distribution consistency built upon historical production data and offline simulation results. This rule base records the corresponding safe threshold for cavity pressure distribution dispersion under different mold geometric features, material properties, and typical cooling parameter ranges. In specific implementation, based on the current mold geometric features, raw material batch characteristics, and key features of candidate parameter combinations (such as total flow range and flow ratio range of each mold), the rule base is queried to obtain the corresponding allowable pressure dispersion threshold σ_allowed. Statistical analysis is performed on σ_t at multiple key time points. When its maximum or average value exceeds the queried σ_allowed, it is determined that this parameter combination will lead to inconsistent pressure distribution between molds during the actual molding process, easily causing defects such as flash, shrinkage, or internal stress differences, thus eliminating the parameter combination.

[0101] After completing the material compatibility screening, geometric compatibility screening, and pressure distribution consistency verification, all verified parameter combinations are integrated to form a refined parameter set. Through multi-level screening and filtering, in multi-mold precision injection molding applications, the optimized cooling parameters can be limited to a parameter space that is compatible with the current raw material characteristics, mold structure features, and molding process state. This effectively solves the problem of insufficient parameter applicability caused by relying solely on the output of optimization algorithms, and improves the stability and reliability of the cooling control scheme in actual production.

[0102] S6. Implement cooling control based on the refining parameter set, evaluate the thermal balance state and cooling effect of the mold cooling process, and generate the final adjustment command when the evaluation result does not reach the preset stable threshold.

[0103] In one specific embodiment, the process of performing step S6 may specifically include the following steps:

[0104] Cooling control is implemented for the multi-mold cooling system based on the refining parameter set, and the operating status information of the cooling system is collected in real time. The operating status information includes at least the state of synchronous maintenance of thermal balance of multiple molds, the adaptive flow trigger record of abnormal blockage of cooling channels, and the execution status of pre-compensation for thermal expansion deformation of molds.

[0105] Based on the operating status information, the thermal balance and cooling effect of the mold cooling process are evaluated, and it is determined whether the evaluation results have reached the preset stability threshold.

[0106] If the evaluation results do not reach the preset stability threshold, a final adjustment instruction will be generated.

[0107] Specifically, based on the aforementioned refining parameter set, the cooling control system performs actual control of the multi-mold cooling system. During implementation, the refining parameter set serves as the target setting parameter for cooling control, providing the cooling system with control targets such as the setpoint for coolant flow rate, setpoint for coolant temperature, and zoned flow distribution ratio for each mold. The cooling control system adjusts the cooling execution unit according to the set parameters and in conjunction with real-time feedback information.

[0108] The cooling control does not infer the control quantity by calculating the coolant flow rate in real time. Instead, it uses the coolant flow rate and zone allocation ratio given in the refining parameter set as the control target input, and the control algorithm adjusts and executes the control based on the deviation. In one embodiment, for each mold's cooling channel, the control system dynamically adjusts the coolant flow rate and coolant temperature based on the mold's corresponding geometric characteristic parameters, equivalent heat flux evaluation value, and current cooling conditions. The adjustment process aims to converge the mold's equivalent heat flux evaluation value towards the target value corresponding to the refining parameter set, rather than calculating the coolant flow rate backward based on the inlet and outlet temperature difference. The cooling control algorithm minimizes the deviation between the actual equivalent heat flux evaluation value and the target equivalent heat flux evaluation value. Its control inputs are coolant flow rate, coolant temperature, or a combination thereof, and its control outputs are adjustments to valve opening, pump speed, or flow allocation ratio. The control algorithm can be implemented using proportional-integral-derivative control, fuzzy control, or model-based control algorithms, and its function is to execute the cooling control target determined by the refining parameter set.

[0109] During the cooling control process, the cooling system does not use real-time calculation of equivalent heat flux evaluation values ​​to inversely calculate the coolant flow rate. Instead, it uses the coolant flow rate, flow distribution ratio, and temperature setpoint given in the refining parameter set as the control targets. The control algorithm adjusts the execution based on the deviation between the actual operating state and the target state. The equivalent heat flux evaluation value is only used for control effect evaluation and parameter update decision-making, and does not directly participate in the inverse calculation of the underlying execution quantities.

[0110] During the implementation of cooling control, the system collects real-time operating status information of the cooling system. 1. Thermal balance synchronization maintenance: By monitoring the surface temperature of each mold using sensors and combining it with the inlet and outlet temperatures of the coolant, the system calculates the actual equivalent heat flux evaluation value for each mold and compares it with the target equivalent heat flux evaluation value determined in S4 to determine the thermal balance synchronization status of the multi-mold system. When the deviation between the equivalent heat flux evaluation values ​​of different molds exceeds the preset range, the thermal balance synchronization status is deemed abnormal. 2. Adaptive flow-around triggering record for abnormal cooling channel blockage: Flow and pressure sensors installed in the cooling system monitor changes in flow rate and pressure drop within the cooling channels in real time. When abnormal fluctuations that are significantly inconsistent with the flow rate set value corresponding to the refining parameter set are detected, it is determined that there may be cooling channel blockage or abnormal flow resistance, and the trigger status of the adaptive flow-around mechanism is recorded. 3. Execution status of mold thermal expansion deformation pre-compensation: By monitoring changes in mold surface temperature and combining it with the thermal expansion coefficient of the mold material, the actual execution effect of the thermal expansion compensation amount set based on the refining parameter set is evaluated to determine whether the compensation is insufficient or excessive.

[0111] Based on the implementation of cooling control and the collection of status information, a comprehensive evaluation of the thermal balance state and cooling effect of the mold cooling process is conducted. 1. Thermal Balance State Evaluation: By comprehensively analyzing the equivalent heat flow evaluation value of each mold, the mold wall temperature distribution, and the deviation between the actual flow rate and the target flow rate of the coolant, it is determined whether the multi-mold system has reached the expected thermal balance state. Specifically, based on the equivalent heat flow evaluation model established in S4, the current equivalent heat flow evaluation value of each mold is calculated and compared with the target equivalent heat flow evaluation value corresponding to the refining parameter set to obtain a thermal balance deviation index characterizing the degree of heat balance matching in the multi-mold system. When the thermal balance deviation index exceeds the preset allowable range, it is determined that the thermal balance under the current multi-mold cooling state has not met the expected requirements. 2. Cooling Effect Evaluation: The criteria for evaluating the cooling effect include the uniformity of the mold surface temperature and the molding quality risks that may be caused by uneven temperature distribution. The cooling effect evaluation can be obtained through mold surface temperature distribution measurement and surface temperature difference analysis. The mold surface temperature difference, as one of the cooling effect evaluation results, also serves as an auxiliary input for calculating the thermal balance deviation index, reflecting the actual impact of cooling control on the overall thermal state of the mold. When the temperature difference on the mold surface exceeds the preset tolerance range, it is determined that the current cooling effect does not meet the production requirements.

[0112] During the comprehensive evaluation process, the thermal balance deviation index obtained from the thermal balance state assessment, combined with the mold surface temperature difference evaluation result obtained from the cooling effect assessment, is compared with a preset stability threshold to determine whether the evaluation result has reached the preset stability threshold. The preset stability threshold is used to limit the maximum allowable thermal balance deviation range and the allowable range of mold surface temperature difference under multi-mold collaborative cooling control. When the thermal balance deviation index exceeds the preset stability threshold, or the mold surface temperature difference evaluation result exceeds the corresponding tolerance range, it is determined that the evaluation result has not reached the preset stability threshold. At this time, the current cooling process is considered to be in an unstable state, and the subsequent cooling parameter adjustment process is triggered.

[0113] Based on the deviation indicators obtained from the thermal balance state assessment, the cooling effect assessment results, and the anomaly types recorded in the operating status information, various deviations are classified and attributed, and the generation method of adjustment commands is determined accordingly. The adjustment commands are parameter update-triggered control commands, used to indicate the direction, intensity, and priority of subsequent cooling parameter updates, without directly limiting the adjustment method of specific execution layer control quantities. For example, when the evaluation results indicate that the thermal balance deviation between multiple molds exceeds the allowable range, the cooling parameter correction amount is calculated based on the equivalent heat flow deviation value corresponding to each mold, and the equivalent heat flow deviation is mapped to the correction direction and correction magnitude of the temperature gradient compensation parameter. This is used to trigger the iterative update of the real-time temperature gradient compensation parameter of the mold surface, thereby indirectly guiding the coolant flow rate and its partition distribution relationship to converge towards the target state determined by the refining parameter set. When the evaluation results indicate that the local mold temperature fluctuation is related to the blockage of the cooling channel or abnormal flow resistance, a cooling circuit switching or bypass opening command is generated based on the adaptive flow trigger record. This command is parsed into a trigger condition for reconstructing the heat conduction path configuration between the mold groups, so as to change the effective heat exchange path between the molds by adjusting the controllable heat conduction structure or thermal coupling relationship. When the evaluation results indicate that the mold thermal expansion compensation amount is insufficient or excessive, a compensation correction command for the coolant temperature or cooling duration is generated based on the deviation between the predicted thermal expansion value and the actual executed value. This compensation correction command is converted into an adjustment basis for the local activation response intensity of the cooling phase change material, so as to quickly compensate the mold thermal state by changing the phase change heat absorption or release mode.

[0114] The final adjustment command is formed by integrating the above-mentioned parameter update trigger commands according to a preset priority. Among them, different types of control commands correspond to temperature gradient compensation parameter updates, phase change material activation response intensity adjustments, and heat conduction path configuration reconstruction, respectively. They are output as a unified high-level control command to the cooling control execution unit to drive the update of cooling parameter configuration in step S7.

[0115] S7. Update the cooling parameter configuration according to the final adjustment instruction, and generate a cooling parameter control scheme for global multi-mold collaborative cooling control.

[0116] In one specific embodiment, the process of performing step S7 may specifically include the following steps:

[0117] Based on the final adjustment instructions, the real-time compensation parameters for the temperature gradient on the mold surface are iteratively updated using an optimization method.

[0118] Based on the final adjustment instructions, and in conjunction with the updated temperature gradient compensation parameters and current cooling requirements, the local activation response intensity of the cooling phase change material is adjusted.

[0119] Based on the final adjustment instructions, and combined with the updated real-time temperature gradient compensation parameters and the activation response intensity of the phase change material, the configuration of the heat conduction path between the mold groups is dynamically reconstructed.

[0120] The updated temperature gradient compensation parameters, phase change material activation response intensity, and heat conduction path configuration are integrated to generate a global cooling parameter control scheme for multi-mold collaborative cooling control.

[0121] Specifically, the temperature gradient compensation parameter describes the allowable intensity of thermal regulation applied per unit time to address uneven temperature distribution on the mold surface. Its update process uses the thermal balance deviation index obtained in step S6 and the corresponding equivalent heat flow deviation of the mold as input. For example, the temperature gradient compensation parameter for the i1th mold can be denoted as G. i1 Its update method is expressed in recursive form as: G i1 r+1 = G i1 r + α i1 × Δq i1 Among them, G i1 r Δq represents the compensation parameter value after the r-th update. i1 α represents the deviation between the current equivalent heat flux evaluation value and the target equivalent heat flux evaluation value of the mold. i1 This represents the compensation step size coefficient, whose value is limited based on the thermal response characteristics of the mold material and the allowable temperature fluctuation range. Upper limit constraints and saturation limits can be set to prevent divergence or oscillation during parameter updates, thus ensuring the stable operation of the cooling control system. Through this iterative update process, the mold surface temperature gradient compensation parameters gradually converge with changes in thermal balance deviation, thereby solving the problem of uneven overall heat distribution under parallel cooling conditions for multiple molds.

[0122] After updating the temperature gradient compensation parameters, the local activation response intensity of the cooling phase change material is adjusted according to the final adjustment command, combined with the updated temperature gradient compensation parameters and current cooling requirements. The phase change material is located in a local area of ​​the mold or near the cooling channel, and its activation response intensity characterizes the effective proportion of heat absorption or release involved in the phase change per unit time. Physically, this is achieved by adjusting the output power of the micro-electric heating element thermally coupled to the phase change material, or by controlling the flow rate of the cooling medium through the area using a proportional control valve, to precisely change the temperature field distribution of the microenvironment where the phase change material is located, thereby triggering or inhibiting its phase change process. For example, the activation response intensity of the phase change material in the i1th mold region can be expressed as R. i1 Its adjustment method can be expressed as R i1 = R base,i1 × (1 + β i1 × G i1 ), where R base,i1 β represents the basic activation intensity of this region. i1 G represents the phase change response adjustment coefficient. i1 These are the updated temperature gradient compensation parameters. When G i1 When R is negative, it indicates that the local cooling intensity needs to be reduced. i1 The calculation result will be less than R. base,i1 The corresponding physical meaning is to reduce the activation amount or triggering ratio of the phase change material in that region.

[0123] After the phase change material activation response intensity is adjusted, the heat exchange relationship between the mold groups is regulated according to the final adjustment command and in conjunction with the updated temperature gradient real-time compensation parameters and the phase change material activation response intensity. The goal of the regulation is to achieve dynamic reconstruction of the equivalent heat conduction path to overcome the mismatch between thermal coupling and dynamic production requirements under a fixed physical structure. The heat conduction path configuration includes the opening and closing state of the controllable thermally conductive connection structure between the molds, the thermal conductivity level, or the configuration state of the thermal isolation structure. For example, when the equivalent heat flux evaluation value of a certain mold is consistently higher than that of other molds and the phase change material activation intensity has reached its upper limit, some heat is guided to transfer to the mold with excess heat capacity by enhancing the effective thermal conductivity of the heat conduction path between this mold and the adjacent low-load mold; when thermal coupling between molds is detected to amplify temperature fluctuations, a physical thermal barrier is intervened or the flow rate of the heat conduction fluid is reduced to increase the effective thermal resistance and weaken unfavorable heat transfer.

[0124] Preferably, after the phase change material activation response intensity is adjusted, the thermal interaction management strategy between mold groups is decided and optimized based on the final adjustment command and the updated real-time temperature gradient compensation parameters and phase change material activation response intensity. The thermal interaction management strategy aims to overcome the mismatch between thermal coupling relationships under a fixed physical structure and dynamic production demands by actively intervening in the heat exchange process between molds. Its implementation does not rely on real-time changes to the physical connections or thermal conductivity of the mold body, but rather on decisions made by the control system to drive peripheral thermal management accessories or adjust process settings to achieve equivalent thermal path modulation. In engineering implementation, a graded response strategy is adopted: Level 1 adjustment (core means): Prioritizing the adjustment of coolant flow distribution and phase change material activation to balance the heat load; this is the most direct and fastest dynamic thermal balance adjustment method. Level 2 adjustment (auxiliary and backup means): When the Level 1 adjustment means have reached their adjustment limits (e.g., flow valve saturation, complete phase change material activation) or when the assessment deems parasitic thermal coupling between molds to be the main source of interference, active management or process compensation of the external thermal environment between molds is initiated. This level of adjustment, depending on the specific production line configuration, can manifest as one or more of the following executable operations: 1. Controlling peripheral thermal management devices: Generating instructions to control auxiliary heat exchange equipment installed between or around molds, such as turning on or off independent air-cooled fan groups pointing to specific molds, or adjusting the power of auxiliary cooling units. 2. Adjusting process parameters for thermal compensation: In subsequent production cycles, for adjacent "low-temperature" molds with high heat loads, inject melt at a slightly higher temperature (within the process allowable range) in advance, or fine-tune their holding pressure parameters to pre-compensate for the expected heat input, thus adjusting the thermal balance from the process source. Specifically, calculate the melt temperature compensation amount based on the predicted temperature difference value and send it to the injection molding machine control system for execution. 3. Triggering alarms and prompts: When the system determines that the thermal imbalance exceeds the automatic control range, or detects a persistent problem that may be caused by a rigid thermal short circuit between molds, a maintenance alarm is triggered, and specific intervention suggestions are given on the human-machine interface (such as checking the status of the mold spacer heat pads, suggesting stopping the machine and installing heat insulation plates, etc.), guiding operators to perform physical repairs or configuration optimization. Through the above-mentioned multi-level strategies, from internal adjustment to external intervention, from automatic control to human-machine collaboration, the most feasible path can be adaptively selected to optimize the overall thermal state of the mold cluster according to the actual working conditions and hardware conditions, thereby achieving and maintaining a stable cooling balance in a complex multi-mold parallel production environment.

[0125] After updating the temperature gradient compensation parameters, adjusting the phase change material activation response intensity, and reconstructing the heat conduction path configuration, the updated parameters are integrated to generate a global cooling parameter control scheme for multi-mold collaborative cooling control. This global cooling parameter control scheme is stored as a parameter set, which includes at least the temperature gradient compensation parameters, phase change material activation response intensity parameters, and heat conduction path configuration parameters between mold groups for each mold. This set is then sent to the cooling control execution unit as a unified control target to coordinate coolant flow rate adjustment, phase change cooling response, and heat exchange behavior between molds during subsequent injection molding cycles, thereby achieving global collaborative control of the multi-mold precision injection cooling process.

[0126] To verify the effectiveness of the technical solution of this invention, a comparative experiment was conducted between the basic fixed cooling scheme and the graded response control scheme of this invention under the same production conditions. A certain type of four-cavity precision injection mold was used in the experiment, the injection material was polycarbonate (PC), and the temperature fluctuation range of the production environment was 22±3℃. Figure 5 As shown, the solution of this invention significantly outperforms traditional fixed-base cooling solutions in all key performance indicators. Specifically:

[0127] 1. Regarding cooling uniformity, the solution of this invention reduces the maximum temperature difference between mold A and mold B from 5.8℃ to 0.9℃, achieving an improvement rate of 84.5%. This indicates that the present invention effectively achieves thermal balance synchronization among multiple molds through multi-source interference data fusion and collaborative optimization.

[0128] 2. In terms of response speed, the thermal equilibrium establishment time was significantly reduced from 420 seconds to 115 seconds, an improvement rate of 72.6%; the cooling loop response delay was reduced from 18.5 seconds to 3.2 seconds, an improvement rate of 82.7%. This proves that the present invention, based on the feedforward compensation mechanism of the key interference influence vector and the closed-loop adaptive adjustment strategy, significantly improves the system's response capability to dynamic operating conditions;

[0129] 3. Regarding product quality, the warpage of the product decreased from 0.12mm to 0.03mm, an improvement rate of 75.0%; the dimensional deviation rate of the product decreased from 2.5% to 0.2%, an improvement rate as high as 92.0%. This verifies that the present invention, through multi-mold collaborative cooling control, can effectively reduce product defects caused by uneven cooling and significantly improve the dimensional accuracy and consistency of precision injection molded products.

[0130] The above experimental results show that the cooling system control method for precision injection molds provided by the present invention can effectively solve the problem of uneven cooling in parallel production of multiple molds, improve the system's adaptability and response speed, and ultimately significantly improve the molding quality of the products.

[0131] The cooling system control method for precision injection molds in the embodiments of this application has been described above. The cooling system control device for precision injection molds in the embodiments of this application is described below. Please refer to [link / reference]. Figure 6 The present application provides a schematic diagram of a cooling system control device for precision injection molds, the device comprising:

[0132] The data acquisition module is used to acquire multi-source interference data affecting the mold cooling process in real time from the workshop environment and injection equipment of precision injection molding production, and to build a preliminary interference dataset.

[0133] The data fusion module is used to perform weighted fusion processing on the initial interference dataset to obtain a weighted interference information matrix.

[0134] The interference analysis module is used to analyze the correlation between various interference factors and their impact on the cooling effect based on the weighted interference information matrix, and to determine the key interference influence vector.

[0135] The collaborative optimization module is used to collaboratively optimize the cooling parameters of the mold cooling system with the key interference influence vector as the optimization input and the constraint of achieving cooling balance among multiple molds, and generate an optimized candidate set of cooling parameters.

[0136] The parameter filtering module is used to perform adaptability screening and filtering on the candidate set of cooling parameters, eliminate parameter combinations that do not meet the current production conditions, and obtain the refined parameter set.

[0137] The effect evaluation module is used to implement cooling control based on the refining parameter set, evaluate the thermal balance state and cooling effect of the mold cooling process, and generate the final adjustment instruction when the evaluation result does not reach the preset stability threshold.

[0138] The global control module is used to update the cooling parameter configuration according to the final adjustment command and generate a cooling parameter control scheme for global multi-mold collaborative cooling control.

[0139] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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 of the technical features. Such 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 this application.

Claims

1. A cooling system control method for a precision injection mold, characterized by, The method includes: S1. Obtain multi-source interference data affecting the mold cooling process in real time from the workshop environment and injection equipment of precision injection molding production, and construct a preliminary interference dataset; S2. Perform weighted fusion processing on the preliminary interference dataset to obtain a weighted interference information matrix; S3. Based on the weighted interference information matrix, analyze the correlation between each interference factor and its impact on the cooling effect to determine the key interference influence vector; S4. Using the key interference influence vector as the optimization input and achieving cooling balance among multiple molds as the constraint, the cooling parameters of the mold cooling system are collaboratively optimized to generate an optimized set of cooling parameter candidates. S5. Perform adaptability screening and filtering on the candidate set of cooling parameters, eliminate parameter combinations that do not meet the current production conditions, and obtain the set of refining parameters; S6. Implement cooling control based on the refining parameter set, evaluate the thermal balance state and cooling effect of the mold cooling process, and generate a final adjustment command when the evaluation result does not reach the preset stability threshold. S7. Update the cooling parameter configuration according to the final adjustment instruction to generate a cooling parameter control scheme for global multi-mold collaborative cooling control; S4 includes: Using the key interference influence vector as the optimization input, under the preset thermal balance constraint between multiple molds, the particle swarm optimization method is used to iteratively optimize the cooling parameters of the mold cooling system. The cooling parameters include at least the total flow rate of the coolant and the partition flow distribution parameters of each mold cooling channel. In the particle swarm optimization process, a multi-mold thermal flux coupling model is established to simulate the dynamic adjustment process of heat flux density equalization adjustment and coolant flow independent zone control between molds, and to generate dynamic adjustment paths. Extract the cooling parameter combination that satisfies the thermal balance constraint between multiple molds from the dynamic adjustment path, and generate the optimized cooling parameter candidate set; S5 includes: For each parameter combination in the candidate set of cooling parameters, determine whether it matches the batch heat capacity difference and melt flow index change of the current injection molding raw material, and eliminate mismatched parameter combinations. For the matched parameter combinations, similarity comparison and filtering are performed based on the geometric feature parameters of each mold cavity, and parameter combinations with geometric similarity lower than the preset similarity threshold are eliminated; For parameter combinations selected through geometric features, the consistency of cavity pressure distribution among multiple molds is verified, and parameter combinations that cause pressure distribution inconsistency to exceed the preset allowable range are eliminated. Integrate the verified parameter combinations to generate a refined parameter set; S6 includes: Based on the refining parameter set, cooling control is implemented for the multi-mold cooling system, and the operating status information of the cooling system is collected in real time. The operating status information includes at least the multi-mold thermal balance synchronous maintenance status, the adaptive flow trigger record of abnormal cooling channel blockage, and the execution status of the pre-compensation amount for mold thermal expansion deformation. Based on the operating status information, the thermal balance state and cooling effect of the mold cooling process are evaluated, and it is determined whether the evaluation result reaches the preset stability threshold. If the evaluation result does not reach the preset stability threshold, the final adjustment instruction is generated; S7 includes: Based on the final adjustment instruction, the real-time compensation parameters for the temperature gradient on the mold surface are iteratively updated using an optimization method. Based on the final adjustment instruction, and in conjunction with the updated temperature gradient compensation parameters and current cooling requirements, adjust the local activation response intensity of the cooling phase change material; Based on the final adjustment instruction, and combined with the updated temperature gradient real-time compensation parameters and the activation response intensity of the phase change material, the heat conduction path configuration between the mold groups is dynamically reconstructed. The updated temperature gradient compensation parameters, phase change material activation response intensity, and heat conduction path configuration are integrated to generate a global cooling parameter control scheme for multi-mold collaborative cooling control.

2. The method of claim 1, wherein, S1 includes: Sensors deployed in the workshop and equipment are used to collect real-time data on ambient temperature fluctuations, increased thermal resistance of cooling channels due to equipment aging, and temperature gradient values ​​on the mold surface. Simultaneously monitor batch-to-batch heat capacity differences and melt flow index changes of raw materials used in injection molding, and determine the set of dynamic interference factors caused by changes in material properties. The ambient temperature fluctuation data, thermal resistance increase data, temperature gradient values, and dynamic interference factor set are time-aligned and integrated to generate the preliminary interference dataset.

3. The method of claim 1, wherein S2 include: Based on the freshness of the acquisition timestamp of each interference data in the preliminary interference dataset and the calibration reliability of the corresponding sensor, a corresponding weight coefficient is assigned to each interference data. The weighted interference information matrix is ​​generated by weighting each interference data in the preliminary interference dataset based on the weighting coefficients.

4. The method of claim 1, wherein, S3 includes: Extract data sequences of at least two interference factors from the weighted interference information matrix, and calculate the correlation strength between the interference factors through correlation analysis; Based on the temperature gradient data sequence in the weighted interference information matrix, the comprehensive influence of each interference factor on the cooling state of the multi-mold is calculated through a preset evaluation model. Based on the correlation strength and the degree of comprehensive influence, the core interference factors that play a dominant role in the cooling effect are identified, and a key interference influence vector is generated.

5. Cooling system control device for precision injection moulds for implementing the method according to any one of claims 1 to 4, characterized in that, The device includes: The data acquisition module is used to acquire multi-source interference data affecting the mold cooling process in real time from the workshop environment and injection equipment of precision injection molding production, and to build a preliminary interference dataset. The data fusion module is used to perform weighted fusion processing on the preliminary interference dataset to obtain a weighted interference information matrix; The interference analysis module is used to analyze the correlation between various interference factors and their impact on the cooling effect based on the weighted interference information matrix, and to determine the key interference influence vector. The collaborative optimization module is used to perform collaborative optimization of the cooling parameters of the mold cooling system with the key interference influence vector as the optimization input and the constraint of achieving cooling balance among multiple molds as the constraint, and to generate an optimized candidate set of cooling parameters. The parameter filtering module is used to perform adaptability screening and filtering on the candidate set of cooling parameters, eliminate parameter combinations that do not meet the current production conditions, and obtain the refined parameter set. An effect evaluation module is configured to implement cooling control according to the set of refining parameters, evaluate a heat balance state and a cooling effect of a mold cooling process, and generate a final adjustment instruction when an evaluation result does not reach a preset stable threshold. A global regulation module is configured to update a cooling parameter configuration according to the final adjustment instruction, and generate a cooling parameter regulation scheme for global multi-mold collaborative cooling control.

Images