Optimization design method and system for injection mold

Abstract

The invention discloses an optimal design method and system for an injection mold, and relates to the technical field of injection mold optimizing.By constructing a cooling time statistical model of a qualified forming period, aggregation statistics is conducted on cooling lag time of all mold cavities and point locations in qualified samples, representative cooling time length distribution is established, and the optimal design of the injection mold is achieved. And a data basis is provided for cooling control. Through the output value of the cooling window function, the cooling process can be divided into a reasonable interval and an unreasonable interval, and traceable labels are formed in data recording and quality tracing. Meanwhile, in combination with the deviation degree of the upper bound and the lower bound, a cooling abnormal degree grading mechanism can be extended, and support is provided for intelligent adjustment strategy selection. An abnormal cooling lag point is identified through the effective cooling attenuation stage starting point Teff; through the normalized WS data, the abnormal deviation trend point of the temperature recovery curve can be visually found, and an accurate cooling deviation input basis is provided for a subsequent module, so that differential cooling control suggestions are supported.

Description

Technical Field

[0001] This invention relates to the field of injection mold optimization technology, specifically to an injection mold optimization design method and system. Background Technology

[0002] In the injection molding process, the rationality of mold design directly affects product quality, production efficiency, and molding stability, thus falling under the specific technical direction of the intersection of mold engineering and thermal control design. Within this sub-field, the design of the cooling system in injection molds, as one of the most critical aspects of process parameter setting and structural layout, not only determines cooling efficiency but also directly affects the molding cycle and the internal stress distribution of the part, making it a core component for achieving high-quality molding.

[0003] Currently, the design and parameter tuning of injection mold cooling systems in industry generally rely on rules of thumb or steady-state heat transfer analysis, neglecting the dynamic thermal behavior evolution of the mold under different production cycles, different material batches, and environmental changes. Especially during the production stage, the cooling time is often set to a fixed value, which cannot be adaptively adjusted based on historical qualified samples and real-time thermal behavior, easily leading to insufficient or excessive cooling.

[0004] The root cause of these problems lies in the fact that existing systems fail to establish a feedback path between cooling behavior, historical data, and qualified samples. They cannot capture the dynamic temperature recovery of different areas within the mold in real time, nor have they established a cooling window model based on time series offsets. Without a feedback mechanism and intelligent parameter tuning, cooling parameter settings are often in a state of blind adjustment. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an injection mold optimization design method and system, which solves the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solution: an injection mold optimization design system, including a temperature recovery data acquisition module, a temperature normalization and recovery rate calculation module, an optimal cooling window function construction module, a nonlinear cooling trajectory fitting and trend prediction module, a cooling time dynamic adjustment suggestion module, and a feedback recording and continuous learning module; The temperature recovery data acquisition module collects data from each mold cavity during the mold cooling stage after injection molding and fits it into a temperature recovery set WD; The temperature normalization and recovery rate calculation module cleans and normalizes the temperature recovery set WD, extracts the temperature recovery rate and its lag time, and obtains the temperature dataset WS. The optimal cooling window function construction module is based on the temperature recovery set WS and extracts the optimal cooling time range based on historical qualified molding cycles to construct the cooling window function LQH; The nonlinear cooling trajectory fitting and trend prediction module reconstructs the thermal decay process using a double exponential model, performs nonlinear curve fitting on the temperature recovery trajectory at each point, and predicts the cooling completion time yT at each point. The dynamic adjustment suggestion module for cooling time combines the cooling completion time yT with the cooling window function LQH to calculate and obtain the matching coverage Pco; The feedback recording and continuous learning module detects, records, and provides feedback on mold forming quality deviations based on the cooling completion time yT and the matching coverage Pco.

[0007] Preferably, the temperature recovery data acquisition module collects temperature change information at a location point in the mold cavity during the cooling stage after injection molding by sampling sensors, including temperature value T, stable temperature of the cooling medium Ten, and cooling start time ts, and fits it into a temperature recovery set WD; The temperature value T is acquired by an embedded temperature sensor, specifically the temperature value T(i,j)t at the j-th observation point in the i-th mold cavity at time t. The stable temperature of the cooling medium, Ten, is acquired through the temperature control output interface of the chiller unit. The cooling start time ts is acquired by reading the trigger signal indicating completion of filling from the injection molding machine controller.

[0008] Preferably, the temperature normalization and recovery rate calculation module includes a temperature data cleaning and dynamic stability window extraction unit and a thermal behavior normalization expression unit; The temperature data cleaning and dynamic stabilization window extraction unit performs preliminary cleaning on the temperature time series in the temperature recovery set WD, identifies non-cooling intervals and sampling anomalies in the temperature data, and extracts the starting point Teff for each position entering the effective cooling decay stage. The cleaning method is as follows: For each temperature value T, first calculate the first derivative approximation: The approximate method for obtaining the first derivative of the temperature value T is as follows: First, take the difference between the temperature value T(i,j)t of the j-th observation point in the i-th cavity at time t+Δt and the temperature value T(i,j)t of the j-th observation point in the i-th cavity at time t+Δt, and then divide it by the time interval Δt to obtain the result. The effective cooling decay phase starting point Teff is obtained using the following formula: Teff(i,j) = min{t | first derivative of temperature value T < -λ duration greater than ΔT}; In the formula, λ represents the minimum rate of decrease for the cooling trend to start, ΔT represents the minimum time window for the decrease to continue, and Teff(i,j) represents the effective cooling decay stage starting point of the j-th observation point in the i-th cavity; The thermal behavior normalization representation unit normalizes the cleaned temperature recovery set WD to obtain the temperature dataset WS; The temperature dataset WS is obtained using the following formula: ; In the formula, WSo(t) represents the o-th data in the temperature dataset WS at time t, WDo(t) represents the o-th data in the temperature recovery set WD at time t, minWDo represents the valley value of the o-th data in the temperature recovery set WD, and maxWDo represents the peak value of the o-th data in the temperature recovery set WD.

[0009] Preferably, the optimal cooling window function construction module includes a qualified sample cooling time statistics unit and a cooling window function generation and expression unit; The qualified sample cooling time statistics unit selects qualified molding cycles from all historical molding records, extracts the cooling lag time of each cycle at all mold cavity points, calculates statistical characteristics including the average cooling time, cooling mean and cooling standard deviation of each cycle, and establishes a cooling time reference distribution model. The average cooling time is obtained as follows: First, in the molding cycle, the mold contains multiple cavities, assuming a total of M cavities, and N observation points in each cavity; for each cavity in the cycle, from the 1st to the Mth cavity, and for each observation point, from the 1st to the Nth observation point, the cooling lag time is calculated, which is the time taken for the temperature at that point to drop to near the cooling medium from the start of cooling; the cooling lag times of all cavities and all points in the cycle are added together to obtain a total cooling time; then, this is divided by the number of all points, which is M×N, to obtain the average cooling time; The cooling average is obtained by summing the average cooling times of all molding cycles to get the average cooling time, and then dividing it by the total number of qualified samples to get the cooling average. The cooling standard deviation is obtained by subtracting the average cooling time from the cooling mean for each molding cycle to obtain the deviation value. Then, the deviation value is squared to obtain the squared deviation for each cycle. Finally, the sum of all the squared deviations is divided by the number of cycles, and the square root is taken to obtain the cooling standard deviation.

[0010] Preferably, the cooling window function generation and expression unit uses the results of statistical features to construct the cooling window function LQH and identifies whether the current predicted cooling time falls within a reasonable range; The cooling window function LQH is obtained using the following formula: ; In the formula, LQH(t) represents the cooling window function at time t; Tmin represents the lower extreme boundary of the window, and Tmax represents the upper extreme boundary of the window; When the cooling window function LQH=1, it means that the current predicted cooling time falls within a reasonable range; When the cooling window function LQH=0, it means that the current predicted cooling time does not fall within a reasonable range; The formulas for obtaining the lower extreme boundary Tmin and the upper extreme boundary Tmax of the window are as follows: Tmin = μcool - co1 × σcool; Tmax = μcool + co2 × σcool; In the formula, μcool represents the cooling mean, σcool represents the cooling standard deviation, co1 represents the lower bound bandwidth coefficient, and co2 represents the upper bound bandwidth coefficient.

[0011] Preferably, the nonlinear cooling trajectory fitting and trend prediction module includes a double exponential decay model fitting unit and a cooling completion time prediction unit; The double exponential decay model fitting unit fits the normalized temperature curve of each cavity observation point (i, j) and uses the double exponential function model to express the entire process of temperature decrease over time. The formula for constructing the double exponential function model is as follows: ; In the formula, AS(i,j)t represents the normalized temperature curve value of the j-th observation point in the i-th mold cavity at time t, A1(i,j) represents the initial intensity of the first exponential decay term at the j-th observation point in the i-th mold cavity, A2(i,j) represents the initial intensity of the second exponential decay term at the j-th observation point in the i-th mold cavity, e represents a constant, k1 represents the decay coefficient of the cooling rate in the first stage, and k2 represents the decay coefficient of the cooling rate in the second stage. The cooling completion time prediction unit is based on a double exponential function model to calculate and obtain the cooling completion time yT for each point. The cooldown completion time yT is obtained using the following formula: ; In the formula, yT(i,j) represents the cooling completion time of the j-th observation point in the i-th mold cavity, Tend represents the preset cooling completion time, ts represents the cooling completion judgment threshold, and min{...} represents the earliest time that meets the condition among all times after the start time.

[0012] Preferably, the dynamic adjustment suggestion module for cooling time includes a matching judgment and coverage calculation unit and an adjustment suggestion generation unit; The matching judgment and coverage calculation unit compares the predicted cooling completion time yT of all cavity observation points with the optimal cooling window function LQH point by point to determine whether it falls within the window, and counts the proportion of all points that meet the conditions to obtain the matching coverage Pco. The judgment method is as follows: traverse all the mold cavity observation points in the mold cavity, obtain the predicted cooling completion time yT of the mold cavity observation point; then, substitute this time into the cooling window function for judgment. When this time point is within the optimal cooling time range, the cooling window function will output a result of 1; If this time point is not within the optimal range, i.e. too early or too late, the function will output a result of 0; Match coverage Pco is obtained using the following formula: ; In the formula, M represents the total number of mold cavities, and N represents the number of observation points in each mold cavity. This represents a Boolean conditional function.

[0013] Preferably, the adjustment suggestion generation unit determines whether the current cooling configuration is reasonable based on the evaluation result of the matching coverage Pco, and provides a suggested cooling duration Tncool when it deviates from the optimal range, and outputs a structured suggestion; The suitability of the cooling configuration is determined by matching the following methods: When the matching coverage Pco ≥ 0.8, it indicates that the observation point has cooled down sufficiently and no adjustment is needed; When the matching coverage Pco < 0.8, it indicates a wide deviation, so the cooldown time is extended and a suggested cooldown time Tncool is obtained; The recommended cooldown time Tncool can be obtained using the following formula: ; Structured recommendations include status labels, such as sufficient or insufficient cooldown time, matching coverage values, recommended cooldown time, and optional hints, such as suggesting an extension by seconds.

[0014] Preferably, the feedback recording and continuous learning module archives and labels the cooling behavior parameters cooling completion time yT, quality status judgment function Qua, and matching coverage Pco, obtains the feedback deviation index Drec, and establishes a feedback path to adjust the future cooling strategy. The quality status determination function Qua is obtained as follows: When the matching coverage Pco ≥ 0.8, the quality status judgment function Qua = 1; When the matching coverage Pco < 0.8, the quality status judgment function Qua = 0; The feedback deviation index Drec is obtained using the following formula: ; In the formula, Drec(tk) represents the feedback deviation index of the k-th cooling process, yT(i,j)tk represents the cooling completion time of the j-th observation point in the i-th cavity of the k-th cooling process, and yTopt represents the theoretical ideal cooling completion time. The future cooling strategy will be adjusted as follows: When the feedback deviation index Drec < 0.3, it indicates a normal situation, the cooling behavior is stable, and the original strategy should be maintained. When 0.3 ≤ feedback deviation index Drec ≤ 0.7, it indicates a slight deviation, and it is recommended to adjust the total cooling time. When 0.7 < feedback deviation index Drec, it indicates a serious deviation, requiring forced adjustment of the cooling window or activation of abnormal path feedback.

[0015] A method for optimizing the design of injection molds includes the following steps: Step 1: The temperature recovery data acquisition module collects data from each mold cavity during the mold cooling stage after injection molding and fits it into a temperature recovery set WD. Step 2: The temperature normalization and recovery rate calculation module cleans and normalizes the temperature recovery set WD, extracts the temperature recovery rate and its lag time, and obtains the temperature dataset WS. Step 3: The optimal cooling window function construction module is based on the temperature recovery set WS and extracts the optimal cooling time range based on the historical qualified molding cycle to construct the cooling window function LQH; Step 4: The nonlinear cooling trajectory fitting and trend prediction module reconstructs the thermal decay process using a double exponential model, performs nonlinear curve fitting on the temperature recovery trajectory at each point, and predicts the cooling completion time yT at each point. Step 5: The dynamic adjustment suggestion module for cooling time combines the cooling completion time yT with the cooling window function LQH to calculate and obtain the matching coverage Pco. Step Six: The feedback recording and continuous learning module detects, records, and provides feedback on mold forming quality deviations based on the cooling completion time yT and matching coverage Pco.

[0016] This invention provides a method and system for optimizing the design of injection molds, which has the following beneficial effects: (1) During system operation, the temperature recovery behavior of the mold during the cooling process is collected at high frequency and multiple points through the embedded temperature sensor, and supplemented by the recording of the stable temperature of the cooling medium Ten and the cooling start time ts, thus constructing a full-process dataset WD of the cooling behavior, which can accurately reflect the cooling behavior and thermal inertial response at each location point.

[0017] The starting point, ending point, and rate of change of the original temperature curves often differ between different cavities or different stages, making direct comparison difficult to form a standardized description of cooling characteristics. By using first-order derivative approximation calculation and a descent persistence identification algorithm, the effective cooling decay stage starting point Teff of each point is automatically identified. Based on this, normalization is performed to form a standard-scale temperature dataset WS, making the cooling behavior at each point comparable and fitable, supporting subsequent cooling window discrimination and fitting prediction.

[0018] (2) By constructing a statistical model of cooling time for a qualified molding cycle, the cooling lag time of all mold cavities and points in the qualified samples is aggregated and statistically analyzed to establish a representative cooling time distribution, providing data basis for cooling control rather than empirical values. Through the output value of the cooling window function LQH, the cooling process can be divided into reasonable and unreasonable intervals, forming traceable tags in data recording and quality traceability. At the same time, combined with the degree of deviation between the upper and lower bounds, a cooling anomaly degree classification mechanism can be derived, providing support for the selection of intelligent adjustment strategies.

[0019] (3) The current cooling strategy is evaluated as a whole by matching the coverage rate Pco, avoiding the bias of judging the overall state based on a few local cooling points. When the matching coverage rate Pco meets the standard, the original strategy is automatically maintained; when the matching coverage rate Pco deviates, the system can automatically determine that there are extensive unreasonable areas in the current cooling configuration, thereby triggering adjustment suggestions. This mechanism breaks away from the traditional manual experience-based strategy setting method and transforms into a unified measurement and triggering mechanism for the entire mold cavity, which has stronger objectivity and interpretability.

[0020] The adjustment suggestion generation unit no longer provides only a single cooling time suggestion, but combines structured tags to form a complete output framework for decision-making reference. At the same time, a feedback recording mechanism is introduced to archive information such as cooling completion time, product quality status, and coverage during each molding process, and to calculate a feedback deviation index to quantify the degree of deviation between the current cooling strategy and the theoretical ideal state.

[0021] (4) By constructing the optimal cooling window function module, the system extracts the cooling completion time interval from historical qualified samples and establishes a cooling window function as a reference template for the cooling behavior of the mold in the current molding cycle. This method does not rely on manual experience threshold settings, has automatic adaptability, and supports cross-batch consistency analysis, which is conducive to long-term accumulation and optimization. The exponential model reconstructs the thermal decay process, which can more realistically reflect the nonlinear characteristics of rapid decline in the early stage of cooling and slow change in the later stage, making the cooling completion time prediction of each mold cavity observation point closer to the real thermal field evolution process, thereby improving the overall performance of the system in terms of prediction timeliness and spatial resolution.

[0022] By comparing the predicted cooling time with the cooling window function, the system can obtain a set of overall matching coverage indicators for the mold cavity observation points. This indicator has spatial statistical significance and is the core criterion for measuring the global adaptability of the current cooling strategy. It supports the judgment entry point for generating subsequent adjustment suggestions and opens up the evaluation-feedback path. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the block diagram of an injection mold optimization design system according to the present invention; Figure 2 This is a schematic diagram illustrating the steps of an injection mold optimization design method according to the present invention; Figure 3 This is a schematic diagram of the cooling completion time acquisition process of the present invention; Figure 4 This is a trend chart of the feedback deviation index of the present invention. Detailed Implementation

[0024] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. Example 1

[0025] This invention provides an injection mold optimization design system. Please refer to [link / reference]. Figures 1 to 4 It includes a temperature recovery data acquisition module, a temperature normalization and recovery rate calculation module, an optimal cooling window function construction module, a nonlinear cooling trajectory fitting and trend prediction module, a cooling time dynamic adjustment suggestion module, and a feedback recording and continuous learning module. The temperature recovery data acquisition module collects data from each mold cavity during the mold cooling stage after injection molding and fits it into a temperature recovery set WD; The temperature normalization and recovery rate calculation module cleans and normalizes the temperature recovery set WD, extracts the temperature recovery rate and its lag time, and obtains the temperature dataset WS. The optimal cooling window function construction module is based on the temperature recovery set WS and extracts the optimal cooling time range based on historical qualified molding cycles to construct the cooling window function LQH; The nonlinear cooling trajectory fitting and trend prediction module reconstructs the thermal decay process using a double exponential model, performs nonlinear curve fitting on the temperature recovery trajectory at each point, and predicts the cooling completion time yT at each point. The dynamic adjustment suggestion module for cooling time combines the cooling completion time yT with the cooling window function LQH to calculate and obtain the matching coverage Pco; The feedback recording and continuous learning module detects, records, and provides feedback on mold forming quality deviations based on the cooling completion time yT and the matching coverage Pco.

[0026] In this embodiment, instead of relying on traditional empirical settings or static thermal balance estimation methods, the actual temperature recovery process of each mold cavity is dynamically acquired through a temperature recovery data acquisition module and a temperature normalization and recovery rate calculation module. This includes cooling rate, hysteresis response, and cooling curve, comprehensively reflecting the thermal inertia characteristics of the mold under different operating conditions. This achieves full-process capture of the mold's thermal field behavior and establishes a data-driven understanding of temperature control.

[0027] The cooling time dynamic adjustment suggestion module calculates the overlap between the predicted cooling time and the ideal window to form a matching coverage index. The feedback recording and continuous learning module then correlates each cooling behavior with the molding quality and records the feedback. This mechanism not only identifies abnormal cooling behavior but also gradually accumulates adjustment data, enabling continuous optimization and intelligent learning as production conditions evolve. It supports independent adjustment and differentiated control strategies across multiple mold cavities in the future. Example 2

[0028] This embodiment is an explanation based on Embodiment 1. Please refer to it. Figure 1 and Figure 3 Specifically: the temperature recovery data acquisition module collects temperature change information at locations within the mold cavity during the cooling stage after injection molding by sampling sensors, including temperature value T, stable temperature of the cooling medium Ten, and cooling start time ts, and fits it into a temperature recovery set WD; The temperature value T is acquired by an embedded temperature sensor, specifically the temperature value T(i,j)t at the j-th observation point in the i-th mold cavity at time t. The stable temperature of the cooling medium, Ten, is acquired through the temperature control output interface of the chiller unit. The cooling start time ts is acquired by reading the trigger signal indicating completion of filling from the injection molding machine controller.

[0029] The temperature normalization and recovery rate calculation module includes a temperature data cleaning and dynamic stability window extraction unit and a thermal behavior normalization expression unit; The temperature data cleaning and dynamic stabilization window extraction unit performs preliminary cleaning on the temperature time series in the temperature recovery set WD, identifies non-cooling intervals and sampling anomalies in the temperature data, and extracts the starting point Teff for each position entering the effective cooling decay stage. The cleaning method is as follows: For each temperature value T, first calculate the first derivative approximation: The approximate method for obtaining the first derivative of the temperature value T is as follows: First, take the difference between the temperature value T(i,j)t of the j-th observation point in the i-th cavity at time t+Δt and the temperature value T(i,j)t of the j-th observation point in the i-th cavity at time t+Δt, and then divide it by the time interval Δt to obtain the result. The effective cooling decay phase starting point Teff is obtained using the following formula: Teff(i,j) = min{t | first derivative of temperature value T < -λ duration greater than ΔT}; In the formula, λ represents the minimum rate of decrease for the cooling trend to start, ΔT represents the minimum time window for the decrease to continue, and Teff(i,j) represents the effective cooling decay stage starting point of the j-th observation point in the i-th cavity; The thermal behavior normalization representation unit normalizes the cleaned temperature recovery set WD to obtain the temperature dataset WS; The temperature dataset WS is obtained using the following formula: ; In the formula, WSo(t) represents the o-th data in the temperature dataset WS at time t, WDo(t) represents the o-th data in the temperature recovery set WD at time t, minWDo represents the valley value of the o-th data in the temperature recovery set WD, and maxWDo represents the peak value of the o-th data in the temperature recovery set WD.

[0030] In this embodiment, an embedded temperature sensor is used to collect temperature recovery behavior during the mold cooling process at high frequency and multiple points. This is supplemented by recording the stable temperature of the cooling medium (Ten) and the cooling start time (ts), thereby constructing a full-process dataset (WD) of the cooling behavior, which can accurately reflect the cooling behavior and thermal inertial response at each location point.

[0031] The starting point, ending point, and rate of change of the original temperature curves often differ between different cavities or different stages, making direct comparison difficult to form a standardized description of cooling characteristics. By using first-order derivative approximation calculation and a descent persistence identification algorithm, the effective cooling decay stage starting point Teff of each point is automatically identified. Based on this, normalization is performed to form a standard-scale temperature dataset WS, making the cooling behavior at each point comparable and fitable, supporting subsequent cooling window discrimination and fitting prediction.

[0032] In this embodiment, abnormal cooling lag points are identified by using the effective cooling decay stage starting point Teff; through the normalized WS data, points where the temperature recovery curve deviates abnormally from the trend can be intuitively identified, providing accurate cooling deviation input basis for subsequent modules, thereby supporting differentiated cooling control suggestions, alarms, or strategy adjustments. Example 3

[0033] This embodiment is an explanation based on Embodiment 2. Please refer to it. Figure 1 Specifically: the optimal cooling window function construction module includes a qualified sample cooling time statistics unit and a cooling window function generation and expression unit; The qualified sample cooling time statistics unit selects qualified molding cycles from all historical molding records, extracts the cooling lag time of each cycle at all mold cavity points, calculates statistical characteristics including the average cooling time, cooling mean and cooling standard deviation of each cycle, and establishes a cooling time reference distribution model. The average cooling time is obtained as follows: First, in the molding cycle, the mold contains multiple cavities, assuming a total of M cavities, and N observation points in each cavity; for each cavity and each observation point in the cycle, the cooling lag time is calculated; the cooling lag times of all cavities and all points in the cycle are added together to obtain a total cooling time; then, the sum is divided by the number of all points to obtain the average cooling time. The cooling average is obtained by summing the average cooling times of all molding cycles to get the average cooling time, and then dividing it by the total number of qualified samples to get the cooling average. The cooling standard deviation is obtained by subtracting the average cooling time from the cooling mean for each molding cycle to obtain the deviation value. Then, the deviation value is squared to obtain the squared deviation for each cycle. Finally, the sum of all the squared deviations is divided by the number of cycles, and the square root is taken to obtain the cooling standard deviation.

[0034] The cooling window function generation and expression unit uses the results of statistical features to construct the cooling window function LQH and identifies whether the current predicted cooling duration falls within a reasonable range. The cooling window function LQH is obtained using the following formula: ; In the formula, LQH(t) represents the cooling window function at time t; Tmin represents the lower extreme boundary of the window, and Tmax represents the upper extreme boundary of the window; When the cooling window function LQH=1, it means that the current predicted cooling time falls within a reasonable range; When the cooling window function LQH=0, it means that the current predicted cooling time does not fall within a reasonable range; The formulas for obtaining the lower extreme boundary Tmin and the upper extreme boundary Tmax of the window are as follows: Tmin = μcool - co1 × σcool; Tmax = μcool + co2 × σcool; In the formula, μcool represents the cooling mean, σcool represents the cooling standard deviation, co1 represents the lower bound bandwidth coefficient, and co2 represents the upper bound bandwidth coefficient.

[0035] In this embodiment, by constructing a statistical model of cooling time for a qualified molding cycle, the cooling lag time of all mold cavities and points in the qualified samples is aggregated and statistically analyzed to establish a representative cooling time distribution, providing data basis for cooling control rather than empirical values.

[0036] Different batches, materials, or cooling conditions may lead to deviations in cooling behavior, which traditional control systems cannot determine as reasonable. This embodiment introduces a cooling window function (LQH) mechanism. By constructing upper and lower extreme boundaries, it determines whether the current cooling duration falls within a statistically reasonable range. This enables online compliance assessment, abnormal cooling flagging, and strategy adjustment triggering, supporting the establishment of subsequent feedback and adjustment mechanisms. In scenarios involving material changes, cooling system aging, or environmental temperature variations, the system can automatically adjust its judgment benchmark in real time based on historical qualified samples, ensuring the stability of the cooling process and the consistency of product molding.

[0037] By using the output value of the cooling window function LQH, the cooling process can be divided into reasonable and unreasonable ranges, forming traceable tags in data recording and quality traceability. Furthermore, by combining the degree of deviation between the upper and lower bounds, a cooling anomaly severity grading mechanism can be derived, providing support for the selection of intelligent adjustment strategies. Example 4

[0038] This embodiment is an explanation based on Embodiment 3. Please refer to it. Figure 3 Specifically: the nonlinear cooling trajectory fitting and trend prediction module includes a double exponential decay model fitting unit and a cooling completion time prediction unit; The double exponential decay model fitting unit fits the normalized temperature curve of each cavity observation point (i, j) and uses the double exponential function model to express the entire process of temperature decrease over time. The formula for constructing the double exponential function model is as follows: ; In the formula, AS(i,j)t represents the normalized temperature curve value of the j-th observation point in the i-th mold cavity at time t, A1(i,j) represents the initial intensity of the first exponential decay term at the j-th observation point in the i-th mold cavity, A2(i,j) represents the initial intensity of the second exponential decay term at the j-th observation point in the i-th mold cavity, e represents a constant, k1 represents the decay coefficient of the cooling rate in the first stage, and k2 represents the decay coefficient of the cooling rate in the second stage. The cooling completion time prediction unit is based on a double exponential function model to calculate and obtain the cooling completion time yT for each point. The cooldown completion time yT is obtained using the following formula: ; In the formula, yT(i,j) represents the cooling completion time of the j-th observation point in the i-th mold cavity, Tend represents the preset cooling completion time, ts represents the cooling completion judgment threshold, and min{...} represents the earliest time that meets the condition among all times after the start time.

[0039] The dynamic adjustment suggestion module for cooldown time includes a matching judgment and coverage calculation unit and an adjustment suggestion generation unit; The matching judgment and coverage calculation unit compares the predicted cooling completion time yT of all cavity observation points with the optimal cooling window function LQH point by point to determine whether it falls within the window, and counts the proportion of all points that meet the conditions to obtain the matching coverage Pco. The judgment method is as follows: traverse all the mold cavity observation points in the mold cavity, obtain the predicted cooling completion time yT of the mold cavity observation point; then, substitute this time into the cooling window function for judgment. When this time point is within the optimal cooling time range, the cooling window function will output a result of 1; If this time point is not within the optimal range, the function will output a result of 0; Match coverage Pco is obtained using the following formula: ; In the formula, M represents the total number of mold cavities, and N represents the number of observation points in each mold cavity. This represents a Boolean conditional function.

[0040] In this embodiment, traditional cooling process modeling is often simplified to linear or single exponential decay, which cannot accurately reflect the complex cooling dynamics of the mold cavity during actual injection molding due to factors such as material heat conduction, environmental differences, and mold structure differences. This embodiment constructs a double exponential decay model to describe the initial rapid cooling stage and the later slow heat dissipation stage, respectively. This makes the fitting of the temperature change trend at each observation point in the mold cavity closer to the real thermal process, which helps to enhance the timing recognition accuracy of the cooling process.

[0041] This embodiment not only models the current cooling state but also predicts the cooling completion time for each observation point based on the fitted model. By dynamically outputting this time point and combining it with a judgment threshold for truncation identification, it can identify cooling delay areas or early cooling areas in advance, providing quantifiable predictive support for subsequent cooling duration settings and molding cycle control, thus enhancing the system's responsiveness and proactive control.

[0042] This embodiment introduces the matching coverage ratio (Pco) metric, comparing the predicted cooling time of all cavity observation points with the cooling window function point by point, statistically analyzing the proportion of reasonable matches, and forming a global quantitative expression of cooling compliance. As an indicator of cooling uniformity, the coverage ratio (Pco) not only reflects whether the current cooling strategy meets overall requirements but also provides a structural reference for adjusting the cooling cycle time.

[0043] This mechanism uses observation points of all mold cavities as the basic unit, combining model predictions and window function outputs to effectively identify mold cavity areas where cooling is completed prematurely or delayed abnormally. Based on this, the adjustment suggestion generation unit can output specific cooling duration adjustment directions, thus forming differentiated control paths for different mold cavities. This promotes a more balanced cooling process throughout the mold, helping to improve internal stress distribution, surface defect probability, and dimensional consistency in the product. Example 5

[0044] This embodiment is an explanation based on Embodiment 4. Please refer to it. Figure 3 and Figure 4 Specifically: the adjustment suggestion generation unit determines whether the current cooling configuration is reasonable based on the evaluation results of the matching coverage Pco, and provides a suggested cooling duration Tncool when it deviates from the optimal range, and outputs structured suggestions; The suitability of the cooling configuration is determined by matching the following methods: When the matching coverage Pco ≥ 0.8, it indicates that the observation point has cooled down sufficiently and no adjustment is needed; When the matching coverage Pco < 0.8, it indicates a wide deviation, so the cooldown time is extended and a suggested cooldown time Tncool is obtained; The recommended cooldown time Tncool can be obtained using the following formula: ; The structured recommendations include status labels, matching coverage values, suggested cooldown time, and optional tips.

[0045] The feedback recording and continuous learning module archives and labels the cooling behavior parameters, cooling completion time yT, quality status judgment function Qua, and matching coverage Pco, obtains the feedback deviation index Drec, and establishes feedback paths to adjust future cooling strategies. The quality status determination function Qua is obtained as follows: When the matching coverage Pco ≥ 0.8, the quality status judgment function Qua = 1; When the matching coverage Pco < 0.8, the quality status judgment function Qua = 0; The feedback deviation index Drec is obtained using the following formula: ; In the formula, Drec(tk) represents the feedback deviation index of the k-th cooling process, yT(i,j)tk represents the cooling completion time of the j-th observation point in the i-th cavity of the k-th cooling process, and yTopt represents the theoretical ideal cooling completion time. The future cooling strategy will be adjusted as follows: When the feedback deviation index Drec < 0.3, it indicates a normal situation, the cooling behavior is stable, and the original strategy should be maintained. When 0.3 ≤ feedback deviation index Drec ≤ 0.7, it indicates a slight deviation, and it is recommended to adjust the total cooling time. When 0.7 < feedback deviation index Drec, it indicates a serious deviation, requiring forced adjustment of the cooling window or activation of abnormal path feedback.

[0046] In this embodiment, the current cooling strategy is evaluated holistically using the matching coverage rate Pco, avoiding the bias of judging the overall state based on a few local cooling points. When the matching coverage rate Pco meets the target, the original strategy is automatically maintained; when the matching coverage rate Pco deviates, the system can automatically determine that there are widespread unreasonable areas in the current cooling configuration, thereby triggering adjustment suggestions. This mechanism breaks away from the traditional manual experience-based strategy setting method and transforms into a unified measurement and triggering mechanism for the entire mold cavity, possessing stronger objectivity and interpretability.

[0047] The system is configured with three feedback response levels, corresponding to operation paths such as maintaining the original strategy, suggesting adjustments to the cooling duration, and forcibly reconstructing the cooling window or path feedback, enabling the system to have differentiated adjustment capabilities. Especially under severe deviation conditions, the system can proactively enter the abnormal path response mechanism to achieve structural-level reconstruction of the cooling strategy, significantly improving the system's adaptability and stability in the face of complex operating conditions. Example 6

[0048] A method for optimizing injection mold design; please refer to [reference needed]. Figure 2 Specifically, it includes the following steps: Step 1: The temperature recovery data acquisition module collects data from each mold cavity during the mold cooling stage after injection molding and fits it into a temperature recovery set WD. Step 2: The temperature normalization and recovery rate calculation module cleans and normalizes the temperature recovery set WD, extracts the temperature recovery rate and its lag time, and obtains the temperature dataset WS. Step 3: The optimal cooling window function construction module is based on the temperature recovery set WS and extracts the optimal cooling time range based on the historical qualified molding cycle to construct the cooling window function LQH; Step 4: The nonlinear cooling trajectory fitting and trend prediction module reconstructs the thermal decay process using a double exponential model, performs nonlinear curve fitting on the temperature recovery trajectory at each point, and predicts the cooling completion time yT at each point. Step 5: The dynamic adjustment suggestion module for cooling time combines the cooling completion time yT with the cooling window function LQH to calculate and obtain the matching coverage Pco. Step Six: The feedback recording and continuous learning module detects, records, and provides feedback on mold forming quality deviations based on the cooling completion time yT and matching coverage Pco.

[0049] In this embodiment, by using a temperature recovery data acquisition module and a normalization analysis module, fine-grained parameters such as temperature recovery rate and hysteresis time are introduced for the first time during the mold cooling stage, enabling precise characterization of the thermal response behavior of each mold cavity. This refined modeling approach breaks through the traditional coarse-grained control mode that uses total cooling time as the sole indicator, laying the foundation for the subsequent formulation of personalized cooling strategies.

[0050] By constructing an optimal cooling window function module, the system extracts the cooling completion time interval from historical qualified samples and establishes a cooling window function as a reference template for the cooling behavior of the mold in the current molding cycle. This method does not rely on manual experience threshold settings, has automatic adaptability, and supports cross-batch consistency analysis, which is conducive to long-term accumulation and optimization. The exponential model reconstructs the thermal decay process, which can more realistically reflect the nonlinear characteristics of rapid decline in the early stage of cooling and slow change in the later stage. This makes the predicted cooling completion time of each mold cavity observation point closer to the actual thermal field evolution process, thereby improving the overall performance of the system in terms of prediction timeliness and spatial resolution.

[0051] By comparing the predicted cooling time with the cooling window function, the system can obtain a set of overall matching coverage indicators for the mold cavity observation points. This indicator has spatial statistical significance and is the core criterion for measuring the global adaptability of the current cooling strategy. It supports the judgment entry point for generating subsequent adjustment suggestions and opens up the evaluation-feedback path.

[0052] The final step involves quality labeling of molding behavior using a quality state function and matching coverage, and calculating a feedback deviation index based on the degree of cooling deviation, forming a feedback database with historical accumulation value. This mechanism can support continuous learning and optimization of subsequent mold design or molding process parameters, realizing the transition of cooling strategies from static settings to dynamic evolution.

[0053] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended technical solutions and their equivalents.

Claims

1. An injection mold optimization design system, characterized in that: It includes a temperature recovery data acquisition module, a temperature normalization and recovery rate calculation module, an optimal cooling window function construction module, a nonlinear cooling trajectory fitting and trend prediction module, a cooling time dynamic adjustment suggestion module, and a feedback recording and continuous learning module. The temperature recovery data acquisition module collects data from each mold cavity during the mold cooling stage after injection molding and fits it into a temperature recovery set WD; The temperature normalization and recovery rate calculation module cleans and normalizes the temperature recovery set WD, extracts the temperature recovery rate and its lag time, and obtains the temperature dataset WS. The optimal cooling window function construction module is based on the temperature recovery set WS and extracts the optimal cooling time range based on historical qualified molding cycles to construct the cooling window function LQH; The nonlinear cooling trajectory fitting and trend prediction module reconstructs the thermal decay process using a double exponential model, performs nonlinear curve fitting on the temperature recovery trajectory at each point, and predicts the cooling completion time yT at each point. The dynamic adjustment suggestion module for cooling time combines the cooling completion time yT with the cooling window function LQH to calculate and obtain the matching coverage Pco; The feedback recording and continuous learning module detects, records, and provides feedback on mold forming quality deviations based on the cooling completion time yT and the matching coverage Pco.

2. The injection mold optimization design system according to claim 1, characterized in that: The temperature recovery data acquisition module collects temperature change information at locations within the mold cavity during the cooling stage after injection molding using sampling sensors. This includes temperature value T, stable temperature of the cooling medium Ten, and cooling start time ts, and fits it into a temperature recovery set WD. The temperature value T is acquired by an embedded temperature sensor, specifically the temperature value T(i,j)t at the j-th observation point in the i-th mold cavity at time t. The stable temperature of the cooling medium, Ten, is acquired through the temperature control output interface of the chiller unit. The cooling start time ts is acquired by reading the trigger signal indicating completion of filling from the injection molding machine controller.

3. The injection mold optimization design system according to claim 2, characterized in that: The temperature normalization and recovery rate calculation module includes a temperature data cleaning and dynamic stability window extraction unit and a thermal behavior normalization expression unit; The temperature data cleaning and dynamic stabilization window extraction unit performs preliminary cleaning on the temperature time series in the temperature recovery set WD, identifies non-cooling intervals and sampling anomalies in the temperature data, and extracts the starting point Teff for each position entering the effective cooling decay stage. The cleaning method is as follows: For each temperature value T, first calculate the first derivative approximation: The approximate method for obtaining the first derivative of the temperature value T is as follows: First, take the difference between the temperature value T(i,j)t of the j-th observation point in the i-th cavity at time t+Δt and the temperature value T(i,j)t of the j-th observation point in the i-th cavity at time t+Δt, and then divide it by the time interval Δt to obtain the result. The effective cooling decay phase starting point Teff is obtained using the following formula: Teff(i,j) = min{t | first derivative of temperature value T < -λ duration greater than ΔT}; In the formula, λ represents the minimum rate of decrease for the cooling trend to start, ΔT represents the minimum time window for the decrease to continue, and Teff(i,j) represents the effective cooling decay stage starting point of the j-th observation point in the i-th cavity; The thermal behavior normalization representation unit normalizes the cleaned temperature recovery set WD to obtain the temperature dataset WS; The temperature dataset WS is obtained using the following formula: ； In the formula, WSo(t) represents the o-th data in the temperature dataset WS at time t, WDo(t) represents the o-th data in the temperature recovery set WD at time t, minWDo represents the valley value of the o-th data in the temperature recovery set WD, and maxWDo represents the peak value of the o-th data in the temperature recovery set WD.

4. The injection mold optimization design system according to claim 3, characterized in that: The optimal cooling window function construction module includes a qualified sample cooling time statistics unit and a cooling window function generation and expression unit; The qualified sample cooling time statistics unit selects qualified molding cycles from all historical molding records, extracts the cooling lag time of each cycle at all mold cavity points, calculates statistical characteristics including the average cooling time, cooling mean, and cooling standard deviation of each cycle, and establishes a cooling time reference distribution model.

5. The injection mold optimization design system according to claim 4, characterized in that: The cooling window function generation and expression unit uses the results of statistical features to construct the cooling window function LQH and identifies whether the current predicted cooling duration falls within a reasonable range. The cooling window function LQH is obtained using the following formula: ； In the formula, LQH(t) represents the cooling window function at time t; Tmin represents the lower extreme boundary of the window, and Tmax represents the upper extreme boundary of the window; When the cooling window function LQH=1, it means that the current predicted cooling time falls within a reasonable range; When the cooling window function LQH=0, it means that the current predicted cooling time does not fall within a reasonable range; The formulas for obtaining the lower extreme boundary Tmin and the upper extreme boundary Tmax of the window are as follows: Tmin = μcool - co1 × σcool; Tmax = μcool + co2 × σcool; In the formula, μcool represents the cooling mean, σcool represents the cooling standard deviation, co1 represents the lower bound bandwidth coefficient, and co2 represents the upper bound bandwidth coefficient.

6. The injection mold optimization design system according to claim 5, characterized in that: The nonlinear cooling trajectory fitting and trend prediction module includes a double exponential decay model fitting unit and a cooling completion time prediction unit. The double exponential decay model fitting unit fits the normalized temperature curve of each cavity observation point (i, j) and uses the double exponential function model to express the entire process of temperature decrease over time. The formula for constructing the double exponential function model is as follows: ； In the formula, AS(i,j)t represents the normalized temperature curve value of the j-th observation point in the i-th mold cavity at time t, A1(i,j) represents the initial intensity of the first exponential decay term at the j-th observation point in the i-th mold cavity, A2(i,j) represents the initial intensity of the second exponential decay term at the j-th observation point in the i-th mold cavity, e represents a constant, k1 represents the decay coefficient of the cooling rate in the first stage, and k2 represents the decay coefficient of the cooling rate in the second stage. The cooling completion time prediction unit is based on a double exponential function model to calculate and obtain the cooling completion time yT for each point. The cooldown completion time yT is obtained using the following formula: ； In the formula, yT(i,j) represents the cooling completion time of the j-th observation point in the i-th mold cavity, Tend represents the preset cooling completion time, ts represents the cooling completion judgment threshold, and min{...} represents the earliest time that meets the condition among all times after the start time.

7. The injection mold optimization design system according to claim 6, characterized in that: The dynamic adjustment suggestion module for cooldown time includes a matching judgment and coverage calculation unit and an adjustment suggestion generation unit; The matching judgment and coverage calculation unit compares the predicted cooling completion time yT of all cavity observation points with the optimal cooling window function LQH point by point to determine whether it falls within the window, and counts the proportion of all points that meet the conditions to obtain the matching coverage Pco. The judgment method is as follows: traverse all the mold cavity observation points in the mold cavity, obtain the predicted cooling completion time yT of the mold cavity observation point; then, substitute this time into the cooling window function for judgment. When this time point is within the optimal cooling time range, the cooling window function will output a result of 1; If this time point is not within the optimal range, the function will output a result of 0; Match coverage Pco is obtained using the following formula: ； In the formula, M represents the total number of mold cavities, and N represents the number of observation points in each mold cavity. This represents a Boolean conditional function.

8. The injection mold optimization design system according to claim 7, characterized in that: The adjustment suggestion generation unit determines whether the current cooling configuration is reasonable based on the evaluation results of the matching coverage Pco, and provides a suggested cooling duration Tncool when it deviates from the optimal range, and outputs structured suggestions; The suitability of the cooling configuration is determined by matching the following methods: When the matching coverage Pco ≥ 0.8, it indicates that the observation point has cooled down sufficiently and no adjustment is needed; When the matching coverage Pco < 0.8, it indicates a wide deviation, so the cooldown time is extended and a suggested cooldown time Tncool is obtained; The recommended cooldown time Tncool can be obtained using the following formula: ； The structured recommendations include status labels, matching coverage values, suggested cooldown time, and optional tips.

9. The injection mold optimization design system according to claim 8, characterized in that: The feedback recording and continuous learning module archives and labels the cooling behavior parameters, cooling completion time yT, quality status judgment function Qua, and matching coverage Pco, obtains the feedback deviation index Drec, and establishes feedback paths to adjust future cooling strategies. The quality status determination function Qua is obtained as follows: When the matching coverage Pco ≥ 0.8, the quality status judgment function Qua = 1; When the matching coverage Pco < 0.8, the quality status judgment function Qua = 0; The feedback deviation index Drec is obtained using the following formula: ； In the formula, Drec(tk) represents the feedback deviation index of the k-th cooling process, yT(i,j)tk represents the cooling completion time of the j-th observation point in the i-th cavity of the k-th cooling process, and yTopt represents the theoretical ideal cooling completion time. The future cooling strategy will be adjusted as follows: When the feedback deviation index Drec < 0.3, it indicates a normal situation, the cooling behavior is stable, and the original strategy should be maintained. When 0.3 ≤ feedback deviation index Drec ≤ 0.7, it indicates a slight deviation, and it is recommended to adjust the total cooling time. When 0.7 < feedback deviation index Drec, it indicates a serious deviation, requiring forced adjustment of the cooling window or activation of abnormal path feedback.

10. A method for optimizing the design of an injection mold, applied to the injection mold optimization design system according to any one of claims 1 to 9, characterized in that: Includes the following steps: Step 1: The temperature recovery data acquisition module collects data from each mold cavity during the mold cooling stage after injection molding and fits it into a temperature recovery set WD. Step 2: The temperature normalization and recovery rate calculation module cleans and normalizes the temperature recovery set WD, extracts the temperature recovery rate and its lag time, and obtains the temperature dataset WS. Step 3: The optimal cooling window function construction module is based on the temperature recovery set WS and extracts the optimal cooling time range based on the historical qualified molding cycle to construct the cooling window function LQH; Step 4: The nonlinear cooling trajectory fitting and trend prediction module reconstructs the thermal decay process using a double exponential model, performs nonlinear curve fitting on the temperature recovery trajectory at each point, and predicts the cooling completion time yT at each point. Step 5: The dynamic adjustment suggestion module for cooling time combines the cooling completion time yT with the cooling window function LQH to calculate and obtain the matching coverage Pco. Step Six: The feedback recording and continuous learning module detects, records, and provides feedback on mold forming quality deviations based on the cooling completion time yT and matching coverage Pco.

Images