Testing method of TPU sheath production mold and related equipment

Abstract

The invention discloses a TPU sheath production mold test method and related equipment, and the method comprises the steps: in a production process, collecting production process data corresponding to each mold number, and arranging the data into a time sequence associated with a cavity or runner identifier according to a time sequence; selecting a reference datum from the time sequence, and generating stage deviation data organized according to the process stage; associating the stage deviation data to a mold structure position according to a corresponding rule of the process stage and the mold structure position, and obtaining a position association result; within a range limited by the position correlation result, extracting quantization parameters related to the surface state based on the corresponding time sequence, and generating a state index of each position according to the quantization parameters; and organizing the state indexes of all the positions according to the mold number to obtain spatial distribution of the state of the inner surface of the mold and trend evaluation data changing along with the mold number. According to the scheme, online identification and quantitative evaluation of the surface deposition or wear state of the mold can be realized, and targeted maintenance is supported.

Description

Technical Field

[0001] This invention relates to the field of online inspection and condition assessment technology for injection molds, and in particular to a testing method and related equipment for TPU sheath production molds. Background Technology

[0002] Thermoplastic polyurethane (TPU) casing production molds are specialized tooling for injection molding TPU protective shells. Their cavity structure includes a main runner, branch runner systems, and a complex molding cavity. Molten TPU material enters the mold's runner system, fills the cavity under injection pressure, and then cools and solidifies. These molds typically incorporate functional features such as button island areas, mirrored display areas, and reinforcing ribs to meet the diverse needs of electronic product protective shells. The surface condition of the mold directly affects the flow characteristics and heat transfer efficiency of the TPU melt, as well as the surface quality and dimensional accuracy of the final product.

[0003] During the continuous production of TPU sheaths, a deposition layer gradually forms on the mold surface due to the migration of additives, pigment deposition, and accumulation of molecular chain degradation products in the TPU material. Simultaneously, the scouring effect caused by high-speed melt flow also leads to wear on the mold surface. These surface changes require detection to determine the timing of cleaning and maintenance. In existing technologies, the detection of mold surface conditions mainly relies on direct observation during periodic shutdowns and mold disassembly, or indirect judgment through random sampling of product appearance quality. While periodic maintenance strategies have clear operational specifications and stable execution processes, their maintenance cycles are often based on experience and are difficult to adapt to differences in deposition rates under different material formulations and production conditions. Although random sampling of product quality can detect mold problems that have already affected the product, it has a significant lag; by the time quality defects are discovered, a batch of defective products has often already been produced. Furthermore, neither of these methods can obtain quantitative information such as the deposition distribution, thickness changes, and development trends on the mold surface in real time during production, making it impossible to accurately predict the remaining service life of the mold and the optimal cleaning time. This lack of detection capability makes production planning difficult to optimize, resulting in either premature shutdowns for cleaning leading to capacity losses or delayed maintenance leading to batch quality incidents. Therefore, there is currently a lack of effective methods for quantitatively detecting and predicting the trend of deposition on the inner surface of a mold without interrupting production or disassembling the mold. Summary of the Invention

[0004] The main objective of this invention is to achieve non-destructive testing of the deposition state on the mold surface by analyzing the response characteristics of injection molding process parameters during continuous production, thereby solving the technical problem that existing technologies cannot quantitatively assess changes in mold surface quality and predict their development trends online.

[0005] To achieve the above objectives, this application provides a testing method for TPU sheath production molds, including: Without interrupting production or disassembling the mold, collect production process data corresponding to each mold batch and organize them into a time series associated with cavity or flow channel identifiers in chronological order. A reference baseline is selected from the time series to generate stage deviation data organized by process stage; According to the correspondence rules between process stages and mold structure positions, the stage deviation data is associated with the mold structure positions to obtain position association results; Within the range defined by the location association results, quantitative parameters related to the surface state are extracted based on the corresponding time series, and state indicators for each location are generated accordingly. The state indicators at each location are organized according to the mold cycle to obtain spatial distribution and trend evaluation data of the state of the inner surface of the mold as a function of the mold cycle.

[0006] To achieve the above objectives, this application also proposes a testing device for TPU sheath production molds, comprising: The data acquisition module is used to collect production process data corresponding to each mold batch without interrupting production or disassembling the mold, and organize it into a time series associated with the cavity or flow channel identifier in chronological order. The reference module is used to select a reference reference from the time series and generate stage deviation data organized by process stage; The association module is used to associate the stage deviation data with the mold structure position according to the correspondence rules between the process stage and the mold structure position and obtain the position association result; The feature extraction module is used to extract quantitative parameters related to the surface state based on the corresponding time series within the range defined by the location association results, and generate state indicators for each location accordingly. The evaluation module is used to organize the state indicators at each location according to the mold cycle to obtain evaluation data on the spatial distribution of the state of the inner surface of the mold and the trend of change with the mold cycle.

[0007] To achieve the above objectives, this application also proposes a testing device for TPU sheath production molds, comprising: a memory and at least one processor, wherein the memory stores instructions, and the memory and the at least one processor are interconnected via a circuit; the at least one processor invokes the instructions in the memory to cause the testing device for TPU sheath production molds to execute the steps of the above-described testing method for TPU sheath production molds.

[0008] To achieve the above objectives, embodiments of this application also propose a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the steps of the above-described testing method for TPU sheath production molds.

[0009] The technical solution provided in this application directly collects production process data corresponding to each mold cycle without interrupting production or disassembling the mold. This data is then organized chronologically into a time series associated with cavity or flow channel identifiers, enabling continuous monitoring throughout the entire mold operation cycle. This approach bypasses the limitations of traditional inspection methods, which require machine shutdown, mold disassembly, or reliance on product sampling, ensuring real-time and complete information collection. After obtaining the time series, a reference benchmark is established by selecting stable data segments. Subsequent mold cycle data is segmented according to process stages such as filling, holding pressure, and cooling. This allows for comparison of data changes under the same stage conditions, identifying subtle shifts caused by deposition or wear.

[0010] To enable spatial localization of the detection results, the solution utilizes the correspondence between process stages and mold structural positions, linking segmented deviation data to specific cavity or runner locations to construct a spatial distribution map of deposition or wear. Based on this, the analysis scope is limited to data at matched locations, and quantitative parameters directly reflecting changes in mold surface condition, such as pressure curve characteristics and temperature change rates, are extracted and converted into state indicators for each location. These indicators, organized by time series according to mold batches, form a spatial distribution and trend assessment of mold surface condition, not only showing the degree of deposition or wear at different locations but also displaying its trajectory as production progresses. This organic combination of continuous monitoring, stage comparison, spatial localization, and trend analysis allows for the acquisition of quantitative information on deposition distribution and development trends in advance during production, supporting targeted maintenance and determination of optimal cleaning times. This solves the core problems of previous detection methods, such as lag, reliance on downtime, and lack of quantitative prediction. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0012] Figure 1 This is a schematic diagram of an embodiment of the testing method for TPU sheath production molds in this invention; Figure 2 This is a schematic diagram of one embodiment of the testing device for TPU sheath production molds in this invention; Figure 3 This is a schematic diagram of one embodiment of the testing equipment for the TPU sheath production mold in this invention. Detailed Implementation

[0013] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0014] A thermoplastic polyurethane (TPU) sheath production mold is a specialized tooling used for injection molding TPU protective shells. Its structure typically includes a main runner, branch runners, and a molding cavity. The main runner introduces molten TPU material into the mold, while the branch runners distribute the melt to different molding areas. The molding cavity shapes the specific form of the product. Common functional structures within the molding cavity include: button island areas, which are localized raised structures corresponding to the side buttons of electronic products, offering flexible, pressable characteristics while requiring sufficient structural strength to maintain shape stability; mirror display areas, which are areas requiring a smooth surface with high gloss or transparency, typically used for display windows or decorative surfaces, and are sensitive to flow marks, scratches, and minor defects; and rib reinforcement structures, which are slender rib structures formed on the inner surface of the product, enhancing its rigidity and resistance to deformation with a smaller material usage. These different functional areas exhibit significant differences in melt flow resistance, heat transfer rate, and surface quality requirements.

[0015] In the injection molding process of TPU sheaths, TPU granules are heated to a molten state and then injected under high pressure into the runners along the main runner. They then fill the functional areas sequentially through the gate, followed by a holding pressure stage to stabilize the product shape. Finally, the product solidifies and is demolded during the cooling stage. As production continues, additives, colorants, and degradation products in the melt gradually deposit at runner nodes and on the cavity surface. Simultaneously, high-speed melt flow and repeated filling cause abrasion on the cavity surface. These deposits and abrasions alter the local dimensions, roughness, and heat transfer performance of the cavity, affecting the melt filling time, holding pressure stability, and cooling uniformity, ultimately leading to a decline in product surface quality. Existing detection methods mostly rely on direct observation during machine shutdown and mold removal or random sampling of product appearance. The former interrupts production and is time-consuming, while the latter suffers from detection lag, and both methods struggle to obtain quantitative information such as deposition location, thickness changes, and development trends during production. This solution collects process data related to the mold structure and location without interrupting production or disassembling the mold. Combined with process stage analysis and location correlation estimation, it extracts quantitative parameters reflecting changes in surface condition in real time and organizes and evaluates them in the spatiotemporal dimensions, thereby achieving continuous monitoring and trend prediction of the deposition and wear state on the mold surface.

[0016] One embodiment of this application provides a testing method for TPU sheath production molds. Figure 1 A flowchart illustrating a testing method for a TPU sheath production mold provided in one embodiment of this application. In this embodiment, the method includes: Please see Figure 1 Under the condition of not interrupting production and not disassembling the mold, the production process data corresponding to each mold batch is collected and organized into a time series associated with the cavity or flow channel identifier in chronological order. In one embodiment of the present invention, the step of collecting production process data corresponding to each module and organizing it chronologically into a time series associated with cavity or flow channel identifiers includes: Time alignment of data from different acquisition channels is performed based on identifiable feature events in the process. The collected data is bound to the corresponding cavity or runner position using the mold structure design information. When direct binding is not possible, the position is estimated by combining the runner distribution and process records. The position is then solidified after the consistency of data from consecutive mold cycles is verified. Data that does not conform to stable mass production conditions or is discontinuous in terms of modules are removed to obtain the time series.

[0017] The following is a detailed description of the steps involved in the above embodiments: Specifically, identifiable characteristic events refer to events or curve features that can stably mark the boundaries of process stages within a single cycle, such as injection start (screw speed jumps from zero and continues), injection to holding pressure transition (control signal switching while mold cavity pressure plateaus), cooling start (cooling circuit valve opening / closing signal or mold temperature curve slope change), mold opening and ejection completion; acquisition channels refer to process data channels from different sources, such as screw position / speed, injection pressure, mold cavity pressure, clamping force, mold temperature, cooling flow / temperature, cycle timing and timestamps, etc. The time alignment implementation process is as follows: using the timestamp of the equipment control system (Programmable Logic Controller, PLC) as a coarse alignment reference, opening windows in the event neighborhood on each channel (e.g., 50–200 ms before and after the event, or 1%–5% of the cycle time) to detect trigger points; using threshold + hysteresis for step events, and using derivative extrema or cross-correlation peaks to determine the arrival time for continuous curves; calculating the offset of each channel relative to the main anchor point, and performing translation correction on the channel time axis. If the channel sampling rates are inconsistent, resampling to a unified frequency (not lower than the highest original sampling rate) and using linear or cubic spline interpolation are performed to avoid phase errors. For channels with integer delays, the event phase residuals are minimized across multiple modules to eliminate fixed communication time differences. The criteria for successful alignment are: the time differences of the three events—injection start, pressure holding, and cooling start—on each channel all fall within a preset tolerance (e.g., ±5–20ms) and remain stable within adjacent modules. This step anchors multi-source data to the same module time axis, eliminating trigger delays and sampling drift, making subsequent segmented comparisons comparable. An equivalent implementation can use template curve registration: using the cavity pressure during the quality stabilization period as a template, the optimal alignment point is searched through a sliding window with minimum mean square error to achieve alignment accuracy comparable to event anchoring.

[0018] Mold structure design information refers to data including the numbers, topological relationships, and geometric parameters of runners / gates / cavities (2D assembly drawings, 3D structures, or electronic bill of materials (BOM)); position binding refers to establishing a one-to-one correspondence between aligned data channels and specific cavity or runner identifiers; position estimation is used for channels that cannot be directly bound; continuous mold data consistency verification refers to checking whether the key timing and morphology of the curve corresponding to the same position identifier are stable within N consecutive mold cycles (e.g., N=20); fixed position identifiers refer to using the verified binding relationships as fixed mappings for subsequent processing. During implementation, the mold structure design information is read to form a mapping table of "branch sequence—effective length—gate number," and channels with clearly marked cavity numbers are directly bound. For channels without location identifiers, the order of responses during the filling stage is matched with the branch sequence: the filling start time and the pressure holding platform establishment time of each channel are calculated to obtain the time sequence between channels, which is then matched with the expected sequence of "upstream branches before downstream, and short branches before long branches"; when multiple solutions occur, the pressure holding platform establishment speed and cooling section slope are introduced as secondary criteria. After initial binding is completed, consistency verification is performed: within N consecutive modules, the event sequence and curve shape difference (such as pressure holding platform length and filling end slope) of the same channel are required to remain within the threshold (e.g., the sequence is not reversed, and the shape difference does not exceed ±k·MAD (Median Absolute Deviation) of the median), where k is the outlier judgment coefficient, used to adjust the strictness of outlier screening, and its value is set according to the fluctuation tolerance of the production process. In the inspection scenario of TPU sheath production molds, the typical value of k ranges from 2 to 3.5. A smaller k value can improve the detection sensitivity of subtle anomalies, while a larger k value can reduce the probability of false rejection due to normal fluctuations. Once the target is met, the location is fixed; if the target is not met, the process reverts to the previous step and re-estimates. This process establishes a stable mapping between time-domain signals and spatial locations, avoiding unclear attribution during subsequent deviation tracing. An equivalent implementation method is micro-perturbation verification: within permissible limits, a very small step is applied to the injection speed or holding pressure to verify whether the differential response of adjacent candidate locations conforms to the estimated flow channel level, thereby accelerating the verification process.

[0019] Stable mass production conditions refer to process parameters being within the process card limit window, material batches and color matching records remaining unchanged, and equipment status being in mass production operation; discontinuous mold cycles refer to time axis breaks caused by machine stoppage, mold cleaning, material change, or data acquisition interruption; time series refers to multi-channel data sequences arranged in chronological order and associated with cavity or flow channel identifiers. The data screening process is as follows: Non-mass production cycles (such as trial molding, machine adjustment, cleaning, and color masterbatch switching) are eliminated based on the status tags and process cards of the Manufacturing Execution System (MES) / Programmable Logic Controller (PLC), and verified using key parameter thresholds (melt temperature, mold temperature, metering volume, cycle time, clamping force, and back pressure). Channel integrity is verified, eliminating cycles with out-of-order timestamps, missing fields, or lost packets. Within the acceptable window, outliers are eliminated using robust statistical methods with a sliding window (e.g., median ± k·MAD or interquartile range for filling time, holding pressure, and cooling time) to ensure sequence continuity and comparability. For the first batch of cycles after a long shutdown, a "re-preheating" exemption segment can be set, and the cycles are included in the sequence starting from the specified cycle. Finally, the aligned data of the retained cycles are organized by cavity / flow channel identifier and timestamp to form a time series for subsequent segmented comparison and positional association. This step ensures that the input data reflects a stable mass production process rather than systemic fluctuations caused by start-up / shutdown transitions or formula changes, thus allowing subsequent identification to focus more on gradual changes caused by surface deposition / wear. An equivalent implementation can introduce quality gating as a parallel condition, for example, excluding modules whose part weight or online appearance score is below a threshold from the time series, to improve data purity.

[0020] Please continue reading. Figure 1 A reference baseline is selected from the time series to generate stage deviation data organized by process stage; In one embodiment of the present invention, the step of selecting a reference baseline from the time series and generating stage deviation data organized by process stage includes: Select continuous data segments that remain stable within a set tolerance range from the time series as reference benchmarks, and establish them separately according to the mold functional areas; Based on the unique physical responses of the filling, holding, and cooling stages in different functional areas, the data of subsequent modules are segmented and compared with the corresponding stages of the reference benchmark to calculate the relative deviation value. In the comparison results, single-mode instantaneous outliers caused by non-deposition factors are removed, and continuous deviation trends consistent with deposition or wear patterns are retained as stage deviation data.

[0021] The following is a detailed description of the steps involved in the above embodiments: Specifically, the reference benchmark refers to a set of stable segments used to compare data from the same process stage in subsequent mold cycles; the mold functional area refers to the region within the molding cavity that is distinguished by geometry and purpose (such as the button structure area, optical appearance area, reinforcing rib area, etc.), and different functional areas have significantly different responses to filling, holding pressure, and cooling. In implementation, the input is a time series that has been time-aligned and position-bound. On the mold cycle axis of continuous mass production, it is first grouped by "part number / color scheme / equipment status" and limited to stable mass production conditions; then, within each mold functional area, representative quantities for the three stages of filling, holding pressure, and cooling are calculated (e.g., filling time, steady-state mean of holding pressure platform pressure, slope of cooling section temperature or pressure), and stable and continuous segments are selected using robust statistics: the representative quantity of the stage is within a set tolerance range (which can be determined by the median of the historical stable period ± k·absolute median difference, where k is the outlier determination coefficient) and its length is not less than the preset minimum number of consecutive mold cycles L (e.g., 50–200 cycles, used to suppress occasional fluctuations and retain stable operating conditions). For each functional area and each stage, a reference baseline is established (which can be stored as the median, interquartile range, and timestamp range) as a comparison for subsequent comparisons. The significance of this approach is to establish a baseline within the functional area that matches its physical characteristics, avoiding the masking of local differences with a global average, thus making subsequent deviations interpretable and comparable. An equivalent implementation can use a percentile envelope baseline: a tolerance band is formed using P25–P75 of the historical stable period, and the median is taken as the reference baseline, serving the same purpose as a robust baseline of median ± k·absolute median difference.

[0022] After obtaining the reference baseline, the data for subsequent cycles needs to be segmented according to process stages and compared with the reference baseline to obtain relative deviation values. Stage boundaries can be determined by aligned event anchor points (injection start, pressure holding transition, cooling start) or by curve shape (e.g., the switching moment corresponding to the zero-crossing point of the first derivative of the mold cavity pressure). In each functional area and each stage, the representative quantity related to the process is calculated and normalized and differnted with the corresponding statistics of the reference baseline to obtain a deviation measure of the same dimension. For example, the relative deviation can be calculated as Δr=(xb) / |b|, where x is the representative quantity of the cycle in that stage, and b is the representative quantity of the reference baseline (e.g., median). To reduce the influence of noise, a sliding median or low-pass filter can be applied within a short window before outputting the deviation sequence. Taking the button structure area as an example, if the representative quantity of the filling stage is the filling time or the final pressure slope, a continuously positive Δr indicates that the flow resistance increases with each mold cycle. Taking the optical appearance area as an example, if the representative quantity of the holding pressure stage is the steady-state average of the platform pressure or the platform establishment time, an increased deviation reflects obstructed holding pressure feeding or thermal management imbalance. In the reinforcing rib area, a decrease in the cooling section slope corresponds to localized heat transfer obstruction. The significance of this step is to align the changes in the same stage and the same area to the same benchmark, eliminating confusion caused by incomparability across stages and regions. An equivalent implementation method can adopt a multi-index weighted difference form: after dimensionless transformation of the representative quantities of multiple stages, the weights are summed according to their weights. The weights can be determined by inversely proportional to the variance of the historical stable period to reduce the influence of high-noise indicators.

[0023] To avoid interference from non-deposition factors, single-mode instantaneous outliers need to be removed from the comparison results, and only continuous deviation trends consistent with deposition or wear patterns should be retained as stage deviation data. Non-deposition factors include, but are not limited to, short-term fluctuations in melt temperature, short-term jitter in clamping force, and instantaneous disturbances in cooling media, which can be determined jointly through equipment logs and process parameters. In implementation, robust outlier detection is first performed on the deviation sequence: within a sliding window, thresholds are set using the median and median absolute deviation (MAD) or the interquartile range (IQR). Isolated points outside the median ± k·MAD or [Q1-k·IQR, Q3+k·IQR] are considered single-mode instantaneous outliers and removed. Then, the continuous deviation trend is determined: within a running window of length W, at least M deviations must be in the same direction and their amplitude must exceed the threshold θ, or their cumulative deviations must be continuously increasing / decreasing; a cumulative sum statistic (CUSUM) can also be used. Parameters W, M, θ, and k can be set based on historical stable period data and engineering tolerance. For example, W can be 20–50 cycles, M can be 50%–70% of W, and θ can correspond to a multiple of the reference variance. Through this process, isolated process noise is eliminated, while slow, unidirectional, and sustainable offsets caused by deposition or wear accumulation are retained, thus obtaining stage deviation data for subsequent location correlation and risk assessment. An equivalent implementation can use a combination of run-length criterion and change point detection: first, the shortest run-length is used to screen unidirectional segments, and then a change point algorithm based on robust variance is used to confirm the trend starting point, which can obtain a trend set consistent with the above method.

[0024] In one embodiment of the present invention, the single-mold instantaneous abnormal value includes at least the filling stage deviation caused by melt temperature fluctuation, the holding stage deviation caused by short-term clamping force jitter, and the cooling stage deviation caused by cooling medium disturbance.

[0025] It should be noted that in the processing of stage deviation data, instantaneous outliers in a single mold cycle refer to short-term, isolated deviations caused by non-deposition or non-wear factors within a single mold cycle. These deviations are not directly related to the formation of a deposit layer or wear on the mold surface, and if not removed, they will interfere with the judgment of long-term trends. Typical types include three situations: First, deviations in the filling stage caused by melt temperature fluctuations. During injection molding, the melt temperature is controlled by the barrel heating system, but due to uneven preheating of the raw material, short-term instability of the heating zone, or sudden changes in the moisture content of the raw material, the flow viscosity in the filling stage will change instantaneously, manifesting as abnormal changes in characteristic quantities such as filling time and filling pressure slope. Second, deviations in the holding pressure stage caused by short-term clamping force fluctuations. The clamping force is provided by a hydraulic or electric system. Short-term hydraulic fluctuations, motor response delays, or mechanical structure resonance can cause abnormal fluctuations in the holding pressure curve, resulting in deviations from the normal range for indicators such as single-mold cycle platform pressure and holding pressure build-up time. Third, deviations in the cooling stage caused by cooling medium disturbances. In circulating water or oil cooling systems, the temperature and flow rate of the cooling medium are affected by the switching of external heat exchange equipment or the transient change in pump speed, which can cause sudden changes in the cooling rate, thereby affecting the slope of the pressure or temperature drop curve during the cooling stage.

[0026] These three types of parameters were chosen as the focus for outlier detection because they have a high probability of occurrence in actual production, short duration, and negligible contribution to long-term trends. Their corresponding physical causes can be cross-validated through equipment operation logs and sensor data, enabling rapid identification and removal during data processing. This is significant in preventing short-term process fluctuations from being misjudged as deposition trends and ensuring that stage deviation data reflects the cumulative changes in the mold surface condition.

[0027] Without deviating from the core principles of this solution, other transient factors unrelated to the mold's condition can also be included in the outlier range. Examples include first-mold deviations during raw material batch changes, abnormal mold closing speeds due to uneven lubrication of the mold opening and closing mechanism, and localized pressure anomalies caused by momentary obstruction of the mold's venting system. These extended types are handled in the same way as the three types of factors mentioned above in the detection logic; they can all be determined and eliminated through parameter fluctuation characteristics and equipment records, thereby ensuring the stability and reliability of the trend analysis.

[0028] Please continue reading. Figure 1 According to the correspondence rules between process stages and mold structure positions, the stage deviation data is associated with the mold structure positions to obtain position association results; In one embodiment of the present invention, the step of associating the stage deviation data with the mold structure position and obtaining the position association result according to the correspondence rule between the process stage and the mold structure position includes: Establish the correspondence between each process stage and the response patterns of different functional areas within the cavity under the influence of deposition; Based on the flow channel distribution, branch sequence, and time difference of functional area response, the stage deviation data is matched to the mold structure position. When the matching conditions are not fully met, the position is estimated by using the correlation between flow channel length and response delay, and deviation sources that do not conform to the deposition mechanism are eliminated. Positions that can be physically explained as being caused by deposition are retained as position correlation results.

[0029] The following is a detailed description of the steps involved in the above embodiments: During deposition state analysis, the response patterns of each process stage and different functional zones within the cavity need to be defined one-to-one. Functional zones refer to areas within the cavity with specific geometries and uses, such as button island areas, mirror display areas, and reinforcing rib structures. The formation rate, thickness distribution, and impact on the filling, holding, and cooling stages of the deposition layer differ in these areas. Therefore, typical response characteristic parameters of functional zones under deposition influence (such as filling time delay, holding pressure plateau deformation, and cooling curve slope changes) can be summarized to form a correspondence table. In practice, existing production data or pre-set experimental data can be used as a benchmark, and patterns matching this correspondence can be identified in newly acquired data. This method can correlate stage deviation data with the physical changes of specific functional zones, ensuring the physical interpretability of the analysis results. This step transforms deposition state judgment from an abstract analysis relying solely on time-series data to a physical determination directly related to the specific cavity structure and material flow path, thereby improving the accuracy and reliability of deposition positioning. The equivalent approach can simulate the effects of deposition on flow and heat transfer using finite element simulation software, and use the simulation results as a response mode reference to achieve the same physical correlation effect.

[0030] Based on the established correspondence, the stage deviation data can be matched to the specific structural location of the mold by utilizing the flow channel distribution, branching sequence, and time difference of functional area response. Flow channel distribution refers to the path structure of the melt within the mold from the main channel to each branch channel and the final cavity; branching sequence describes the order in which the melt arrives at each functional area; and functional area response time difference reflects the order of response of each functional area during filling, holding, or cooling processes. By comparing these characteristics with the time points of occurrence of the stage deviation data, the deviation data can be accurately matched to the corresponding location. When direct matching conditions cannot be met due to missing data or signal interference, the possible location of the deviation can be estimated using the linear or nonlinear correlation between flow channel length and response delay. The longer the flow channel and the greater the flow resistance, the more significant the corresponding response delay; this physical characteristic can be used as a basis for estimation. This step ensures relatively accurate location positioning even when some signals are interfered with. An equivalent implementation can use the peak propagation time of the filling pressure waveform instead of the temperature response delay for estimation, still achieving the positioning effect based on the flow path.

[0031] After completing location matching or estimation, it is necessary to eliminate sources of deviation that do not conform to the deposition mechanism. The deposition mechanism refers to the physicochemical process by which additives, pigments, or degradation products in the melt gradually adhere to and accumulate on the cavity surface to form a deposited layer under specific production conditions. Deviations that do not conform to this mechanism include anomalies caused by mechanical vibration, hydraulic fluctuations, and changes in ambient temperature. Although these factors can cause data fluctuations, they do not cause actual deposition or wear. By eliminating these interfering factors, only deviation locations that can be physically explained as being caused by deposition are retained, ultimately forming the location correlation results. This approach establishes the deposition trend analysis on the basis of real physical causes, reducing the risk of misjudgment. Equivalently, the elimination criteria can also be extended to include long-term stable structural features that are unrelated to deposition (such as constant response differences caused by insufficient local cooling), ensuring the relevance and reliability of the final results.

[0032] In one embodiment of the present invention, the stage deviation data is matched to the mold structure position based on the flow channel distribution, branch sequence, and time difference of functional area response. When the matching conditions are not fully met, the position is estimated using the correlation between flow channel length and response delay, and deviation sources inconsistent with the deposition mechanism are eliminated. Positions that can be physically explained as being caused by deposition are retained as the position correlation result, including: Based on the mold's three-dimensional structural data and process history, a time difference pattern for each functional area during the filling, holding, and cooling stages is established, and the stage deviation data is compared with the time difference pattern one by one to determine the candidate positions. For candidate locations that fail to meet all matching conditions, the expected response delay value is calculated using the effective length of the flow channel, the melt viscosity range, and the injection speed, and then fitted with the measured delay to estimate the most likely location. The matched and estimated locations are compared with the sedimentation mechanism model, and locations that do not conform to the sedimentation formation law due to flow velocity or temperature conditions are eliminated, retaining only the location results that conform to the physical interpretation.

[0033] The following is a detailed description of the steps involved in the above embodiments: Specifically, the time difference model is the stable pattern of the response sequence and relative time intervals of each functional area in the three stages of filling, holding pressure, and cooling. Inputs include time series with completed time alignment and position binding, mold 3D structural data (Computer-Aided Design, CAD), and process records from historical stable periods. The implementation involves extracting the filling start time (e.g., screw displacement inflection point or cavity pressure rise point), the holding pressure plateau establishment time (time when the pressure plateau reaches steady state), and the cooling linear segment start time (time when the temperature / pressure drop curve enters the steady state segment) within each functional area. These times are then calculated as a ternary time difference relative to the same reference event (e.g., nozzle pressure rise point). A tolerance band is formed using the median of historical stable periods and robust outlier scales (e.g., absolute median difference), resulting in a "functional area × stage" time difference table. When online data is received, the occurrence time of the corresponding event in the stage deviation data is compared one by one with the time difference table. Functional areas where at least two stages simultaneously fall within the tolerance zone and the overall deviation is minimized are recorded as candidate positions, along with a matching score (a combination of the number of matching stages and the deviation amount). This method narrows down time-domain anomalies to a few spatial candidates, reducing mismatches caused by indiscriminate searches. If the field prefers a quantile interval representation, the absolute median can be replaced by P25–P75 to construct the tolerance zone, achieving the same effect and facilitating parameter maintenance.

[0034] When candidate locations do not meet all matching conditions or have multiple solutions, a response delay estimation based on flow path and physical property parameters is introduced. Here, the effective flow path length refers to the equivalent distance from the main flow channel collector along the centerline to the target gate inlet, with length correction coefficients set according to abrupt changes in cross-section, corners, and flow splitting losses. The melt viscosity range is taken from the Technical Data Sheet (TDS) or values ​​obtained from an online rheometer within the process temperature and shear rate range. The injection speed is directly provided by the equipment process record. The implementation process is as follows: A field calibration is performed using samples from the historical stable period showing "known location – known delay," forming a reference relationship curve of "effective length, viscosity, injection speed → expected delay." During online operation, the three parameters of the current batch are substituted into this reference relationship to obtain the expected delay, and residual calculation is performed with the measured delay. The location with the smallest residual is selected as the most probable location. If multiple residuals are close, a small number of parallel candidates are retained for the next stage. The rationale for these parameter selections is that the greater the length, the higher the viscosity, and the lower the injection speed, the later the filling and holding pressure characteristics arrive, resulting in a greater expected delay. If a rheometer is not available, the viscosity can be approximated by a combination of melt temperature and back pressure, or the volumetric flow rate (metered volume / filling time) can be used instead of the injection speed to ensure the putative chain is feasible.

[0035] To avoid mistaking anomalies unrelated to deposition for final location, physical consistency verification is required. The deposition mechanism model here refers to a set of criteria used to determine whether a location meets the necessary conditions for deposition accumulation under mass production conditions. It includes at least three types of quantities: local average flow velocity (converted from volumetric flow rate and gate / local cross-sectional area), wall temperature range (estimated from mold temperature and the temperature difference between the inlet and outlet of the cooling circuit), and geometric retention characteristics (identified by CAD based on abrupt changes in cross-section, dead corners at the ends, and small-radius corners). During implementation, the above three types of quantities are calculated for each candidate location and compared with the preset deposition criteria: locations with flow velocities in the low-velocity, easily deposited range, wall temperatures falling within the material adhesion-prone zone, and exhibiting geometric retention patterns are retained; those that do not meet the criteria are discarded. The criteria are set based on stable mass production statistics of the same mold and the TDS process window, and can be calibrated on a batch-by-batch basis. To improve robustness, it is recommended to confirm retention only after meeting two or more criteria within W consecutive mold runs (e.g., 20–50). An equivalent implementation method is a risk scoring table: scores are assigned to flow rate, wall temperature, and geometric retention, and thresholds are set. Candidates below the threshold are eliminated, and the output is a location result consistent with the criterion comparison method. Through this verification, only locations that can be physically explained as being caused by deposition are retained, reducing false positives and improving the reliability of subsequent trend assessments.

[0036] Please continue reading. Figure 1 Within the range defined by the location association results, quantitative parameters related to the surface state are extracted based on the corresponding time series, and state indicators for each location are generated accordingly. In one embodiment of the present invention, the step of extracting quantization parameters related to the surface state based on the corresponding time series within the range defined by the location association results, and generating state indicators for each location accordingly, includes: Within the functional area defined by the location association results, quantitative parameters that can characterize the physical effects of surface deposition or wear in the functional area are extracted from the time series. The quantitative parameters include at least one of the pressure change curve characteristics during the filling stage, the pressure stability characteristics during the holding stage, and the temperature change rate characteristics during the cooling stage. Non-depositional factors are eliminated from the quantification parameters, including excluding instantaneous anomalies caused by melt temperature fluctuations, clamping force jitter, or cooling medium disturbances. The quantization parameters after interference removal are normalized, and state indicators are generated to characterize the degree of surface state change in each functional area.

[0037] The following is a detailed description of the steps involved in the above embodiments: Within the functional area defined by the location correlation results, quantitative parameters refer to the features extracted from the time series that can numerically characterize the physical effects caused by surface deposition or wear at each process stage. In specific implementation, each module is first divided into three stages—filling, holding, and cooling—based on the completed event anchor points, and stage features are extracted by opening windows in each functional area. In the filling stage, the time interval from the start to the end of filling, the pressure slope at the end, the peak pressure, and the arrival delay of the flow front can be calculated; these quantities are sensitive to local flow resistance and the effective flow area of ​​the cross section, and deposition thickening is often manifested as a longer time and an increased slope. In the holding stage, the steady-state mean of the platform pressure, the variance of the steady-state section, the platform establishment time, and the platform decay rate can be calculated; deposition or micro-leakage will change the feed path, resulting in establishment delay, mean increase, or increased steady-state fluctuation. In the cooling stage, the rate of temperature or pressure decrease over time, the fitted equivalent time constant, and the moment of entering the linear heat dissipation section can be calculated; changes in surface roughness and increases in interfacial thermal resistance will slow down the rate of decrease and increase the time constant. To suppress measurement noise, median filtering or bandpass denoising can be used within each stage window before outputting the quantized parameter vector for that functional area in that module. Taking the optical appearance area as an example, using a combination of the final pressure slope and plateau variance can reflect the changes in flow resistance and increased pressure fluctuations caused by minute deposits earlier, thus forming an input quantity directly related to the surface state. An equivalent implementation can replace pressure and temperature parameters with dimensionless energy indicators, such as the square integral of pressure per unit time or the average absolute gradient of temperature change, which are equally sensitive to morphological changes within the stage and are more robust to differences in sampling rates.

[0038] When eliminating non-depositional interferences in quantization parameters, the goal is to exclude transient anomalies caused by melt temperature fluctuations, clamping force jitter, or cooling medium disturbances, preventing them from being misjudged as surface state evolution. The implementation involves cross-validation of two parallel paths. Path one is source gating: reading the status tags and key process quantities of the Programmable Logic Controller (PLC) and Manufacturing Execution System (MES), any instances of material temperature exceeding limits, instantaneous clamping force drops, or sudden changes in cooling circuit flow or inlet / outlet temperature differences within a given module are flagged, and the corresponding quantization parameters are temporarily stored. Path two is robust statistical screening: within a sliding window centered on the module, threshold rules based on the median and absolute median difference (MAD) or interquartile range (IQR) are applied to each quantization parameter. Points falling within the median ± k·MAD or exceeding [Q1-k·IQR, Q3+k·IQR] and occurring only in a single module are identified as transient anomalies and eliminated. To reduce the probability of false rejection, points identified as anomalies can be reviewed over a short window. If adjacent preceding and following modules do not show the same direction of shift and the equipment tag shows corresponding fluctuations, the rejection is maintained; otherwise, it is retained. The parameter k is recommended to be set between 2 and 3.5 based on the fluctuation tolerance of historical stable periods to achieve a balance between sensitivity and robustness. This process ensures that the data entering subsequent evaluations primarily reflects cumulative changes rather than process noise. An equivalent implementation can use Exponentially Weighted Moving Average (EWMA) to detect short-term anomalies and cross-confirm with PLC / MES tags, achieving the same screening effect as MAD / IQR.

[0039] The quantified parameters, after interference removal, are normalized to generate state indices, aiming to eliminate batch-to-batch material differences and equipment drift, making them comparable across different functional areas and batches. In implementation, a two-stage transformation is used, referencing the aforementioned benchmark. The first stage is dimensionless transformation within a stage, subtracting the corresponding benchmark statistic for each quantified parameter from the relevant functional area's statistic and scaling it using a scale, commonly in the form of relative difference or z-score. If the parameter's natural lower bound is zero and its distribution is skewed, a minimum-maximum interval mapping or logarithmic transformation is used to compress the long tail. The second stage is multi-parameter fusion to generate a single state index. Fusion can use inverse variance weighting, giving higher weights to parameters with smaller fluctuations and higher stability; alternatively, it can be based on sensitivity calibration, setting the final pressure slope, plateau variance, and cooling time constant—the parameters most sensitive to deposition—as the primary weights, with the rest as secondary weights. To improve temporal stability, exponentially weighted smoothing (EWMA) or cumulative sum statistics (CUSUM) can be applied to the state index sequence to suppress high-frequency noise and highlight slow drift. The output is a one-dimensional status index for each functional area and each module; the larger the value, the more significant the deviation from the reference benchmark. In engineering applications, tiered thresholds can be set to classify the status index into three levels: operable, maintenance recommended, and requiring shutdown. The thresholds are determined jointly based on the percentile of historical stable periods and quality thresholds. An equivalent implementation can replace the normalized benchmark with the percentile envelope within the functional area instead of the overall statistic, and fuse them using piecewise linear weights to obtain comparability and sensitivity equivalent to the above scheme.

[0040] Please continue reading. Figure 1 The state indicators at each location are organized according to the mold cycle to obtain spatial distribution and trend evaluation data of the state of the inner surface of the mold as a function of the mold cycle.

[0041] In one embodiment of the present invention, the step of organizing the state indicators at each location according to the mold cycle to obtain spatial distribution and trend evaluation data of the inner surface state of the mold with each mold cycle includes: At all locations of the location association results, the corresponding state indices are arranged in modular order and mapped to a spatial coordinate system corresponding to the mold structure. Based on the state index sequence arranged by module, a time trend curve of state change at each position is constructed, and non-deposition anomalies caused by melt temperature fluctuation, clamping force jitter or cooling medium disturbance are removed during the trend construction process. For locations not directly monitored, their state values ​​are estimated using spatial interpolation algorithms based on the state indicators of adjacent locations to complete the spatial distribution of the state on the inner surface of the mold.

[0042] The following is a detailed description of the steps involved in the above embodiments: At all locations defined by the location association results, the status indicators corresponding to each location need to be organized in module order and projected onto the mold space coordinate system. The mold space coordinate system refers to the three-dimensional coordinate system determined by the Computer-Aided Design (CAD) file, with the mold base reference plane and reference hole as the origin and axial source. During implementation, the location identifier and functional area affiliation are first obtained from the location association results. Then, the status indicators of that location in each module are extracted from the time series and arranged in ascending order by module timestamp to form a location-module sequence object. Subsequently, the cavity or flow channel geometry in the CAD is called to obtain the three-dimensional coordinates or surface parameter coordinates of the location, and a mapping table from location identifier to coordinates is established. For segments involving batch switching or shutdown restarts, they are stored in segments according to production batches to avoid cross-segment misconnections. The output is a spatiotemporal dataset with coordinate information, which can be directly used for spatial rendering and trend analysis. This processing ensures a one-to-one correspondence between numerical results and specific geometric locations, facilitating the determination of the spatial expansion path of deposition or wear within the mold. An equivalent implementation can use two-dimensional unfolded coordinates (parameterized unfolding of the cavity surface) to store the position, and then map it back to three dimensions during rendering, keeping the data structure consistent with the analysis process.

[0043] After obtaining the state index sequence sorted by module, time trend curves are constructed for each location, and non-depositional anomalies are eliminated. Anomaly detection employs two types of cross-referenced information: one type comes from event labels from the Equipment and Manufacturing Execution System (MES) and Programmable Logic Controller (PLC), such as material temperature exceeding limits, instantaneous drop in clamping force, sudden changes in cooling flow rate or inlet / outlet temperature difference; the other type is a robust statistical threshold, setting a threshold within a sliding window using the median and absolute median difference. Points that exceed limits and occur only in a single module are judged as instantaneous anomalies and eliminated. Trend construction uses Exponentially Weighted Moving Average (EWMA) to smooth high-frequency noise, and Cumulative Sum (CUSUM) is used to detect the starting point and intensity of slow drift; if necessary, the sequence is segmented and fitted across batches to avoid the step effect introduced by batch changes. The output shows the smoothing trend and change point information for each location, providing a clear view of the deviation rate and persistence, which facilitates the determination of the maintenance window. An equivalent implementation can use piecewise linear regression to replace CUSUM for change point detection, and with the same anomaly removal and smoothing strategies, consistent trend results can be obtained.

[0044] For locations not directly monitored, a spatial interpolation algorithm is used to estimate state values ​​based on state indices of adjacent locations to complete the spatial distribution of the mold's internal surface conditions. The interpolation domain is limited to the same functional area, with the cavity surface as a geometric constraint. The distance metric uses geodesic distance along the surface rather than three-dimensional straight-line distance to avoid crossing non-physical nearest neighbors caused by steel thickness. During implementation, the functional area surface is triangulated, and the geodesic distance between each known location is calculated. Inverse Distance Weighting (IDW) is used, with the power exponent set to 2 to 3 to balance smoothness and locality. Boundary constraints are added near the gate or at the boundary to ensure that the interpolation does not cross the gate boundary or abrupt changes in rib positions. For sparse areas or slender ribs with occlusion, natural neighborhood interpolation (interpolation within the Voronoi neighborhood of the surface triangulation) can be switched to reduce the bias caused by extrapolation. The output is a complete spatial distribution map for each mold pass, which can be overlaid with three-dimensional geometry for display. The parameter selection is based on a balance between engineering robustness and implementation complexity: geodesic distance ensures physical adjacency, and an IDW power exponent of 2 to 3 can balance detail and noise reduction. An equivalent implementation can use thin-plate splines for interpolation within the surface parameter domain, while still maintaining constraints and boundary treatment within the functional area, achieving a similar spatial completion effect.

[0045] The testing method for the TPU sheath production mold in the embodiments of the present invention has been described above. The testing apparatus for the TPU sheath production mold in the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 2 One embodiment of the testing device for TPU sheath production molds of the present invention includes: The acquisition module 101 is used to acquire production process data corresponding to each mold without interrupting production or disassembling the mold, and organize it into a time series associated with the cavity or flow channel identifier in chronological order. Reference module 102 is used to select a reference reference from the time series and generate stage deviation data organized by process stage; The association module 103 is used to associate the stage deviation data with the mold structure position according to the correspondence rules between the process stage and the mold structure position and obtain the position association result; The feature extraction module 104 is used to extract quantitative parameters related to the surface state based on the corresponding time series within the range defined by the location association result, and generate state indicators for each location accordingly. The evaluation module 105 is used to organize the state indicators at each location according to the mold sequence to obtain evaluation data on the spatial distribution of the state of the inner surface of the mold and the trend of change with the mold sequence.

[0046] above Figure 2The testing device for the TPU sheath production mold in this embodiment of the invention will be described in detail from the perspective of modular functional entities. The testing equipment for the TPU sheath production mold in this embodiment of the invention will be described in detail from the perspective of hardware processing.

[0047] Figure 3 This is a schematic diagram of a testing device for a TPU sheath production mold according to an embodiment of the present invention. The testing device 200 for the TPU sheath production mold can vary significantly due to different configurations or performance characteristics. It may include one or more processors 210 (e.g., one or more processors) and a memory 220, and one or more storage media 230 (e.g., one or more mass storage devices) for storing application programs 233 or data 232. The memory 220 and storage media 230 can be temporary or persistent storage. The program stored in the storage media 230 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the testing device 200 for the TPU sheath production mold. Furthermore, the processor 210 may be configured to communicate with the storage media 230 and execute the series of instruction operations in the storage media 230 on the testing device 200 for the TPU sheath production mold to implement the steps of the aforementioned testing method for the TPU sheath production mold.

[0048] The testing equipment 200 for TPU sheath production molds may also include one or more power supplies 240, one or more wired or wireless network interfaces 250, one or more input / output interfaces 260, and / or one or more operating systems 231, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 3 The test equipment structure of the TPU sheath production mold shown does not constitute a limitation on the test equipment of the TPU sheath production mold provided by the present invention. It may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0049] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the test method for the TPU sheath production mold.

[0050] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0051] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0052] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. All equivalent structural transformations made by using the contents of the present invention description and drawings under the inventive concept of the present invention, or direct / indirect application in other related technical fields are included in the patent protection scope of the present invention.

Claims

1. A testing method for a TPU sheath production mold, characterized in that, include: Without interrupting production or disassembling the mold, collect production process data corresponding to each mold batch and organize them into a time series associated with cavity or flow channel identifiers in chronological order. A reference baseline is selected from the time series to generate stage deviation data organized by process stage; According to the correspondence rules between process stages and mold structure positions, the stage deviation data is associated with the mold structure positions to obtain position association results; Within the range defined by the location association results, quantitative parameters related to the surface state are extracted based on the corresponding time series, and state indicators for each location are generated accordingly. The state indicators at each location are organized according to the mold cycle to obtain spatial distribution and trend evaluation data of the state of the inner surface of the mold as a function of the mold cycle.

2. The testing method for the TPU sheath production mold according to claim 1, characterized in that, The collected production process data corresponding to each module is organized chronologically into a time series associated with cavity or flow channel identifiers, including: Time alignment of data from different acquisition channels is performed based on identifiable feature events in the process. The collected data is bound to the corresponding cavity or runner position using the mold structure design information. When direct binding is not possible, the position is estimated by combining the runner distribution and process records. The position is then solidified after the consistency of data from consecutive mold cycles is verified. Data that does not conform to stable mass production conditions or is discontinuous in terms of modules are removed to obtain the time series.

3. The testing method for the TPU sheath production mold according to claim 1, characterized in that, The step of selecting a reference baseline from the time series and generating stage deviation data organized by process stage includes: Select continuous data segments that remain stable within a set tolerance range from the time series as reference benchmarks, and establish them separately according to the mold functional areas; Based on the unique physical responses of the filling, holding, and cooling stages in different functional areas, the data of subsequent modules are segmented and compared with the corresponding stages of the reference benchmark to calculate the relative deviation value. In the comparison results, single-mode instantaneous outliers caused by non-deposition factors are removed, and continuous deviation trends consistent with deposition or wear patterns are retained as stage deviation data.

4. The testing method for the TPU sheath production mold according to claim 3, characterized in that, The single-mold instantaneous abnormal values ​​include at least the filling stage deviation caused by melt temperature fluctuations, the holding stage deviation caused by short-term clamping force jitter, and the cooling stage deviation caused by cooling medium disturbance.

5. The testing method for the TPU sheath production mold according to claim 1, characterized in that, The step of associating the stage deviation data with the mold structure position according to the correspondence rule between the process stage and the mold structure position and obtaining the position association result includes: Establish the correspondence between each process stage and the response patterns of different functional areas within the cavity under the influence of deposition; Based on the flow channel distribution, branch sequence, and time difference of functional area response, the stage deviation data is matched to the mold structure position. When the matching conditions are not fully met, the position is estimated by using the correlation between flow channel length and response delay, and deviation sources that do not conform to the deposition mechanism are eliminated. Positions that can be physically explained as being caused by deposition are retained as position correlation results.

6. The testing method for the TPU sheath production mold according to claim 5, characterized in that, Based on the flow channel distribution, branch sequence, and time difference of functional area response, the stage deviation data is matched to the mold structure position. When the matching conditions are not fully met, the position is estimated using the correlation between flow channel length and response delay, and deviation sources inconsistent with the deposition mechanism are eliminated. Positions that can be physically explained as being caused by deposition are retained as the position correlation results, including: Based on the mold's three-dimensional structural data and process history, a time difference pattern for each functional area during the filling, holding, and cooling stages is established, and the stage deviation data is compared with the time difference pattern one by one to determine the candidate positions. For candidate locations that fail to meet all matching conditions, the expected response delay value is calculated using the effective length of the flow channel, the melt viscosity range, and the injection speed, and then fitted with the measured delay to estimate the most likely location. The matched and estimated locations are compared with the sedimentation mechanism model, and locations that do not conform to the sedimentation formation law due to flow velocity or temperature conditions are eliminated, retaining only the location results that conform to the physical interpretation.

7. The testing method for the TPU sheath production mold according to claim 1, characterized in that, Within the scope defined by the location association results, quantitative parameters related to the surface state are extracted based on the corresponding time series, and state indicators for each location are generated accordingly, including: Within the functional area defined by the location association results, quantitative parameters that can characterize the physical effects of surface deposition or wear in the functional area are extracted from the time series. The quantitative parameters include at least one of the pressure change curve characteristics during the filling stage, the pressure stability characteristics during the holding stage, and the temperature change rate characteristics during the cooling stage. Non-depositional factors are eliminated from the quantification parameters, including excluding instantaneous anomalies caused by melt temperature fluctuations, clamping force jitter, or cooling medium disturbances. The quantization parameters after interference removal are normalized, and state indicators are generated to characterize the degree of surface state change in each functional area.

8. The testing method for the TPU sheath production mold according to claim 1, characterized in that, The process of organizing the state indicators at each location according to the mold cycle to obtain spatial distribution and trend evaluation data of the mold inner surface state with each mold cycle includes: At all locations of the location association results, the corresponding state indices are arranged in modular order and mapped to a spatial coordinate system corresponding to the mold structure. Based on the state index sequence arranged by module, a time trend curve of state change at each position is constructed, and non-deposition anomalies caused by melt temperature fluctuation, clamping force jitter or cooling medium disturbance are removed during the trend construction process. For locations not directly monitored, their state values ​​are estimated using spatial interpolation algorithms based on the state indicators of adjacent locations to complete the spatial distribution of the state on the inner surface of the mold.

9. A testing device for a TPU sheath production mold, characterized in that, include: The data acquisition module is used to collect production process data corresponding to each mold batch without interrupting production or disassembling the mold, and organize it into a time series associated with the cavity or flow channel identifier in chronological order. The reference module is used to select a reference reference from the time series and generate stage deviation data organized by process stage; The association module is used to associate the stage deviation data with the mold structure position according to the correspondence rules between the process stage and the mold structure position and obtain the position association result; The feature extraction module is used to extract quantitative parameters related to the surface state based on the corresponding time series within the range defined by the location association results, and generate state indicators for each location accordingly. The evaluation module is used to organize the state indicators at each location according to the mold cycle to obtain evaluation data on the spatial distribution of the state of the inner surface of the mold and the trend of change with the mold cycle.

10. A testing device for a TPU sheath production mold, characterized in that, The testing equipment for the TPU sheath production mold includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor invokes the instructions in the memory to cause the testing equipment for the TPU sheath production mold to perform the steps of the testing method for the TPU sheath production mold as described in any one of claims 1 to 8.

Images