An online quality determination method based on indentation mechanics in an injection molding process

Abstract

The application discloses a kind of quality online determination methods based on indentation mechanics in injection molding process, comprising the following steps: (1) system configuration: build quality online determination system;(2) benchmark establishment: pre-construct qualified sample mechanics feature library and store in qualified sample feature library unit, as quality determination benchmark;(3) system deployment: complete quality online determination system and the linkage debugging verification of injection molding machine, ensure that detection action and injection cycle are synchronous;(4) injection molding and synchronous indentation test and original data acquisition;(5) data preprocessing and mechanics feature extraction;(6) quality determination;(7) closed-loop feedback and feature library dynamic optimization.The application can realize indentation detection and injection cycle synchronization, signal acquisition anti-interference, internal quality full detection determination, and form "detection-determination-feedback-optimization" production line closed loop, adapt to the rhythm of injection molding mass continuous production.

Description

Technical Field

[0001] This invention relates to the field of injection molding process and production line online inspection technology, and in particular to an online quality judgment method based on indentation mechanics during injection molding. Background Technology

[0002] Currently, online quality inspection methods for injection molding products can be mainly categorized as follows:

[0003] One is online inspection of appearance and dimensions based on machine vision, laser contouring, or 3D scanning, which identifies appearance defects such as missing material, shortage, warping, burrs, and flashing material through image / contour comparison.

[0004] Secondly, the process parameters are collected in real time by relying on in-mold or machine sensors (such as cavity pressure, injection speed, and melt temperature), and the product quality is indirectly inferred through empirical formulas or statistical models.

[0005] Third, offline mechanical property testing (tensile, bending, hardness, micro-indentation, etc.) is conducted on samples after demolding or leaving the factory to obtain material mechanical properties and durability.

[0006] Fourth, instrumented indentation is used for materials research or quality assessment, but it is mostly used in laboratory / offline conditions to record load-displacement curves for calculating material parameters such as hardness and elastic modulus.

[0007] While the above methods are effective in their respective fields, they still have significant shortcomings in online quality control for large-scale continuous injection molding production:

[0008] 1) Appearance / dimension inspection can only identify surface and geometric anomalies, and cannot directly reflect the inherent mechanical defects of the material such as porosity, density or internal stress;

[0009] 2) Indirect inference based on process parameters relies on a stable parameter-quality mapping relationship, which is affected by factors such as raw material batches, mold wear, and ambient temperature, resulting in a high probability of misjudgment or omission.

[0010] 3) Although offline mechanical testing can directly reflect material properties, it has the characteristics of lag, strong sampling, and inability to inspect all materials, and is prone to missing hidden defects in mass production.

[0011] 4) Existing indentation tests are mostly conducted offline, lacking an in-mold inspection scheme synchronized with the injection molding cycle; directly implementing the indentation method in the mold or at the moment of demolding faces engineering challenges such as high temperature of the mold, vibration, short time window, non-destructive loading and mechanical integration, making it difficult to achieve full inspection and online judgment of the indentation method on the production line.

[0012] In addition, industrial field signal acquisition is significantly affected by mold working conditions: high temperature causes sensor thermal drift, mold vibration generates interference noise, and positioning errors between the pressure head and the workpiece will reduce the reliability of the judgment. At the same time, the production line cycle has strict requirements on the detection delay, and the detection device must complete the test and return the signal without increasing or only slightly increasing the cycle. Summary of the Invention

[0013] The technical problem to be solved by the present invention is to provide an online quality judgment method based on indentation mechanics in the injection molding process. This online quality judgment method based on indentation mechanics in the injection molding process can realize the synchronization of indentation detection and injection cycle, anti-interference of signal acquisition, and full inspection and judgment of internal quality, and form a closed loop of "detection-judgment-feedback-optimization" production line, which is suitable for the rhythm of large-scale continuous injection molding production.

[0014] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0015] A method for online quality assessment based on indentation mechanics during injection molding, characterized by the following steps:

[0016] (1) System configuration: Build an online quality judgment system that includes a qualified sample feature library unit, an indentation detection execution unit, a force-displacement signal acquisition unit, a data processing unit, a quality judgment unit, and a production line closed-loop feedback unit; each unit is an independent hardware module and data communication is achieved through the injection molding machine PLC control system;

[0017] (2) Benchmark establishment: A qualified sample mechanical feature library is pre-constructed and stored in the qualified sample feature library unit as a quality judgment benchmark;

[0018] (3) System deployment: Complete the linkage debugging and verification of the online quality judgment system and the injection molding machine to ensure that the detection action is synchronized with the injection cycle;

[0019] (4) Injection molding and synchronous indentation test and raw data acquisition: The injection molding machine completes the injection molding of the injection part. At the same time, the indentation detection execution unit receives the synchronous signal from the injection molding machine PLC control system to perform the indentation test. The force-displacement signal acquisition unit collects the force value and displacement data in real time during the indentation process, records the acquisition timestamp synchronously, generates the raw dataset containing force value, displacement, and timestamp and the force-displacement real-time acquisition time sequence curve, and transmits the raw dataset to the data preprocessing module in real time.

[0020] (5) Data preprocessing and mechanical feature extraction: The data processing unit performs noise removal and temperature drift correction on the original dataset to obtain a standardized force-displacement curve. Then, the core mechanical features are extracted from the standardized force-displacement curve and combined to form the mechanical feature vector of the test piece, which is transmitted to the quality judgment unit in real time.

[0021] (6) Quality judgment: The quality judgment unit retrieves the benchmark data from the mechanical feature library of qualified samples in step (2), compares the mechanical feature vector of the test piece in step (5) with the qualified threshold range, and outputs the quality judgment result.

[0022] (7) Closed-loop feedback and dynamic optimization of feature library: The production line closed-loop feedback unit transmits the quality judgment result to the injection molding machine PLC control system, and links the production line to perform diversion, alarm or process adjustment actions. At the same time, it transmits the mechanical feature vector of qualified parts back to the qualified sample feature library unit, dynamically optimizes the qualified sample mechanical feature library, and forms a complete adaptive closed loop of detection, judgment, production line action and benchmark optimization.

[0023] The aforementioned qualified sample feature library unit is used to store, retrieve, and update the qualified sample mechanical feature library.

[0024] The aforementioned indentation detection execution unit includes a servo drive module, an indenter, an in-mold positioning stage, a miniature pressure sensor, and a micro-displacement sensor, used to perform in-mold / demolding instantaneous indentation tests.

[0025] The force-displacement signal acquisition unit is connected to the indentation detection execution unit, with an acquisition frequency of ≥1kHz, and is used to acquire raw indentation force-displacement data in real time.

[0026] The aforementioned data processing unit is used for the purification, standardization, and mechanical feature extraction of the raw data.

[0027] The aforementioned quality assessment unit is used to quickly compare the features of the test piece with those of the qualified piece.

[0028] The aforementioned production line closed-loop feedback unit is used to convert the quality judgment results into production line actions and feed them back to the qualified sample feature library unit.

[0029] Step (2) above can establish a unified quality judgment benchmark, solve the problem of no accurate reference and fuzzy feature threshold in the existing internal quality judgment of injection molded parts, provide a clear comparison standard for subsequent online judgment, and cover the normal fluctuation range of the process, thereby improving the rationality of the judgment.

[0030] The above step (3) can ensure that the entire online quality judgment system is stably connected to the injection molding production line, realize the coordinated linkage between the six functional units and the injection molding machine, eliminate equipment errors and timing deviations, and lay the foundation for subsequent online real-time detection.

[0031] In step (4) above, the force-displacement signal acquisition unit acquires force and displacement data in real time at a frequency of ≥1kHz during the indentation process, and records the acquisition timestamp synchronously to ensure the temporal integrity of the data. This enables simultaneous "injection molding-indentation detection" to obtain unbiased original detection data, avoiding detection errors caused by warping or deformation of the injection molded part after demolding. At the same time, the indentation test status is monitored through real-time charts (force-displacement time-series curves) to promptly detect acquisition anomalies. The aforementioned real-time force-displacement acquisition time-series curve is displayed on the production line monitoring screen for real-time monitoring of the acquisition status.

[0032] The above step (5) can purify the raw data, extract effective mechanical features, and provide accurate and reliable input data for subsequent quality judgment. At the same time, the core mechanical features are marked by visual charts (standardized force-displacement curves), which makes it convenient for operators to understand the mechanical properties of the test piece.

[0033] The above step (6) can quickly and accurately complete the quality judgment, identify hidden defective products that have no surface abnormalities but whose internal mechanical properties do not meet the standards, and adapt to the high-speed injection molding production rhythm.

[0034] The above step (7) can transform the quality judgment result into actual production line action, reduce the generation of batch defective products, and at the same time, through the dynamic update of qualified part data, allow the qualified sample feature library to adapt to the slow changes in the process (such as slight wear of molds and fine adjustment of raw material batches), continuously improve the accuracy of subsequent judgment, and realize the adaptive quality control of the production line.

[0035] In the preferred embodiment, the specific steps for constructing the qualified sample mechanical feature library in step (2) are as follows:

[0036] (2-1) Input the batch of qualified samples, preset indentation parameters, and parameters for the homogeneous detection area;

[0037] (2-2) Batch indentation test: Each qualified sample is placed in the testing station in the simulated mold in sequence, and a micro indentation test is performed according to the preset indentation parameters. The original force-displacement curves of each qualified sample in the three stages of "loading-retention-unloading" are collected synchronously through the indentation detection execution unit.

[0038] (2-3) Data purification and standardization: The original force-displacement curves of "loading-detention-unloading" for each qualified sample in step (2-2) are subjected to noise removal and temperature drift correction to obtain smooth and standardized force-displacement curves.

[0039] The specific operation for noise removal is as follows: After the data processing unit receives the original dataset of the indentation test in step (4), it removes the noise caused by mold vibration and electromagnetic interference through wavelet filtering algorithm.

[0040] The specific steps for temperature drift correction are as follows: The sensor temperature drift error caused by the mold operating at high temperatures of 80℃-150℃ is eliminated using a linear temperature correction algorithm. The correction formula for the linear temperature correction algorithm is as follows: ,in To standardize the indentation depth, The original indentation depth was collected by the sensor. The correction factor is used for the trial molding calibration. The actual ambient temperature during mold testing / inspection. The reference mold temperature;

[0041] (2-4) Feature extraction and statistics: Four types of core mechanical features were extracted from the standardized force-displacement curve: peak force Residual depth The rebound slope (k) and absorbed energy (W) are used to calculate the mean values ​​of various core mechanical characteristics using statistical algorithms. Based on these mean values, a fluctuation range of ±10% to ±15% of the mean is set as the acceptable threshold range for each type of core mechanical characteristic, forming a qualified sample mechanical characteristic library. Among these, peak power... The maximum value during the loading stage reflects the local stiffness and density of the material; residual depth The remaining displacement after unloading reflects the proportion of plastic deformation in the material; the springback slope k is the slope of the linear segment of the unloading curve, reflecting the material's elastic recovery ability; the absorbed energy W is the area enclosed by the curve, reflecting the material's resistance to deformation.

[0042] (2-5) Output the mechanical feature library of qualified samples and enter the mechanical feature library of qualified samples into the database of the injection molding machine PLC control system as a quality judgment benchmark.

[0043] The qualified injection molded part samples in step (2-1) above are tested offline using a universal testing machine (tensile / bending test) and a Rockwell hardness tester. At the same time, visual inspection equipment is used to confirm that there are no surface defects. Samples with stable performance and no internal micropores are selected to cover the normal qualified fluctuation range under the same injection molding process.

[0044] The preset indentation parameters in step (2-1) above are determined through trial molding to ensure that the indentation test does not damage the appearance and performance of the injection molded part, and can accurately reflect the inherent properties of the material. Preset indentation parameters include loading speed, maximum loading force, and residence time.

[0045] In the above step (2-1), the homogeneous region is generally selected as the inner plane of the shell with uniform wall thickness, no weld lines, and which is neither visible nor under stress.

[0046] In a further optimized scheme, in steps (2-4), the qualified threshold range for each type of core mechanical feature is uniformly set to the mean ± 10%, specifically:

[0047] Phonak The acceptable threshold range: ,in The mean peak intensity of the qualified samples;

[0048] Residual depth The acceptable threshold range: ,in The mean residual depth of qualified samples;

[0049] The acceptable threshold range for the rebound slope k: ,in The mean rebound slope of the qualified samples;

[0050] The acceptable threshold range for absorbed energy W: ,in This represents the average absorbed energy of the qualified samples.

[0051] In a further optimized scheme, in step (4), when the injection molding machine is molding the injection part, the injection molding machine PLC control system sends out injection raw material, injection molding parameters, and injection cycle synchronization signals in real time; the injection molding machine performs melt filling, holding pressure, and cooling processes according to the preset indentation parameters in step (2-2) to complete the molding of the injection part. This injection molding method can ensure that the injection part has no obvious surface defects such as missing material or flash. The above-mentioned injection cycle synchronization signal is automatically sent by the injection molding machine PLC control system after the injection part has cooled down, marking the detection trigger node. The above-mentioned injection molding parameters include melt temperature, filling speed, and holding pressure.

[0052] In a further preferred embodiment, the specific steps of the indentation detection execution unit performing the indentation test in step (4) are as follows: When the injection molding machine PLC control system sends a synchronization signal, the indentation detection execution unit is immediately triggered. According to the preset indentation parameters and the detection homogeneous area parameters in step (2-1), the indenter in the mold cavity is driven to contact the preset homogeneous area of ​​the injection molded part and perform the three-stage action of "loading-dwelling-unloading". During the injection molding cycle (in-mold / demolding moment), micro-indentation is applied and the three-stage standardized test of "loading-dwelling-unloading" is completed, so that the indentation detection is strictly synchronized with the injection cycle, thereby eliminating the influence of deformation after demolding on the judgment result, and ensuring that the detection is completed without increasing or only slightly increasing the production cycle.

[0053] During the injection molding cycle, since the product is still constrained by the mold or is in a stable time window after demolding, a controlled micro-indentation load is applied to the surface of the product, and the force-displacement response of the entire loading-dwelling-unloading process is collected. This response is used as a statistical judgment signal to characterize the internal molding quality of the injection molded part.

[0054] During injection molding, the internal quality state of the product (such as material density, uniformity of cooling shrinkage, and distribution of internal stress) directly affects its deformation and springback behavior under short-term loading conditions. By conducting micro-indentation tests in the mold or at the moment of demolding, the transient mechanical response of the product can be captured before the deformation is fully released. This makes the obtained force-displacement curve more sensitive to molding defects and avoids interference from factors such as long-term relaxation after demolding and changes in ambient temperature on the judgment results.

[0055] Specifically, during the controlled indentation stage, the relationship between indentation load and indentation depth reflects the local stiffness and material density of the product; during the retention stage, load decay and displacement stability indicate the short-term creep and internal stress release characteristics of the material; during the unloading stage, the springback slope and residual indentation depth reflect the elastic recovery capacity and plastic deformation ratio of the product. These stages collectively constitute a time-consistent force-displacement response characteristic, and its overall morphology and statistical properties can be used to distinguish between normally molded and abnormally molded products.

[0056] In a further optimized scheme, the specific steps of data preprocessing and mechanical feature extraction in step (5) are as follows:

[0057] (5-1) Data preprocessing: The wavelet filtering algorithm and linear temperature correction algorithm, which are completely consistent with those in step (2-3), are used to perform noise removal and temperature drift correction on the original force-displacement curve of the test piece collected in step (4) to eliminate signal errors caused by high temperature, vibration and electromagnetic interference of the mold, and thus obtain a smooth and standardized force-displacement curve that can be compared with qualified samples.

[0058] (5-2) Extraction of core mechanical features: Extract four types of core mechanical features consistent with step (2-4) from the standardized force-displacement curve: peak force Residual depth The springback slope k and absorbed energy W are combined to form the mechanical characteristic vector of the test piece, and transmitted to the quality judgment unit in real time.

[0059] In a further preferred embodiment, the specific steps for comparing the mechanical feature vector of the test piece with the acceptable threshold range in step (6) are as follows:

[0060] (6-1) Reference data retrieval: After receiving the mechanical feature vector of the test piece, the quality judgment unit automatically retrieves the corresponding reference data from the qualified sample mechanical feature library in step (2) to determine the qualified threshold range of each of the four core mechanical features.

[0061] (6-2) Feature-by-feature numerical comparison:

[0062] If the peak strength to be measured If so, the peak power index is deemed qualified;

[0063] If the peak strength to be measured or If so, the peak power index is deemed unqualified;

[0064] If the residual depth to be measured If so, the residual depth index is deemed qualified;

[0065] If the residual depth to be measured or If so, the residual depth index is deemed unqualified;

[0066] If the rebound slope to be measured If so, the rebound slope index is deemed qualified;

[0067] If the rebound slope to be measured or If so, the rebound slope index is deemed unqualified;

[0068] If the absorbed energy to be measured If so, the energy absorption index is deemed qualified;

[0069] If the absorbed energy to be measured or If so, the energy absorption index is deemed unqualified;

[0070] (6-3) Overall quality assessment:

[0071] Based on the comparison results of various core mechanical features in step (6-2) above, the following judgment rules are applied:

[0072] If all four core mechanical characteristics meet their respective qualification threshold range requirements, the injection molded part is determined to be a qualified part.

[0073] If any one or more core mechanical characteristics exceed the corresponding qualified threshold range, the injection molded part is determined to be a part that needs to be re-inspected.

[0074] (6-4) Output the quality judgment result and transmit it to the production line closed-loop feedback unit; at the same time, generate a radar chart comparing the mechanical characteristics of the test part and the qualified sample, and display it on the system human-machine interface to facilitate operators to quickly judge the cause of the defect.

[0075] In a further optimized scheme, the specific steps of production line closed-loop feedback and feature library dynamic optimization in step (7) are as follows:

[0076] (7-1) Closed-loop feedback action of the production line:

[0077] (7-1.1) If the injection molded part is a qualified part, then control the production line to perform normal diversion and enter the subsequent process;

[0078] (7-1.2) If the injection molded part is a part that needs to be re-inspected, control the production line to divert it to the re-inspection station and trigger a slight audible and visual alarm to remind the operator to pay attention;

[0079] (7-1.3) If three consecutive injection molded parts are determined to be parts that need to be re-inspected, a high-frequency audible and visual alarm will be triggered, and process adjustment suggestions will be output.

[0080] (7-1.4) Mark the nodes in the production line that are judged as qualified parts, parts that need to be re-inspected, and parts that need to be re-inspected for 3 consecutive parts, and generate a closed-loop feedback node marking diagram;

[0081] (7-2) Dynamically optimize the mechanical feature library of qualified samples:

[0082] (7-2.1) Data screening: The mechanical feature vectors of qualified parts in the production line are sent back to the qualified sample feature library unit, and the mechanical feature vectors of parts that need to be re-inspected are removed;

[0083] (7-2.2) Periodic update: For every 1,000 injection molded parts produced, the qualified sample feature library unit re-statistically analyzes all qualified part data in the library using statistical algorithms, updates the mean and qualified threshold range of the four core mechanical features, and realizes periodic optimization of the feature library;

[0084] (7-2.3) Data support: After the feature library is updated, the process adjustment timing curve of the mean change of the core mechanical features is automatically generated and stored in the database of the injection molding machine PLC control system. The process adjustment timing curve is associated with the change of injection molding process parameters, providing accurate data support for subsequent process optimization and mold maintenance.

[0085] The aforementioned quality judgment results drive diversion, alarms, or process adjustment suggestions in real time. Through a controlled dynamic update mechanism of qualified part data screening and batch periodic updates, the qualified sample feature library is optimized to achieve a continuous adaptive closed loop of detection, judgment, feedback, and benchmark updates.

[0086] In the preferred embodiment, the specific steps for the linkage debugging and verification of the online quality judgment system and the injection molding machine in step (3) are as follows:

[0087] (3-1) Debugging preparation: Obtain the hardware parameters of the qualified sample feature library unit, indentation detection execution unit, force-displacement signal acquisition unit, data processing unit, quality judgment unit, and production line closed-loop feedback unit; communication parameters of the injection molding machine PLC control system; and debugging samples;

[0088] (3-2) Module communication setup: Establish wired communication connections between the qualified sample feature library unit, the indentation detection execution unit, the force-displacement signal acquisition unit, the data processing unit, the quality judgment unit, and the production line closed-loop feedback unit and the injection molding machine PLC control system, determine the data flow between each unit, and ensure stable data transmission and no packet loss;

[0089] (3-3) Accuracy and timing debugging:

[0090] The force-displacement signal acquisition unit is calibrated to ensure that the force error is ≤ ±0.1N and the displacement error is ≤ ±0.001mm, thus eliminating the influence of temperature drift of the sensor.

[0091] Adjust the triggering timing of the indentation detection execution unit to ensure accurate triggering when the mold opens to the preset position, synchronized with the injection cycle, without adding extra production time;

[0092] (3-4) Functional verification: Send the debugging sample into the injection molding machine PLC control system to verify the accuracy of the judgment logic and the effectiveness of the closed-loop feedback action, and troubleshoot communication interruption and data delay problems;

[0093] (3-5) System acceptance: The online quality judgment system is debugged and qualified, and a system unit linkage timing diagram is generated and stored in the maintenance interface of the injection molding machine PLC control system for easy subsequent fault location and module upgrade.

[0094] The hardware parameters of each unit in step (3-1) above include the acquisition frequency and transmission rate, which are generally provided by the manufacturer or obtained through conventional hardware calibration.

[0095] The communication parameters of the injection molding machine PLC control system in step (3-1) above include baud rate, address code, and communication protocol (such as Modbus), which can be adapted to the communication standards of existing production lines.

[0096] The test samples in step (3-1) above include qualified parts and known defective parts; qualified parts refer to parts without obvious internal mechanical defects; known defective parts refer to parts prepared by artificially lowering the holding pressure and shortening the cooling time to ensure that there are obvious internal mechanical defects.

[0097] The above system module linkage logic diagram is a visual chart that clarifies the linkage relationship between each unit, facilitating subsequent maintenance and troubleshooting.

[0098] Compared with the prior art, the present invention has the following advantages:

[0099] (1) Full inspection and control to improve the finished product qualification rate: realize the internal quality inspection of each product, with no sampling omissions, and reduce the flow of hidden defective products into subsequent processes from the source.

[0100] (2) Real-time response without affecting production line efficiency: The detection action is synchronized with the injection molding cycle, and data collection and judgment are completed in a very short time, without increasing or slightly increasing the production cycle, and is suitable for large-scale continuous production.

[0101] (3) Closed-loop optimization to continuously improve production stability: Through production line feedback and feature library dynamic optimization, a closed loop of "detection-judgment-adjustment-optimization" is achieved, continuously improving the accuracy of quality control and process stability.

[0102] (4) This invention does not aim to calculate the mechanical parameters of a single material, but regards the entire process of indentation force-displacement as a set of repeatable and quantifiable quality judgment signals. By statistically comparing it with the characteristic distribution of qualified samples established in advance, the internal quality of a single injection molded product can be judged online.

[0103] (5) This invention is mainly applied to the field of high-speed injection molding production. It is suitable for quality inspection and production line closed-loop control of various injection molded parts (especially injection molded products with no obvious surface defects but whose internal mechanical properties need to be precisely controlled). It can be widely used in quality control scenarios of injection molding industries such as automotive parts, electronic housings, and home appliance accessories. Attached Figure Description

[0104] Figure 1 This is an execution flowchart of a specific embodiment of the present invention;

[0105] Figure 2 This is a point-by-point display diagram of the peak force mean feature library in a specific embodiment of the present invention;

[0106] Figure 3 This is a point-by-point display diagram of the residual depth feature library in a specific embodiment of the present invention;

[0107] Figure 4 This is a point-by-point display diagram of the rebound slope feature library in a specific embodiment of the present invention;

[0108] Figure 5 This is a point-by-point display diagram of the energy absorption feature library in a specific embodiment of the present invention;

[0109] Figure 6 This is a system unit linkage timing diagram of a specific embodiment of the present invention;

[0110] Figure 7 This is a time-series curve of real-time force-displacement acquisition according to a specific embodiment of the present invention;

[0111] Figure 8 This is a preprocessed standardized force-displacement curve diagram of a specific embodiment of the present invention;

[0112] Figure 9 This is a radar image comparing the mechanical characteristics of the test piece and a qualified sample according to a specific embodiment of the present invention.

[0113] Figure 10 This is a process adjustment timing curve diagram and a closed-loop feedback node annotation diagram of a specific embodiment of the present invention. Detailed Implementation

[0114] The following description, in conjunction with the accompanying drawings and preferred embodiments of the present invention, will provide further details.

[0115] This embodiment verifies the feasibility and effectiveness of the method of the present invention in the continuous injection molding production scenario of PP plastic household appliance shells. The production cycle is 30 seconds / piece, a horizontal injection molding machine is used, and an in-mold indentation detection execution unit (including a micro pressure sensor and a micro displacement sensor) is provided. The injection molding machine PLC control system is a Siemens S7-1200 PLC, and the data processing terminal is an industrial computer.

[0116] Sensor assumed parameters: Miniature pressure sensor range 0-200N, accuracy ±0.1N; micro-displacement sensor range 0-0.5mm, accuracy ±0.001mm, acquisition frequency 1kHz; all parameter settings are adapted to the elastic-plastic properties of PP plastic.

[0117] like Figure 1 As shown in this embodiment, the online quality determination method based on indentation mechanics during injection molding includes the following steps:

[0118] (1) System configuration: Build an online quality judgment system that includes a qualified sample feature library unit, an indentation detection execution unit, a force-displacement signal acquisition unit, a data processing unit, a quality judgment unit, and a production line closed-loop feedback unit; each unit is an independent hardware module and data communication is achieved through the injection molding machine PLC control system;

[0119] (2) Benchmark establishment: A qualified sample mechanical feature library is pre-constructed and stored in the qualified sample feature library unit as a quality judgment benchmark;

[0120] (3) System deployment: Complete the linkage debugging and verification of the online quality judgment system and the injection molding machine to ensure that the detection action is synchronized with the injection cycle;

[0121] (4) Injection molding and synchronous indentation test and raw data acquisition: The injection molding machine completes the injection molding of the injection part. At the same time, the indentation detection execution unit receives the synchronous signal from the injection molding machine PLC control system to perform the indentation test. The force-displacement signal acquisition unit collects the force value and displacement data in the indentation process in real time, generates the raw dataset containing force value, displacement, and timestamp, as well as the force-displacement real-time acquisition time sequence curve, and transmits the raw dataset to the data preprocessing module in real time.

[0122] (5) Data preprocessing and mechanical feature extraction: The data processing unit performs noise removal and temperature drift correction on the original dataset to obtain a standardized force-displacement curve. Then, the core mechanical features are extracted from the standardized force-displacement curve and combined to form the mechanical feature vector of the test piece, which is transmitted to the quality judgment unit in real time.

[0123] (6) Quality judgment: The quality judgment unit retrieves the benchmark data from the mechanical feature library of qualified samples in step (2), compares the mechanical feature vector of the test piece in step (5) with the qualified threshold range, and outputs the quality judgment result.

[0124] (7) Closed-loop feedback and dynamic optimization of feature library: The production line closed-loop feedback unit transmits the quality judgment result to the injection molding machine PLC control system, and links the production line to perform diversion, alarm or process adjustment actions. At the same time, it transmits the mechanical feature vector of qualified parts back to the qualified sample feature library unit, dynamically optimizes the qualified sample mechanical feature library, and forms a complete adaptive closed loop of detection, judgment, production line action and benchmark optimization.

[0125] In step (2), the specific steps for constructing the qualified sample mechanical feature library are as follows:

[0126] (2-1) Input 50 qualified samples of PP appliance shells, preset indentation parameters (loading speed 0.3mm / min, maximum loading force 100N, residence time 2s), and parameters of the homogeneous area to be tested (homogeneous plane with a wall thickness of 3mm on the inner side of the PP appliance shell (non-appearance / non-stress area)).

[0127] (2-2) Batch indentation test: Place each qualified sample in the test station in the simulated mold in sequence, and perform micro indentation test according to the preset indentation parameters. The force-displacement original curve of each qualified sample in the three stages of "loading-retention-unloading" is collected synchronously through micro pressure sensor / displacement sensor.

[0128] (2-3) Data purification and standardization: The original force-displacement curves of "loading-detention-unloading" for each qualified sample in step (2-2) are subjected to noise removal and temperature drift correction to obtain smooth and standardized force-displacement curves.

[0129] The specific operation for noise removal is as follows: After the data processing unit receives the original dataset of the indentation test in step (4), it removes the noise caused by mold vibration and electromagnetic interference through wavelet filtering algorithm.

[0130] The specific steps for temperature drift correction are as follows: The sensor temperature drift error caused by the mold operating at high temperatures of 80℃-150℃ is eliminated using a linear temperature correction algorithm. The correction formula for the linear temperature correction algorithm is as follows: ,in To standardize the indentation depth, The original indentation depth acquired by the sensor, with a correction factor. , The actual ambient temperature during mold testing / inspection, the reference temperature. ;

[0131] (2-4) Feature extraction and statistics: Four types of core mechanical features were extracted from the standardized force-displacement curve: peak force Residual depth The rebound slope (k) and absorbed energy (W) are used to calculate the mean values ​​of various core mechanical characteristics using statistical algorithms. Based on these mean values, a fluctuation range of ±10% to ±15% of the mean is set as the acceptable threshold range for each type of core mechanical characteristic, forming a qualified sample mechanical characteristic library. Among these, peak power... The maximum value during the loading stage reflects the local stiffness and density of the material; residual depth The remaining displacement after unloading reflects the proportion of plastic deformation in the material; the springback slope k is the slope of the linear segment of the unloading curve, reflecting the material's elastic recovery ability; the absorbed energy W is the area enclosed by the curve, reflecting the material's resistance to deformation.

[0132] The arithmetic mean of the core mechanical characteristics of the same type for 50 qualified samples was taken, and uniformly set as ±10% of the mean, specifically:

[0133] Mean peak intensity of qualified samples: acceptable threshold range ;

[0134] Mean residual depth of qualified samples: acceptable threshold range ;

[0135] Mean rebound slope of qualified samples: acceptable threshold range ;

[0136] Mean energy absorbed by qualified samples: acceptable threshold range ;

[0137] (2-5) Output a qualified sample mechanical feature library (e.g.) Figure 2-5 As shown in the figure, the distribution of the four core mechanical characteristics of 50 qualified PP appliance shell samples is displayed. The characteristic values ​​of all samples fluctuate slightly around the preset values ​​and are within the qualified threshold range. The mechanical characteristic library of qualified samples is entered into the database of the injection molding machine PLC control system as a quality judgment benchmark.

[0138] In step (3), the specific steps for the linkage debugging and verification of the online quality judgment system and the injection molding machine are as follows:

[0139] (3-1) Debugging preparation: Obtain the hardware parameters (acquisition frequency 1kHz, transmission rate ≥1Mbps) of the qualified sample feature library unit, indentation detection execution unit, force-displacement signal acquisition unit, data processing unit, quality judgment unit, and production line closed-loop feedback unit; communication parameters of the injection molding machine PLC control system (Modbus protocol, baud rate 9600, address code 01); 15 debugging samples (10 qualified parts and 5 known defective parts; defective parts were prepared by lowering the holding pressure to 40MPa to ensure the presence of micropore defects);

[0140] (3-2) Module communication setup: Establish wired communication connections between the qualified sample feature library unit, indentation detection execution unit, force-displacement signal acquisition unit, data processing unit, quality judgment unit, and production line closed-loop feedback unit and the injection molding machine PLC control system, determine the data flow between each unit, and ensure stable data transmission and no packet loss (test communication packet loss rate ≤ 0.1%).

[0141] (3-3) Accuracy and timing debugging:

[0142] The force-displacement signal acquisition unit is calibrated to ensure that the force error is ≤ ±0.1N and the displacement error is ≤ ±0.001mm, thus eliminating the influence of temperature drift of the sensor.

[0143] - Debugging the core parameters of the preprocessing algorithm: Use db4 wavelet for noise filtering. The optimal filtering threshold was determined to be 0.02 through multiple sets of comparative experiments and recorded in the accuracy calibration report;

[0144] - Perform temperature drift calibration: Test the sensor temperature drift pattern within the mold temperature range of 80-150℃ and determine the linear temperature correction coefficient. Temperature drift of sensors under high mold temperatures is eliminated through a temperature compensation algorithm;

[0145] The timing of the indentation detection execution unit is bound to the preset position of the mold opening, ensuring that the indentation detection is triggered ≤10ms after the mold opening and positioning is completed, and the time for a single detection is ≤2s, without increasing the production cycle by 30s / piece.

[0146] (3-4) Functional verification: 15 test samples were sent into the injection molding machine PLC control system to verify the effectiveness of the judgment logic (100% defective part identification rate and 0% qualified part misjudgment rate) and closed-loop feedback action (diversion response time ≤1s and alarm trigger delay ≤0.5s), and to investigate communication interruption and data delay problems.

[0147] (3-5) System Acceptance: Determine the system's quality online to ensure it passes debugging, and generate system unit linkage timing diagrams (e.g., Figure 6As shown in the figure, the data is stored in the maintenance interface of the injection molding machine PLC control system, which facilitates subsequent fault location and module upgrades.

[0148] In step (4), when the injection molding machine is molding the injection part, the injection molding machine PLC control system sends out the injection raw material (PP plastic particles), injection molding parameters (melt temperature 220℃, mold filling speed 50mm / s, holding pressure 60MPa), and injection cycle synchronization signal in real time. The injection molding machine executes the melt filling, holding pressure, and cooling processes (cooling time 15s, production cycle 30s / piece) according to the preset indentation parameters in step (2-2) to complete the molding of the injection part. This injection molding method can ensure that the injection part has no obvious surface defects such as missing material or flash. The above-mentioned injection cycle synchronization signal is automatically sent by the injection molding machine PLC control system after the injection part has cooled down, marking the detection trigger node. The above-mentioned injection molding parameters include melt temperature, mold filling speed, and holding pressure.

[0149] In step (4), the specific steps of the indentation detection execution unit performing the indentation test are as follows: When the injection molding machine PLC control system sends a synchronization signal, the indentation detection execution unit is immediately triggered. According to the preset indentation parameters (loading speed 0.3mm / min, maximum loading force 100N, residence time 2s) in step (2-1) and the parameters of the homogeneous detection area (the homogeneous detection area is a homogeneous plane with an inner wall thickness of 3mm), the indenter in the mold cavity is driven to contact the preset homogeneous area of ​​the injection molded part and perform the three-stage action of "loading-retention-unloading"; at the same time, the force-displacement signal acquisition unit collects the force value and displacement data in the indentation process in real time, records the acquisition timestamp synchronously, and generates the original dataset containing force value, displacement, and timestamp, as well as the force-displacement real-time acquisition time sequence curve (e.g. Figure 7 As shown, the raw dataset is transmitted to the data preprocessing module in real time. Micro-indentations are applied during the injection molding cycle (in-mold / at the moment of demolding) and a standardized three-stage test of "loading-retention-unloading" is completed. This ensures that indentation detection is strictly synchronized with the injection molding cycle, thereby eliminating the influence of deformation after demolding on the judgment results and guaranteeing that the detection is completed without increasing or only slightly increasing the production cycle time.

[0150] like Figure 7 The force-displacement real-time acquisition time-series curve shown reconstructs the dynamic process of the indentation in the three stages of "loading-retention-unloading".

[0151] Figure 7 The upper half of the graph shows that the indentation force increases linearly to 100N from 0 to 20s (loading), remains at 100N from 20 to 22s (holding), and decreases linearly to 0 from 22 to 37s (unloading).

[0152] Figure 7 The lower half of the figure shows that the displacement increases with loading, stabilizes during the dwell phase, and exhibits a slight rebound during the unloading phase.

[0153] In step (5), the specific steps of data preprocessing and mechanical feature extraction are as follows:

[0154] (5-1) Data preprocessing: The wavelet filtering algorithm and linear temperature correction algorithm, which are completely consistent with those in step (2-3), are used to perform noise removal and temperature drift correction on the original force-displacement curve of the test piece collected in step (4) to eliminate signal errors caused by high temperature, vibration and electromagnetic interference of the mold, and thus obtain a smooth and standardized force-displacement curve that can be compared with qualified samples.

[0155] (5-1.1) Wavelet filtering algorithm for noise removal:

[0156] After receiving the raw dataset, the data preprocessing module uses db4 wavelet to perform a three-level decomposition of the force-displacement data, removing high-frequency noise (corresponding to mold vibration and electromagnetic interference signals) with a threshold of 0.02. The signal-to-noise ratio of the processed data is improved from 25dB to 42dB, resulting in a smooth, standardized force-displacement curve, as shown below. Figure 8 As shown. Figure 8 The upper left side of the image shows a comparison between the original force value and the pre-processed force value, with the time points of the loading stage and the residence stage marked. Figure 8 The enlarged view on the upper right side of the image shows the loading stage from 15 to 20 seconds in the left image. The thin red line is the original force curve containing high-frequency noise and temperature drift, which shows obvious fluctuations. The thick dark blue line is the force curve after noise reduction and temperature drift correction, which is smooth and stable. The two form a sharp contrast, which more clearly highlights the high-frequency fluctuations of the original curve and the smoothness of the preprocessed curve, demonstrating the effect of preprocessing on signal quality improvement. Figure 8 The lower half of the image shows the preprocessed standardized force-displacement curve, which marks three core mechanical features: orange dots represent peak force, red dots represent residual depth, and green filled areas represent absorbed energy.

[0157] (5-1.2) Linear temperature correction algorithm for temperature drift correction:

[0158] The error index is based on the inherent accuracy of the sensor and the assumptions of the correction logic: it is assumed that under the reference condition of 100℃, the ideal indentation depth is 0.24mm; when the temperature rises to 120℃, the sensor will read the true value of 0.24mm as 0.3mm due to temperature drift.

[0159] Based on the real-time temperature of the mold (120℃ in this experiment), the sensor initially collected the indentation depth. Correction factor The temperature drift coefficient calibrated in step (3-3) is the actual ambient temperature during mold testing / inspection. reference temperature Through a linear temperature correction algorithm:

[0160] The measured value of 0.3 mm at high temperature was corrected to an equivalent value of 0.24 mm under the reference condition of 100℃, eliminating the sensor temperature drift error caused by high temperature. The corrected displacement error is ≤ ±0.001 mm, which is completely aligned with the detection conditions of the qualified sample mechanical feature library in step (2). The correction principle of the force value is the same as above, and the corrected force value error is ≤ ±0.1 N.

[0161] (5-2) Extraction of core mechanical features: Extract four types of core mechanical features consistent with step (2-4) from the corrected standardized force-displacement curve:

[0162] Phonak : Maximum force value during the loading stage (the value extracted in this experiment is 81.2N);

[0163] Residual depth : Displacement stability value after unloading (the value extracted in this experiment is 0.081 mm);

[0164] Springback slope k: The slope of the linear segment of the unloading curve (the value extracted in this experiment is 202.3 N / mm).

[0165] Absorbed energy W: Area enclosed by the force-displacement curve (the value extracted in this experiment is 3.24 N·mm);

[0166] The mechanical feature vector of the test piece is formed by combining the features of the qualified samples in step (2) and is transmitted to the quality judgment unit in real time.

[0167] In step (6), the specific steps for comparing the mechanical feature vector of the test piece with the acceptable threshold range are as follows:

[0168] (6-1) Reference data retrieval: After receiving the mechanical feature vector of the test piece, the quality judgment unit automatically retrieves the corresponding reference data from the qualified sample mechanical feature library in step (2) to determine the mean value and qualified threshold range of the four core mechanical features, as follows:

[0169] Phonak The mean of the qualified samples is The acceptable threshold range is [72, 88]. ;

[0170] Residual depth The mean of the qualified samples is The acceptable threshold range is [0.072, 0.088] mm;

[0171] The mean of the acceptable sample of rebound slope k is The acceptable threshold range is [180, 220]. ;

[0172] The mean of the qualified sample of absorbed energy W is The acceptable threshold range is [2.88, 3.52]. ;

[0173] (6-2) Feature-by-feature numerical comparison:

[0174] Based on the four core mechanical characteristics of the part to be tested ( , , , Taking as an example, compare each data point with the baseline data in step (6-1):

[0175] -Phone If 72 ≤ 82 ≤ 88, then the peak power index is considered qualified;

[0176] -Residual Depth If 0.071 < 0.072, then the residual depth index is deemed unqualified.

[0177] - If the springback slope k is 180≤195≤220, then the springback slope index is considered qualified;

[0178] - If the absorbed energy W: 4.2 > 3.52, then the absorbed energy index is deemed unqualified;

[0179] (6-3) Overall quality assessment:

[0180] Based on the comparison results of various core mechanical features in step (6-2) above, the following judgment rules are applied:

[0181] If all four core mechanical characteristics meet their respective qualification threshold range requirements, the injection molded part is determined to be a qualified part.

[0182] If any one or more core mechanical characteristics exceed the corresponding qualified threshold range, the injection molded part is determined to be a part that needs to be re-inspected.

[0183] (6-4) The output quality judgment result is "The part to be tested is a part that needs to be re-inspected", which is transmitted to the closed-loop feedback unit of the production line; at the same time, a radar chart comparing the mechanical characteristics of the part to be tested and the qualified sample is generated (e.g., Figure 9 As shown in the figure, the feature deviation is displayed on the system's human-machine interface, which intuitively presents the characteristic deviation, making it easy for operators to quickly determine the cause of the problem.

[0184] In step (7), the specific steps of production line closed-loop feedback and feature library dynamic optimization are as follows:

[0185] (7-1) Closed-loop feedback action of the production line:

[0186] (7-1.1) If the injection molded part is a qualified part, then control the production line to perform normal diversion and enter the subsequent process;

[0187] (7-1.2) If the injection molded part is a part that needs to be re-inspected, control the production line to divert it to the re-inspection station and trigger a slight audible and visual alarm to remind the operator to pay attention;

[0188] (7-1.3) If three consecutive injection molded parts are determined to be parts that need to be re-inspected, a high-frequency audible and visual alarm will be triggered, and process adjustment suggestions will be output.

[0189] (7-1.4) Mark the nodes in the production line that are judged as qualified parts, parts that need to be re-inspected, and parts that require re-inspection for three consecutive parts, and generate a closed-loop feedback node marking diagram (e.g.) Figure 10 (as shown)

[0190] (7-2) Dynamically optimize the mechanical feature library of qualified samples:

[0191] (7-2.1) Data screening: The mechanical feature vectors of qualified parts in the production line are sent back to the qualified sample feature library unit, and the mechanical feature vectors of parts that need to be re-inspected are removed;

[0192] (7-2.2) Periodic update: For every 1,000 injection molded parts produced, the qualified sample feature library unit re-statistically analyzes all qualified part data in the library using statistical algorithms, updates the mean and qualified threshold range of the four core mechanical features, and realizes periodic optimization of the feature library;

[0193] (7-2.3) Data Support: After the feature library is updated, the process adjustment timing curve of the mean change of the core mechanical features is automatically generated (e.g., Figure 10 As shown in the figure, the data is stored in the database of the injection molding machine PLC control system. The process adjustment timing curve is associated with the changes in injection molding process parameters, providing accurate data support for subsequent process optimization and mold maintenance.

[0194] The aforementioned quality judgment results drive diversion, alarms, or process adjustment suggestions in real time. Through a controlled dynamic update mechanism of qualified part data screening and batch periodic updates, the qualified sample feature library is optimized to achieve a continuous adaptive closed loop of detection, judgment, feedback, and benchmark updates.

[0195] Furthermore, it should be noted that the names of the various parts of the specific embodiments described in this specification may differ. All equivalent or simple variations made to the structure, features, and principles described in this invention are included within the scope of protection of this invention. Those skilled in the art can make various modifications or additions to the described specific embodiments or use similar methods to replace them, as long as they do not deviate from the structure of this invention or exceed the scope defined in these claims, all of which should fall within the scope of protection of this invention.

Claims

1. A method for online quality determination based on indentation mechanics during injection molding, characterized in that... Includes the following steps: (1) System configuration: Build an online quality judgment system that includes a qualified sample feature library unit, an indentation detection execution unit, a force-displacement signal acquisition unit, a data processing unit, a quality judgment unit, and a production line closed-loop feedback unit; each unit is an independent hardware module and data communication is achieved through the injection molding machine PLC control system; (2) Benchmark establishment: A qualified sample mechanical feature library is pre-constructed and stored in the qualified sample feature library unit as a quality judgment benchmark; (3) System deployment: Complete the linkage debugging and verification of the online quality judgment system and the injection molding machine to ensure that the detection action is synchronized with the injection cycle; (4) Injection molding and synchronous indentation test and raw data acquisition: The injection molding machine completes the injection molding of the injection part. At the same time, the indentation detection execution unit receives the synchronous signal from the injection molding machine PLC control system to perform the indentation test. The force-displacement signal acquisition unit collects the force value and displacement data in real time during the indentation process, records the acquisition timestamp synchronously, generates the raw dataset containing force value, displacement, and timestamp and the force-displacement real-time acquisition time sequence curve, and transmits the raw dataset to the data preprocessing module in real time. (5) Data preprocessing and mechanical feature extraction: The data processing unit performs noise removal and temperature drift correction on the original dataset to obtain a standardized force-displacement curve. Then, the core mechanical features are extracted from the standardized force-displacement curve and combined to form the mechanical feature vector of the test piece, which is transmitted to the quality judgment unit in real time. (6) Quality judgment: The quality judgment unit retrieves the benchmark data from the mechanical feature library of qualified samples in step (2), compares the mechanical feature vector of the test piece in step (5) with the qualified threshold range, and outputs the quality judgment result. (7) Closed-loop feedback and dynamic optimization of feature library: The production line closed-loop feedback unit transmits the quality judgment result to the injection molding machine PLC control system, and links the production line to perform diversion, alarm or process adjustment actions. At the same time, it transmits the mechanical feature vector of qualified parts back to the qualified sample feature library unit, dynamically optimizes the qualified sample mechanical feature library, and forms a complete adaptive closed loop of detection, judgment, production line action and benchmark optimization.

2. The online quality determination method based on indentation mechanics during injection molding as described in claim 1, characterized in that: In step (2), the specific steps for constructing the qualified sample mechanical feature library are as follows: (2-1) Input the batch of qualified samples, preset indentation parameters, and parameters for the homogeneous detection area; (2-2) Batch indentation test: Each qualified sample is placed in the testing station in the simulated mold in sequence, and a micro indentation test is performed according to the preset indentation parameters. The original force-displacement curves of each qualified sample in the three stages of "loading-retention-unloading" are collected synchronously through the indentation detection execution unit. (2-3) Data purification and standardization: The original force-displacement curves of "loading-detention-unloading" for each qualified sample in step (2-2) are subjected to noise removal and temperature drift correction to obtain smooth and standardized force-displacement curves. The specific operation for noise removal is as follows: After the data processing unit receives the original dataset of the indentation test in step (4), it removes the noise caused by mold vibration and electromagnetic interference through wavelet filtering algorithm. The specific steps for temperature drift correction are as follows: The sensor temperature drift error caused by the mold operating at high temperatures of 80℃-150℃ is eliminated using a linear temperature correction algorithm. The correction formula for the linear temperature correction algorithm is as follows: ,in To standardize the indentation depth, The original indentation depth was collected by the sensor. The correction factor is used for the trial molding calibration. The actual ambient temperature during mold testing / inspection. The reference mold temperature; (2-4) Feature extraction and statistics: Four types of core mechanical features were extracted from the standardized force-displacement curve: peak force Residual depth The rebound slope (k) and absorbed energy (W) are used to calculate the mean values ​​of various core mechanical characteristics using statistical algorithms. Based on these mean values, a fluctuation range of ±10% to ±15% of the mean is set as the acceptable threshold range for each type of core mechanical characteristic, forming a qualified sample mechanical characteristic library. Among these, peak power... The maximum value during the loading stage reflects the local stiffness and density of the material; residual depth The remaining displacement after unloading reflects the proportion of plastic deformation in the material; the springback slope k is the slope of the linear segment of the unloading curve, reflecting the material's elastic recovery ability; the absorbed energy W is the area enclosed by the curve, reflecting the material's resistance to deformation. (2-5) Output the mechanical feature library of qualified samples and enter the mechanical feature library of qualified samples into the database of the injection molding machine PLC control system as a quality judgment benchmark.

3. The online quality judgment method based on indentation mechanics during injection molding as described in claim 2, characterized in that: In steps (2-4), the qualified threshold range for each type of core mechanical feature is uniformly set to the mean ± 10%, specifically: Phonak The acceptable threshold range: ,in The mean peak intensity of the qualified samples; Residual depth The acceptable threshold range: ,in The mean residual depth of qualified samples; The acceptable threshold range for the rebound slope k: ,in The mean rebound slope of the qualified samples; The acceptable threshold range for absorbed energy W: ,in This represents the average absorbed energy of the qualified samples.

4. The online quality determination method based on indentation mechanics during injection molding as described in claim 2, characterized in that: In step (4), when the injection molding machine performs injection molding on the injection molded part, the injection molding machine PLC control system sends injection raw material, injection molding parameters, and injection cycle synchronization signals in real time; the injection molding machine performs melt filling, pressure holding, and cooling processes according to the preset indentation parameters in step (2-2) to complete the injection molding of the part.

5. The online quality determination method based on indentation mechanics during injection molding as described in claim 4, characterized in that: In step (4), the specific steps of the indentation detection execution unit to perform the indentation test are as follows: When the injection molding machine PLC control system sends a synchronization signal, the indentation detection execution unit is immediately triggered. According to the preset indentation parameters and detection homogeneous area parameters in step (2-1), the pressure head in the mold cavity is driven to contact the preset homogeneous area of ​​the injection molded part and perform the three-stage action of "loading-dwelling-unloading".

6. The online quality determination method based on indentation mechanics during injection molding as described in claim 5, characterized in that: In step (5), the specific steps of data preprocessing and mechanical feature extraction are as follows: (5-1) Data preprocessing: The wavelet filtering algorithm and linear temperature correction algorithm, which are completely consistent with those in step (2-3), are used to perform noise removal and temperature drift correction on the original force-displacement curve of the test piece collected in step (4) to eliminate signal errors caused by high temperature, vibration and electromagnetic interference of the mold, and thus obtain a smooth and standardized force-displacement curve that can be compared with qualified samples. (5-2) Extraction of core mechanical features: Extract four types of core mechanical features consistent with step (2-4) from the standardized force-displacement curve: peak force Residual depth The springback slope k and absorbed energy W are combined to form the mechanical characteristic vector of the test piece, and transmitted to the quality judgment unit in real time.

7. The online quality determination method based on indentation mechanics during injection molding as described in claim 6, characterized in that: In step (6), the specific steps for comparing the mechanical feature vector of the test piece with the acceptable threshold range are as follows: (6-1) Reference data retrieval: After receiving the mechanical feature vector of the test piece, the quality judgment unit automatically retrieves the corresponding reference data from the qualified sample mechanical feature library in step (2) to determine the qualified threshold range of each of the four core mechanical features. (6-2) Feature-by-feature numerical comparison: If the peak strength to be measured If so, the peak power index is deemed qualified; If the peak strength to be measured or If so, the peak power index is deemed unqualified; If the residual depth to be measured If so, the residual depth index is deemed qualified; If the residual depth to be measured or If so, the residual depth index is deemed unqualified; If the rebound slope to be measured If so, the rebound slope index is deemed qualified; If the rebound slope to be measured or If so, the rebound slope index is deemed unqualified; If the absorbed energy to be measured If so, the energy absorption index is deemed qualified; If the absorbed energy to be measured or If so, the energy absorption index is deemed unqualified; (6-3) Overall quality assessment: Based on the comparison results of various core mechanical features in step (6-2) above, the following judgment rules are applied: If all four core mechanical characteristics meet their respective qualification threshold range requirements, the injection molded part is determined to be a qualified part. If any one or more core mechanical characteristics exceed the corresponding qualified threshold range, the injection molded part is determined to be a part that needs to be re-inspected. (6-4) Output the quality judgment result and transmit it to the production line closed-loop feedback unit; at the same time, generate a radar chart comparing the mechanical characteristics of the test part and the qualified sample, and display it on the system human-machine interface to facilitate operators to quickly judge the cause of the defect.

8. The online quality determination method based on indentation mechanics during injection molding as described in claim 7, characterized in that: In step (7), the specific steps of production line closed-loop feedback and feature library dynamic optimization are as follows: (7-1) Closed-loop feedback action of the production line: (7-1.1) If the injection molded part is a qualified part, then control the production line to perform normal diversion and enter the subsequent process; (7-1.2) If the injection molded part is a part that needs to be re-inspected, control the production line to divert it to the re-inspection station and trigger a slight audible and visual alarm to remind the operator to pay attention; (7-1.3) If three consecutive injection molded parts are determined to be parts that need to be re-inspected, a high-frequency audible and visual alarm will be triggered, and process adjustment suggestions will be output. (7-1.4) Mark the nodes in the production line that are judged as qualified parts, parts that need to be re-inspected, and parts that need to be re-inspected for 3 consecutive parts, and generate a closed-loop feedback node marking diagram; (7-2) Dynamically optimize the mechanical feature library of qualified samples: (7-2.1) Data screening: The mechanical feature vectors of qualified parts in the production line are sent back to the qualified sample feature library unit, and the mechanical feature vectors of parts that need to be re-inspected are removed; (7-2.2) Periodic update: For every 1,000 injection molded parts produced, the qualified sample feature library unit re-statistically analyzes all qualified part data in the library using statistical algorithms, updates the mean and qualified threshold range of the four core mechanical features, and realizes periodic optimization of the feature library; (7-2.3) Data support: After the feature library is updated, the process adjustment timing curve of the mean change of the core mechanical features is automatically generated and stored in the database of the injection molding machine PLC control system. The process adjustment timing curve is associated with the change of injection molding process parameters, providing accurate data support for subsequent process optimization and mold maintenance.

9. The online quality determination method based on indentation mechanics during injection molding as described in claim 1, characterized in that: In step (3), the specific steps for the linkage debugging and verification of the online quality judgment system and the injection molding machine are as follows: (3-1) Debugging preparation: Obtain the hardware parameters of the qualified sample feature library unit, indentation detection execution unit, force-displacement signal acquisition unit, data processing unit, quality judgment unit, and production line closed-loop feedback unit; communication parameters of the injection molding machine PLC control system; and debugging samples; (3-2) Module communication setup: Establish wired communication connections between the qualified sample feature library unit, the indentation detection execution unit, the force-displacement signal acquisition unit, the data processing unit, the quality judgment unit, and the production line closed-loop feedback unit and the injection molding machine PLC control system, determine the data flow between each unit, and ensure stable data transmission and no packet loss; (3-3) Accuracy and timing debugging: The force-displacement signal acquisition unit is calibrated to ensure that the force error is ≤ ±0.1N and the displacement error is ≤ ±0.001mm, thus eliminating the influence of temperature drift of the sensor. Adjust the triggering timing of the indentation detection execution unit to ensure accurate triggering when the mold opens to the preset position, synchronized with the injection cycle, without adding extra production time; (3-4) Functional verification: Send the debugging sample into the injection molding machine PLC control system to verify the accuracy of the judgment logic and the effectiveness of the closed-loop feedback action, and troubleshoot communication interruption and data delay problems; (3-5) System acceptance: The online quality judgment system is debugged and qualified, and a system unit linkage timing diagram is generated and stored in the maintenance interface of the injection molding machine PLC control system for easy subsequent fault location and module upgrade.

Images