Precise injection mold accessory production quality traceability management method and system

Abstract

The invention provides a precision injection mold accessory production quality traceability management method and system, and relates to the technical field of intelligent manufacturing, and the method comprises the steps: collecting material, process, equipment and environment data in real time through a distributed sensor; fusing multi-source data based on dynamic material characteristic parameters and process stability indexes, and quantifying melt flow and process fluctuation characteristics; constructing a mixed kernel function anomaly detection model to identify quality deviation; establishing a cross-process association map to reveal the space-time relationship among the raw materials, the process and the finished product; generating a three-dimensional traceability identifier containing the material hash, the process compression code and the block chain address; a block chain enhanced database is adopted to realize tamper-proof storage; and generating a visual traceability report through reverse analysis. Through dynamic modeling, cross-process association and block chain technologies, the problems of data isolation, detection lag and low traceability credibility in a traditional method are solved, the quality traceability efficiency and precision are remarkably improved, and precise injection molding full life cycle management is supported.

Description

Technical Field

[0001] The present invention relates to the field of intelligent manufacturing technology, and in particular to a method and system for tracing the production quality of precision injection mold accessories. Background Art

[0002] The production of precision injection mold components presents multiple technical challenges for quality control and traceability management. Traditional monitoring systems typically rely on the collection and analysis of a single data source, focusing solely on process parameters or finished product inspection results. This results in ineffective integration of multi-dimensional data such as material properties, equipment status, and environmental conditions. This data silo phenomenon makes it difficult to fully capture key factors influencing the production process, making it impossible to establish a dynamic correlation model between material behavior and process parameters. This ultimately leads to inefficient tracing of the root causes of quality issues.

[0003] Existing anomaly detection technologies are mostly based on offline spot checks or fixed threshold judgments, lacking the ability to dynamically adapt to the time-varying characteristics of molten materials. During the injection molding process, nonlinear changes in parameters such as melt flow rate and viscosity are often overlooked, making it difficult for anomaly detection models to identify early quality deviations in real time. This results in defective products being discovered only after mass production, significantly increasing costs and wasting resources. Furthermore, traditional detection methods lack the ability to analyze the effects of multi-physics coupling, further reducing the accuracy and timeliness of quality warnings.

[0004] When it comes to quality data management, existing systems typically use centralized databases to store information, which carries the risk of data tampering and hinders trusted sharing across departments and supply chains. Key data such as raw material batches, process parameter sets, and equipment operation logs lack temporal and spatial correlation, making it difficult to quickly pinpoint the responsible link when quality issues occur. For example, dimensional deviations in finished products can be caused by a variety of factors, including material fluctuations, mold wear, or process parameter drift, but traditional methods cannot effectively trace the specific causes.

[0005] Furthermore, existing traceability technologies often rely on one-dimensional codes or simple labels, which have limited information capacity and are easily forged. Complex characteristics of process parameters (such as pressure curves and temperature gradients) are difficult to compress and store, resulting in high redundancy in traceability data and low query efficiency. Furthermore, visual report generation relies on manual compilation and empirical judgment, lacking automated reconstruction capabilities, making it difficult to meet the modern manufacturing industry's demand for real-time traceability and decision support. Summary of the Invention

[0006] In order to solve the technical problems in the prior art, the present invention provides a method and system for tracing the production quality of precision injection mold accessories.

[0007] The technical solutions provided by the present invention are as follows:

[0008] First aspect:

[0009] The present invention provides a method for tracing the quality of production of precision injection mold parts, comprising:

[0010] S1. Real-time collection of multi-source heterogeneous data during mold production through a distributed sensor network, including material property data, process parameter data, equipment status data, and environmental monitoring data;

[0011] S2. Fusion processing of the raw data based on dynamic material property parameters and process stability index. The dynamic material property parameters (DMP) reflect the time-varying characteristics of the molten material during the injection molding process, and the process stability index (PSI) characterizes the degree of fluctuation of key process parameters between consecutive production batches.

[0012] S3, uses a multi-dimensional anomaly detection model to perform online quality analysis on the fused data to identify potential quality deviation patterns;

[0013] S4. Establish a cross-process quality data association map to temporally and spatially associate raw material batches, process parameter sets, equipment operation logs, and finished product quality test results;

[0014] S5. Generate a unique three-dimensional traceability code, wherein the code includes a material feature vector, a process feature matrix, and a detection feature sequence;

[0015] S6. Build a blockchain-enhanced quality database to store traceability data in distributed nodes in timestamp order;

[0016] S7. Based on user query requests, the complete quality history is reconstructed through the reverse parsing engine and a visual traceability report is generated.

[0017] Second aspect:

[0018] The present invention provides a precision injection mold parts production quality traceability management system, comprising:

[0019] A distributed sensing module includes embedded temperature sensors, pressure sensors, and visual detection units; an edge computing gateway equipped with a data fusion processor and a local anomaly detection unit; a cloud-based analysis platform deployed with a quality correlation graph engine and blockchain nodes; a user interaction terminal equipped with an augmented reality (AR) display device and a voice control interface; and a secure communication module including an industrial Ethernet switch that uses the quantum key distribution (QKD) protocol.

[0020] The beneficial effects brought about by the technical solution provided by the present invention include at least:

[0021] (1) In this invention, a multi-physics field coupled quality evaluation model is constructed by collecting material properties, process parameters, equipment status, and environmental data in real time through a distributed sensor network. This model combines the dynamic material property parameter (DMP) with the process stability index (PSI) for fusion calculation. This approach breaks through the limitations of traditional single data dimensions and accurately quantifies the time-varying characteristics of melt flow behavior through dynamic weight allocation and feature extraction. This significantly improves the sensitivity and accuracy of anomaly detection, enabling early warning and intervention of quality issues.

[0022] (2) In the present invention, a quality data association graph is constructed based on a graph neural network, and raw material batches, process parameter sets, and equipment operation logs are mapped into spatiotemporal association nodes to reveal the propagation path of quality problems. Combined with layered encrypted three-dimensional traceability coding (material feature hash, process matrix compression, spatiotemporal block address), unique identification and efficient storage of all factor data are achieved. This technology effectively breaks down information barriers between processes, supports full-link traceability from raw materials to finished products, and significantly shortens the time to locate quality problems.

[0023] (3) In this invention, a blockchain database with a hybrid consensus mechanism, PBFT+PoW, is used to encrypt and store production process data in distributed nodes according to timestamps. Data verification rules are automatically executed through a hash chain structure and smart contracts. This design not only meets the efficiency requirements of real-time data writing, but also ensures the immutability of historical data, building a reliable quality data storage system. At the same time, quantum key distribution technology strengthens communication security and prevents man-in-the-middle attacks and information leakage during industrial data transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0025] Figure 1 A schematic diagram of a process for tracing the quality of precision injection mold parts provided by an embodiment of the present invention;

[0026] Figure 2 A flow chart of a real-time melt flow rate testing method for a method for traceability management of production quality of precision injection mold parts provided by an embodiment of the present invention;

[0027] Figure 3 A schematic flow chart of an orthogonal test method for the production quality traceability management method of a precision injection mold component provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0029] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0030] In the embodiments of the present invention, the terms "image" and "picture" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same. The terms "of," "corresponding," and "corresponding" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same.

[0031] In the embodiments of the present invention, sometimes a subscript such as W1 may be mistakenly written as a non-subscript form such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0032] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0033] Reference Manual Figure 1 , which shows a flow chart of a method for tracing the production quality of precision injection mold accessories provided by an embodiment of the present invention.

[0034] An embodiment of the present invention provides a method for tracing the production quality of precision injection mold parts. The method can be implemented by a device for tracing the production quality of precision injection mold parts. The device can be a terminal or a server. The processing flow of the method for tracing the production quality of precision injection mold parts can include the following steps:

[0035] S1. Real-time collection of multi-source heterogeneous data in the mold production process through a distributed sensor network, including material property data, process parameter data, equipment status data, and environmental monitoring data.

[0036] It is understandable that by deploying multi-type sensor networks at key positions of the mold, multi-source heterogeneous data such as material melting state, process dynamic parameters, equipment operation indicators and environmental conditions can be collected in real time. This step aims to eliminate the data blind spots of the traditional monitoring system and provide comprehensive and real-time underlying data support for subsequent analysis. The integrity and consistency of information collection can be ensured through complementary verification of multi-sensor data.

[0037] S2. The original data is fused and processed based on dynamic material characteristic parameters and process stability index. The dynamic material characteristic parameters reflect the time-varying characteristics of the molten material during the injection molding process, and the process stability index represents the degree of fluctuation of key process parameters between consecutive production batches.

[0038] It is understandable that the time-varying characteristics of melt flow are analyzed through dynamic material characteristic parameters, the steady-state characteristics of the production process are evaluated in combination with the process stability index, and the multi-dimensional raw data are converted into process quality indicators that can be quantified and analyzed using a feature fusion algorithm. The core of this is to establish a mapping relationship between material behavior and process parameters, and to solve the problem of weak data correlation caused by the coupling of multiple physical fields in the injection molding process.

[0039] S3. Use a multi-dimensional anomaly detection model to perform online quality analysis on the fused data to identify potential quality deviation patterns.

[0040] It is understandable that it is necessary to build a multi-level anomaly detection framework that integrates statistical analysis, pattern recognition and machine learning, conduct real-time monitoring and cross-validation of the integrated process quality indicators, identify potential defect characteristics such as material rheological anomalies and parameter drift through a dynamic learning mechanism, and achieve early warning of quality deviations, effectively avoiding batch quality problems caused by the lag of traditional detection methods.

[0041] S4. Establish a cross-process quality data association map to temporally and spatially associate raw material batches, process parameter sets, equipment operation logs, and finished product quality inspection results.

[0042] It is understandable that by building a cross-process association network based on a graph-structured data model, discrete raw material properties, process settings, equipment status and other data are mapped into time-space association nodes, and the transmission path of process parameter fluctuations to finished product quality is revealed through edge weight calculation, breaking through the information barriers of traditional stage-by-stage quality control, and establishing a causal reasoning system covering the entire production chain.

[0043] S5. Generate a unique three-dimensional traceability code, which includes a material feature vector, a process feature matrix, and a detection feature sequence.

[0044] It is understandable that layered encryption technology is used to generate a three-dimensional composite code that integrates material characteristics, process fingerprints and test results. Each code uniquely identifies the full-factor quality characteristics of a single product, and the traceability information is ensured to be tamper-proof through cross-binding of multi-dimensional characteristics. At the same time, data compression technology is used to reduce storage and transmission loads.

[0045] S6. Build a blockchain-enhanced quality database and store traceability data in distributed nodes in timestamp order.

[0046] It is understandable that the quality database architecture based on distributed ledger technology is designed, the complete production process data with timestamps is encrypted and written into the blockchain node, the hash chain structure is used to detect data tampering, and the data writing efficiency and anti-counterfeiting security are taken into account through a hybrid consensus mechanism to build a trustworthy quality data storage system.

[0047] S7. Based on user query requests, the complete quality history is reconstructed through the reverse parsing engine and a visual traceability report is generated.

[0048] It is understandable that developing a quality traceability engine based on reverse parsing of code can quickly locate related data nodes in the distributed database according to user requests, reconstruct the entire process quality event chain from raw material storage to finished product inspection, and generate a visual interactive report including process parameter evolution and defect correlation analysis, which significantly improves the timeliness and explainability of quality problem tracing.

[0049] In a possible implementation, S2 specifically includes:

[0050] The calculation method of dynamic material characteristic parameter DMP is:

[0051]

[0052] Among them, MFR t is the real-time melt flow rate (MFR), the unit is g / 10min, the test method complies with ISO 1133 standard, ΔT is the temperature gradient in the mold cavity, the unit is ℃, CTE is the material thermal expansion coefficient (CTE, Coefficient of Thermal Expansion), the unit is μm / m·℃, P m is the real-time pressure in the mold cavity, in MPa, η is the melt viscosity, in Pa·s, η0 is the reference viscosity, in Pa·s, α, β, and γ are weight coefficients determined by the orthogonal test method and satisfy α+β+γ=1.

[0053] Furthermore, if Figure 2 As shown, the real-time melt flow rate MFR t The test method is as follows:

[0054] S101. Install an online melt flow rate tester at the melt flow channel of the injection molding machine. The tester includes a constant temperature control module and an automatic cutting device.

[0055] S102. Set test conditions according to ISO 1133 standard, including:

[0056] Test temperature: set according to the material type (such as PP material set to 230±0.5℃),

[0057] Load weight: 2.16kg standard weight,

[0058] Barrel diameter: 9.55±0.05mm;

[0059] S103, Testing process:

[0060] a) Preheat the barrel to the set temperature and keep it for 10 minutes,

[0061] b) After loading the pellets into the barrel, apply a standard load.

[0062] c) When the melt flows out of the standard die, start the automatic cutting device to cut the sample every 10 seconds.

[0063] d) Weigh the average mass of 5 samples and calculate the melt flow rate:

[0064]

[0065] in, is the average mass of a single sample, in g; t is the cutting time interval, in s.

[0066] Furthermore, if Figure 3 As shown in Figure 2, the process of determining the weight coefficients α, β, and γ by the orthogonal test method is as follows:

[0067] S111. Determine the experimental factors and levels:

[0068] Factor 1: Temperature gradient ΔT (level: 10°C, 15°C, 20°C),

[0069] Factor 2: Cavity pressure P m (Level: 50MPa, 70MPa, 90MPa),

[0070] Factor 3: Material type (level: PP, ABS, PC);

[0071] S112. Select L9(3^4) orthogonal array to arrange the experiment, and repeat each experiment 3 times;

[0072] S113. Collect quality indicator data: product dimensional accuracy (μm), surface defect rate (%), mechanical property deviation (%);

[0073] S114. Calculate the contribution rate of each factor:

[0074]

[0075] Among them, K jm is the sum of the indicators at the mth level of the jth factor, n is the number of repetitions for each level, and N is the total number of experiments;

[0076] S115. Allocate weight coefficients based on contribution rate:

[0077]

[0078] Finally, normalization is performed to satisfy α+β+γ=1.

[0079] In a possible implementation, S2 specifically includes:

[0080] The calculation method of process stability index PSI is:

[0081] PSI=1-[w1×CV(P)+w2×CV(T)+w3×CV(t)]

[0082] Among them, CV(P), CV(T), and CV(t) are the coefficients of variation of pressure, temperature, and holding time between consecutive production batches, respectively. w1, w2, and w3 are dynamic weights, which are calculated using the improved entropy weight method:

[0083]

[0084] Among them, E i is the information entropy value of each parameter, in nat.

[0085] Furthermore, the process of calculating dynamic weights using the improved entropy weight method is as follows:

[0086] S201, construct an evaluation matrix: CV(P), CV(T), CV(t) of 10 consecutive production batches form a 3×10 matrix X;

[0087] S202, standardization processing:

[0088]

[0089] S203. Calculate information entropy:

[0090]

[0091] in,;

[0092] S204, calculate the correction weight:

[0093]

[0094] Among them, δ=0.01 is a smoothing factor to prevent the denominator from being zero.

[0095] In a possible implementation, S3 specifically includes:

[0096] The multi-dimensional anomaly detection model uses a comprehensive quality score Q for judgment:

[0097] Q = λ·DMP+μ·PSI+θ·F(t)

[0098] Among them, λ, μ, and θ are adaptive weight coefficients determined by the hierarchical analysis method, and F(t) is the time correction factor, F(t) = e -kt , k is the material aging rate constant, unit is s -1 , calculated using the Arrhenius equation.

[0099] Furthermore, the process of determining the adaptive weight coefficient by the hierarchical analysis method is as follows:

[0100] S301. Build a hierarchical structure:

[0101] Target layer: quality score Q is optimal,

[0102] Criterion layer: DMP, PSI, F(t),

[0103] Solution layer: different product types;

[0104] S302. Establish a judgment matrix (taking auto parts as an example):

[0105]

[0106] S303, calculate the eigenvector:

[0107] a) Calculate the geometric mean of each row,

[0108] b) Normalize to get the weight vector [λ,μ,θ],

[0109] c) Consistency test: CR = CI / RI < 0.1;

[0110] S304. Adjustment based on product type: Precision parts: λ weight increased by 20%, Structural parts: μ weight increased by 15%.

[0111] Furthermore, the process of calculating the material aging rate constant through the Arrhenius equation is as follows:

[0112] S311. Design accelerated aging test: set 3 temperature points: T1, T2, T3 (usually 10°C apart), and test 5 groups of specimens at each temperature point;

[0113] S312. Measure performance decay time: record the time t when the elastic modulus of the material drops to 95% at each temperature. i ;

[0114] S313, Fitting the Arrhenius equation:

[0115]

[0116] Among them, E a is the activation energy in J / mol, R is the gas constant, 8.314 J / (mol·K), and A is the pre-exponential factor;

[0117] S314. Calculate the aging rate constant:

[0118]

[0119] Among them, T process is the actual production process temperature.

[0120] In a possible implementation, S3 further includes:

[0121] The anomaly detection model uses a hybrid kernel function support vector machine algorithm, and the kernel function is defined as:

[0122] K(x i ,x j )=exp(-γ||x i -x j || 2 +α·(x i ·x j ) d )

[0123] Among them, γ is the radial basis function parameter, α is the polynomial kernel coefficient, d is the polynomial order, which is a positive integer. The kernel parameters are optimized and determined by the grid search method.

[0124] Furthermore, the process of optimizing kernel parameters by grid search method is as follows:

[0125] S321. Define parameter search space:

[0126] γ:10 -3 ,10 -2 ,...,10 2 (Log uniform distribution):

[0127] α:0.1,0.5,1.0

[0128] d:2,3,4;

[0129] S322. Generate parameter grid:

[0130] Use Cartesian product to generate 3×3×3=27 sets of parameter combinations;

[0131] S323, cross validation training:

[0132] a) Divide the training set into 5 subsets

[0133] b) Calculate the average classification accuracy for each set of parameters:

[0134]

[0135] Among them, TP k is the number of true positive samples, TN k is the number of true negative samples, N k is the total number of subset samples;

[0136] S324. Select optimal parameters:

[0137] With the constraints of accuracy > 95% and computation time < 200ms, the Pareto optimal solution is selected.

[0138] In a possible implementation, S4 further includes:

[0139] The quality data association graph is constructed through a graph neural network. The node features include equipment ID, process time window, and process parameter set. The edge weight is determined by the weighted sum of the process transfer probability and quality correlation.

[0140] Furthermore, the edge weight calculation process is as follows:

[0141] S401. Calculate the process transition probability:

[0142]

[0143] Among them, N transition (i→j) represents the number of historical transfers from process i to process j, N total (i) represents the total number of executions of process i;

[0144] S402, calculate quality correlation:

[0145]

[0146] Where Δμ i→j is the mean change of key quality indicators when switching from process i to j, μ base is the mean value of the benchmark quality index;

[0147] S403, composite edge weight:

[0148] W edge =0.7·P ij +0.3·Q ij

[0149] Among them, W edge is the edge weight, and the coefficient is determined through regression analysis of historical data.

[0150] In a possible implementation, S5 further includes:

[0151] The first dimension: material feature hash value based on the SHA-256 algorithm;

[0152] Second dimension: Singular value decomposition SVD compression coding of process parameter feature matrix;

[0153] The third dimension: composite encoding of time-space stamp and blockchain address.

[0154] Furthermore, the implementation process of three-dimensional traceability coding is as follows:

[0155] S501, first dimension (material feature hash value):

[0156] a) Extract material batch number, supplier code, and DMP parameter sequence,

[0157] b) Calculate the SHA-256 hash value after concatenating the strings,

[0158] c) Output 64-bit hexadecimal code;

[0159] S502, second dimension (SVD compression coding):

[0160] a) Construct process parameter matrix X m×n (m is the number of processes, n is the parameter type),

[0161] b) Perform singular value decomposition:

[0162] X=U∑V T

[0163] c) retain the first three singular values to form the compressed code σ1, σ2, σ3;

[0164] S503, the third dimension (space-time composite coding):

[0165] a) Timestamp encoding: Convert UNIX time to a 12-digit hexadecimal string

[0166] b) Spatial coding: Extract the last 6 decimal places of the injection molding machine GPS coordinates

[0167] c) Blockchain address: the first 8 digits of the transaction hash value

[0168] d) Combined format: [timestamp]@[longitude],[latitude]#[block address].

[0169] For example, 3K7Z9W4F6L2@121.473702,31.230416#5a3bd8e9

[0170] In a possible implementation, S6 further includes:

[0171] The blockchain-enhanced database adopts a hybrid consensus mechanism that combines the Practical Byzantine Fault Tolerance (PBFT) algorithm and the Proof-of-Work (PoW) algorithm. The PoW difficulty value is dynamically adjusted according to the frequency of data updates.

[0172] Furthermore, the hybrid consensus mechanism implementation process is as follows:

[0173] S601, data classification processing:

[0174] Real-time data (<1s delay): uses PBFT consensus, batch data (>1min): uses PoW consensus;

[0175] S602. Dynamic Difficulty Adjustment Algorithm:

[0176]

[0177] Among them, Base is the basic difficulty value, which is initially set to 1,000,000, F update The number of data updates in the past 10 minutes;

[0178] S603, PBFT execution process:

[0179] a) The client sends a request to the master node

[0180] b) The master node broadcasts to all backup nodes

[0181] c) Nodes execute the three-phase protocol (pre-prepare, prepare, commit)

[0182] d) The block is written after more than 2 / 3 of the nodes pass the verification.

[0183] In a possible implementation, S7 further includes:

[0184] When generating a visual traceability report, the multidimensional scaling analysis (MDS) algorithm is used to reduce the dimensionality of quality features, and the visual layout is optimized through the pressure function.

[0185] Furthermore, the process of MDS dimensionality reduction and visualization optimization is as follows:

[0186] S701. Construct quality feature matrix:

[0187] The DMP, PSI, and Q values of 100 samples were collected to form a 100 × 3 matrix;

[0188] S702, calculate the dissimilarity matrix:

[0189]

[0190] S703, MDS iterative solution:

[0191] a) Initialize the two-dimensional coordinate Y (0)

[0192] b) Calculate the stress function:

[0193]

[0194] c) Use gradient descent method to minimize the Stress value;

[0195] S704, pressure function layout optimization:

[0196]

[0197] Iteratively adjust the layout until the combined force is balanced.

[0198] The present invention also provides a precision injection mold parts production quality traceability management system, which is applied to the precision injection mold parts production quality traceability management method, comprising:

[0199] It includes a distributed sensing module with embedded temperature sensors, pressure sensors and visual detection units; an edge computing gateway equipped with a data fusion processor and a local anomaly detection unit; a cloud analysis platform deployed with a quality correlation graph engine and blockchain nodes; a user interaction terminal equipped with an augmented reality (AR) display device and a voice control interface; and a secure communication module including an industrial Ethernet switch using the quantum key distribution (QKD) protocol.

[0200] A multi-type sensor cluster is deployed at key workstations in the injection mold, including temperature, pressure, displacement, and visual inspection units. The temperature sensing unit utilizes high-precision thermocouples covering the processing temperature range of common engineering plastics to ensure reliable melt temperature monitoring. The pressure sensing unit utilizes a piezoelectric sensor that meets the maximum clamping force requirements of the injection molding machine, with a sampling rate that fully captures dynamic pressure changes. The displacement sensing unit is equipped with a high-resolution laser displacement sensor to detect subtle mold deformation. The visual inspection unit utilizes an industrial camera combined with a ring light source to achieve real-time scanning of appearance defects under high-speed production. The sensor network is connected via an industrial Ethernet switch, supporting multi-protocol communication, and the data acquisition latency meets the real-time requirements of the injection molding process.

[0201] The edge computing gateway is equipped with a multi-core processor and dedicated hardware acceleration modules, along with memory and storage components. The processor performs real-time calculations of dynamic material properties and process stability indices, while the hardware acceleration module optimizes signal processing algorithms to eliminate mechanical interference. The memory capacity supports a sliding time window data processing mechanism, and the storage space can cache periodic production data. The anomaly detection unit integrates a hybrid kernel function machine learning model to perform online assessments of fused quality indicators and trigger a graded alarm mechanism to prevent the continuous production of defective products. The communication interface uses a lightweight protocol to control uplink bandwidth usage to avoid network congestion.

[0202] The cloud-based analytics platform utilizes a graph-based quality correlation engine to store material attributes, equipment status, and process time series data. Node relationship definitions include causal edges reflecting parameter influence and transfer edges marking process time dependencies. The blockchain-enhanced database utilizes a distributed architecture, with master nodes working collaboratively with external validation nodes. Smart contracts define key threshold rules to automatically trigger data protection mechanisms. The consensus mechanism combines real-time response with tamper-proofing requirements, adapting differentiated strategies to different data types.

[0203] The user interaction terminal is equipped with an augmented reality display, which uses optical perspective technology to visualize process parameters overlaid on the work area and supports gesture interaction to access relevant data. The voice control interface integrates a noise reduction module, ensuring stable recognition of operating commands even in noisy industrial environments. The secure communication module utilizes dynamic key distribution technology, combined with a high-speed network architecture to create an end-to-end encrypted channel. When the systems work together, the end-to-end response efficiency from data collection to report generation meets production traceability requirements. Anomaly detection capabilities are validated using historical data, and the blockchain network supports high-concurrency data writing. Overall system stability meets industrial-grade operating standards.

[0204] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0205] (1) In this invention, a multi-physics field coupled quality evaluation model is constructed by collecting material properties, process parameters, equipment status, and environmental data in real time through a distributed sensor network, combined with the fusion calculation of dynamic material property parameters (DMP) and process stability index (PSI). This solution breaks through the limitations of traditional single data dimensions and accurately quantifies the time-varying characteristics of melt flow behavior through dynamic weight allocation and feature extraction, significantly improving the sensitivity and accuracy of anomaly detection, and achieving early warning and intervention of quality issues.

[0206] (2) In the present invention, a quality data association graph is constructed based on a graph neural network, and raw material batches, process parameter sets, and equipment operation logs are mapped into spatiotemporal association nodes to reveal the propagation path of quality problems. Combined with layered encrypted three-dimensional traceability coding (material feature hash, process matrix compression, spatiotemporal block address), unique identification and efficient storage of all factor data are achieved. This technology effectively breaks down information barriers between processes, supports full-link traceability from raw materials to finished products, and significantly shortens the time to locate quality problems.

[0207] (3) In this invention, a blockchain database with a hybrid consensus mechanism (PBFT+PoW) is used to encrypt and store production process data in distributed nodes according to timestamps, and data verification rules are automatically executed through a hash chain structure and smart contracts. This design not only meets the efficiency requirements of real-time data writing, but also ensures the immutability of historical data, and builds a reliable quality data storage system. At the same time, quantum key distribution technology strengthens communication security and prevents man-in-the-middle attacks and information leakage during industrial data transmission.

[0208] The above content is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

[0209] There are a few points to note:

[0210] (1) The drawings of the embodiments of the present invention only relate to the structures related to the embodiments of the present invention. Other structures may refer to conventional designs.

[0211] (2) For the sake of clarity, the thickness of layers or regions in the drawings used to describe the embodiments of the present invention are exaggerated or reduced, that is, these drawings are not drawn to scale. It is understood that when an element such as a layer, film, region, or substrate is referred to as being "on" or "under" another element, the element may be "directly" "on" or "under" the other element or intervening elements may be present.

[0212] (3) In the absence of conflict, the embodiments of the present invention and the features therein may be combined with each other to form new embodiments.

[0213] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. The protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for tracing the production quality of precision injection mold parts, characterized in that: include: S1. Real-time collection of multi-source heterogeneous data during mold production through a distributed sensor network, including material property data, process parameter data, equipment status data, and environmental monitoring data; S2. Fusion processing of the raw data based on dynamic material characteristic parameters and a process stability index, wherein the dynamic material characteristic parameters reflect the time-varying characteristics of the molten material during the injection molding process, and the process stability index characterizes the degree of fluctuation of key process parameters between consecutive production batches; S3, uses a multi-dimensional anomaly detection model to perform online quality analysis on the fused data to identify potential quality deviation patterns; S4. Establish a cross-process quality data association map to temporally and spatially associate raw material batches, process parameter sets, equipment operation logs, and finished product quality test results; S5. Generate a unique three-dimensional traceability code, wherein the code includes a material feature vector, a process feature matrix, and a detection feature sequence; S6. Build a blockchain-enhanced quality database to store traceability data in distributed nodes in timestamp order; S7. Based on user query requests, the complete quality history is reconstructed through the reverse parsing engine and a visual traceability report is generated.

2. A method for tracing the production quality of precision injection mold parts according to claim 1, characterized in that: The S2 specifically includes: The calculation method of the dynamic material characteristic parameter DMP is: Among them, MFR t is the real-time melt flow rate, in g / 10min, the test method complies with ISO 1133 standard, ΔT is the temperature gradient in the mold cavity, in °C, CTE is the thermal expansion coefficient of the material, in μm / m·°C, P m is the real-time pressure in the mold cavity, in MPa, η is the melt viscosity, in Pa·s, η0 is the reference viscosity, in Pa·s, α, β, and γ are weight coefficients determined by the orthogonal test method and satisfy α+β+γ=1.

3. A method for tracing the production quality of precision injection mold parts according to claim 2, characterized in that: Said S2 specifically includes: The calculation method of the process stability index PSI is: PSI=1-[w1×CV(P)+w2×CV(T)+w3×CV(t)] Among them, CV(P), CV(T), and CV(t) are the coefficients of variation of pressure, temperature, and holding time between consecutive production batches, respectively. w1, w2, and w3 are dynamic weights, which are calculated using the improved entropy weight method: Among them, E i is the information entropy value of each parameter, in nat.

4. A method for tracing the production quality of precision injection mold parts according to claim 3, characterized in that: The S3 specifically includes: The multi-dimensional anomaly detection model uses a comprehensive quality score Q for judgment: Q = λ·DMP+μ·PSI+θ·F(t) Among them, λ, μ, and θ are adaptive weight coefficients determined by the hierarchical analysis method, and F(t) is the time correction factor, F(t) = e -kt , k is the material aging rate constant, unit is s -1 , calculated using the Arrhenius equation.

5. A method for tracing the production quality of precision injection mold parts according to claim 4, characterized in that: Said S3 further comprises: The anomaly detection model adopts a hybrid kernel function support vector machine algorithm, and the kernel function is defined as: K(x i ,x j )=exp(-γ||x i -x j || 2 +α·(x i ·x j ) d ) Among them, γ is the radial basis function parameter, α is the polynomial kernel coefficient, d is the polynomial order, which is a positive integer. The kernel parameters are optimized and determined by the grid search method.

6. A method for tracing the production quality of precision injection mold parts according to claim 5, characterized in that: Said S4 further comprises: The quality data association graph is constructed through a graph neural network. The node features include equipment ID, process time window, and process parameter set. The edge weight is determined by the weighted sum of the process transition probability and the quality correlation.

7. A method for tracing the production quality of precision injection mold parts according to claim 6, characterized in that: Said S5 further comprises: The first dimension: material feature hash value based on the SHA-256 algorithm; Second dimension: Singular value decomposition SVD compression coding of process parameter feature matrix; The third dimension: composite encoding of time-space stamp and blockchain address.

8. A method for tracing the production quality of precision injection mold parts according to claim 7, characterized in that: Said S6 further comprises: The blockchain-enhanced database adopts a hybrid consensus mechanism that combines the Practical Byzantine Fault Tolerance (PBFT) algorithm and the Proof-of-Work (PoW) algorithm. The PoW difficulty value is dynamically adjusted according to the frequency of data updates.

9. A method for tracing the production quality of precision injection mold parts according to claim 8, characterized in that: The S7 further includes: When generating a visual traceability report, the multidimensional scaling analysis (MDS) algorithm is used to reduce the dimensionality of quality features, and the visual layout is optimized through the pressure function.

10. A precision injection mold parts production quality traceability management system, characterized by: include: A distributed sensing module containing embedded temperature sensors, pressure sensors, and visual detection units; An edge computing gateway equipped with a data fusion processor and a local anomaly detection unit; A cloud-based analysis platform deployed with a quality correlation graph engine and blockchain nodes; a user interaction terminal equipped with an augmented reality (AR) display device and a voice control interface; and a secure communication module including an industrial Ethernet switch that uses the quantum key distribution (QKD) protocol.

Images