Directional detection method and system for production quality of 200-grade direct-welding enameled wire

Abstract

The invention discloses a production quality directional detection method of a 200-grade direct-welding enameled wire, and relates to the technical field of enameled wire quality detection.The method comprises the steps that production simulation is carried out on the basis of a kth production control deviation matrix of a kth process node and by combining enameled wire defect parameters of a (k-1) th node, the enameled wire defect parameters of the kth node are obtained, k is an integer, and k is an integer; n is the total number of enameled wire process nodes, and the (k-1) th node is an upstream process node of the kth node; when k is equal to N, integrating the defect parameters of the enameled wire, and constructing a defect triggering time sequence matrix; performing defect spreading type combination on the defect triggering time sequence matrix to generate a defect spreading map; traversing the defect spreading map to carry out confidence evaluation to obtain a confidence coefficient; and extracting union set defect types of the defect spreading maps with confidence greater than or equal to a confidence threshold, and sending the union set defect types to a quality inspection end to execute directional quality inspection. According to the invention, the technical problem of low detection efficiency of the 200-level directly-welded enameled wire in the prior art is solved.

Description

Technical Field

[0001] The invention relates to the technical field of enameled wire quality detection, and in particular to a production quality directional detection method and system for a 200-level direct-weldable enameled wire. Background Art

[0002] In the field of 200-level direct-weld enameled wire manufacturing, the strict temperature resistance and direct-weldability requirements of 200-level direct-weld enameled wire make it possible for minor defects to lead to performance failure, and the multi-process complex process chain further amplifies the risk of defect transmission. Therefore, quality inspection is particularly important to ensure that the electrical performance, mechanical strength and direct-weldability of the enameled wire meet the standards. Existing technologies usually rely on random sampling for quality control, that is, samples are selected from the production batch, and then a comprehensive test is conducted on the many defect types of each sample, including but not limited to uneven insulation thickness, coating defects, thermal aging instability and other items. Although this full coverage inspection can discover various potential problems, it inherently leads to significantly low inspection efficiency: on the one hand, all defect types, regardless of their frequency of occurrence, need to be checked equally, consuming a lot of testing resources and time; on the other hand, with the multi-node connection of the production process, from the stages of wire laying, annealing, painting, baking to wire winding, the slight deviation of the upstream process may affect the downstream through the propagation of defects, forming a complex time-series propagation relationship, and the traditional sampling method treats the defects of each node in isolation, and cannot capture this dynamic propagation law, resulting in poor targeting and resource redundancy. In addition, when faced with large-scale production, efficiency bottlenecks are particularly prominent. Insufficient sampling frequency may miss key defects, and simply increasing the amount of testing will increase the cost burden and be inefficient.

[0003] In the prior art, the problem of being unable to track the defect transmission chain across processes and low detection efficiency in the field of 200-level direct-weld enameled wire manufacturing needs to be solved urgently. Summary of the invention

[0004] The present application provides a production quality directional detection method and system for 200-level direct-weld enameled wire, which is used to solve the technical problems in the prior art of being unable to track the cross-process defect transmission chain and having low detection efficiency.

[0005] In view of the above problems, the present application provides a method and system for directional detection of production quality of 200-level direct-weld enameled wire.

[0006] In a first aspect, the present application provides a production quality directional detection method for a 200-level direct-weld enameled wire, the method comprising: based on a kth production control deviation matrix of a kth process node, combined with a k-1th node enameled wire defect parameter, performing a production simulation to obtain a kth node enameled wire defect parameter, k is an integer, N≥k≥1, N represents the total number of enameled wire process nodes, and the k-1th node is an upstream process node of the kth process node; Among them, when k = 1, the enameled wire defect parameter of the k-1 node is empty; When k is equal to N, integrate the enameled wire defect parameters from the first node to the Nth node to construct a defect trigger timing matrix; Perform a combination of defect spread types on the defect trigger timing matrix to generate a number of defect spread maps; Traverse the several defect spread maps for confidence evaluation to obtain several confidence levels, where the confidence level is positively correlated with the defect spread map trigger frequency; According to the several confidence levels, extract the union defect types of the defect spread maps with a confidence level greater than or equal to the confidence level threshold and send them to the quality inspection end for targeted quality inspection.

[0007] In a second aspect, the present application provides a production quality targeted detection system for 200-level directly solderable enameled wire, including: A control deviation matrix module, configured to perform production simulation based on the kth production control deviation matrix of the kth process node, combined with the enameled wire defect parameter of the k-1 node, to obtain the enameled wire defect parameter of the kth node, where k is an integer, N≥k≥1, N represents the total number of enameled wire process nodes, and the k-1 node is the upstream process node of the kth process node; Among them, when k = 1, the enameled wire defect parameter of the k-1 node is empty; When k is equal to N, integrate the enameled wire defect parameters from the first node to the Nth node to construct a defect trigger timing matrix; A defect spread map module, configured to perform a combination of defect spread types on the defect trigger timing matrix to generate a number of defect spread maps; A confidence evaluation module, configured to traverse the several defect spread maps for confidence evaluation to obtain several confidence levels, where the confidence level is positively correlated with the defect spread map trigger frequency; A defect detection summary module, configured to extract the union defect types of the defect spread maps with a confidence level greater than or equal to the confidence level threshold according to the several confidence levels and send them to the quality inspection end for targeted quality inspection.

[0008] One or more technical solutions provided in the present application have at least the following technical effects or advantages: This application proposes a production quality directional detection method and system for 200-level directly weldable enameled wires. By constructing a multi-node defect transfer time series model and a confidence-driven spread map screening mechanism, the targeting of defect detection and the full-process quality prevention and control ability are significantly improved. Compared with traditional methods, the technical solution provided by this application significantly overcomes the tracking blind spots of the cross-process defect transfer chain. By dynamically simulating the cascading impact of upstream defects on downstream processes, early identification of defect germination nodes is achieved. At the same time, based on a confidence evaluation system weighted by historical frequencies, low-probability defect combinations are accurately eliminated, focusing detection resources on high-incidence defect clusters and significantly reducing the cost of ineffective screening.

[0009] This application achieves the technical effect of covering the largest quality risk area with the minimum detection cost, tracking the cross-process defect transfer chain, and improving the detection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0011] Figure 1 It is a schematic flow chart of a production quality directional detection method for 200-level directly weldable enameled wires provided by an embodiment of this application; Figure 2 It is a schematic structural diagram of a production quality directional detection system for 200-level directly weldable enameled wires provided by an embodiment of this application.

[0012] In the drawings, the components represented by each label are described as follows: Control deviation matrix module 100, defect spread map module 200, confidence evaluation module 300, defect detection summary module 400. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0013] This application provides a production quality directional detection method and system for 200-level directly weldable enameled wires to solve the technical problems in the prior art of being unable to track the cross-process defect transfer chain and having low detection efficiency in the manufacturing field of 200-level directly weldable enameled wires.

[0014] The following will clearly and completely describe the technical solutions in the embodiments of this application with reference to the drawings in the embodiments of this application. Obviously, the described embodiments are only some embodiments of this application, not all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of this application without creative efforts belong to the scope of protection of this application.

[0015] It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units need not be limited to those steps or units clearly listed, but may include other steps or modules not clearly listed or inherent to these processes, methods, products, or devices.

[0016] Example 1, as Figure 1 shown, the present application provides a method for directional detection of the production quality of 200-level directly weldable enameled wire, wherein the method includes: S10: Based on the kth production control deviation matrix of the kth process node, combined with the enameled wire defect parameters of the (k - 1)th node, perform production simulation to obtain the enameled wire defect parameters of the kth node, where k is an integer, 1 ≤ k ≤ N, N represents the total number of enameled wire process nodes, and the (k - 1)th node is the upstream process node of the kth process node; wherein, when k = 1, the enameled wire defect parameters of the (k - 1)th node are empty; When k is equal to N, integrate the enameled wire defect parameters from the first node to the enameled wire defect parameters of the Nth node to construct a defect trigger timing matrix.

[0017] In the continuous production of enameled wire with multiple processes, in the traditional method, due to the isolation of detection data between processes, it is impossible to establish a quantitative model for the impact of upstream defects on downstream. Especially when the defect originates from the first process, there is a lack of an initial parameter benchmark, resulting in the interruption of the simulation; while the downstream nodes only rely on the control deviation analysis of this process and ignore upstream defects, such as the cascading effect of annealing oxidation caused by abnormal wire feeding tension, resulting in the missed judgment of the root cause of the defect and the breakage of the transmission path. In addition, the discrete node detection data is difficult to integrate into a global defect evolution view, hindering the pre-judgment of the whole process risk and resulting in low detection efficiency.

[0018] The step S10 in the method provided by the embodiment of the present application includes: Collect the kth production control deviation record matrix set of the kth process node and the enameled wire defect record parameters of the (k - 1)th node as input data, collect the enameled wire defect record parameters of the kth node as true value data, and construct the training data of the kth process node production simulator; wherein, the kth process node includes: Wire feeding process node, annealing process node, painting process node, baking process node, cooling process node, wire winding process node; Based on the training data of the kth process node production simulator, train several basic kth process node simulators, wherein when k = 1, the enameled wire defect record parameters of the (k - 1)th node are empty; Perform union integration on the outputs of the several k-th process node simulators to obtain the k-th process node production simulator. Based on the k-th production control deviation matrix of the k-th process node, combined with the enameled wire defect parameters of the (k - 1)-th node, perform production simulation to obtain the enameled wire defect parameters of the k-th node; When k < N, transmit the enameled wire defect parameters of the k-th node to the (k + 1)-th process node production simulator; When k = N, output the enameled wire defect parameters of the first node until the enameled wire defect parameters of the N-th node.

[0019] Exemplarily, collect the k-th production control deviation record matrix set and the enameled wire defect record parameters of the (k - 1)-th node included in the k-th process node as input data. The k-th process node includes a wire pay-off process node, an annealing process node, a painting process node, a baking process node, a cooling process node, and a wire take-up process node, and N = 6. The production control deviation record matrix set is a two-dimensional matrix of the deviation set between the production control parameters (such as temperature, speed, tension, etc.) of the k-th process node and the standard values. The rows of the matrix are the time batch sequences, such as the batch at 10:00 AM on June 5th, and the columns are the absolute values of the differences between the specific production control parameters and the standard values. For example, the paint liquid temperature deviation, the wire pay-off length deviation, etc. The data unit is the specific production control parameter, such as °C, mm, etc. Exemplarily, for the paint thickness deviation, the paint liquid temperature is collected using a measuring ruler. The production control parameter of the painting process node = |actual detected thickness value - standard value|. For example, the standard value of the paint thickness is 5 mm, and the detected value is 6 mm, and it is the batch at 10:00 AM on June 5th. Then the production control parameter of the painting process node = |6 - 5| = 1. This production control parameter is represented as (6.5.10:00 AM, 1) in the matrix.

[0020] Collect the enameled wire of the k-th node as the true value data. Exemplarily, collect whether the paint thickness does not meet the standard as the true value data of the painting process node.

[0021] Construct several k-th process node production simulators. Exemplarily, use a neural network to construct the k-th process node production simulator. The process node production simulator has a three-layer structure, namely an input layer, a hidden layer, and an output layer. The input layer has 2 nodes to input the k-th production control deviation record matrix set of the k-th process node and the enameled wire defect record of the (k - 1)-th node. The hidden layer has a three-layer structure. The first fully connected layer has 128 neurons, and the activation function uses ReLU, aiming to fuse the defect and control deviation features. The dropout rate parameter of the Dropout layer is set to 0.3 to prevent overfitting. The second fully connected layer has 64 neurons, and the activation function uses ReLU, aiming to extract high-order non-linear relationships. The output layer has 1 node to output the simulated enameled wire defect record parameters of the k-th node. Among them, when k = 1, the enameled wire defect record parameters of the (k - 1)-th node are empty.

[0022] Train a number of production simulators for the k-th process node, perform union integration on the output data of the production simulators for several process nodes. Exemplarily, train 5 production simulators for the painting process node, combine the enameled wire defect parameters of the annealing process node, perform production simulation, obtain the enameled wire defect parameters of the painting process node, and take the union of the defect record parameters output by the production simulator for the painting process node for output, such as excessive painting thickness, insufficient painting thickness, etc.

[0023] When k < N, transmit the enameled wire defect parameters of the k-th node to the production simulator of the k + 1 process node. Exemplarily, transmit the defect parameters of the annealing process node to the painting process node.

[0024] When k = N, output the enameled wire defect parameters of the first node until the enameled wire defect parameters of the N-th node. Exemplarily, when k = N = 6, output the enameled wire defect parameters of the first node until the enameled wire defect parameters of the sixth node, such as the defect parameters from the wire feeding process node to the wire winding process node. Construct a defect trigger timing matrix, where the rows of the matrix are k process nodes and the columns are defect types.

[0025] In the embodiment of the present application, by dynamically coupling the production control deviation matrix of the k-th node with the defect parameters of the k - 1 node, a closed-loop simulation chain for defect transfer between processes is constructed. When k = 1, empty parameters are automatically initialized to ensure the integrity of the starting point of the entire chain; node-by-node simulation realizes the full-process tracking of defects from germination to evolution, such as scratches at the annealing node may cause bubbles at the painting node. Finally, the timing matrix generated by integrating the data of N nodes is structured to present the defect transfer trajectory across processes, laying a core data foundation for cross-process correlation analysis.

[0026] S20: Perform combination of defect spread types on the defect trigger timing matrix to generate a number of defect spread maps; Although the defect timing matrix records the defect types of each node, it is difficult to automatically identify highly correlated cross-process combinations from a large amount of discrete data. Manually enumerating potential transfer paths, such as the situation where the annealing temperature deviation leads to uneven baking and curing, is actually inefficient, and it is easy to miss low-frequency but highly harmful hidden links such as the weak correlation between cooling microcracks and early wire feeding scratches. Existing methods cannot transform the defect spatio-temporal relationship in the matrix into a computable topological network, resulting in insufficient depth of key defect cluster mining.

[0027] Step S20 in the method provided by the embodiment of the present application includes: Extract the first node defect type set from the defect trigger timing matrix, perform i-item combination enumeration to obtain the first node defect combination set, where 1 ≤ i ≤ Q, i is an integer, and Q is the total number of first node defect types; Among them, extract the first node defect type set from the defect trigger timing matrix, perform i-item combination enumeration, and obtain the first node defect combination set, including: Perform i-item enumeration on the first node defect type set to obtain the first node i-item initial combination set; Traverse the first node i-item initial combination set, and retrieve the detection frequency ratio set of each i-item initial combination in the historical production samples of the first node; According to the detection frequency ratio set, extract the i-item initial combinations whose detection frequency ratio is greater than or equal to the detection frequency ratio threshold, and set them as the first node i-item combination set, with the initial value of i equal to 1; If i = Q, add the first node one-item combination set to the first node Q-item combination set to the first node defect combination set; If i < Q, increment i and execute the loop.

[0028] Until the Nth node defect type set is extracted from the defect trigger timing matrix, perform j-item combination enumeration to obtain the Nth node defect combination set, where 1 ≤ j ≤ Y, j is an integer, and Y is the total number of defect types of the Nth node; For the first node defect combination set to the Nth node defect type set, randomly select one combination for each node, and construct the first defect propagation graph according to the node timing connection, and add it to the several defect propagation graphs; After the defect propagation graph enumeration is completed, output the several defect propagation graphs.

[0029] In the embodiments of the present application, the first node defect type set is extracted from the defect trigger timing matrix, where i-item enumeration is performed on the first node defect type set to obtain the first node i-item initial combination set, 1 ≤ i ≤ Q, i is an integer, and Q is the total number of defect types of the first node. Exemplarily, if Q is 3, then 1 ≤ i ≤ 3. When i = 1, enumerate all single defects, such as scratches. The enumeration is implemented using the itertools.combinations function and conditional filtering is performed using filter.

[0030] Traverse the first node i-item initial combination set, and retrieve the detection frequency ratio set of each i-item initial combination in the historical production samples of the first node. For example, if a scratch appears 200 times in 1000 productions, then the frequency = number of occurrences ÷ total number of productions = 200 ÷ 1000 = 0.2.

[0031] According to the detection frequency ratio set, extract the i-item initial combinations whose detection frequency ratio is greater than or equal to the detection frequency ratio threshold, and set them as the first node i-item combination set, with the initial value of i equal to 1; the detection frequency threshold is pre-set data reflecting the occurrence frequency of the defect type. Exemplarily, the detection frequency threshold is set to 0.05, then all initial combinations with a detection frequency greater than or equal to 0.05 are extracted.

[0032] When i = Q, add the combination sets from the first node's first combination set to the first node's Qth combination set into the first node's defect combination set, obtaining a defect type set with an extraction detection frequency ratio greater than or equal to the detection frequency ratio threshold. When i < Q, increment i by one and execute the loop again until i = Q.

[0033] Repeat the enumeration process until the defect type set of the Nth node is extracted from the defect trigger timing matrix. Execute the enumeration of j combinations to obtain the defect combination set of the Nth node. Among them, 1 ≤ j ≤ Y, j is an integer, and Y is the total number of defect types of the Nth node. Traverse the initial combination set of j items of the Nth node, and retrieve the detection frequency ratio set of each initial combination of j items in the historical production samples of the Nth node. For example, if the paint is too thick and appears 100 times in 1000 productions, then the frequency = number of occurrences ÷ total number of productions = 100 ÷ 1000 = 0.1. According to the detection frequency ratio set, extract the initial combinations of j items with a detection frequency ratio greater than or equal to the detection frequency ratio threshold, and set them as the combination set of j items of the Nth node. The initial value of j is equal to 1, and the detection frequency threshold is set to 0.05 to obtain the defect combination set of the Nth node.

[0034] For the defect combination set of the first node to the defect type set of the Nth node, randomly select one combination for each node, and construct the first defect spread graph according to the node timing connection. For example, for the first node, select wire bending, for the second node, select oxidation spots, for the third node, select too thick paint, for the fourth node, select uneven curing, for the fifth node, select insufficient cooling, and for the sixth node, select too large diameter. Then, connect the six defects according to the node timing to construct the first defect spread graph. Repeat to generate directed defect spread graphs until the combination selection is exhausted, obtaining several defect spread graphs. Among them, the random selection is implemented using the random.choice function.

[0035] After the enumeration of the defect spread graphs is completed, output several defect spread graphs.

[0036] In the embodiment of the present application, based on the node defect type set of the timing matrix, an automated combination enumeration is used to generate a defect spread graph set covering the entire process chain. Each graph represents a possible defect transfer path. For example, for the second node, select oxidation spots, for the third node, select too thick paint, and for the fourth node, select uneven curing. Convert the discrete defect data into a visual directed relationship network, visually presenting the possible potential defect transfer paths. This mechanism breaks through the limitations of manual experience, systematically presents the explicit and implicit defect associations, and provides a complete path space for subsequent confidence evaluation.

[0037] S30: Traverse the several defect spread graphs for confidence evaluation to obtain several confidence levels, where the confidence level is positively correlated with the defect spread graph trigger frequency; The authenticity assessment of defect propagation maps has long relied on expert experience and lacks objective quantitative basis. High-frequency but low-hazard paths such as wire-drawing scratches may over-occupy detection resources, while low-frequency but high-risk hidden links such as annealing oxidation-induced cooling fractures are easily ignored due to insufficient samples. Existing methods cannot balance the weights of path frequency and hazard, and short-chain maps such as maps with only 2 nodes have a high probability of misjudging reliability due to data sparsity.

[0038] Step S30 in the method provided by the embodiment of the present application includes: Extract the first defect propagation map from the several defect propagation maps; Split the first defect propagation map to obtain the first two-node propagation map until the first N-node propagation map; Retrieve the proportion of the first detection frequency of the first two-node propagation map in the quality inspection log of 200-level straight-soldering enameled wire; Until retrieving the proportion of the (N-1)th detection frequency of the first N-node propagation map in the quality inspection log of 200-level straight-soldering enameled wire; Calculate the mean value of the proportion of the first detection frequency until the proportion of the (N-1)th detection frequency, set it as the confidence level of the first defect propagation map, and add it to the several defect propagation maps; Among them, calculating the mean value of the proportion of the first detection frequency until the proportion of the (N-1)th detection frequency includes: Taking the ratio of 2 to N as the weight of the proportion of the first detection frequency, until taking the ratio of N to (N-1) as the weight of the proportion of the (N-1)th detection frequency; Based on the weight of the proportion of the first detection frequency until the weight of the proportion of the (N-1)th detection frequency, calculate the mean value of the proportion of the first detection frequency until the proportion of the (N-1)th detection frequency.

[0039] In the embodiment of the present application, the first defect propagation map is randomly extracted from several propagation maps, and the random extraction is implemented by using the random.choice function.

[0040] Split the first defect propagation map to obtain the first two-node propagation map, the first three-node propagation map until the first N-node propagation map.

[0041] Retrieve the proportion of the first detection frequency of the first two-node propagation map in the quality inspection log of 200-level straight-soldering enameled wire. The proportion of the first detection frequency = the number of times the first two-node map appears in the quality inspection log ÷ the total number of quality inspection log entries. For example, if the number of times the first two-node map appears in the quality inspection log is 300 times and the total number of quality inspection log entries is 1000, then the proportion of the first detection frequency = 300 ÷ 1000 = 0.3.

[0042] Repeat the retrieval until retrieving the proportion of the (N-1)th detection frequency of the first N-node propagation map in the quality inspection log of 200-level straight-soldering enameled wire.

[0043] Take the ratio of 2 to N as the weight of the first detection frequency ratio, the ratio of 3 to N as the weight of the second detection frequency ratio, until the ratio of N - 1 to N is the weight of the (N - 1)th detection frequency ratio. Exemplarily, when N = 6, the weight of the first detection frequency ratio = 2÷6 = 0.33, the weight of the second detection frequency ratio = 3÷6 = 0.5, and the weight of the (N - 1)th detection frequency ratio = 5÷6 = 0.83.

[0044] Based on the weight of the first detection frequency ratio until the weight of the (N - 1)th detection frequency ratio, calculate the mean value of the first detection frequency ratio until the (N - 1)th detection frequency ratio. The mean value of the first detection frequency = Σ (the ratio of the first to the Nth frequency × the weight of the first to the Nth) / Σ the weight of the first to the Nth. Set the calculated first detection frequency as the first confidence level and add it to the defect propagation map. Calculate using the same calculation method to obtain the first to the (N - 1)th confidence levels.

[0045] In the embodiment of the present application, through the frequency statistics of the historical quality inspection logs, a confidence value positively correlated with the trigger frequency is given to each propagation map, replacing subjective experience with objective data. The time - series weighting mechanism, such as the increasing weight of subsequent nodes, can strengthen the evaluation rigor of long - chain maps and avoid the false high - confidence phenomenon of short - chain paths. For example, the map of "scratch → oxidation → bubble → micro - crack" containing 4 nodes obtains a higher weight due to the complete transfer chain, while the sub - path of "bubble → micro - crack" containing only 2 nodes has a lower weight, ensuring that the evaluation result truly reflects the global reliability of the defect chain.

[0046] S40: According to the several confidence levels, extract the union defect types of the defect propagation maps with confidence levels greater than or equal to the confidence level threshold and send them to the quality inspection end for targeted quality inspection; The traditional full - process detection mode rigidly allocates resources and cannot flexibly focus on the dynamically changing high - risk defect chains. The ineffective screening of low - confidence maps, such as low - frequency isolated defects, causes waste of screening costs and manpower, while high - hazard defects may not be covered and discovered due to sparse distribution. There is a lack of a closed - loop mechanism in the prior art to convert the map analysis conclusion into a detection instruction, resulting in poor connection and integration between prediction and execution and failing to achieve a good early warning effect.

[0047] In the embodiments of the present application, according to the calculated confidence level, the union defect types of the defect spread maps with a confidence level greater than or equal to the confidence threshold are extracted. The confidence threshold is a pre-set value reflecting the accuracy of the confidence level. Exemplarily, if the confidence level is set to 0.85, then the defect spread maps with a confidence level greater than or equal to 0.85 are screened, and the union defect types of all defect spread maps with a confidence level greater than or equal to 0.85 are taken and sent to the quality inspection terminal. Exemplarily, the Internet of Things is used to send the union defect types. After receiving them, the quality inspection terminal performs text display and ringtone playback to remind of targeted quality inspection.

[0048] In the embodiments of the present application, in this step, high-reliability maps are automatically screened through the confidence threshold, and their union defect types such as oxidation spots, paint film bubbles, uneven curing, etc. are extracted and pushed to the quality inspection terminal. This mechanism can achieve three breakthroughs: First, dynamically lock high-frequency and high-risk defect clusters, and accurately allocate detection resources to the maximum risk points; second, avoid repeated detection through the union operation and compress redundant monitoring; third, establish a "prediction-execution" closed loop, so that the defect analysis conclusions across processes directly drive the quality inspection behavior, and ultimately achieve the core goal of covering the maximum risk surface with the minimum detection input to complete accurate detection.

[0049] Embodiment 2, as Figure 2 shown, based on the same inventive concept as the production quality directional detection method of a 200-class directly weldable enameled wire provided in Embodiment 1, the embodiments of the present invention further provide a production quality directional detection system for a 200-class directly weldable enameled wire, including: A control deviation matrix module 100, configured to perform production simulation based on the kth production control deviation matrix of the kth process node, in combination with the defect parameters of the enameled wire at the (k - 1)th node, to obtain the defect parameters of the enameled wire at the kth node, where k is an integer, 1 ≤ k ≤ N, N represents the total number of enameled wire process nodes, and the (k - 1)th node is the upstream process node of the kth process node; Among them, when k = 1, the defect parameters of the enameled wire at the (k - 1)th node are empty; When k is equal to N, the defect parameters of the enameled wire from the first node to the Nth node are integrated to construct a defect trigger timing matrix A defect spread map module 200, configured to perform defect spread type combinations on the defect trigger timing matrix to generate a number of defect spread maps; A confidence evaluation module 300, configured to traverse the number of defect spread maps for confidence evaluation to obtain a number of confidence levels, where the confidence level is positively correlated with the defect spread map trigger frequency; A defect detection summary module 400, configured to extract the union defect types of the defect spread maps with a confidence level greater than or equal to the confidence threshold according to the number of confidence levels and send them to the quality inspection terminal to perform targeted quality inspection.

[0050] In one embodiment, the control deviation matrix module 100 is further configured to: Collect the kth production control deviation record matrix set of the kth process node and the enameled wire defect record parameters of the (k - 1)th node as input data, collect the enameled wire defect record parameters of the kth node as true value data, and construct the training data of the kth process node production simulator; Wherein, the kth process node includes a wire pay-off process node, an annealing process node, a painting process node, a baking process node, a cooling process node, and a wire take-up process node; Based on the training data of the kth process node production simulator, train a number of base kth process node simulators. When k = 1, the enameled wire defect record parameters of the (k - 1)th node are empty; Perform union integration on the outputs of the aforenamed base kth process node simulators to obtain the kth process node production simulator. Based on the kth production control deviation matrix of the kth process node and in combination with the enameled wire defect parameters of the (k - 1)th node, perform production simulation to obtain the enameled wire defect parameters of the kth node; When k < N, transmit the enameled wire defect parameters of the kth node to the (k + 1)th process node production simulator; When k = N, output the enameled wire defect parameters of the first node until the enameled wire defect parameters of the Nth node.

[0051] In one embodiment, the defect spread map module 200 is further configured to: Extract the first node defect type set from the defect trigger timing matrix, perform i-item combination enumeration to obtain the first node defect combination set, where 1 ≤ i ≤ Q, i is an integer, and Q is the total number of first node defect types; Among them, extracting the first node defect type set from the defect trigger timing matrix and performing i-item combination enumeration to obtain the first node defect combination set includes: Perform i-item enumeration on the first node defect type set to obtain the first node i-item initial combination set; Traverse the first node i-item initial combination set, and retrieve the detection frequency proportion set of each i-item initial combination in the historical production samples of the first node; According to the detection frequency proportion set, extract the i-item initial combinations with a detection frequency proportion greater than or equal to the detection frequency proportion threshold, and set them as the first node i-item combination set, with the initial value of i equal to 1; If i = Q, add the first node one-item combination set to the first node Q-item combination set to the first node defect combination set; If i < Q, increment i and execute the loop.

[0052] Until the N-node defect type set is extracted from the defect trigger timing matrix, perform j-item combination enumeration to obtain the N-node defect combination set, where 1 ≤ j ≤ Y, j is an integer, and Y is the total number of N-node defect types; For each of the first node defect combination set to the N-node defect type set, randomly select a combination for each node, and construct the first defect propagation map according to the node timing connection, and add it to the several defect propagation maps; When the enumeration of the defect propagation map is completed, output the several defect propagation maps.

[0053] In one embodiment, the confidence evaluation module 300 is further configured to: Extract the first defect propagation map from the several defect propagation maps; Split the first defect propagation map to obtain the first two-node propagation map until the first N-node propagation map; Retrieve the first detection frequency proportion of the first two-node propagation map in the quality inspection log of 200-level direct solderable enameled wire; Until retrieving the (N - 1)th detection frequency proportion of the first N-node propagation map in the quality inspection log of 200-level direct solderable enameled wire; Calculate the mean of the first detection frequency proportion until the (N - 1)th detection frequency proportion, and set it as the confidence of the first defect propagation map, and add it to the several defect propagation maps; Among them, calculating the mean of the first detection frequency proportion until the (N - 1)th detection frequency proportion includes: Taking the ratio of 2 to N as the weight of the first detection frequency proportion, until taking the ratio of N to (N - 1) as the weight of the (N - 1)th detection frequency proportion; Based on the weight of the first detection frequency proportion until the weight of the (N - 1)th detection frequency proportion, calculate the mean of the first detection frequency proportion until the (N - 1)th detection frequency proportion.

[0054] In summary, the embodiments of the present application at least have the following technical effects: The present application proposes a method and system for targeted detection of the production quality of 200-level direct solderable enameled wire. By constructing a multi-node defect transfer timing model and a confidence-driven propagation map screening mechanism, the targeting of defect detection and the full-process quality prevention and control ability are significantly improved. Compared with traditional methods, the technical solution provided by the present application significantly overcomes the tracking blind area of the cross-process defect transfer chain, realizes the early identification of defect germination nodes by dynamically simulating the cascading effect of upstream defects on downstream processes; at the same time, based on the confidence evaluation system weighted by historical frequencies, low-probability defect combinations are accurately eliminated, and the detection resources are focused on high-incidence defect clusters, greatly reducing the cost of ineffective screening.

[0055] This application achieves the technical effects of covering the largest quality risk area with the minimum detection cost, tracking the cross-process defect transfer chain, and improving the detection efficiency.

[0056] It should be noted that the above-mentioned sequence of embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above describes specific embodiments of this specification. Additionally, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0057] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application.

[0058] This specification and the drawings are only exemplary descriptions of the present application and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. A method for directional detection of the production quality of Class 200 directly weldable enameled wire, characterized in that, Including: Based on the kth production control deviation matrix of the kth process node, combined with the enameled wire defect parameters of the (k - 1)th node, perform production simulation to obtain the enameled wire defect parameters of the kth node, where k is an integer, N≥k≥1, N represents the total number of enameled wire process nodes, and the (k - 1)th node is the upstream process node of the kth process node; Among them, when k = 1, the enameled wire defect parameters of the (k - 1)th node are empty; When k is equal to N, integrate the enameled wire defect parameters from the first node to the enameled wire defect parameters of the Nth node to construct a defect trigger timing matrix; Perform defect spread type combination on the defect trigger timing matrix to generate several defect spread maps; Traverse the several defect spread maps for confidence evaluation to obtain several confidence levels, where the confidence level is positively correlated with the defect spread map trigger frequency; According to the several confidence levels, extract the union defect types of the defect spread maps with confidence levels greater than or equal to the confidence level threshold and send them to the quality inspection end for targeted quality inspection.

2. The method according to claim 1, wherein Based on the kth production control deviation matrix of the kth process node, combined with the enameled wire defect parameters of the (k - 1)th node, perform production simulation to obtain the enameled wire defect parameters of the kth node, including: Collect the kth production control deviation record matrix set of the kth process node and the enameled wire defect record parameters of the (k - 1)th node as input data, collect the enameled wire defect record parameters of the kth node as true value data, and construct the training data of the kth process node production simulator; Based on the training data of the kth process node production simulator, train several basic kth process node simulators, where when k = 1, the enameled wire defect record parameters of the (k - 1)th node are empty; Perform union integration on the outputs of the several basic kth process node simulators to obtain the kth process node production simulator, based on the kth production control deviation matrix of the kth process node, combined with the enameled wire defect parameters of the (k - 1)th node, perform production simulation to obtain the enameled wire defect parameters of the kth node; When k < N, transfer the enameled wire defect parameters of the kth node to the production simulator of the (k + 1)th process node; When k = N, output the enameled wire defect parameters from the first node to the enameled wire defect parameters of the Nth node.

3. The method according to claim 2, wherein The kth process node includes a wire feeding process node, an annealing process node, a painting process node, a baking process node, a cooling process node, and a wire winding process node.

4. The method according to claim 1, wherein Perform defect spread type combination on the defect trigger timing matrix to generate several defect spread maps, including: Extract the first node defect type set from the defect trigger timing matrix, perform i-item combination enumeration to obtain the first node defect combination set, where 1≤i≤Q, i is an integer, and Q is the total number of first node defect types; Until extracting the Nth node defect type set from the defect trigger timing matrix, perform j-item combination enumeration to obtain the Nth node defect combination set, where 1≤j≤Y, j is an integer, and Y is the total number of Nth node defect types; For the first node defect combination set to the Nth node defect type set, randomly select one combination for each node, and connect them according to the node timing to construct the first defect spread map and add it to the several defect spread maps; After the enumeration of the defect propagation maps is completed, output the several defect propagation maps.

5. The method according to claim 4, wherein Extract the first node defect type set from the defect trigger timing matrix, perform i-item combination enumeration, and obtain the first node defect combination set, including: Perform i-item enumeration on the first node defect type set to obtain the first node i-item initial combination set; Traverse the first node i-item initial combination set, and retrieve the detection frequency proportion set of each i-item initial combination in the historical production samples of the first node; According to the detection frequency proportion set, extract the i-item initial combinations whose detection frequency proportion is greater than or equal to the detection frequency proportion threshold, and set them as the first node i-item combination set, with the initial value of i equal to 1; If i = Q, add the first node one-item combination set to the first node Q-item combination set to the first node defect combination set; If i < Q, increment i and execute the loop.

6. The method according to claim 1, characterized in that Traverse the several defect propagation maps for confidence evaluation to obtain several confidence levels, including: Extract the first defect propagation map from the several defect propagation maps; Split the first defect propagation map to obtain the first two-node propagation map until the first N-node propagation map; Retrieve the first detection frequency proportion of the first two-node propagation map in the quality inspection log of 200-level straight solderable enameled wire; Until retrieving the (N - 1)th detection frequency proportion of the first N-node propagation map in the quality inspection log of 200-level straight solderable enameled wire; Calculate the mean value of the first detection frequency proportion until the (N - 1)th detection frequency proportion, and set it as the confidence level of the first defect propagation map, and add it to the several defect propagation maps.

7. The method according to claim 6, wherein Calculating the mean value of the first detection frequency proportion until the (N - 1)th detection frequency proportion includes: Taking the ratio of 2 to N as the weight of the first detection frequency proportion, until taking the ratio of (N - 1) to N as the weight of the (N - 1)th detection frequency proportion; Based on the first detection frequency proportion weight until the (N - 1)th detection frequency proportion weight, calculate the mean value of the first detection frequency proportion until the (N - 1)th detection frequency proportion.

8. A production quality directional detection system for 200-grade directly weldable enameled wire, characterized in that, For implementing the production quality directional detection method of the 200-level straight solderable enameled wire according to any one of claims 1-7, the system includes: A control deviation matrix module, configured to perform production simulation based on the kth production control deviation matrix of the kth process node, in combination with the enameled wire defect parameters of the (k - 1)th node, to obtain the enameled wire defect parameters of the kth node, where k is an integer, N ≥ k ≥ 1, N represents the total number of enameled wire process nodes, and the (k - 1)th node is the upstream process node of the kth process node; When k = 1, the enameled wire defect parameters of the (k - 1)th node are empty; When k is equal to N, integrate the enameled wire defect parameters of the first node until the enameled wire defect parameters of the Nth node to construct a defect trigger timing matrix; A defect propagation map module, configured to perform defect propagation type combination on the defect trigger timing matrix to generate several defect propagation maps; A confidence evaluation module, configured to traverse the several defect propagation maps for confidence evaluation to obtain several confidence levels, where the confidence level is positively correlated with the defect propagation map trigger frequency; A defect detection summary module is used to extract the union defect types of defect spread maps with confidence levels greater than or equal to a confidence threshold based on the several confidence levels and send them to the quality inspection end for performing targeted quality inspection.

Images