A working state abnormality early warning method and system of a wire cutting device

Abstract

The present application relates to cutting equipment monitoring technical field, especially to a kind of line cutting equipment's working state abnormal early warning method and system, method includes: obtaining current electrical data and working condition parameter;Obtain historical working condition parameter, and obtain multiple working condition clusters by clustering historical working condition parameter, current working condition parameter is located in one of working condition clusters;Obtain the latest multiple historical electrical data sections in current working condition cluster, process multiple historical electrical data sections to obtain dynamic reference template;The difference of current electrical data and dynamic reference template is calculated, and original deviation degree is obtained;Instantaneous machining stability factor is calculated;Original deviation degree is corrected using instantaneous machining stability factor, and local sensitive deviation degree is obtained;In response to local sensitive deviation degree exceeds preset early warning threshold, early warning prompt is sent.The present application realizes high-precision adaptive monitoring to sudden broken wire and gradual failure.

Description

Technical Field

[0001] This invention relates to the field of cutting equipment monitoring technology, and in particular to a method and system for early warning of abnormal working status of wire cutting equipment. Background Technology

[0002] Electrical discharge wire cutting is one of the core technologies of modern precision machining, mainly used in mold manufacturing, aerospace and other fields. Its machining process is a complex electro-mechanical-hydraulic coupling process. The machining stability directly determines the workpiece quality and efficiency, and is related to whether serious failures such as electrode wire breakage will occur.

[0003] Currently, monitoring methods for wire EDM equipment mainly suffer from the following problems: First, they have poor adaptability to dynamic operating conditions. Traditional methods often rely on setting fixed upper and lower thresholds for key parameters such as discharge voltage and peak current. However, the processing conditions of wire EDM are dynamically changing, such as variations in workpiece thickness, corner curve cutting, and high-speed straight cutting. Under different operating conditions, the normal range of the above parameters also changes. Fixed threshold methods are prone to misinterpreting normal fluctuations during operating condition switching as short circuits or abnormalities, such as speed reduction at corners, leading to a high false alarm rate.

[0004] Secondly, it is not sensitive to gradual failures. Many serious failures (such as wire breakage and surface roughening) do not occur instantaneously, but rather gradually worsen from the accumulation of early, subtle abnormal signs, such as contaminated working fluid, minor loosening of electrode wires, and poor chip removal. These early signs are not significant in global, average statistical indicators, making it difficult for existing monitoring methods to provide timely warnings before failures occur, thus missing maintenance opportunities.

[0005] In addition, existing algorithm models have limitations. Even when using machine learning algorithms such as autoencoders or isolated forests, they are mostly global detectors. They are good at finding isolated points that are significantly different from all normal data, but they are not sensitive to gradual faults that gradually shift from normal to abnormal states, and they have difficulty distinguishing between normal fluctuations caused by changes in operating conditions and abnormal shifts caused by state deterioration. Summary of the Invention

[0006] To address the problem that existing monitoring methods are insensitive to gradual faults, this invention provides a method and system for early warning of abnormal operating conditions of wire cutting equipment.

[0007] In a first aspect, the present invention provides a method for early warning of abnormal operating status of wire cutting equipment, which adopts the following technical solution: A method for early warning of abnormal operating status of wire EDM equipment includes: acquiring current electrical data and operating parameters; acquiring historical operating parameters, clustering the historical operating parameters to obtain multiple operating condition clusters, with the current operating parameters located in one of these clusters; acquiring the latest multiple historical electrical data segments within the current operating condition cluster, processing the multiple historical electrical data segments to obtain a dynamic reference template; calculating the difference between the current electrical data and the dynamic reference template to obtain the original deviation; calculating an instantaneous machining stability factor, which is negatively correlated with the Shannon entropy of the discharge state in the electrical data and positively correlated with the probability of a normal discharge state; using the instantaneous machining stability factor to correct the original deviation to obtain a local sensitive deviation; and issuing an early warning prompt in response to the local sensitive deviation exceeding a preset warning threshold.

[0008] Compared to existing methods that rely on fixed voltage and current thresholds for monitoring, this method clusters historical operating parameters and constructs dynamic reference templates. It can automatically match the most similar historical reference standard based on the current actual processing conditions, solving the problem of false alarms that are prone to occur when the fixed threshold method changes operating conditions. At the same time, it introduces an instantaneous processing stability factor based on the Shannon entropy of the discharge state to correct the original deviation. It uses the entropy value to distinguish between normal violent fluctuations and abnormal fault symptoms, thereby obtaining the local sensitive deviation, improving the detection sensitivity of gradual faults, and avoiding the early weak fault signals being masked by normal fluctuations.

[0009] Preferably, the calculation method for the dynamic reference template includes: performing a time-weighted average of multiple historical electrical data segments to obtain the dynamic reference template. The expression is: ; In the formula, Indicates a cluster of operating conditions The first in N represents the number of historical electrical data segments used for fusion; i represents the index of the historical electrical data segment. Indicates the first The time weight of each historical electrical data segment, and the time weight relative to the current moment. and the The moment of each historical electrical data segment The time difference between them is negatively correlated.

[0010] The dynamic reference template is calculated using a time-weighted average method, and the weights are negatively correlated with the time difference. This allows the reference template to adapt to the slow and natural changes in the equipment status, ensuring that the baseline always matches the current equipment health status and preventing false alarms caused by normal aging and drift of the equipment.

[0011] The preferred method for calculating the original deviation is as follows: Calculate the morphological deviation between the current electrical data and the dynamic reference template. The morphological deviation is the dynamic time warping distance between the two. Calculate the mean and standard deviation of the current electrical data to construct the first eigenvector. Calculate the mean and standard deviation of the dynamic reference template to construct the second eigenvector. Use the Euclidean distance between the first and second eigenvectors as the statistical deviation. Sum the morphological deviation and the statistical deviation by weight to obtain the original deviation.

[0012] The original deviation is calculated by combining morphological deviation and statistical deviation. Morphological deviation can capture physical process anomalies in waveform shape, while statistical deviation can capture changes in energy stability. The weighted fusion of the two can comprehensively quantify the degree of anomaly in electrical data from both microscopic waveform and macroscopic statistical levels, thus improving the robustness of feature extraction.

[0013] Preferably, the formula for calculating the instantaneous processing stability factor is: ; in, Indicates time Instantaneous processing stability factor; This indicates the probability of a normal discharge state within a set time window; Indicates time Shannon entropy of the discharge state; To prevent extremely small positive numbers with a denominator of zero.

[0014] By using Shannon entropy in the discharge state to quantify the disorder of the processing, the deviation signal is amplified when the processing is stable, making the system sensitive to minor anomalies; when the processing is unstable, the deviation signal is suppressed, which plays the role of adaptive gain control and effectively reduces the false alarm rate under harsh working conditions.

[0015] Preferably, the anomaly warning method further includes: calculating a comprehensive warning index and issuing a warning based on the comprehensive warning index; the comprehensive warning index is calculated as follows: normalizing the local sensitivity deviation to obtain an instantaneous risk index; performing exponential smoothing moving average processing on the local sensitivity deviation to obtain a cumulative risk index; calculating the standard deviation of the local sensitivity deviation within a set window to obtain a volatility risk index; and weighting and summing the instantaneous risk index, the cumulative risk index, and the volatility risk index to obtain a comprehensive warning index.

[0016] The instantaneous index is used to capture sudden strong short circuits, the cumulative index is used to capture the trend of gradual faults, and the fluctuation index is used to reflect the impact of processing stability on workpiece quality. By constructing a comprehensive early warning index that includes the instantaneous risk index, the cumulative risk index, and the fluctuation risk index, all-round monitoring of different types of faults is achieved.

[0017] Preferably, the weighting coefficients for calculating the comprehensive early warning index are determined based on the working condition clusters. The method is as follows: when the working condition cluster is high-speed straight rough machining, the weighting coefficients of the instantaneous risk index and the cumulative risk index are increased; when the working condition cluster is low-speed fine machining, the weighting coefficient of the fluctuation risk index is increased; when the working condition cluster is corner machining, the weighting coefficient of the instantaneous risk index is increased.

[0018] The weighting coefficients of the three risk indices are dynamically adjusted according to different working conditions. For example, the focus is on preventing wire breakage during rough machining and on surface quality during finishing, so that the early warning system is more in line with actual processing needs and maximizes processing efficiency and quality while ensuring processing safety.

[0019] Preferably, the operating parameters include: electrode wire feed speed, X-axis feed speed of the worktable, Y-axis feed speed of the worktable, and pulse width setting value.

[0020] Preferably, the method for obtaining multiple operating condition clusters is as follows: collect historical operating condition parameters, and use the K-Means algorithm to cluster the historical operating condition parameters to obtain multiple operating condition clusters.

[0021] By using the K-Means algorithm to cluster historical operating conditions, a large amount of historical processing data can be divided into several typical standard operating condition patterns, thereby improving the automation level of the system.

[0022] Preferably, issuing an early warning based on a comprehensive early warning index includes: setting a first early warning threshold. Second warning threshold ,and When the comprehensive early warning index When, it is judged as a normal state; when When, issue a warning at the attention level; when At that time, a danger level warning will be issued.

[0023] By setting dual thresholds, a graded early warning mechanism is implemented. Combined with the threshold determined by the ROC (Receiver Operating Characteristic) curve, it can prompt maintenance in the early stage of failure to prevent deterioration, and promptly alarm and shut down at the critical point of failure to prevent serious accidents such as wire breakage, thus optimizing the operation and maintenance strategy of the equipment.

[0024] Secondly, the present invention provides an early warning system for abnormal operating status of wire cutting equipment, which adopts the following technical solution: An abnormal working state early warning system for a wire cutting device includes a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the abnormal working state early warning method for a wire cutting device described above is implemented.

[0025] The above-mentioned method for early warning of abnormal working status of wire cutting equipment is generated into a computer program and stored in a memory so that it can be loaded and executed by a processor. Thus, a system is made based on the memory and processor for convenient use.

[0026] The present invention has the following technical effects: The K-Means clustering algorithm is used to identify the current processing condition and a time-weighted dynamic reference template is constructed to solve the false alarm problem under variable conditions. By introducing an instantaneous processing stability factor based on the Shannon entropy of the discharge state, the deviation is dynamically corrected. False alarms can be suppressed in roughing with drastic fluctuations, and weak fault signals can be amplified in stable finishing. Combined with instantaneous, cumulative, and fluctuation multi-dimensional risk indices, high-precision adaptive monitoring of sudden wire breakage and gradual faults is achieved. Attached Figure Description

[0027] Figure 1 This is a flowchart of an abnormal working status early warning method for wire cutting equipment according to the present invention.

[0028] Figure 2 This is a schematic diagram illustrating the changes in the original deviation and local sensitive deviation of the present invention over time.

[0029] Figure 3 This is a schematic diagram illustrating the monitoring based on a comprehensive early warning index according to the present invention. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] This invention discloses a method for early warning of abnormal operating status of wire cutting equipment, referring to... Figure 1 As shown, it includes the following steps: S1: Obtain electrical parameter data and operating condition parameter data.

[0032] Acquire electrical parameter data and operating condition parameter data. The electrical parameter data includes pulse voltage and pulse current. The acquisition frequency is 1Hz, which means that the data is acquired once per second.

[0033] Operating parameters include: electrode wire feed speed Unit: m / s; Table x-axis feed speed Unit: mm / min; Y-axis feed rate of the worktable Unit: mm / min; Pulse width setting value Unit: μs. Z-score standardization was performed on the operating condition parameter data to eliminate the influence of dimensions. For ease of subsequent processing, the current time... The operating condition parameter data are combined into an operating condition vector. ,Right now The vector It objectively describes the current processing conditions of the equipment.

[0034] S2: Build a dynamic reference template for operating conditions.

[0035] To identify different typical operating conditions, operating condition vectors accumulated over a recent period are retrieved from the historical database. For example, the most recent period is the past hour. Using the K-Means clustering algorithm, the historical working condition vectors are clustered to obtain multiple clusters, thereby realizing the identification of typical working conditions. In the K-Means algorithm, the K value is determined based on experience or indicators such as the profile coefficient.

[0036] For example, here is set Calculate each cluster center Synthetic feed rate modulus The expression is: ,in, For the first x-axis feed velocity in the vector of cluster centers For the first The y-axis feed velocity in the vector of each cluster center, for Sort the clusters in descending order to obtain the sorted cluster sequence.

[0037] Define the working condition type based on the sorting results: The first cluster, with the highest processing speed, is defined as high-speed straight-line roughing; the second cluster, with the second highest speed (the machine tool usually automatically slows down at corners), is defined as medium-speed corner machining; the third cluster is defined as low-speed finishing; the fourth cluster is defined as the slow start / end phase; and the fifth cluster, with extremely low speed, is defined as other states, such as standby. Through clustering, the complex continuous working space can be divided into five clusters: Cluster 1 represents high-speed straight-line roughing; Cluster 2 represents medium-speed corner machining; Cluster 3 represents low-speed finishing; Cluster 4 represents the slow start / end phase; and Cluster 5 represents other specific machining states.

[0038] During real-time monitoring, the current operating condition vector is determined. The cluster of operating conditions is closest to which the condition is classified into. , In the relevant operating condition cluster From historical data, retrieve data that is time-remote of the current moment. Recent A historical electrical parameter data segment, for example, retrieving the most recent Each historical electrical parameter data segment contains multiple consecutive data points, specifically a historical electrical data segment within 100 seconds.

[0039] In order to construct a dynamic reference template that can reflect ideal performance under recent and similar operating conditions. This is achieved by using a time-weighted average method to integrate these... This involves a historical electrical parameter data segment. The weighted average is introduced to allow the reference template to adapt to slow changes in equipment status, such as the gradual wear of electrode wires. Historical data more recent to the current moment should have higher reference value; this is the dynamic reference template. The expression is: ; ; in, Indicates the current moment Operating Condition Cluster The dynamic reference template obtained from the calculation is, for example, an ideal voltage and current waveform with a length of 100ms; This indicates the number of historical electrical parameter data segments used for fusion, in this embodiment. ; This represents the i-th historical electrical parameter data segment in the operating condition cluster k; Indicates the first Time weighting of historical electrical parameter data segments; The timestamp represents the current moment; Indicates the first Timestamps of historical electrical parameter data segments ; This represents the time decay constant, used to control the decay rate of historical data weights. The larger the value, the more persistent the impact of historical data; The smaller the value, the more sensitive the template is to recent data changes. The value is set manually based on the actual situation; for example... It can be set to 30 minutes, and exp represents an exponential function with base e.

[0040] By using time weighting, more recent historical electrical parameter data segments are assigned higher weights, while more distant historical electrical parameter data segments are assigned lower weights. The final calculated... It is a smooth, dynamically updated benchmark, meaning the dynamic reference template changes over time.

[0041] S3: Calculate the original deviation.

[0042] To quantify the difference between the currently acquired actual electrical parameters and the dynamic reference template, it is necessary to calculate the raw deviation, which includes morphological deviation and statistical deviation. For the electrical parameter data at the current time t, a time window is constructed along the historical direction, with a length of [missing information]. Obtain the electrical waveform at the current time t. For example, the pulse voltage waveform at the current time t or the pulse current waveform at the current time t.

[0043] Morphological deviation Used to measure electrical waveforms With template waveform Similarity in shape. In electrical discharge machining (EDM), the waveform morphology contains rich information about the physical process. Therefore, a dynamic time warping algorithm is used to calculate the morphological deviation. Dynamic time warping is an algorithm that can calculate the similarity between two time series by non-linearly warping the time axis to find the optimal alignment path between the two series. That is and The minimum DTW (Dynamic Time Warping) distance between them. The larger this value, the greater the difference in morphology between the electrical waveform at the current time t and the dynamic reference template. It should be noted that the morphological deviation... Electrical waveform Medium pulse voltage sequence and template waveform The DTW (Dynamic Time Warping) distance of the pulse voltage, or the electrical waveform. Medium pulse current sequence and template waveform The DTW (Dynamic Time Warping) distance of the pulse current.

[0044] Statistical deviation Used to measure electrical waveforms With template waveform Differences in statistical characteristics. Calculation of electrical waveforms. The mean and standard deviation are calculated, and the first eigenvector is constructed. The template waveform is then calculated. The mean and standard deviation of the two eigenvectors are calculated, and a second eigenvector is constructed. The Euclidean distance between the two eigenvectors is used as the statistical deviation. The larger this value, the more it indicates that the energy stability of the actual processing deviates from the ideal state.

[0045] The method for calculating the original deviation is as follows:

[0046] in, Indicates the current time The original deviation, Indicates the deviation of the electrical waveform from the template waveform in terms of shape; Indicates the statistical deviation between the electrical waveform and the template waveform; This represents the preset weighting coefficients used to balance the importance of the morphological and statistical dimensions. For example, based on experience, both can be set to be equally important, i.e. , .

[0047] The differences between morphological deviation and statistical deviation are integrated into a unified original deviation index. .when A larger value indicates that the current operating condition deviates from the ideal state. This initially suggests that there may be an abnormality in the working condition of the cutting equipment.

[0048] S4: Processing stability assessment and deviation correction.

[0049] The calculated raw deviation cannot distinguish between normal deviations caused by changes in operating conditions and abnormal deviations caused by deterioration in condition. For example, during corner machining, the raw deviation becomes very large due to drastic fluctuations in machining conditions, leading to false alarms; while during stable finishing, the raw deviation of a small, gradual early sign of a fault may be significantly reduced. It may be overwhelmed, leading to missed reports.

[0050] Therefore, this step introduces an instantaneous processing stability factor to dynamically correct the original deviation. First, through analysis... and Waveform, real-time determination of the state of each discharge pulse The method involves setting voltage and current thresholds to categorize each pulse into three states: {normal discharge, short circuit, open circuit}. The discharge state sequence within the time window is then acquired. And calculate the probability of each of the three states occurring within the time window: probability of normal discharge state. Short circuit probability and open-circuit state probability .

[0051] Calculate the Shannon entropy of the discharge state:

[0052] in, Indicates time Shannon entropy of the discharge state at time, This represents the probability of the i-th state. Represents the logarithm to base 2, when When the value is 0, It is not included in the calculation.

[0053] When the process is very stable, for example, when it is in a completely normal discharge state ( ),at this time A smaller entropy value indicates a more stable state; when the process is highly unstable, such as when the three states frequently alternate and are evenly distributed, , , The value is approximately At this time, Shannon entropy A larger value indicates a more chaotic state.

[0054] Calculate the instantaneous processing stability factor based on Shannon entropy:

[0055] in, The instantaneous processing stability factor at time t; Indicates time Shannon entropy of the discharge state; To represent a very small positive number, This is used to prevent the denominator from being zero. This represents the probability of a normal discharge state within a set time window.

[0056] When the processing is stable The value approaches 1, the Shannon entropy approaches 0.01, and the instantaneous processing stability factor approaches 100; when a fault occurs, such as a continuous short circuit, When the value is 0, the instantaneous machining stability factor is equal to 0; when machining is unstable... The value is approximately 0.33, the Shannon entropy is approximately 1.58, and the instantaneous processing stability factor approaches 0.2.

[0057] To amplify minor anomalies in the stable region and suppress normal disturbances in the fluctuating region, an instantaneous processing stability factor is used to adjust the original deviation. After correction, the local sensitivity deviation is obtained, expressed as:

[0058] in, This represents the local sensitivity deviation at time t; Indicates time The original deviation; The instantaneous processing stability factor at time t.

[0059] Combination Figure 2 As shown, when the machining is in a steady state, the instantaneous machining stability factor The deviation is relatively large. If a minor equipment malfunction occurs at this time, such as contamination of the working fluid causing a smaller initial deviation value, the initial deviation... It will also be amplified, making it more sensitive to gradual faults in a steady state.

[0060] When machining is in a normal unstable state, such as corner machining, the instantaneous machining stability factor is... The value is low, even if the original deviation is low. Because of these dramatic fluctuations, the local sensitivity deviation becomes very large, but it will eventually be suppressed by the transient processing stability factor, thus reducing the local sensitivity deviation. It remains at a relatively low level.

[0061] S5: Construct multi-dimensional risk indices and dynamically integrate them.

[0062] S51: Calculate the instantaneous risk index.

[0063] Local sensitivity deviation Min-Max normalization is performed to obtain the instantaneous risk index, which is used to capture sudden and severe abnormal events, such as instantaneous strong short circuits.

[0064] S52: Calculate the cumulative risk index.

[0065] In order to quantify In terms of recent cumulative trends, the cumulative risk index is calculated using the Exponential Moving Average (EMA). EMA is a weighted average algorithm that assigns higher weight to recent data.

[0066]

[0067] in, This represents the cumulative risk index at time t. This represents the cumulative risk index at time t-1. This represents the local sensitivity deviation at time t. Represents the smoothing coefficient ( ), The closer the curve is to 1, the stronger the dependence on historical data, and the smoother the curve. The closer to 0, the faster the response to the current data. For example, This is to focus on a medium- to long-term cumulative effect. If A sustained increase indicates that the equipment is in a process of continuous deterioration, such as the accumulation of gradual failures.

[0068] S53: Calculate the volatility risk index.

[0069] By calculating the local sensitivity deviation At the most recent H time points (e.g. The standard deviation over a 50-second period is used as the volatility risk index at time t. This index is particularly important for finishing processes. When the local sensitivity deviation fluctuates greatly, even if the mean value is not high, it may lead to a decrease in the surface quality of the workpiece.

[0070] S54: Calculate the comprehensive early warning index.

[0071] The three risk indices are dynamically weighted and fused to obtain the final comprehensive early warning index. A linear normalization algorithm is then used to normalize the instantaneous risk index, cumulative risk index, and volatility risk index respectively. The expression for the comprehensive early warning index is as follows:

[0072] in, The comprehensive early warning index at time t is represented. This represents the weighting coefficient of the preset instantaneous risk index. This represents the weighting coefficient of the preset cumulative risk index. This represents the weighting coefficient of the preset volatility risk index. The instantaneous risk index represents the risk level at time t. This represents the cumulative risk index at time t. The volatility risk index represents the volatility risk index at time t.

[0073] The weighting coefficients are set manually based on the actual situation. For example, when the working condition is high-speed linear roughing, the biggest concerns are efficiency and the risk of wire breakage. Therefore, the instantaneous risk of short circuits and the cumulative risk of wire breakage are important considerations. .

[0074] When the working condition is low-speed precision finishing, the biggest concern is affecting the surface quality of the workpiece. Therefore, fluctuations in the machining process are a critical concern. .

[0075] When machining at corners: Under these conditions, the cutting equipment itself is unstable, so the primary concern is whether a short circuit will occur. Therefore, the instantaneous risk has the highest weighting. .

[0076] The higher the value of the comprehensive early warning index, the greater the risk that the equipment is in an abnormal state.

[0077] S6: Use a comprehensive early warning index to monitor the operating status of wire cutting equipment.

[0078] By analyzing historical fault data, a first warning threshold is set. Second warning threshold This is to determine the working status of the cutting equipment.

[0079] For example, in combination Figure 3 As shown, through ROC (Receiver Operating Characteristic) curve analysis, the following parameters were set: , ,when When, it is determined to be in normal processing condition; when When this indicates that early abnormal signs have been detected or that the cumulative risk is rising, a warning at the attention level is issued; when When a serious fault is detected as imminent or has already occurred, such as an imminent wire breakage, a danger warning is issued.

[0080] This invention also discloses an abnormal working state early warning system for wire cutting equipment, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a method for early warning of abnormal working state of wire cutting equipment according to the present invention is implemented.

[0081] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0082] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for early warning of abnormal operating status of wire cutting equipment, characterized in that, The anomaly warning method includes: acquiring current electrical data and operating parameters; acquiring historical operating parameters, clustering the historical operating parameters to obtain multiple operating condition clusters, with the current operating parameters located in one of these clusters; acquiring the latest historical electrical data segments within the current operating condition cluster, processing these segments to obtain a dynamic reference template; calculating the difference between the current electrical data and the dynamic reference template to obtain the original deviation; calculating the instantaneous processing stability factor, which is negatively correlated with the Shannon entropy of the discharge state in the electrical data and positively correlated with the probability of a normal discharge state; using the instantaneous processing stability factor to correct the original deviation to obtain the local sensitive deviation; and issuing a warning when the local sensitive deviation exceeds a preset warning threshold.

2. The method for early warning of abnormal working status of wire cutting equipment according to claim 1, characterized in that, The calculation method for the dynamic reference template includes: performing a time-weighted average of multiple historical electrical data segments to obtain the dynamic reference template. The expression is: ； In the formula, Indicates a cluster of operating conditions The first in N represents the number of historical electrical data segments used for fusion; i represents the index of the historical electrical data segment. Indicates the first The time weight of each historical electrical data segment, and the time weight relative to the current moment. and the The moment of each historical electrical data segment The time difference between them is negatively correlated.

3. The method for early warning of abnormal working status of wire cutting equipment according to claim 1, characterized in that, The method for calculating the original deviation is as follows: Calculate the morphological deviation between the current electrical data and the dynamic reference template. The morphological deviation is the dynamic time warping distance between the two. Calculate the mean and standard deviation of the current electrical data to construct the first eigenvector. Calculate the mean and standard deviation of the dynamic reference template to construct the second eigenvector. Use the Euclidean distance between the first and second eigenvectors as the statistical deviation. Sum the morphological deviation and the statistical deviation by weight to obtain the original deviation.

4. The method for early warning of abnormal working status of wire cutting equipment according to claim 1, characterized in that, The formula for calculating the instantaneous processing stability factor is: ； in, Indicates time Instantaneous processing stability factor; This indicates the probability of a normal discharge state within a set time window; Indicates time Shannon entropy of the discharge state; To prevent extremely small positive numbers with a denominator of zero.

5. The method for early warning of abnormal working status of wire cutting equipment according to claim 1, characterized in that, The abnormal early warning method also includes: calculating a comprehensive early warning index and issuing an early warning based on the comprehensive early warning index; the calculation method of the comprehensive early warning index is as follows: normalizing the local sensitivity deviation to obtain the instantaneous risk index; performing exponential smoothing moving average processing on the local sensitivity deviation to obtain the cumulative risk index; calculating the standard deviation of the local sensitivity deviation within the set window to obtain the volatility risk index; and weighting and summing the instantaneous risk index, the cumulative risk index, and the volatility risk index to obtain the comprehensive early warning index.

6. The method for early warning of abnormal working status of wire cutting equipment according to claim 5, characterized in that, The weighting coefficients for calculating the comprehensive early warning index are determined based on the working condition clusters. The method is as follows: when the working condition cluster is high-speed straight rough machining, the weighting coefficients of the instantaneous risk index and the cumulative risk index are increased; when the working condition cluster is low-speed fine machining, the weighting coefficient of the fluctuation risk index is increased; when the working condition cluster is corner machining, the weighting coefficient of the instantaneous risk index is increased.

7. The method for early warning of abnormal working status of wire cutting equipment according to claim 1, characterized in that, Operating parameters include: electrode wire feed speed, X-axis feed speed of the worktable, Y-axis feed speed of the worktable, and pulse width setting.

8. The method for early warning of abnormal working status of wire cutting equipment according to claim 1, characterized in that, The method for obtaining multiple operating condition clusters is as follows: collect historical operating condition parameters, use the K-Means algorithm to cluster the historical operating condition parameters, and obtain multiple operating condition clusters.

9. A method for early warning of abnormal operating status of a wire cutting equipment according to claim 5, characterized in that, Warnings are issued based on a comprehensive early warning index, including: setting a first early warning threshold. Second warning threshold ,and When the comprehensive early warning index When, it is judged as a normal state; when When, issue a warning at the attention level; when At that time, a danger level warning will be issued.

10. An abnormal working status early warning system for wire cutting equipment, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement a method for early warning of abnormal operating status of a wire cutting device as described in any one of claims 1-9.

Images