Statistical analysis method for production energy consumption data of automobile parts

Abstract

The invention discloses an automobile part production energy consumption data statistical analysis method, which comprises the following steps: carrying out specific value extraction on a welding energy consumption peak value according to a corrected effective area of a stirring area and material extrusion quality, generating unit mass connection energy consumption, and constructing a standardized energy consumption acquisition record; according to a change curve of shaft shoulder pressure and unit mass connection energy consumption in the standardized energy consumption acquisition record, identifying the lowest point position of energy consumption, and extracting a starting boundary of an optimal matching interval; starting from the starting boundary of the optimal matching interval, acquiring continuous shaft shoulder pressure acquisition records, grouping the continuous shaft shoulder pressure acquisition records into a high-resistance cluster and a low-resistance cluster, and extracting the center of each cluster as a reference threshold value of the plastic flow resistance of the material; and performing difference value evaluation on the current shaft shoulder pressure and the cluster center, identifying an adjustment demand when the difference value exceeds a reference threshold value, and generating a downward adjustment amount of the shaft shoulder pressure and an adjusted pressure control amount.

Description

Technical Field

[0001] This invention relates to the field of next-generation information technology, and in particular to a method for statistical analysis of energy consumption data in automotive parts production. Background Technology

[0002] In the automotive parts manufacturing sector, the statistics and analysis of energy consumption data are crucial for improving manufacturing efficiency and reducing costs. Particularly in welding processes, energy consumption directly impacts the economics and environmental friendliness of the production line, making it a key focus for the industry. However, accurately assessing and optimizing energy consumption under complex process conditions remains a critical challenge that urgently needs to be addressed. Currently, while many methods attempt to estimate energy consumption by monitoring equipment operating parameters, these approaches often overlook the dynamic changes in material states during the process, leading to significant discrepancies between the analysis results and actual production conditions. Especially in scenarios involving the interaction of multiple physical factors, relying solely on fixed parameters or static standards makes it difficult to capture the core variables affecting energy consumption, thus failing to provide an effective basis for optimization control. A deeper technical challenge lies in the highly complex impact of shoulder pressure, a key factor in the welding process, on energy consumption. Shoulder pressure not only determines the amount of energy input but also directly affects material flow behavior, influencing the effective range of the welding area. Improper pressure settings can lead to excessive material extrusion, reducing the actual part involved in the connection and significantly increasing energy consumption per unit of material. The matching problem between this pressure and material flow resistance has become a core obstacle to energy consumption analysis and process optimization. Specifically, in the actual operation of friction stir welding, because this process uses a solid-state joining method, the frictional heat generated by the rotation of the stirring head softens the material and achieves plastic flow, forming a continuous weld rather than traditional discrete weld points. If the shoulder pressure exceeds the material's tolerance range, material in the weld area will be squeezed to the outside, resulting in waste, while the quality of the actually joined part deteriorates due to insufficient material, thus reducing energy utilization efficiency. Therefore, accurately identifying the optimal balance point between shoulder pressure and material flow resistance during the dynamically changing welding process, and optimizing energy consumption statistics and process control based on this, has become a critical problem that urgently needs to be solved. Summary of the Invention

[0003] In view of the shortcomings of the existing technology, the purpose of this invention is to provide a method for statistical analysis of energy consumption data in automotive parts production, which solves the above-mentioned problems in the existing technology.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] Real-time acquisition of shoulder pressure, material extrusion quality, and peak welding energy consumption during the welding process of automotive parts; combined with the captured effective stirring zone boundary contour image, the stirring zone area is extracted; and the lower limit of the material yield threshold range is obtained from the automotive parts material process library.

[0006] The area of ​​the mixing zone is compared with the lower limit of the material yield threshold range to identify the area deviation when the area of ​​the mixing zone is lower than the lower limit of the material yield threshold range. The area of ​​the mixing zone is then compensated based on the area deviation to obtain the corrected effective area of ​​the mixing zone.

[0007] Based on the corrected effective area of ​​the mixing zone and the material extrusion mass, the ratio of the peak welding energy consumption is extracted to generate the unit mass connection energy consumption and construct a standardized energy consumption collection record.

[0008] Based on the variation curve of shoulder pressure and unit mass connection energy consumption in the standardized energy consumption collection records, the location of the lowest energy consumption point is identified, and the starting boundary of the best matching interval is extracted.

[0009] Starting from the initial boundary of the optimal matching interval, continuous shoulder pressure acquisition records are obtained, grouped into high resistance clusters and low resistance clusters, and the center of each cluster is extracted as a reference threshold for the plastic flow resistance of the material.

[0010] The difference between the current shoulder pressure and the cluster center is evaluated to identify adjustment needs where the difference exceeds the reference threshold, and the downward adjustment amount of the shoulder pressure and the adjusted pressure control amount are generated.

[0011] Furthermore, the real-time acquisition of shoulder pressure, material extrusion quality, and welding energy consumption peaks during the automotive parts welding process, combined with the captured effective stirring zone boundary contour image, extracts the stirring zone area. Simultaneously, it obtains the lower limit of the material yield threshold range from the automotive parts material process library, including:

[0012] Continuous pressure monitoring is performed on the shoulder of the friction stir welding equipment according to a preset sampling frequency. The shoulder pressure data and corresponding timestamps are recorded. At the same time, the material extrusion quality data is acquired. By reading the instantaneous power value of the welding equipment power meter, the maximum value in the power value sequence is identified as the peak welding energy consumption. The boundary contour image of the stirring zone is acquired. The boundary contour image is processed by the grayscale threshold segmentation method to identify the boundary line. The area value of the stirring zone is calculated based on the closed area enclosed by the boundary line and matched with the timestamp of the shoulder pressure data. The current welding material grade information is queried from the automotive parts material process library, the corresponding material yield strength data table is extracted, and the lower limit of the material yield threshold range is obtained.

[0013] Furthermore, the step of comparing the area of ​​the mixing zone with the lower limit of the material's yield threshold range, identifying the area deviation when the area of ​​the mixing zone is lower than the lower limit of the material's yield threshold range, and compensating the area of ​​the mixing zone based on the area deviation to obtain the corrected effective area of ​​the mixing zone includes:

[0014] Calculate the difference between the area of ​​the mixing zone and the lower limit of the material yield threshold range. If the area of ​​the mixing zone is less than the lower limit of the material yield threshold range, the area deviation is determined as the difference. Divide the area deviation by the lower limit of the material yield threshold range to obtain the deviation ratio. Based on the deviation ratio, find the compensation coefficient in the compensation mapping relationship. Use the compensation coefficient to perform a product operation on the original area of ​​the mixing zone to obtain the corrected effective area of ​​the mixing zone.

[0015] Furthermore, the ratio of the peak welding energy consumption to the effective area of ​​the corrected mixing zone and the material extrusion mass is extracted to generate the unit mass connection energy consumption, and a standardized energy consumption collection record is constructed, including:

[0016] The material flow density is determined by the ratio of the material extrusion mass to the corrected effective area of ​​the mixing zone; the area energy consumption density is obtained by dividing the peak welding energy consumption by the corrected effective area of ​​the mixing zone, and the area energy consumption density is corrected according to the material flow density to obtain the unit mass connection energy consumption; the unit mass connection energy consumption is combined with the shoulder pressure data, material grade information and timestamp organization data to form a standardized energy consumption collection record.

[0017] Furthermore, the step of identifying the location of the lowest energy consumption point and extracting the starting boundary of the optimal matching interval based on the variation curve of shoulder pressure and unit mass connection energy consumption in the standardized energy consumption collection records includes:

[0018] The shoulder pressure sequence and the corresponding unit mass connection energy consumption sequence are extracted from the standardized energy consumption collection records. The data is smoothed using the moving average method to obtain the smoothed pressure-energy consumption sequence, and a pressure-energy consumption change curve is constructed. The difference between adjacent data points on the change curve is calculated to obtain the rate of change of each segment of the curve. When a turning point where the rate of change changes from negative to positive is detected, the point is determined as a candidate position for the lowest energy consumption point. The position with the minimum energy consumption is selected as the lowest energy consumption point. With the lowest energy consumption point as the center, a search is conducted in the direction of decreasing pressure. When the energy consumption value increases relative to the lowest point by more than a preset threshold, the position is determined as the starting boundary of the best matching interval, and the shoulder pressure value corresponding to the starting boundary is recorded.

[0019] Furthermore, starting from the initial boundary of the optimal matching interval, continuous shoulder pressure acquisition records are obtained, grouped into high-resistance clusters and low-resistance clusters, and the center of each cluster is extracted as a reference threshold for the plastic flow resistance of the material, including:

[0020] Starting from the initial boundary position of the optimal matching interval, subsequent shoulder pressure acquisition records are read in chronological order to obtain a continuous pressure data sequence within a preset time period. The pressure data sequence is then clustered using the K-means clustering algorithm to divide the data into two groups. Initial cluster centers are selected based on the pressure values. The distance from each data point to the cluster center is iteratively calculated and the center position is updated until convergence, forming a high-resistance data group and a low-resistance data group. The mean of the high-resistance data group is calculated as the high-resistance cluster center value, and the mean of the low-resistance data group is calculated as the low-resistance cluster center value. The high-resistance cluster center value is determined as the high threshold of the material's plastic flow resistance, and the low-resistance cluster center value is determined as the low threshold of the material's plastic flow resistance.

[0021] Furthermore, the step of evaluating the difference between the current shoulder pressure and the cluster center, identifying adjustment needs where the difference exceeds a reference threshold, and generating a downward adjustment amount for the shoulder pressure and an adjusted pressure control amount includes:

[0022] Calculate the difference between the current shoulder pressure value and the high and low thresholds of the material's plastic flow resistance. If the current pressure value exceeds the high threshold, a pressure adjustment requirement is identified. Calculate the deviation between the current pressure value and the low threshold. Find the adjustment coefficient corresponding to the deviation based on the adjustment mapping relationship. Multiply the deviation by the adjustment coefficient to obtain the downward adjustment amount of the shoulder pressure. Subtract this downward adjustment amount from the current pressure value to obtain the adjusted pressure control amount.

[0023] Furthermore, the method also includes: sending the adjusted pressure control value to the friction stir welding equipment to perform pressure adjustment, re-acquiring the boundary contour image of the stirring zone, and extracting the updated stirring zone area for the next round of automotive parts production energy consumption statistics. Further, sending the adjusted pressure control value to the friction stir welding equipment to perform pressure adjustment, re-acquiring the boundary contour image of the stirring zone, and extracting the updated stirring zone area for the next round of automotive parts production energy consumption statistics includes:

[0024] The adjusted pressure control value is sent to the control port of the friction stir welding equipment. The equipment performs a pressure adjustment operation and obtains a pressure adjustment completion signal. Based on the pressure adjustment completion signal, the welding area is re-photographed to obtain the boundary contour image of the stirring zone. The updated stirring zone area value is calculated by grayscale threshold segmentation and boundary extraction methods. The updated stirring zone area value is combined with the adjusted pressure control value, the real-time collected material extrusion quality and welding energy consumption value to construct a new round of standardized energy consumption collection records.

[0025] Compared to existing technologies, this invention offers the following advantages: a method for statistical analysis of energy consumption data in automotive parts production. During welding, insufficient stirring zone area often leads to inadequate plastic flow of materials, resulting in excessively high energy consumption peaks and unstable connection quality. This invention collects shoulder pressure, material extrusion mass, and welding energy consumption peaks in real time, extracts the stirring zone area using boundary imaging, compares it with the lower limit of the material yield threshold, calculates the area deviation, and compensates to obtain the corrected effective stirring zone area. Based on this, it generates the unit mass connection energy consumption and constructs a standardized energy consumption record. By analyzing the lowest energy consumption point to determine the optimal matching interval, clustering to extract the material plastic flow resistance reference threshold, and real-time evaluating the current pressure difference to generate a downward adjustment amount, driving the equipment to perform pressure regulation, and using the updated stirring zone area to optimize the next round of production in a closed loop, this significantly reduces welding energy consumption, ensures sufficient and stable material flow, and improves the welding quality and production efficiency of automotive parts. Attached Figure Description

[0026] Figure 1 This is a flowchart of the method for statistical analysis of energy consumption data in automotive parts production in this embodiment of the solution;

[0027] Figure 2 This is a schematic diagram of the statistical analysis method for energy consumption data in automotive parts production in this embodiment of the solution;

[0028] Figure 3 This is another schematic diagram of the statistical analysis method for energy consumption data in automotive parts production in this embodiment of the solution. Detailed Implementation

[0029] The present invention will now be described in detail through specific embodiments:

[0030] like Figures 1-3 This embodiment of a method for statistical analysis of energy consumption data in automotive parts production may specifically include:

[0031] S101. Real-time acquisition of shoulder pressure, material extrusion quality, and welding energy consumption peaks during the welding process of automotive parts. Combined with the captured effective stirring zone boundary contour image, the stirring zone area is extracted. At the same time, the lower limit of the material yield threshold range is obtained from the automotive parts material process library.

[0032] The pressure sensing unit continuously monitors the shoulder of the friction stir welding equipment at a preset sampling frequency. When the shoulder pressure value exceeds a preset pressure threshold, it records the shoulder pressure data and corresponding timestamp at that moment. Simultaneously, it acquires the material extrusion mass data from the collection tank outside the welding area. By reading the instantaneous power value from the welding equipment's power meter, the maximum value in the power value sequence is identified as the peak welding energy consumption. The boundary imaging unit uses an industrial camera to vertically photograph the welding area, acquiring the boundary contour image of the stirring zone. The boundary contour image is processed using a grayscale threshold segmentation method to identify the boundary line formed by the plastic flow of the material. The area value of the stirring zone is calculated based on the closed area enclosed by the boundary line, and the stirring zone area value is matched with the timestamp of the shoulder pressure data. The unit queries the current welding material grade information from the automotive parts material process library, extracts the corresponding material yield strength data table based on the material grade, obtains the material yield threshold range under welding temperature conditions, and reads the lower limit value of the material yield threshold range. The effective area value of the mixing zone is determined by comparing the area value of the mixing zone with the lower limit of the material yield threshold range. The shoulder pressure data, material extrusion quality, welding energy consumption peak value and effective area value are summarized to form a basic dataset for energy consumption statistics in the welding process of automotive parts.

[0033] In one implementation, energy consumption data acquisition during the welding process of automotive parts is achieved through the collaborative operation of multiple sensors. Three key components—a pressure sensing unit, a boundary imaging unit, and a material processing library—work together to complete the entire process from data acquisition to effective area extraction. The pressure sensing unit employs a high-precision piezoelectric sensor, mounted on the shoulder support structure of the friction stir welding equipment, to monitor pressure changes in real time during the welding process. The boundary imaging unit captures visual information of the welding area using an industrial camera, providing image data for subsequent area calculations.

[0034] Specifically, the sampling frequency of the pressure sensing unit is set to 1000Hz to ensure the capture of transient pressure changes during the welding process. Once welding begins, the sensor continuously records the pressure applied by the shoulder to the workpiece surface, and these values ​​are transmitted to the control system via a data acquisition card. The preset pressure threshold is determined based on the material type and plate thickness. For example, for welding aluminum alloy automotive door panels, the threshold is set to 0.8 times the material's yield strength σy, i.e., 200MPa, as the lower limit of the yield threshold range. When the detected pressure value exceeds the preset threshold, the system automatically records the pressure data and precise timestamp at that moment. Simultaneously, the power meter monitors the input power of the welding equipment in real time through a current transformer and a voltage sensor, with a sampling interval of 10 milliseconds. A sliding window algorithm with a window size of 5 and a step size of 1 is used to identify local maxima in the power sequence. The algorithm input is the power sequence, and the process involves point-by-point sliding calculation. When the power values ​​of five consecutive sampling points show a trend of first increasing and then decreasing, the intermediate point is identified as the peak welding energy consumption and output. The material extrusion quality is collected through an annular collection trough around the welding area. A high-precision weighing sensor is installed at the bottom of the collection tank, which can monitor the cumulative mass of the extruded material in real time.

[0035] For example, the industrial camera of the boundary imaging unit employs a 5-megapixel CMOS sensor and is equipped with a telecentric lens to eliminate perspective distortion. The camera is vertically mounted above the welding area and maintains timing consistency with pressure acquisition via a synchronous trigger signal. The acquired raw image is first denoised using median filtering to eliminate salt-and-pepper noise. A grayscale thresholding method adaptively determines the segmentation threshold based on Otsu's algorithm, converting the image into a binary image. The plastic flow region of the material appears as a highlighted area in the image. The contour of the largest connected region is extracted through connected component analysis, and the area of ​​the closed contour is calculated using Green's formula.

[0036] In one possible implementation, the calculation of the stirring zone area needs to consider the conversion from pixel coordinates to actual physical dimensions. Using a pre-calibrated pixel equivalent coefficient, the pixel area is converted to the actual physical area, in square millimeters. The automotive parts material process library is built using a relational database management system, containing complete mechanical property parameters for various automotive metal materials. The material grade serves as the primary key, associated with parameters such as yield strength, tensile strength, and elongation at different temperatures. During a query, the corresponding yield strength data table is extracted using SQL statements based on the grade of the current welding material. Considering the temperature effect during welding, the system calculates the material yield threshold range at the current temperature using linear interpolation based on the real-time temperature obtained from a temperature sensor, and reads the lower limit of this range as a reference benchmark for subsequent processing.

[0037] Preferably, the timestamp matching uses millisecond-level precision to ensure the time sequence correspondence of data from different sensors. All sensor data are synchronized using a unified clock source to eliminate latency differences between different acquisition channels.

[0038] In one embodiment, the effective area value is determined through a comparison algorithm. The calculated stirring zone area is compared with the lower limit of the material yield threshold range. If the stirring zone area is less than 90% of the lower limit, it is determined that there is insufficient material flow, and the area value needs to be compensated and corrected. The compensation coefficient is dynamically determined based on pressure data and material properties to ensure that the final effective area value truly reflects the actual connection status of the welded area.

[0039] For example, the construction of the basic dataset organizes the shoulder pressure data, material extrusion quality, welding energy consumption peak and effective area value according to the time series to form a multi-dimensional structured data record.

[0040] S102. Compare the area of ​​the mixing zone with the lower limit of the material yield threshold range, identify the area deviation when the area of ​​the mixing zone is lower than the lower limit of the material yield threshold range, compensate the area of ​​the mixing zone according to the area deviation, and obtain the corrected effective area of ​​the mixing zone.

[0041] The difference between the area of ​​the stirring zone and the lower limit of the material yield threshold range is calculated. If the stirring zone area is less than the lower limit of the material yield threshold range, the area deviation is determined as the difference between the lower limit of the threshold range and the stirring zone area. The area deviation is divided by the lower limit of the threshold range to obtain the deviation ratio. Based on the deviation ratio, the corresponding compensation coefficient is found in a pre-established compensation mapping relationship. The compensation mapping relationship is established using historical welding data and records the correspondence between the deviation ratio and the compensation coefficient. When the deviation ratio is less than a preset threshold, the compensation coefficient takes a baseline value; when the deviation ratio exceeds the preset threshold, the compensation coefficient increases linearly. The original stirring zone area is corrected using the compensation coefficient. The corrected effective stirring zone area is obtained by multiplying the original stirring zone area by the compensation coefficient. The obtained effective area value can be used for subsequent statistical calculations of welding energy consumption.

[0042] In one implementation, the stirring zone area compensation mechanism is achieved by establishing a mapping relationship between the deviation and the compensation coefficient. When the actual stirring zone area during the welding process fails to meet the material yield requirements, the compensation program is automatically activated, and the area value is corrected using compensation rules established from historical data.

[0043] Specifically, the deviation ratio is calculated based on the relative relationship between the area of ​​the mixing zone and the lower limit of the material's yield threshold range. First, the currently measured area of ​​the mixing zone is obtained, and the difference is calculated between this area and the lower limit of the yield threshold range extracted from the material processing library. If the mixing zone area is 800 square millimeters, and the lower limit of the material's yield threshold range is 1000 square millimeters, then the area deviation is 200 square millimeters. The deviation ratio is obtained by dividing the deviation by the lower threshold limit; in this example, the deviation ratio is 0.2, indicating that the actual area deviates from the standard requirement by 20%. The establishment of the compensation mapping relationship relies on historical production data of automotive parts welding, collecting actual connection strength data of different materials and thicknesses of plates under various deviation conditions.

[0044] For example, the determination of the compensation coefficient follows a piecewise linear rule. When the deviation ratio is less than 0.1, the deviation is considered small, and the compensation coefficient is set to a baseline value of 1.05. When the deviation ratio is between 0.1 and 0.3, the compensation coefficient increases linearly, with the coefficient increasing by 0.1 for every 0.1 increase in the deviation ratio. This increasing relationship reflects that as the deviation increases, a greater compensation force is needed to ensure welding quality.

[0045] Preferably, the corrected effective area of ​​the mixing zone is calculated by multiplying the original area by the compensation coefficient. In the example above, the original area of ​​800 square millimeters is multiplied by a compensation coefficient of 1.15 to obtain an effective area value of 920 square millimeters. This effective area value more accurately reflects the actual size of the area involved in material bonding, providing a reliable area parameter for subsequent energy consumption statistics.

[0046] S103. Based on the corrected effective area of ​​the mixing zone and the material extrusion mass, the ratio of the peak welding energy consumption is extracted to generate the unit mass connection energy consumption and construct a standardized energy consumption collection record.

[0047] The corrected effective area of ​​the mixing zone and the mass of extruded material collected during welding are obtained. The peak welding energy consumption at the corresponding timestamp is read. The material flow density is determined based on the ratio of the extruded material mass to the effective area of ​​the mixing zone, where the material flow density reflects the degree of material loss per unit area. The area energy consumption density is obtained by dividing the peak welding energy consumption by the effective area of ​​the mixing zone. This area energy consumption density is then corrected by multiplying the area energy consumption density by a material flow density compensation coefficient, which is obtained by consulting a preset lookup table. This yields the energy consumption per unit mass connection. The preset lookup table is constructed based on historical data and established through linear regression analysis. The input is the material flow density, and the output is the compensation coefficient. The regression equation is... k is the compensation coefficient, D is the material flow density, and a and b are regression coefficients. The compensation coefficient is obtained by substituting D into the equation. Using the unit mass connection energy consumption as the core indicator, combined with real-time shoulder pressure data collected from sensors, material grade information read from the process parameter table, and corresponding collection timestamps, the data is organized according to a predetermined format to form a standardized energy consumption collection record containing welding batch identifiers, energy consumption index values, and process condition records.

[0048] In one implementation, the energy consumption per unit mass of weld is calculated through material flow density correction. By comprehensively considering three key parameters—effective area of ​​the stirring zone, material extrusion mass, and peak welding energy consumption—an energy consumption evaluation index reflecting actual welding efficiency is established. This energy consumption evaluation index overcomes the shortcomings of traditional methods that only consider total energy consumption while neglecting material utilization.

[0049] Specifically, the physical meaning of material flow density lies in quantifying the degree of material loss during the welding process. When the shoulder pressure in friction stir welding is too high, the plastically flowing material is squeezed out of the weld, forming flash or overflow. The material flow density is calculated by dividing the mass of material extruded by the effective area of ​​the stirring zone, and the unit is grams per square millimeter. A higher material flow density indicates that more material is extruded per unit area, resulting in reduced welding efficiency.

[0050] For example, when welding a car door panel, if the extruded mass is 5 grams and the effective area is 1000 square millimeters, the material flow density is 0.005 grams per square millimeter. This value directly reflects the degree of material waste.

[0051] It should be noted that the area energy consumption density value is obtained by dividing the peak welding energy consumption by the effective area of ​​the mixing zone, reflecting the energy consumed per unit area.

[0052] For example, the compensation coefficient reference table is established based on a large amount of historical welding data. Actual welding strength test results under different material flow density conditions are collected, and the correspondence between the compensation coefficient and the material flow density is determined through regression analysis. When the material flow density is in the range of 0.001 to 0.003, the compensation coefficient is 1.1; when the density is in the range of 0.003 to 0.006, the compensation coefficient increases linearly to 1.3. The compensation coefficient is used to correct the area energy consumption density value, obtaining a more accurate unit mass connection energy consumption.

[0053] Preferably, the standardized energy consumption collection records are organized in a unified format, including a welding batch identifier field, an energy consumption index value field, and a process condition record field.

[0054] S104. Based on the variation curve of shoulder pressure and unit mass connection energy consumption in the standardized energy consumption collection record, identify the location of the lowest energy consumption point and extract the starting boundary of the best matching interval.

[0055] The shoulder pressure sequence and the corresponding unit mass connection energy consumption sequence are extracted from standardized energy consumption collection records. A moving average method is used to smooth the data, with the window size set to a preset number of adjacent data points, resulting in a smoothed pressure-energy consumption sequence. A pressure-energy consumption variation curve is constructed. The rate of change between adjacent data points is calculated for this curve. When a turning point where the rate of change changes from negative to positive is detected, this point is identified as a candidate location for the lowest energy consumption point. If multiple candidate locations exist, the location with the lowest energy consumption is selected as the lowest energy consumption point. Using the lowest energy consumption point as the center, a search is performed in the direction of decreasing pressure. When the energy consumption value relative to the lowest point increases by more than a preset threshold, this location is determined as the starting boundary of the optimal matching interval, and the shoulder pressure value corresponding to the starting boundary is recorded. Based on the difference between the starting boundary pressure value and the lowest energy consumption point pressure value, the pressure range of the optimal matching interval is determined. This pressure range can be used for subsequent welding process parameter optimization and control.

[0056] In one implementation, the identification of the lowest energy consumption point and the extraction of the optimal matching interval boundary are achieved through in-depth analysis of standardized energy consumption data collection records. The intrinsic relationship between shoulder pressure and energy consumption is mined from historical welding data, establishing a data foundation for pressure optimization control. This method overcomes the blindness of traditional experience-based adjustments, accurately locating the optimal energy-efficiency operating range through a data-driven approach.

[0057] Specifically, the data extraction process reads welding records within a specific time period from a database of standardized energy consumption data. Each record contains three core fields: timestamp, shoulder pressure value, and energy consumption per unit mass connection. The data is arranged chronologically to form a one-to-one correspondence between the pressure sequence and the energy consumption sequence. For automotive door welding, a typical data acquisition frequency is 10 sample points per second to ensure the capture of detailed features of pressure changes. During data preprocessing, obvious outliers are removed, such as data points where pressure fluctuations exceed the normal range by 50%, to avoid noise interference in subsequent analysis. The pressure sequence and energy consumption sequence constitute the original discrete data pairs, providing the basis for curve construction. The window size for the moving average method is dynamically determined based on the data sampling rate and welding speed, typically selecting a number of data points covering a 0.5-second time span to achieve local smoothing while preserving the overall trend characteristics of the curve.

[0058] For example, differential calculation and inflection point identification constitute the core steps in locating the lowest energy consumption point. The differential value is obtained by calculating the energy consumption difference between adjacent data points, reflecting the instantaneous rate of change of the curve. As the welding process gradually increases the shoulder pressure from an insufficient pressure state, the energy consumption per unit mass connection shows a decreasing trend, and the differential value is negative. As the pressure continues to increase, reaching the ideal state of plastic flow of the material, the energy consumption reaches its minimum value. If the pressure increases further, too much material is squeezed out of the weld, and the energy consumption increases instead, and the differential value turns positive. By detecting the sign change of the differential value through a sliding window, when a positive value appears after a continuous negative value sequence, the inflection point is determined as a candidate location for the lowest energy consumption point. Considering that there may be local fluctuations in the actual data, the algorithm requires at least 3 consecutive negative differential values ​​followed by 2 positive differential values ​​to determine a valid inflection point.

[0059] It should be noted that a global minimum value selection strategy is used to process multiple candidate locations. The energy consumption values ​​corresponding to all candidate locations are compared, and the location with the smallest value is selected as the final point of lowest energy consumption. The shoulder pressure corresponding to this point is the theoretically optimal pressure value.

[0060] In one possible implementation, the search for the initial boundary begins at the point of lowest energy consumption and proceeds point by point in the direction of decreasing pressure. The growth rate of energy consumption at each pressure point relative to the lowest point is calculated, and the location is considered the initial boundary when the growth rate first exceeds a preset threshold. The preset threshold is set considering welding quality requirements and energy consumption tolerance, and is typically set to 10% to 20% of the energy consumption value at the lowest point.

[0061] For example, if the energy consumption at the lowest point is 100 joules per gram, when the search to the left reaches a position where the energy consumption reaches 115 joules per gram, that point is determined as the starting boundary.

[0062] Preferably, the determination of the pressure range considers not only the initial boundary but also a comprehensive evaluation based on material properties and plate thickness parameters. The lower limit of the pressure range is the initial boundary pressure value, and the upper limit is obtained by searching in the direction of increasing pressure using a similar method.

[0063] For example, in the welding application of the inner panel and reinforcing plate of the automotive front door, the yield strength of the aluminum alloy material is approximately 200 MPa, and the plate thickness is 3 mm. Using the above method, the pressure corresponding to the lowest energy consumption point was identified as 15 kN, the initial boundary pressure as 12 kN, and the optimal matching range was 12 to 18 kN. In actual production, controlling the shoulder pressure within this range achieved a 15% reduction in energy consumption.

[0064] Understandably, the method automatically identifies the optimal working range through data analysis.

[0065] S105. Starting from the initial boundary of the optimal matching interval, acquire continuous shoulder pressure acquisition records, group them into high resistance clusters and low resistance clusters, and extract the center of each cluster as a reference threshold for the plastic flow resistance of the material.

[0066] Starting from the initial boundary of the optimal matching interval, subsequent shoulder pressure acquisition records are read sequentially over time to obtain a continuous pressure data sequence within a preset duration. This pressure data sequence contains complete information on the changes in the plastic flow state of the material during welding. The pressure data sequence is then clustered using the K-means clustering algorithm, dividing the data into two groups. The maximum and minimum pressure values ​​are selected as initial cluster centers. The distance from each data point to the cluster center is iteratively calculated and the center position is updated until convergence, forming a high-resistance data group and a low-resistance data group. Each cluster is a data group, and the center value is the converged mean. The mean of the high-resistance data group is calculated as the high-resistance cluster center value, and the mean of the low-resistance data group is calculated as the low-resistance cluster center value. If the difference between the two cluster center values ​​is less than a preset threshold, the acquisition time is increased, and pressure data is reacquired for clustering. If the total time exceeds 30 seconds after the increase, the process stops and the current value is used. The center value of the high-resistance cluster is determined as the high threshold of the plastic flow resistance of the material, and the center value of the low-resistance cluster is determined as the low threshold of the plastic flow resistance of the material. The high threshold and the low threshold are used as reference thresholds for the plastic flow resistance of the material.

[0067] In one implementation, the reference threshold for the plastic flow resistance of the material is extracted through cluster analysis of continuous pressure data. Pressure data is collected starting from the initial boundary of the optimal matching interval to identify the changing patterns of the material flow state during welding, providing a quantitative basis for pressure control.

[0068] Specifically, the acquisition of the pressure data sequence is triggered from the initial boundary position, continuously recording the shoulder pressure values ​​within a preset duration. The preset duration is determined based on the welding speed and weld length, typically covering at least one complete welding cycle. In automotive body side panel welding, at a welding speed of 300 mm / min, the acquisition duration is set to 20 seconds, encompassing approximately 100 mm of weld data. The pressure sensor samples at a frequency of 100 Hz, forming a pressure sequence containing 2000 data points. These data reflect the resistance changes during the material's transition from a solid to a plastic flow state.

[0069] For example, the K-means clustering algorithm is implemented using an iterative optimization approach to process pressure data. The initial cluster centers are selected based on the statistical characteristics of the pressure sequence. First, the maximum and minimum values ​​of all pressure values ​​are calculated and used as the two initial center points. This selection method ensures that the initial centers represent high-resistance and low-resistance states, respectively. During iteration, the algorithm calculates the distance from each pressure data point to the two centers, using the absolute difference as the distance metric. Data points are grouped into clusters with closer proximity. After one round of classification, the average value of all data points within each cluster is recalculated as the new cluster center. Iteration continues until the change in center position between two adjacent iterations is less than a set threshold, typically 0.1% of the pressure range. The two converged clusters correspond to the high-resistance and low-resistance states in the welding process, respectively. The cluster center value is obtained by calculating the arithmetic mean of all data points within each cluster, representing the typical pressure level of the corresponding resistance state.

[0070] In one possible implementation, a difference-based evaluation mechanism is used to assess the effectiveness of the clustering results. When the difference between the center values ​​of high-resistance clusters and low-resistance clusters is too small, it indicates that the pressure changes within the current data segment are not significant enough to effectively distinguish different resistance states. The preset threshold is typically set to 5% of the pressure range. If the difference is less than this threshold, the system automatically extends the data acquisition time, increases the number of sampling points by 50%, and re-executes the clustering process. This mechanism ensures accurate identification of resistance characteristics even during the welding stage where pressure changes are gradual.

[0071] For example, when welding thin aluminum alloy sheets, the material softening temperature is low and the pressure change is relatively gradual, so it may be necessary to extend the collection time to 30 seconds to obtain effective clustering results.

[0072] Preferably, the high and low thresholds correspond to the flow resistance characteristics of the material at different temperatures. A high threshold reflects the flow resistance when the material is at a relatively low temperature, where its plasticity is poor and a larger pressure is required for effective bonding. A low threshold corresponds to the flow resistance after the material has softened sufficiently, characterizing an ideal plastic flow state.

[0073] For example, in welding aluminum alloy sheets for automotive engine hoods, the high resistance threshold obtained using the above method is 18 kN, and the low resistance threshold is 10 kN. During actual welding, when the detected pressure approaches the high threshold, it indicates that the material temperature is too low or the mixing is insufficient, requiring adjustment of process parameters. When the pressure stabilizes near the low threshold, it indicates that the welding is in an ideal state.

[0074] Understandably, the data distribution characteristics of the two clusters also reflect the stability of the welding process. A higher degree of dispersion in the high-resistance cluster indicates an uneven material state or initial welding stage. Conversely, a higher concentration of data in the low-resistance cluster indicates that the welding process has entered a stable phase. By establishing a reference threshold for the plastic flow resistance of the material, real-time judgment and classification of the welding process state are achieved, providing a quantitative standard for subsequent adaptive pressure adjustment.

[0075] S106. Evaluate the difference between the current shoulder pressure and the cluster center, identify adjustment needs where the difference exceeds the reference threshold, and generate the downward adjustment amount of the shoulder pressure and the adjusted pressure control amount.

[0076] The real-time shoulder pressure value during the current welding process is acquired, and the difference between it and the high and low thresholds of the material's plastic flow resistance is calculated. If the current pressure value exceeds the high threshold, a pressure adjustment requirement is identified, and a downward adjustment is determined. Once the pressure adjustment requirement is identified, the deviation between the current pressure value and the low threshold is calculated. Based on an adjustment mapping relationship established from historical welding data, the adjustment coefficient corresponding to the deviation is found. This adjustment mapping relationship records the pressure correction ratio under different degrees of deviation. The downward adjustment amount of the shoulder pressure is obtained by multiplying the adjustment coefficient by the deviation amount. The adjusted pressure control amount is obtained by subtracting this downward adjustment amount from the current pressure value. This pressure control amount is not lower than the low threshold and does not exceed the high threshold.

[0077] In one implementation, dynamic adjustment of the shoulder pressure is achieved through real-time monitoring and feedback control. The current pressure is continuously compared to a reference threshold. When a pressure deviation from the ideal range is detected, the adjustment amount is automatically calculated and pressure correction is performed to maintain the stability of the welding process. The identification of adjustment needs is based on a comparison logic between the pressure value and a threshold. After obtaining the real-time shoulder pressure, the difference between it and the high threshold of the material's plastic flow resistance is first calculated. When the difference is positive, it indicates that the current pressure exceeds the high threshold, the material is in an over-extruded state, and flash and material waste are likely to occur. At this time, the system determines that there is a downward adjustment requirement.

[0078] For example, in the welding of the B-pillar reinforcement plate of a car, the high threshold is set to 16 kN. When the actual pressure is detected to reach 18 kN, the overpressure state of 2 kN is immediately identified and the pressure adjustment program is triggered.

[0079] It should be noted that the adjustment mapping relationship is established by analyzing the pressure adjustment records in historical welding data, statistically analyzing the effective adjustment ratio under different degrees of deviation, and forming a correspondence table between deviation and adjustment coefficient.

[0080] For example, the adjustment coefficient is determined by considering both material response characteristics and equipment control precision. When the deviation is within 10% of the threshold range, a conservative adjustment coefficient of 0.5 is used to avoid over-adjustment causing oscillations. When the deviation exceeds 20%, the adjustment coefficient is increased to 0.8 to achieve rapid correction. The adjustment amount is calculated by multiplying the deviation by the adjustment coefficient to ensure a smooth transition of pressure to the target value.

[0081] Preferably, the pressure control quantity is limited by a dual constraint mechanism. The adjusted pressure value must not fall below the low threshold to avoid insufficient welding strength, nor exceed the high threshold to prevent excessive material consumption. When the calculated control quantity exceeds the range, the system automatically limits it to the boundary value. In automotive front bulkhead welding applications, through the above adjustment mechanism, the pressure control accuracy is improved to ±0.5 kN, material utilization is increased by 8%, and welding quality consistency is significantly improved.

[0082] S107. Send the adjusted pressure control quantity to the friction stir welding equipment to perform pressure adjustment, re-acquire the boundary contour image of the stirring zone, and extract the updated stirring zone area for the next round of automotive parts production energy consumption statistics.

[0083] The adjusted pressure control value is sent to the control port of the friction stir welding equipment. Upon receiving the command, the equipment executes the pressure adjustment operation, monitoring the pressure sensor feedback value until the target pressure value is reached, thus obtaining a pressure adjustment completion signal. Based on this signal, the boundary imaging unit is triggered to re-image the welding area, acquiring an image of the boundary contour of the stirring zone after pressure adjustment. The updated stirring zone area value is calculated using grayscale thresholding and boundary extraction methods. This updated stirring zone area value is combined with the adjusted pressure control value, the real-time collected material mass extruded during welding, and the energy consumed during welding to construct a new round of standardized energy consumption data records. These records serve as input data for the next round of energy consumption statistics for automotive parts production.

[0084] In one implementation, pressure regulation and energy consumption statistics form a closed-loop control mechanism. By adjusting the shoulder pressure in real time and re-collecting the stirring zone area, welding parameters are dynamically optimized. The data after each adjustment is automatically updated in the energy consumption statistics record, providing a data foundation for continuous optimization in the continuous production process.

[0085] Specifically, the pressure control signal is transmitted to the programmable logic controller (PLC) of the friction stir welding equipment via a serial communication protocol. After parsing the command, the PLC drives the proportional valve to adjust the oil supply pressure of the hydraulic cylinder, thereby achieving precise adjustment of the shoulder pressure. The pressure sensor continuously monitors the actual pressure value at a sampling interval of 100 milliseconds. When the deviation between the pressure value at five consecutive sampling points and the target value is less than a set threshold, the pressure adjustment is considered complete. The completion signal is fed back to the main control system through the digital output port. The entire adjustment process is typically completed within 2 seconds.

[0086] It should be noted that the pressure regulation completion signal directly triggers the image acquisition function of the boundary imaging unit, avoiding image blurring during pressure fluctuations and ensuring that images of the welding area are acquired in a stable state.

[0087] For example, image processing uses the same parameter settings as the initial acquisition, including the same exposure time, gain value, and white balance parameters. Gray-scale thresholding is performed using Otsu's method to adaptively determine the segmentation threshold, and boundary extraction employs an eight-connected-domain labeling algorithm to identify the largest connected region. This consistency processing method ensures the comparability of the stirring zone area acquired at different times, eliminating errors caused by differences in image processing.

[0088] Preferably, the construction of the new round of energy consumption statistics records adopts an incremental update method. Historical data from the previous round is retained, and only the two changing parameters—pressure control quantity and stirring zone area—are updated. The material extrusion quality and energy consumption values ​​are obtained through real-time data acquisition. The standardized records are stored in a time-series format, with each record containing a complete set of process parameters, supporting subsequent trend analysis and parameter optimization. In the automotive chassis longitudinal beam welding application, through five consecutive rounds of pressure adjustment and data updates, energy consumption gradually decreases and stabilizes at an ideal level.

[0089] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for statistical analysis of energy consumption data in automotive parts production, characterized in that, The method includes: Real-time acquisition of shoulder pressure, material extrusion quality, and peak welding energy consumption during the welding process of automotive parts; combined with the captured effective stirring zone boundary contour image, the stirring zone area is extracted; and the lower limit of the material yield threshold range is obtained from the automotive parts material process library. The area of ​​the mixing zone is compared with the lower limit of the material yield threshold range to identify the area deviation when the area of ​​the mixing zone is lower than the lower limit of the material yield threshold range. The area of ​​the mixing zone is then compensated based on the area deviation to obtain the corrected effective area of ​​the mixing zone. Based on the corrected effective area of ​​the mixing zone and the material extrusion mass, the ratio of the peak welding energy consumption is extracted to generate the unit mass connection energy consumption and construct a standardized energy consumption collection record. Based on the variation curve of shoulder pressure and unit mass connection energy consumption in the standardized energy consumption collection records, the location of the lowest energy consumption point is identified, and the starting boundary of the best matching interval is extracted. Starting from the initial boundary of the optimal matching interval, continuous shoulder pressure acquisition records are obtained, grouped into high resistance clusters and low resistance clusters, and the center of each cluster is extracted as a reference threshold for the plastic flow resistance of the material. The difference between the current shoulder pressure and the cluster center is evaluated to identify adjustment needs where the difference exceeds the reference threshold, and the downward adjustment amount of the shoulder pressure and the adjusted pressure control amount are generated.

2. The method for statistical analysis of energy consumption data in automotive parts production according to claim 1, characterized in that, The system collects peak values ​​of shoulder pressure, material extrusion quality, and welding energy consumption during the real-time welding process of automotive parts. Combined with the captured effective stirring zone boundary contour image, the stirring zone area is extracted. Simultaneously, the lower limit of the material yield threshold range is obtained from the automotive parts material process library, including: Continuous pressure monitoring is performed on the shoulder of the friction stir welding equipment according to a preset sampling frequency. The shoulder pressure data and corresponding timestamps are recorded. At the same time, the material extrusion quality data is acquired. By reading the instantaneous power value of the welding equipment power meter, the maximum value in the power value sequence is identified as the peak welding energy consumption. The boundary contour image of the stirring zone is acquired. The boundary contour image is processed by the grayscale threshold segmentation method to identify the boundary line. The area value of the stirring zone is calculated based on the closed area enclosed by the boundary line and matched with the timestamp of the shoulder pressure data. The current welding material grade information is queried from the automotive parts material process library, the corresponding material yield strength data table is extracted, and the lower limit of the material yield threshold range is obtained.

3. The method for statistical analysis of energy consumption data in automotive parts production according to claim 1, characterized in that, The step of comparing the area of ​​the mixing zone with the lower limit of the material's yield threshold range, identifying the area deviation when the area of ​​the mixing zone is lower than the lower limit of the material's yield threshold range, and compensating the area of ​​the mixing zone based on the area deviation to obtain the corrected effective area of ​​the mixing zone includes: Calculate the difference between the area of ​​the mixing zone and the lower limit of the material yield threshold range. If the area of ​​the mixing zone is less than the lower limit of the material yield threshold range, the area deviation is determined as the difference. Divide the area deviation by the lower limit of the material yield threshold range to obtain the deviation ratio. Based on the deviation ratio, find the compensation coefficient in the compensation mapping relationship. Use the compensation coefficient to perform a product operation on the original area of ​​the mixing zone to obtain the corrected effective area of ​​the mixing zone.

4. The method for statistical analysis of energy consumption data in automotive parts production according to claim 1, characterized in that, The ratio of peak welding energy consumption to the corrected effective area of ​​the mixing zone and the material extrusion mass is extracted to generate unit mass connection energy consumption, and a standardized energy consumption collection record is constructed, including: The material flow density is determined by the ratio of the material extrusion mass to the corrected effective area of ​​the mixing zone; the area energy consumption density is obtained by dividing the peak welding energy consumption by the corrected effective area of ​​the mixing zone, and the area energy consumption density is corrected according to the material flow density to obtain the unit mass connection energy consumption; the unit mass connection energy consumption is combined with the shoulder pressure data, material grade information and timestamp organization data to form a standardized energy consumption collection record.

5. The method for statistical analysis of energy consumption data in automotive parts production according to claim 1, characterized in that, The process of identifying the location of the lowest energy consumption point and extracting the starting boundary of the optimal matching interval based on the variation curve of shoulder pressure and unit mass connection energy consumption in the standardized energy consumption collection records includes: The shoulder pressure sequence and the corresponding unit mass connection energy consumption sequence are extracted from the standardized energy consumption collection records. The data is smoothed using the moving average method to obtain the smoothed pressure-energy consumption sequence, and a pressure-energy consumption change curve is constructed. The difference between adjacent data points on the change curve is calculated to obtain the rate of change of each segment of the curve. When a turning point where the rate of change changes from negative to positive is detected, the point is determined as a candidate position for the lowest energy consumption point. The position with the minimum energy consumption is selected as the lowest energy consumption point. With the lowest energy consumption point as the center, a search is conducted in the direction of decreasing pressure. When the energy consumption value increases relative to the lowest point by more than a preset threshold, the position is determined as the starting boundary of the best matching interval, and the shoulder pressure value corresponding to the starting boundary is recorded.

6. The method for statistical analysis of energy consumption data in automotive parts production according to claim 1, characterized in that, Starting from the initial boundary of the optimal matching interval, continuous shoulder pressure acquisition records are obtained, grouped into high-resistance clusters and low-resistance clusters, and the center of each cluster is extracted as a reference threshold for the plastic flow resistance of the material, including: Starting from the initial boundary position of the optimal matching interval, subsequent shoulder pressure acquisition records are read in chronological order to obtain a continuous pressure data sequence within a preset time period. The pressure data sequence is then clustered using the K-means clustering algorithm to divide the data into two groups. Initial cluster centers are selected based on the pressure values. The distance from each data point to the cluster center is iteratively calculated and the center position is updated until convergence, forming a high-resistance data group and a low-resistance data group. The mean of the high-resistance data group is calculated as the high-resistance cluster center value, and the mean of the low-resistance data group is calculated as the low-resistance cluster center value. The high-resistance cluster center value is determined as the high threshold of the material's plastic flow resistance, and the low-resistance cluster center value is determined as the low threshold of the material's plastic flow resistance.

7. The method for statistical analysis of energy consumption data in automotive parts production according to claim 1, characterized in that, The step of evaluating the difference between the current shoulder pressure and the cluster center, identifying adjustment needs where the difference exceeds a reference threshold, and generating a downward adjustment amount for the shoulder pressure and a post-adjustment pressure control amount includes: Calculate the difference between the current shoulder pressure value and the high and low thresholds of the material's plastic flow resistance. If the current pressure value exceeds the high threshold, a pressure adjustment requirement is identified. Calculate the deviation between the current pressure value and the low threshold. Find the adjustment coefficient corresponding to the deviation based on the adjustment mapping relationship. Multiply the deviation by the adjustment coefficient to obtain the downward adjustment amount of the shoulder pressure. Subtract this downward adjustment amount from the current pressure value to obtain the adjusted pressure control amount.

8. The method for statistical analysis of energy consumption data in automotive parts production according to claim 1, characterized in that, The method further includes: sending the adjusted pressure control quantity to the friction stir welding equipment to perform pressure adjustment, re-acquiring the boundary contour image of the stirring zone, and extracting the updated stirring zone area for the next round of automotive parts production energy consumption statistics.

9. The method for statistical analysis of energy consumption data in automotive parts production according to claim 8, characterized in that, The process of sending the adjusted pressure control value to the friction stir welding equipment to perform pressure adjustment, re-acquiring the boundary contour image of the stirring zone, and extracting the updated stirring zone area for the next round of automotive parts production energy consumption statistics includes: The adjusted pressure control value is sent to the control port of the friction stir welding equipment. The equipment performs a pressure adjustment operation and obtains a pressure adjustment completion signal. Based on the pressure adjustment completion signal, the welding area is re-photographed to obtain the boundary contour image of the stirring zone. The updated stirring zone area value is calculated by grayscale threshold segmentation and boundary extraction methods. The updated stirring zone area value is combined with the adjusted pressure control value, the real-time collected material extrusion quality and welding energy consumption value to construct a new round of standardized energy consumption collection records.

Images