Cylindrical battery point bottom welding hidden false welding identification method based on resistance and energy parameter correlation analysis
By correlating resistance and energy parameters, a two-dimensional qualified area was established to identify hidden weld defects at the bottom of cylindrical batteries. This solved the problem that existing equipment could not identify hidden weld defects, achieving efficient quality monitoring and interception of hidden weld defects, thus improving product quality and safety.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- NANJING CBAK NEW ENERGY TECH CO LTD
- Filing Date
- 2026-03-25
- Publication Date
- 2026-06-05
AI Technical Summary
Existing welding equipment cannot effectively identify hidden weld defects during the identification of bottom welds in cylindrical batteries, leading to problems such as increased internal resistance, overheating, and capacity decay. Existing monitoring methods have limited ability to detect such defects.
By correlating the average resistance and welding energy, a two-dimensional joint qualified area is established. The two-dimensional coordinates of resistance and energy during the battery welding process are monitored in real time to determine whether they fall within the qualified area. If they do not fall within the qualified area, they are judged as hidden poor welds and a sorting signal is triggered.
It achieves efficient identification of hidden solder joints, reduces the overall risk of solder joint defects, improves product quality consistency and safety, and is low in cost and does not require additional hardware sensors.
Smart Images
Figure CN122141980A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of lithium battery manufacturing technology, specifically to a method for identifying hidden poor welds in the bottom welding of cylindrical batteries based on the correlation analysis of resistance and energy parameters. Background Technology
[0002] In cylindrical battery manufacturing, spot welding is a crucial process for attaching the bottom tabs of the battery cell to the steel casing. Poor welding is one of the main defects in this process, leading to increased internal resistance, overheating, and capacity decay.
[0003] Currently, high-end resistance welding equipment (such as the Miyaki MD-A8000B) is generally equipped with process monitoring functions, which can set upper and lower limits for parameters such as welding current, voltage, and power, and issue alarms when these limits are exceeded. This type of monitoring can effectively identify obvious welding defects caused by abnormal power output.
[0004] However, in actual production, there is another type of "hidden weld defects": due to microscopic contamination, oxidation, or abnormal fit gaps on the bottom of the battery steel casing or the surface of the electrode, the input electrical parameters (current, voltage, etc.) during welding are still within the normal range set by the equipment, but the welding interface does not form a good metallurgical bond, resulting in insufficient weld strength. Existing monitoring methods based on independent judgment of single parameters have limited ability to detect such defects, becoming the main pathway for weld defects to emerge.
[0005] Therefore, there is a need for a monitoring method that can more effectively identify hidden solder joint defects without increasing hardware costs. Summary of the Invention
[0006] The purpose of this invention is to address the problem that existing welding equipment with built-in monitoring functions is insufficient in detecting hidden weld defects.
[0007] To achieve the above objectives, the present invention provides the following technical solution: This invention provides a method for identifying hidden weld defects in the bottom welding of cylindrical batteries based on the correlation analysis of resistance and energy parameters, including the following steps: S100: Under normal production conditions, welding process data of a preset number of qualified samples of bottom welding of cylindrical batteries are collected. The welding process data includes at least the average resistance and welding energy of each welding. S200: Based on the welding process data of the historical qualified samples, perform joint distribution characteristic analysis on the average resistance and the welding energy, and establish a two-dimensional joint qualified region of the average resistance and the welding energy; S300: During the production of cylindrical batteries, the average resistance and welding energy of the bottom welding of the battery under test are acquired in real time to form a two-dimensional coordinate point; S400: Determine whether the two-dimensional coordinate point falls within the two-dimensional joint qualified area: If the two-dimensional coordinate point falls within the two-dimensional joint qualified area, then the current welding is determined to be qualified; If the two-dimensional coordinate point falls outside the two-dimensional joint qualified area, it is determined that there is a hidden risk of poor welding in the current welding, and a sorting signal is triggered.
[0008] Further, in step S100, the method for obtaining the historical qualified samples under normal production conditions includes: selecting a preset production period, and using the cylindrical batteries that are continuously produced within the preset production period and whose welding quality has been confirmed to be stable by subsequent sampling inspections as the historical qualified samples. The welding energy can be obtained by: directly providing it by the spot welding equipment after each welding operation; or by collecting the real-time welding current, real-time welding voltage, and welding time during each spot welding operation, and then calculating the welding energy.
[0009] Furthermore, the acquisition of the welding process data specifically includes: continuously reading the average resistance and the welding energy within the preset production period from the spot welding equipment via a data communication interface.
[0010] Further, in step S200, establishing a two-dimensional joint qualified region of the average resistance and the welding energy specifically includes: Calculate the mean and standard deviation of the average resistance and the mean and standard deviation of the welding energy in the historical qualified samples respectively; With the goal of meeting the maximum process fluctuation range allowed by the battery welding strength requirements, a first process tolerance coefficient for the average resistance and a second process tolerance coefficient for the welding energy are respectively set. Based on the mean, the standard deviation, the first process tolerance coefficient, and the second process tolerance coefficient, an initial two-dimensional parameter boundary is constructed; Based on the correlation characteristics between the average resistance and the welding energy, the initial two-dimensional parameter boundary is further optimized and determined to obtain the closed two-dimensional joint qualified region.
[0011] Furthermore, the values of both the first process tolerance coefficient and the second process tolerance coefficient are greater than or equal to 2; Furthermore, the shape of the closed two-dimensional joint qualified region can be any one of ellipse, rectangle or polygon.
[0012] Furthermore, when the closed two-dimensional joint qualified region is elliptical, its construction method is specifically as follows: Extract the negative correlation feature between the average resistance and the welding energy; The coordinate point formed by the average value of the resistance and the average value of the welding energy is used as the center of the ellipse. The product of the first process tolerance coefficient and the standard deviation of the average resistance, and the product of the second process tolerance coefficient and the standard deviation of the welding energy, are respectively used as the reference axis lengths; Based on the negative correlation characteristics, the two-dimensional joint qualified region is established by fitting an ellipse equation.
[0013] Furthermore, in step S300, the average resistance and welding energy of the current bottom welding point of the battery under test are acquired in real time, specifically including: After the single-point bottom welding cycle of the current battery under test is completed, and before the current battery under test is transferred to the next station; The average resistance and welding energy of the battery under test are read in real time from the spot welding equipment by the production line programmable logic controller or host computer through a data communication interface.
[0014] Furthermore, the formation of two-dimensional coordinate points specifically involves: Using either the real-time reading of the average resistance or the welding energy as the abscissa and the other as the ordinate, a two-dimensional coordinate point is constructed to characterize the coupling relationship between the current welding interface contact state and the total heat input of the system.
[0015] Further, in step S400, determining whether the two-dimensional coordinate point falls within the two-dimensional joint qualified area specifically includes: Calculate the Mahalanobis distance from the two-dimensional coordinate point of the current battery under test to the center of the ellipse; The Mahalanobis distance is compared with a preset distance threshold to determine whether the two-dimensional coordinate point falls within the two-dimensional joint qualified area.
[0016] Furthermore, the preset distance threshold is 1; If the Mahalanobis distance is less than 1, then the two-dimensional coordinate point is determined to fall within the two-dimensional joint qualified area formed by the ellipse, and the current welding is qualified; If the Mahalanobis distance is greater than or equal to 1, it is determined that the two-dimensional coordinate point falls outside the two-dimensional joint qualified area formed by the ellipse, and there is a hidden risk of poor welding in the current welding. The trigger sorting signal specifically includes: A defective product rejection instruction is sent to the sorting execution mechanism of the production line to drive the push rod to remove the battery under test that is determined to have a hidden risk of poor soldering from the normal production flow channel. Beneficial effects
[0017] This invention directly reflects the average resistance and total heat input energy of the welding interface through correlation analysis, and can capture abnormal coupling patterns that cannot be detected by independent monitoring of single parameters, thereby effectively identifying hidden cold welds caused by abnormal workpiece conditions.
[0018] Low implementation cost: It fully utilizes the existing output data of the welding equipment without adding any hardware sensors or changing the existing production layout, making it particularly suitable for upgrading the quality monitoring of existing production lines.
[0019] The principle is clear and the logic is rigorous: the criteria are established based on the physical principles of welding. The average resistance reflects the interface contact state, and the welding energy reflects the total heat input. An abnormal combination of the two directly points to a specific type of welding defect.
[0020] Highly adaptable: By collecting data under normal production line conditions to establish a baseline area, it can adapt to process fluctuations under different product and equipment conditions, and has good universality. Attached Figure Description
[0021] Figure 1 This is a flowchart illustrating the overall process of the present invention for identifying hidden weld defects in cylindrical battery bottom welding based on the correlation analysis of resistance and energy parameters. Figure 2 This is a flowchart illustrating the operation of the method for identifying hidden weld defects in cylindrical batteries based on the correlation analysis of resistance and energy parameters, as described in this invention. Detailed Implementation
[0022] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.
[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but includes other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0024] The present invention will now be described in further detail with reference to the accompanying drawings: Example
[0025] As shown in the figure, the method for identifying hidden weld defects at the bottom of cylindrical batteries based on the correlation analysis of resistance and energy parameters includes the following steps: S100: Under normal production conditions, collect welding process data from a preset number of qualified samples of cylindrical battery bottom welding history. The welding process data includes at least the average resistance and welding energy of each welding. S200: Based on the welding process data of the historical qualified samples, perform joint distribution characteristic analysis on the average resistance and the welding energy, and establish a two-dimensional joint qualified region of the average resistance and the welding energy; S300: During the production of cylindrical batteries, the average resistance and welding energy of the bottom welding of the battery under test are acquired in real time to form a two-dimensional coordinate point; S400: Determine whether the two-dimensional coordinate point falls within the two-dimensional joint qualified area: If the two-dimensional coordinate point falls within the two-dimensional joint qualified area, then the current welding is determined to be qualified; If the two-dimensional coordinate point falls outside the two-dimensional joint qualified area, it is determined that there is a hidden risk of poor welding in the current welding, and a sorting signal is triggered.
[0026] Furthermore, the specific implementation process of step S100 is as follows: In a specific embodiment of the present invention, the reference data acquisition process described in step S100 is explained in detail in conjunction with the actual production scenario of a 32140 model cylindrical battery production line.
[0027] Addressing the technical challenge of current single-parameter monitoring failing to identify hidden weld defects, this embodiment aims to acquire underlying process data without increasing the cost of additional physical sensor hardware. Specifically, this embodiment employs the Miyaki MD-A8000B high-end resistance welding machine to perform spot welding operations (i.e., welding from the bottom tab of the battery cell to the steel shell). This equipment has its own process parameter measurement function, and the system directly and continuously reads the output parameters of the underlying equipment from the production line's host computer by calling the welding machine's built-in data communication interface (e.g., RS-485 industrial communication protocol).
[0028] To ensure that the subsequently constructed two-dimensional joint qualified region has accurate process benchmark value, the historical sample data used for modeling must be highly reliable and pure. Therefore, this embodiment strictly defines "normal production state": the system does not blindly capture historical data, but deliberately selects a continuous production period, and the cylindrical batteries produced during this period must be verified by subsequent destructive physical sampling inspections (such as pull-out force tests) to ensure that their welding strength meets the process standards and that quality fluctuations are in a stable and controllable state.
[0029] After identifying the stable production period confirmed by the sampling inspection, the host computer continuously collects data on the total number of data completed during that period through the aforementioned data communication interface. The process data of each welding operation is used as the baseline sample for modeling. In this embodiment, to ensure statistical significance, a preset sample size is used. The value is 130075. For each spot welding action, the system focuses on extracting two core related parameters that can reflect the physical nature of the hidden cold weld: one is the average resistance value that can directly reflect the micro-contamination or gap state of the welding interface; the other is the welding energy that can reflect the total heat input of the system.
[0030] From a mathematical modeling perspective, the above data collection process ultimately constructs a massive dataset of historical qualified samples. Let for the th The average resistance value read by the system from each historical qualified welding operation is recorded as follows: (Unit: microeuro) The corresponding welding energy is denoted as (Unit: Joules, J), then the basic dataset collected and stored in this step This can be expressed by the following formula: , in, are positive integers and In this embodiment =130075.
[0031] The above dataset Each two-dimensional coordinate pair These data accurately reflect the normal physical fluctuations of qualified solder joints in terms of "contact state" and "heat input" under current equipment and process conditions. This benchmark construction method, driven by massive amounts of measured data, completely eliminates the blindness of manually setting empirical thresholds.
[0032] Furthermore, the specific implementation process of step S200 is as follows: After obtaining the massive benchmark dataset described in step S100 Next, step S200 aims to extract the joint distribution characteristics of parameters under normal welding conditions through a data-driven approach, thereby constructing a two-dimensional joint qualified area that can accurately identify "hidden cold welds".
[0033] Specifically, the system first uses an analysis and judgment unit to extract basic statistical characteristics of the average resistance and welding energy from the historical qualified samples. For samples containing... For each sample dataset, the system calculates the mean of the average resistance values. and standard deviation and the average welding energy and standard deviation The basic calculation formula is as follows: , In the 32140 cylindrical battery production line of this embodiment, the average value of the resistance is calculated by the system. for Standard deviation for Average welding energy for Standard deviation for .
[0034] After obtaining the aforementioned basic statistics, the key innovation of this invention lies in the fact that it does not employ the rigid, fixed statistical multiples (such as the absolute "three standard deviations") found in traditional quality control (SPC). Instead of using principles to define the acceptance line, a process tolerance coefficient strongly correlated with the actual welding process is introduced. The core physical characteristic of hidden weld defects is often high initial contact resistance due to poor microscopic contact, accompanied by low actual effective welding energy as the system tries to maintain a constant output, exhibiting a typical "high resistance - low energy" abnormal coupling state. Simply using a rectangular boundary for single-parameter truncation easily misses such defects located in corner areas.
[0035] Therefore, this embodiment combines the pull-out force welding strength standard of this battery model with the limit of misjudgment rate allowed by the production cycle, and sets independent process tolerance coefficients for the two parameters respectively.
[0036] Preferably, a first process tolerance factor is set for the average resistance value. The second process tolerance coefficient for setting welding energy. .
[0037] Subsequently, based on the extracted mean, standard deviation, and set tolerance coefficient, and considering the objectively existing negative correlation between the average resistance and welding energy in the physical process, the system constructs a closed two-dimensional elliptical qualified region. The center coordinates of this ellipse are anchored to the two-dimensional mean point. ,Right now The baseline distribution range of the ellipse along the corresponding parametric axes (axis length related variables) is respectively determined by... and The decision was made. Through this joint correlation fitting, the constructed two-dimensional elliptical qualified region can not only accommodate reasonable process fluctuations in normal production, but also form a tight inward-shrinking boundary in specific quadrants of the two-dimensional coordinate system (especially the high-resistance, low-energy quadrant). This provides an accurate mathematical criterion for subsequent real-time online interception of hidden poorly soldered products where single parameters are within the "normal range" but interface bonding has actually occurred.
[0038] Furthermore, the specific implementation process of step S300 is as follows: After completing the construction of the two-dimensional joint qualified region as described in step S200, the mathematical model of this region (i.e., the center coordinates and major and minor axis parameters of the ellipse) will be sent and stored in the storage unit of the production line control system (such as a host computer or programmable logic controller PLC). Subsequently, the system proceeds to step S300, which involves high-frequency, real-time online interception of process data for the actual products produced on the production line.
[0039] Specifically, in the continuous automated production process of 32140 cylindrical batteries, the production cycle is tight, requiring that no quality monitoring action should cause production line stagnation. Therefore, in this embodiment, the data acquisition action is strictly constrained to be completed within a millisecond-level time window after the end of the single bottom welding physical cycle of the current battery under test and before the battery flows to the next station with the turntable or conveyor belt.
[0040] The moment the spot welding equipment (Miyaki MD-A8000B welding machine) completes a discharge welding operation and the welding head lifts, the production line PLC immediately sends a data reading command to the welding equipment via the RS-485 industrial communication interface, extracting the real-time average resistance value generated during the welding process (denoted as ) with zero delay. ) and real-time welding energy (denoted as ).
[0041] After obtaining the two real-time parameters mentioned above, the analysis and judgment unit does not perform isolated single-value threshold comparisons on them, but immediately generates a real-time two-dimensional coordinate point representing the current battery welding state in the virtual two-dimensional space of memory. The mathematical expression for this coordinate point is as follows: , In a physical sense, this coordinate point x-coordinate The instantaneous characterization represents the microscopic contact state of the battery under test at the interface between its steel casing and the tabs (e.g., the presence of extremely small oil stains, oxide layers, or abnormal assembly gaps of a few micrometers); while its ordinate... This synchronously maps the actual total effective heat injected into the system by the welding power source when overcoming the contact state. By generating such a physically coupled two-dimensional coordinate point, the system successfully transforms the invisible underlying microscopic metallurgical bonding risk into digital coordinates that can be quickly quantified and judged using geometric and statistical tools in subsequent steps, thus realizing the prerequisites for truly "online" and "non-destructive" monitoring.
[0042] Furthermore, the specific implementation process of step S400 is as follows: In step S300, the two-dimensional coordinates of the battery under test are acquired in real time. ( , After that, the system enters step S400, where the analysis and judgment unit performs online judgment and the control output unit performs the corresponding sorting action.
[0043] Traditional single-parameter threshold determination typically employs Euclidean distance or absolute threshold comparison. However, this method ignores the significant dimensional difference between resistance (on the order of microohms) and energy (on the order of joules) and fails to reflect the process-related correlation between the two. To address this issue, this embodiment innovatively introduces Mahalanobis distance as the core discrimination criterion. Mahalanobis distance effectively handles dimensional inconsistencies and correlations among multiple variables, making it the optimal mathematical tool for measuring the degree to which the current coordinate point deviates from the center of a normal two-dimensional joint distribution.
[0044] The analysis and judgment unit extracts the two-dimensional mean vector established in step S200. And incorporates process tolerance coefficients ( The inverse matrix of the covariance matrix For the current real-time coordinates point The system quickly calculates the squared Mahalanobis distance to the center of the reference ellipse using the following matrix formula (denoted as ). ): , This calculation process is completed instantaneously in the memory of the PLC or host computer within microseconds. Subsequently, the system displays the calculated... The value is compared rigidly with a preset judgment threshold. In the normalized elliptic model established in this embodiment, the judgment benchmark threshold is strictly set to 1.
[0045] The specific judgment and execution logic is as follows: (1) Qualified circulation branches: If the calculation yields In terms of geometric topology, this indicates the current real-time coordinates. The sample must fall precisely within the two-dimensional joint qualified elliptical region constructed in step S200. At this point, the analysis and judgment unit determines that the microscopic contact state of the current point bottom welding interface matches the actual effective heat input well, and the welding quality is qualified (OK). The control output unit does not trigger any abnormal signals, and the battery under test flows normally to the next process (such as liquid injection or sealing process) with the production line conveyor belt.
[0046] (2) Intercept and remove branches: If the calculation yields This indicates the current real-time coordinates. It broke through the envelope boundary of normal process fluctuations and fell outside the elliptical region. This is especially true in specific quadrants where resistance is high and energy is low, even... and The absolute values of the values did not trigger the original single-parameter upper and lower limit alarms of the equipment, and the system was still able to keenly capture its abnormal coupling state through two-dimensional joint analysis.
[0047] At this point, the analysis and judgment unit determines that the current welding process has an extremely high risk of "hidden solder joint" (NG). The control output unit immediately sends a high-level rejection command to the sorting actuator downstream of the production line via the I / O bus. When the downstream photoelectric sensor detects that the abnormal battery has arrived at the sorting station, the high-speed pneumatic pusher moves rapidly, physically pushing the battery with the risk of hidden solder joint from the normal good product conveyor belt and discharging it into the defective product collection tank.
[0048] Example 2: On a certain 32140 cylindrical battery production line, the Miyaki MD-A8000B welding machine is used for spot welding. This equipment can output process parameters such as the average resistance (R_avg) and welding energy (E) for each weld through its data communication interface (such as RS-485).
[0049] Step 1: Establishment of Data-Driven Joint Qualified Regions Baseline data acquisition: Select a production period in which welding quality has been confirmed to be stable by subsequent sampling inspections, and continuously collect the average resistance (R_avg) and welding energy (E) data of 130,075 welds during this period through the host computer as the modeling baseline.
[0050] Data analysis and tolerance coefficient determination: The baseline data were analyzed, and the mean value of R_avg, μ_R, was calculated to be 1363 micro-ohms, and the standard deviation σ_R was 56.3 micro-ohms; the mean value of E, μ_E, was 22.6 joules, and the standard deviation σ_E was 1.169 joules.
[0051] The key to this invention lies in not simply using a fixed statistical multiple (such as three standard deviations), but rather determining the tolerance coefficient based on process requirements. Combining the welding strength standard for this battery model with the allowable error rate during production cycle, the process tolerance coefficients are comprehensively determined as: k_R=3.0, k_E=2.6. This setting aims to ensure a high capture rate for the core cold solder joint pattern of "high resistance - low energy".
[0052] A two-dimensional elliptical qualified region was constructed: Centered on the aforementioned mean points (μ_R, μ_E), with k_R·σ_R and k_E·σ_E as the reference axis lengths, and considering the negative correlation between R_avg and E (observed as an elliptical distribution in the scatter plot), a "two-dimensional joint qualified region" covering approximately 99.5% of the baseline data was established through elliptical equation fitting. The mathematical expression for this elliptical region was extracted and solidified.
[0053] Step 2: Implementation of Online Monitoring and Automatic Sorting The mathematical parameters (center coordinates, rotation angle, major and minor axis radii) of the aforementioned elliptical region are written into the production line's programmable logic controller (PLC). During production, the system executes the following closed-loop control: The PLC reads the R_avg and E values of each battery cell in real time after welding.
[0054] Calculate the Mahalanobis distance D² corresponding to the data point based on the equation of the ellipse.
[0055] Judgment and Execution: If D² < 1, the solder joint is deemed qualified, and the battery continues to be circulated.
[0056] If D²≥1, a hidden risk of poor soldering is identified. The PLC immediately triggers the pneumatic push rod to accurately remove the battery into the defective product collection box and records all parameters of this soldering process.
[0057] Step 3: Effect Verification and Comparison After operating as an independent intelligent checkpoint, the system successfully intercepted abnormal solder joints that traditional methods could not detect. A typical example is as follows: During one welding operation, the welding machine's built-in single-parameter comparator (monitoring current and voltage) did not trigger an alarm. However, the system of this invention captured the following parameters: R_avg = 1610 microohms (far exceeding μ_R + 3σ_R) and E = 19.8 joules (far below μ_E - 2σ_E). This "high resistance - low energy" combination precisely points to a poor weld interface and insufficient heat input, indicating a cold weld defect. The system immediately identifies and rejects this defect as NG (Not Good).
[0058] Offline destructive tear testing confirmed that the weld strength was only 35% of the acceptable standard, thus identifying it as a cold weld. This demonstrates that the present invention, through dual-parameter correlation analysis, achieves accurate identification and interception of hidden cold welds, filling the blind spots of existing single-parameter monitoring technologies and effectively improving the overall safety and consistency of the product.
[0059] Through continuous monitoring, this method can effectively screen out the decline in welding consistency caused by fluctuations in the surface condition of the workpiece and micro-wear of the welding needle. Thus, on the basis of the original equipment monitoring, an additional quality defense line against hidden defects is added, reducing the overall risk of defective welds.
[0060] 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 the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for identifying hidden weld defects in the bottom weld of cylindrical batteries based on the correlation analysis of resistance and energy parameters, characterized in that, Includes the following steps: S100: Under normal production conditions, collect welding process data from a preset number of qualified samples of cylindrical battery bottom welding history. The welding process data includes at least the average resistance and welding energy of each welding. S200: Based on the welding process data of the historical qualified samples, perform joint distribution characteristic analysis on the average resistance and the welding energy, and establish a two-dimensional joint qualified region of the average resistance and the welding energy; S300: During the production of cylindrical batteries, the average resistance and welding energy of the bottom welding of the battery under test are acquired in real time to form a two-dimensional coordinate point; S400: Determine whether the two-dimensional coordinate point falls within the two-dimensional joint qualified area: If the two-dimensional coordinate point falls within the two-dimensional joint qualified area, then the current welding is determined to be qualified; If the two-dimensional coordinate point falls outside the two-dimensional joint qualified area, it is determined that there is a hidden risk of poor welding in the current welding, and a sorting signal is triggered.
2. The method for identifying hidden weld defects in cylindrical battery bottom welding based on resistance and energy parameter correlation analysis according to claim 1, characterized in that, In step S100, the method for obtaining historical qualified samples under normal production conditions includes: selecting a preset production period and using cylindrical batteries that are continuously produced within the preset production period and whose welding quality is confirmed to be stable by subsequent sampling inspections as the historical qualified samples. The welding energy can be obtained by: directly providing it by the spot welding equipment after each welding operation; or by collecting the real-time welding current, real-time welding voltage, and welding time during each spot welding operation, and then calculating the welding energy.
3. The method for identifying hidden weld defects in cylindrical battery bottom welding based on resistance and energy parameter correlation analysis according to claim 2, characterized in that, The acquisition of welding process data specifically includes: continuously reading the average resistance and welding energy within the preset production period from the spot welding equipment via a data communication interface.
4. The method for identifying hidden weld defects in cylindrical battery bottom welding based on resistance and energy parameter correlation analysis according to claim 3, characterized in that, In step S200, establishing a two-dimensional joint qualified region for the average resistance and the welding energy specifically includes: Calculate the mean and standard deviation of the average resistance and the mean and standard deviation of the welding energy in the historical qualified samples respectively; With the goal of meeting the maximum process fluctuation range allowed by the battery welding strength requirements, a first process tolerance coefficient for the average resistance and a second process tolerance coefficient for the welding energy are respectively set. Based on the mean, the standard deviation, the first process tolerance coefficient, and the second process tolerance coefficient, an initial two-dimensional parameter boundary is constructed; Based on the correlation characteristics between the average resistance and the welding energy, the initial two-dimensional parameter boundary is further optimized and determined to obtain the closed two-dimensional joint qualified region.
5. The method for identifying hidden weld defects in cylindrical battery bottom welding based on resistance and energy parameter correlation analysis according to claim 4, characterized in that, The values of both the first process tolerance coefficient and the second process tolerance coefficient are greater than or equal to 2; Furthermore, the shape of the closed two-dimensional joint qualified region can be any one of ellipse, rectangle or polygon.
6. The method for identifying hidden weld defects in cylindrical battery bottom welding based on resistance and energy parameter correlation analysis according to claim 5, characterized in that, When the closed two-dimensional joint qualified region is elliptical, its construction method is as follows: Extract the negative correlation feature between the average resistance and the welding energy; The coordinate point formed by the average value of the resistance and the average value of the welding energy is used as the center of the ellipse. The product of the first process tolerance coefficient and the standard deviation of the average resistance, and the product of the second process tolerance coefficient and the standard deviation of the welding energy, are respectively used as the reference axis lengths; Based on the negative correlation characteristics, the two-dimensional joint qualified region is established by fitting an ellipse equation.
7. The method for identifying hidden weld defects in cylindrical battery bottom welding based on resistance and energy parameter correlation analysis according to claim 6, characterized in that, In step S300, the average resistance and welding energy of the current battery under test welding point are acquired in real time, specifically including: After the single-point bottom welding cycle of the current battery under test is completed, and before the current battery under test is transferred to the next station; The average resistance and welding energy of the battery under test are read in real time from the spot welding equipment by the production line programmable logic controller or host computer through a data communication interface.
8. The method for identifying hidden weld defects in cylindrical battery bottom welding based on resistance and energy parameter correlation analysis according to claim 7, characterized in that, The formation of two-dimensional coordinate points specifically involves: Using either the real-time reading of the average resistance or the welding energy as the abscissa and the other as the ordinate, a two-dimensional coordinate point is constructed to characterize the coupling relationship between the current welding interface contact state and the total heat input of the system.
9. The method for identifying hidden weld defects in cylindrical battery bottom welding based on resistance and energy parameter correlation analysis according to claim 8, characterized in that, In step S400, determining whether the two-dimensional coordinate point falls within the two-dimensional joint qualified area specifically includes: Calculate the Mahalanobis distance from the two-dimensional coordinate point of the current battery under test to the center of the ellipse; The Mahalanobis distance is compared with a preset distance threshold to determine whether the two-dimensional coordinate point falls within the two-dimensional joint qualified area.
10. The method for identifying hidden weld defects in cylindrical battery bottom welding based on resistance and energy parameter correlation analysis according to claim 8, characterized in that, The preset distance threshold is 1; If the Mahalanobis distance is less than 1, then the two-dimensional coordinate point is determined to fall within the two-dimensional joint qualified area formed by the ellipse, and the current welding is qualified; If the Mahalanobis distance is greater than or equal to 1, it is determined that the two-dimensional coordinate point falls outside the two-dimensional joint qualified area formed by the ellipse, and there is a hidden risk of poor welding in the current welding. The trigger sorting signal specifically includes: A defective product rejection instruction is sent to the sorting execution mechanism of the production line to drive the push rod to remove the battery under test that is determined to have a hidden risk of poor soldering from the normal production flow channel.