Lithium battery production safety monitoring method and system

Abstract

The invention belongs to the technical field of safety monitoring, and particularly relates to a lithium battery production safety monitoring method and system.The lithium battery production safety monitoring method comprises the steps that in the liquid injection starting stage, whether liquid injection pressure and flow keep synchronous rising in the pressure increasing process or not is verified, in the pressure maintaining stage, the liquid injection pressure and flow meet a preset association mode, and if any one does not pass, it is judged that abnormity is generated and removed; entering a constant-pressure liquid injection stage, continuously analyzing a preset association mode between the liquid injection pressure and the instantaneous flow rate, if deviation is monitored, judging whether a blockage or leakage risk exists or not, if so, removing the risk, otherwise, entering an ending stage, and predicting whether the moment when the pressure is reduced to a zero point is consistent with a preset liquid injection stopping moment or not; according to the method, through dynamic relation monitoring, whole-process data analysis and self-adaptive parameter adjustment, the problems of static and one-sided analysis of an existing monitoring method are effectively solved.

Description

Technical Field

[0001] This invention belongs to the field of safety monitoring technology, specifically a method and system for monitoring the safety of lithium battery production. Background Technology

[0002] Lithium batteries are widely used in new energy vehicles and energy storage devices due to their high energy density and cycle life. With the increasing market demand, the production scale of lithium batteries continues to expand, and safety and product consistency in the production process have become core issues of concern to the industry.

[0003] In the lithium battery manufacturing process, the liquid injection process is a critical step. The accuracy of liquid injection pressure and flow control, as well as the consistency of liquid injection volume, directly affect the electrochemical performance, safety performance, and service life of the battery.

[0004] Currently, the safety monitoring methods for the lithium battery electrolyte filling process still have the following limitations: 1. Existing monitoring methods lack the ability to comprehensively analyze data from the entire electrolyte filling process. They only perform static detection based on pressure, flow rate, or electrolyte volume data at a single moment, failing to effectively capture the dynamic response relationship between pressure and flow rate during the electrolyte filling process. Therefore, it is difficult to identify potential abnormal risks caused by blockages, leaks, etc. in the early stages.

[0005] 2. In the final stage of injection, the existing technology fails to coordinate the pressure decay characteristics with the injection timing and liquid level requirements, which makes it impossible to predict and guarantee that the zero pressure point matches the injection completion time. This can easily lead to problems such as the liquid level not meeting the standard or over-injection when the injection is stopped. Summary of the Invention

[0006] To overcome the shortcomings of the prior art, this invention provides a method for monitoring the safety of lithium battery production, which can effectively solve the problems mentioned in the prior art.

[0007] The objective of this invention can be achieved through the following technical solutions: In the first aspect, this invention provides a method for monitoring the safety of lithium battery production, comprising: during the liquid injection start-up stage, controlling the liquid injection pressure to increase stepwise to a target value according to a preset strategy and maintaining it thereafter, verifying whether the liquid injection pressure and flow rate maintain a synchronous upward trend during the pressure increase process, and whether they conform to a preset correlation mode during the pressure maintenance process. If both verifications are successful, the constant pressure liquid injection stage is entered; otherwise, the liquid injection is determined to be abnormal and a rejection operation is performed.

[0008] During the constant pressure injection stage, the preset correlation pattern between injection pressure and instantaneous flow rate is continuously analyzed. If the two deviate from this pattern, the risk of blockage or leakage is determined based on the deviation, and a rejection operation is performed. Otherwise, the injection closure stage is entered.

[0009] In the end of the liquid injection stage, whether the time when the pressure drops to zero is consistent with the preset cut-off time of the liquid injection is predicted according to the pressure decay characteristics of the liquid injection process, and if the prediction is inconsistent, the pressure decay rate is controlled by adjusting the liquid injection parameters to make the time consistent.

[0010] The liquid injection amount of the completed battery is monitored, and corresponding adjustment operations are performed on the battery whose liquid injection amount does not meet the standard to make the liquid injection amount reach the set standard.

[0011] In the second aspect, the application further provides a lithium battery production safety monitoring system, comprising: a liquid injection start control module, a constant pressure liquid injection monitoring module, a liquid injection end adjustment module and a liquid injection amount verification module.

[0012] The liquid injection start control module is connected with the constant pressure liquid injection monitoring module, and the liquid injection end adjustment module is connected with the liquid injection amount verification module.

[0013] The liquid injection start control module controls the liquid injection pressure to increase to a target value according to a preset strategy in the liquid injection start stage, verifies whether the liquid injection pressure and flow rate maintain a synchronous rising trend in the pressure increase process, and whether they conform to a preset correlation mode in the pressure maintenance process, if both are verified, the constant pressure liquid injection stage is entered, otherwise, the liquid injection is determined to be abnormal, and the rejection operation is performed.

[0014] The constant pressure liquid injection monitoring module continuously analyzes the preset correlation mode between the liquid injection pressure and the instantaneous flow rate, if the two deviate from the mode, whether there is a risk of blockage or leakage is determined based on the deviation performance, if it is determined that there is, the rejection operation is performed, otherwise, the liquid injection end stage is entered.

[0015] The liquid injection end adjustment module predicts whether the time when the pressure drops to zero is consistent with the preset cut-off time of the liquid injection according to the pressure decay characteristics of the liquid injection process, and if the prediction is inconsistent, the pressure decay rate is controlled by adjusting the liquid injection parameters.

[0016] The liquid injection amount verification module monitors the liquid injection amount of the completed battery, and performs corresponding adjustment operations on the battery whose liquid injection amount does not meet the standard to make the liquid injection amount reach the set standard.

[0017] Compared with the prior art, the embodiments of the application have at least the following advantages or beneficial effects: (1) the liquid injection process is divided into three core stages of start, constant pressure and end, and specific monitoring, verification and control strategies are designed for each stage, the synchronous rising trend and the preset correlation mode of the pressure and flow rate are verified in the start stage, the real-time capture of the dynamic response relationship of the liquid injection process is realized, the pressure-flow dynamic coupling relationship is continuously analyzed in the constant pressure liquid injection stage, the blockage and leakage can be effectively identified and the rejection operation is performed, and in the end stage, the pressure zero point is matched with the cut-off time of the liquid injection by exponential decay fitting and PID dynamic adjustment.

[0018] (2) After the electrolyte injection is completed, the system detects whether the liquid level meets the standard in real time and automatically performs the operation of adding electrolyte or extracting excess electrolyte. Through closed-loop adjustment of the injection volume and early rejection of abnormalities, the pass rate of battery cell products is significantly improved. Attached Figure Description

[0019] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0021] Figure 2 This is a logic diagram for determining whether a blockage or leakage exists in this invention.

[0022] Figure 3 This is a flowchart illustrating the method for determining whether a preset association mode is satisfied according to the present invention.

[0023] Figure 4 This is a module connection diagram of the present invention. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some 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 are within the scope of protection of the present invention.

[0025] Reference Figure 1 As shown, in the first aspect of the present invention, a method for monitoring the safety of lithium battery production is provided, comprising: S1, in the liquid injection start-up stage, controlling the liquid injection pressure to increase stepwise to a target value according to a preset strategy and then maintaining it, verifying whether the liquid injection pressure and flow rate maintain a synchronous upward trend during the pressure increase process, and whether they conform to a preset correlation mode during the pressure maintenance process. If both are verified, the constant pressure liquid injection stage is entered; otherwise, the liquid injection is determined to be abnormal and a rejection operation is performed.

[0026] It should be noted that verifying whether the injection pressure and flow rate maintain a synchronous upward trend during the pressure increase process and whether the pressure maintenance process conforms to the preset correlation mode are prerequisites for the normal start of the injection process.

[0027] Considering that the electrolyte needs to gradually fill the gaps inside the battery during the initial stage of lithium battery electrolyte injection, the injection pressure drives the electrolyte flow, and the corresponding injection flow rate will increase with the increase of pressure. After the pressure reaches the target value and enters the maintenance process, the injection dynamics state should transition from transient to quasi-steady state. At this time, the pressure and flow rate in the system should follow the preset correlation mode calibrated by historical normal production data.

[0028] The aforementioned synchronous upward trend specifically refers to: real-time acquisition of injection flow rate values ​​during the pressure increase process and quantification of their rate of change relative to the previous sampling time.

[0029] Extract the pressure growth rate at each sampling time as specified by the preset strategy.

[0030] It should be noted that the injection pressure and flow rate are collected in real time by using a high-precision flow sensor and a pressure sensor installed on the main pipeline at the inlet of the injection nozzle.

[0031] It should also be noted that the above-mentioned preset strategy is based on the process requirements such as battery model and electrolyte viscosity. The pressure-flow-time correspondence is established through stepped pressure wetting characteristic tests. For example, for a specific battery model, the pressure step increase level is set to 5 levels, with the target pressure value of each level being 0.1MPa, 0.2MPa, 0.3MPa, 0.4MPa, and 0.5MPa. The pressure ramp rate of each level is 0.05MPa / s, and the residence time at each pressure level is 10s. The pressure-flow-time correspondence is established by collecting normal electrolyte injection data through multiple experiments.

[0032] This strategy includes, but is not limited to, setting the number of pressure increments, the target pressure value for each increment, the pressure ramp-up rate for each increment, and the dwell time at each pressure increment. Based on this, the pressure at each sampling moment can be extracted in real time, and the pressure growth rate can be calculated.

[0033] The injection flow rate at the corresponding sampling time is collected in real time by a mass flow meter installed on the main pipeline between the injection pump outlet and the battery injection port. Then, the rate of change of the flow rate relative to the previous sampling time is quantified by difference comparison-time normalization.

[0034] Specifically: a uniform time interval is set as the pressure data sampling time according to a preset strategy, and the injection flow rate is collected at the same sampling time. For the collected flow rate data, the flow rate values ​​of two adjacent samples are compared in chronological order, the difference between the flow rate of the latter and the flow rate of the former are calculated, and the difference between the adjacent flow rates is compared with the set time interval between two adjacent samples to obtain the flow rate change rate.

[0035] Calculate the correlation coefficient between the rate of increase of injection pressure and the rate of change of injection flow rate at each sampling time.

[0036] It is understandable that the injection pressure and flow rate increase synchronously. The correlation coefficient can be used to convert the relationship between the pressure growth rate and the flow rate change rate into a quantitative value, directly reflecting the consistency of their changing trends.

[0037] The correlation coefficient specifically uses the Pearson correlation coefficient to quantify the linear correlation between the pressure growth rate sequence and the flow rate change sequence. The Pearson correlation coefficient is existing technology and will not be elaborated here.

[0038] Verify that the injection pressure and flow rate maintain a synchronous upward trend during the pressure increase process based on any of the following conditions: I. The correlation coefficients at all sampling times during the pressure increase process are within the qualified correlation coefficient range defined by historical samples.

[0039] II. The number of sampling times outside the interval is less than the proportion of the total sampling times, and the maximum absolute deviation in the correlation coefficient outside the interval is less than or equal to the proportion of the interval span.

[0040] Understandably, when lithium battery electrolyte filling is in the startup phase and under normal operating conditions, the coordinated changes in electrolyte filling pressure and flow rate will follow the stable pattern verified by historical normal production, specifically manifested as I. However, during the production process, external factors may interfere with and cause II to occur. For example, temperature changes can cause fluctuations in electrolyte viscosity, and viscosity changes can affect the balance of fluid inertial and viscous forces, causing the pressure response to lag behind or lead the flow rate change. This results in the correlation coefficient between the pressure growth rate and the flow rate change rate at some sampling moments possibly temporarily exceeding the acceptable correlation coefficient range, but overall it is still in a stable upward trend.

[0041] The total sampling time ratio refers to the product of the total sampling time and the first preset ratio, and the interval span ratio refers to the product of the interval span and the second preset ratio. The determination of these two preset ratios is based on the statistical analysis of a large amount of historical normal production batch data, and the determination logic of the second preset ratio is the same as that of the first preset ratio. Taking the first preset ratio as an example, the specific determination process is as follows: Collect historical batch data that are judged to be normal under the same production conditions. For each historical normal batch, calculate the proportion of the number of sampling times in the measured flow sequence that exceed the preset qualified interval during the entire injection pressure maintenance process to the total number of sampling times. Statistically analyze the distribution characteristics of all historical batch data, and use the mode as the first preset ratio.

[0042] The common feature of the above-mentioned qualified correlation coefficient interval and the method for establishing the pressure-flow mapping function is that: by collecting historical normal lithium battery production batch data, production batches with a high degree of matching with the current production conditions are selected as historical samples. The production conditions include electrolyte viscosity, injection pipeline size and battery internal pore structure parameters.

[0043] It should be noted that the higher the viscosity of the electrolyte, the greater the internal friction. Under the same injection pressure, the higher the viscosity of the electrolyte, the smaller the injection flow rate.

[0044] The electrolyte injection line is the channel that transports the electrolyte from the equipment to the inside of the battery. Its size directly affects the resistance of the line. The smaller the inner diameter and the longer the length of the line, the greater the frictional resistance and local resistance when the electrolyte flows. To achieve the same flow rate, a higher injection pressure needs to be applied.

[0045] The internal pore structure parameters of a battery directly determine the filling resistance of the electrolyte inside the battery. The ultimate goal of electrolyte injection is to fill the electrode pores and separator pores inside the battery with electrolyte. The higher the porosity and the larger the pore size, the faster the electrolyte permeates inside the battery and the smaller the filling resistance. Under the same injection pressure, the electrolyte can fill the pores faster, and the corresponding flow rate decay trend is more gradual.

[0046] In summary, selecting parameters that meet the requirements of electrolyte viscosity, injection pipeline size, and internal pore structure of the battery is essential to ensure that the subsequent construction criteria conform to the normal state of the current process.

[0047] In this embodiment, the quantification process for high matching degree is as follows: First, the electrolyte viscosity, injection pipeline diameter and porosity of each production batch are constructed into a three-dimensional feature vector.

[0048] Secondly, to eliminate the influence of different parameter dimensions and orders of magnitude, all feature vectors in the historical dataset need to be normalized before calculation so that the values ​​of each dimension are within the same comparable range.

[0049] Next, the target production conditions of the current batch are defined as the first vector, and the production conditions of a certain batch in the historical batch library are defined as the second vector.

[0050] Finally, the cosine similarity between the two vectors is calculated. The cosine similarity is calculated in the range of (-1, 1). The closer the value is to 1, the more consistent the production conditions of the two batches are in the multidimensional space, that is, the higher the matching degree. For example, in this scheme, based on the clustering analysis results of historical normal batch data, when the cosine similarity is greater than or equal to 0.95, the difference in process parameters between batches is less than 5%. Then the similarity threshold can be set to 0.95, and all batches with a cosine similarity greater than or equal to 0.95 with the current batch condition vector are selected to form the final high-matching historical sample.

[0051] The high-qualification correlation coefficient interval is constructed by extracting the correlation coefficients between the pressure growth rate and the flow change rate during the growth process in the startup phase of historical samples, constructing a correlation coefficient set, dividing the historical samples into preset percentiles, and defining the interval between the preset percentiles as the qualified correlation coefficient interval.

[0052] The preset percentile division is specifically as follows: all values ​​in the correlation coefficient set are arranged in ascending order, outliers are removed by box plot method, and the qualified correlation coefficient interval covers 90% of the normal interval in historical normal samples. Two percentiles are preset as the upper and lower boundaries of the interval. For example, the 5th percentile is set as the lower limit and the 95th percentile is set as the upper limit. The range of values ​​between the 5th percentile and the 95th percentile is defined as the qualified correlation coefficient interval.

[0053] It should also be noted that implementers can customize preset percentiles according to the production requirements for injection stability.

[0054] Reference Figure 3 As shown in a preferred embodiment of the present invention, the preset association mode meets the following conditions: real-time acquisition of injection flow rate values ​​during pressure maintenance to construct a measured flow rate sequence.

[0055] Extract the target pressure value at the same sampling time from the preset strategy, and calculate the expected flow sequence through the preset pressure-flow mapping function calibrated by historical samples.

[0056] The pressure-flow mapping function described above is established using a statistical regression fitting method. Taking linear regression fitting as an example, a univariate linear regression function is constructed as follows: .

[0057] in For the expected traffic, Injection pressure, The slope This is the intercept.

[0058] The coefficients a and intercept b can be solved using the least squares method. The mean injection pressure, mean injection flow rate, covariance of injection pressure and flow rate, and pressure variance can be directly calculated using Python's statistical library.

[0059] The slope The ratio of covariance to force variance, and the intercept. Mean injection flow rate and slope The difference between the product of the injection pressure and the mean injection pressure.

[0060] It should also be noted that if the relationship between pressure and flow shows a rapid growth trend, a nonlinear function should be considered for mapping. For example, an exponential function can be used, which uses a nonlinear least squares fitting algorithm to directly regress the original data and solve for the undetermined parameters in the function.

[0061] The injection pressure at each sampling time is collected, and the corresponding injection flow rate is calculated using a preset pressure-flow rate mapping function to construct the expected flow rate sequence.

[0062] With the sampling time as the horizontal axis and the flow rate as the vertical axis, the discrete points in the expected flow rate sequence and the measured flow rate sequence are connected in chronological order to construct the expected flow rate curve and the measured flow rate curve, respectively.

[0063] Calculate the integral value of the absolute deviation between the expected flow rate curve and the measured flow rate curve during the pressure maintenance process. If the integral value of the absolute deviation is less than or equal to the preset allowable deviation value, it is determined that the preset correlation mode is met.

[0064] The specific formula for the integral value of the absolute deviation is: .

[0065] in, The integral value of the absolute deviation. It is a continuous function of the measured flow rate curve. It is a continuous function of the expected flow rate curve. , These represent the start and end times of the pressure maintenance phase, respectively.

[0066] This formula quantifies the cumulative effect of deviation over the entire time period, avoiding the judgment flaw of only looking at a single point and ignoring the whole process.

[0067] The preset allowable deviation value is pre-calibrated through experiments based on the electrolyte characteristics and injection accuracy requirements. The experimental process includes: selecting key variables, such as injection speed gradient, ambient temperature gradient, and a set of candidate preset allowable deviation values; performing multiple injection operations under each variable combination; and accurately recording the actual injection volume each time.

[0068] The batteries that have completed electrolyte injection are subjected to electrochemical performance tests to establish performance qualification standards. Subsequently, the proportion of electrolyte injection operations that can simultaneously meet the final actual electrolyte injection accuracy requirements and battery performance qualification standards under each candidate preset allowable deviation value is counted as the comprehensive qualification rate of that candidate value.

[0069] Compare the overall pass rates corresponding to all candidate preset allowable deviation values, and select the candidate value with the largest value among all candidate values ​​that can ensure that the overall pass rate is not lower than the preset pass rate threshold, such as 95%, and mark it as the preset allowable deviation value.

[0070] In a preferred embodiment of the present invention, the specific process of determining the injection abnormality and performing the rejection operation is as follows: extract the data sequence of injection pressure and injection flow rate during the abnormality determination period as the sequence to be determined.

[0071] The specific anomaly determination period is as follows: during the pressure increase process, the sampling time when the correlation coefficient first exceeds the qualified correlation coefficient range is taken as the starting point of the anomaly determination period, and the natural end point of the pressure increase process is taken as the end point of the anomaly determination period.

[0072] Considering that the moment when the correlation coefficient first exceeds the range during the pressure increase process in the lithium battery electrolyte injection start-up stage is the critical node when the electrolyte injection process enters an abnormal deviation state from the normal coordinated state, taking the moment when the first exceeds the range as the starting point and the natural end point of the pressure increase process as the end point, it can completely cover the entire process of abnormal initiation-abnormal development-abnormal termination, avoiding the loss of data in the initial stage of the abnormality due to the starting point being too late, and the inability to trace the initial parameter change characteristics of the fault.

[0073] During the pressure maintenance process, the absolute deviation integral value is continuously calculated according to a preset period. The starting point of the period in which the integral value first exceeds the preset allowable deviation value is taken as the starting point of the abnormal judgment period, and the natural end point of the pressure maintenance process is taken as the end point of the abnormal judgment period.

[0074] Considering that flow fluctuations may be intermittent or gradual during the pressure maintenance phase, the integral value of absolute deviation is continuously calculated according to a preset period. This method captures such implicit deviations by accumulating over time rather than by sampling at a single point, ensuring that the integral value reflects the cumulative effect of the deviation within that time period.

[0075] For example, if the deviation gradually increases in the first 5 seconds of the cycle, and the integral value exceeds the limit only in the 8th second, if the starting point is set to the 8th second, the cumulative deviation in the first 5 seconds will be ignored. Therefore, taking the starting point of the cycle in which the integral value first exceeds the preset allowable deviation value as the starting point of the abnormal judgment period can avoid missing the deviation change data in the early stage of the cycle. Taking the natural end point of the pressure maintenance process as the end point of the abnormal judgment period can ensure that the data sequence to be judged is in a process environment with constant pressure throughout the entire process.

[0076] It should be noted that the implementer can set the cycle period to a preset cycle according to the cycle period of the liquid pump speed fine adjustment or electrolyte replenishment operation. This can help to quickly locate the cause of the abnormality. For example, if electrolyte replenishment is performed at the beginning of a certain cycle and the integral value of that cycle exceeds the standard for the first time, it can be preliminarily checked whether the deviation is caused by the fluctuation of electrolyte viscosity after replenishment.

[0077] Calculate the dynamic time-normalized distance between the sequence to be judged and each typical abnormal sequence in the preset abnormal database.

[0078] It should be noted that the preset abnormal database is constructed through the following steps: First, data sequences of injection pressure and injection flow rate during historical lithium battery injection abnormal processes are collected. Each sequence corresponds to a known abnormal type, such as leakage or sensor failure, and the label of the abnormal type is recorded.

[0079] Then, sequence processing is performed, and the collected data sequence is preprocessed, including data cleaning, standardization, and time alignment, to eliminate noise and the influence of units, and to ensure the consistency of the sequence.

[0080] Then, anomaly classification is performed, and the processed data sequence is divided into different typical anomaly sequences according to the anomaly type label.

[0081] Finally, a database is generated, storing the mapping relationship between the typical anomaly categories and their corresponding typical anomaly sequences to form the preset anomaly database.

[0082] If the dynamic time warp distance is less than or equal to the preset distance threshold, the anomaly type is determined to be consistent with the corresponding typical case, and the frequency of occurrence of this type of anomaly and the associated injection equipment parameters are recorded.

[0083] The calculation logic of the above dynamic time warping distance is as follows: construct the cumulative distance matrix of the sequence to be judged and the typical abnormal sequence, and find the optimal curved path through the cumulative distance matrix. This path can minimize the cumulative distance between the two sequences after warping. This minimum cumulative distance is the dynamic time warping distance.

[0084] The preset distance threshold can be specifically determined through experimental calibration. Specifically, historical abnormal data is collected, and the dynamic time-normalized distance between the data and typical sequences of each category in the preset abnormal database is calculated. The distances calculated with typical sequences of the correct category are assigned to the same category distance set, and the distances calculated with typical sequences of the wrong category are assigned to the opposite category distance set.

[0085] Arrange the values ​​in the same-type distance set and the different-type distance set in ascending order, and take the maximum value of the same-type distance set and the minimum value of the different-type distance set. The interval formed by these two values ​​is the safe interval.

[0086] Take the value at one-quarter of the safe interval as the initial threshold, and then use the initial threshold to judge a batch of verification data with known abnormal labels. If the false negative rate is higher than the process allowable value, such as greater than 5%, it means that the initial threshold is too high and needs to be appropriately reduced. If the false positive rate is higher than the process allowable value, such as greater than 5%, it means that the initial threshold is too low and needs to be appropriately increased. Based on the above principles, iterative adjustments are made repeatedly until the false positive rate and false negative rate meet the process requirements. The threshold at this time is the preset distance threshold.

[0087] If the dynamic time warp distance is greater than the preset distance threshold, the abnormal data sequence will be marked as a new type of anomaly. After manual confirmation, its feature data will be included in the preset anomaly database.

[0088] Reference Figure 2 As shown in Figure S2, during the constant pressure injection stage, the preset correlation pattern between injection pressure and instantaneous flow rate is continuously analyzed. If the two deviate from this pattern, the risk of blockage or leakage is determined based on the deviation. If the risk is determined to be present, a rejection operation is performed; otherwise, the injection closure stage is entered.

[0089] Given that the injection process follows the basic laws of fluid mechanics during the constant pressure injection stage, whether the injection is normal will be directly reflected in the dynamic changes of flow resistance and the attenuation characteristics of the flow rate. Therefore, the specific process for determining whether there is a blockage or leakage due to deviation is as follows: Based on the injection pressure and injection flow rate sequence, the instantaneous flow resistance is calculated. The instantaneous flow resistance is the ratio of the injection pressure to the injection flow rate at the same moment, and an instantaneous flow resistance sequence is formed.

[0090] The instantaneous flow resistance sequence is linearly fitted to obtain the first fitting slope.

[0091] The first and second derivatives of the instantaneous injection flow rate sequence are calculated to form the decay rate sequence and decay acceleration sequence, respectively.

[0092] It should be noted that the first derivative reflects the instantaneous rate of change of flow rate over time, i.e., the decay rate. If the derivative is negative, it indicates a decrease in flow rate, which is consistent with the flow rate change characteristics during blockage or leakage. The first derivatives at the beginning and end of the instantaneous injection flow rate sequence are calculated using the forward difference method and the backward difference method, and the other first derivatives are calculated using the central difference method. Arranging the first derivatives at all times in chronological order constitutes the decay rate sequence. The calculation here is existing technology and will not be elaborated upon here.

[0093] The second derivative reflects the rate of change of the decay rate over time, i.e., the decay acceleration. If the second derivative is positive, it means that the decay rate is decreasing, such as when the flow rate decreases more slowly during leakage. If it is negative, it means that the decay rate is increasing, such as when the flow rate decreases more quickly during blockage. Then, based on the decay rate sequence, the second derivative is calculated using the same logic as the first derivative calculation method, and a decay acceleration sequence is constructed.

[0094] The absolute values ​​of the decay rate sequence are linearly fitted to obtain a second fitting slope, and the calculation logic of the second fitting slope is the same as that of the first fitting slope.

[0095] If the first fitting slope is positive and the second fitting slope is negative, and the average flow attenuation acceleration is greater than zero, then a leak is determined to exist.

[0096] If the first fitting slope is positive, the second fitting slope is positive, and the average flow attenuation acceleration is less than zero, then a blockage is determined to exist.

[0097] Considering that both leakage and blockage will reduce the effective cross-section of the electrolyte, resulting in decreased flow rate and increased flow resistance, the slope of the first fitting is a common premise for determining the existence of a fault.

[0098] Once the fault is confirmed, leakage and blockage can be distinguished by analyzing the dynamic characteristics of flow decay. Specifically, if leakage occurs, since the gap at the leak point is usually of a fixed size, as the injection progresses, the flow state gradually stabilizes, and the rate of flow decrease will gradually slow down. That is, the absolute value of the flow decay rate decreases over time, so the second fitting slope is negative. This is one of the key characteristics that distinguishes leakage faults from blockage faults.

[0099] Secondly, the rate of flow decay gradually slows down during leakage, and the change in the rate of decay is positive. For example, if the rate of decay changes from -0.5 mL / s² to -0.3 mL / s², the rate of decay is +0.2 mL / s³. Taking the average of the flow decay acceleration over the entire period, the result is still greater than zero. This indicator mathematically verifies the trend of increasingly slower flow decay, further pointing to a leakage fault.

[0100] In summary, if all the above conditions are met, it can be determined that a leak exists.

[0101] If a blockage occurs, such as if there are impurities in the electrolyte, the impurities may accumulate further as the electrolyte is continuously injected, causing the cross-section of the injection channel to continuously decrease, the electrolyte flow resistance to continuously increase, and the absolute value of the flow rate decay rate to increase over time. This eventually leads to a positive second fitting slope, and the average value of the flow rate decay acceleration over the entire time period will definitely be less than zero. Therefore, if the above conditions are met, it can be determined that a blockage exists.

[0102] S3. In the final stage of injection, based on the pressure decay characteristics of the injection process, predict whether the time when the pressure drops to zero is consistent with the preset injection cutoff time. If the prediction is inconsistent, control the pressure decay rate by adjusting the injection parameters.

[0103] Considering that the time when the pressure drops to zero often deviates from the preset injection cutoff time due to actual factors such as injection leakage, pipeline residual pressure, or injection valve response lag, it is necessary to predict in real time whether the time when the pressure drops to zero is consistent with the preset injection cutoff time. The specific process is as follows: continuously add the injection pressure value collected in real time at the end of the injection stage to the constructed historical pressure sequence, and perform exponential decay fitting based on its decay trend to obtain the instantaneous pressure decay time constant.

[0104] The historical pressure sequence is an initial data sequence formed by using a preset pressure acquisition device to collect pressure values ​​in real time from the injection pipeline or battery injection port in the early stage of the lithium battery liquid injection process, based on the process rhythm of each stage of liquid injection and matching the acquisition frequency. The data is then continuously stored in a correlation format of timestamp and pressure value.

[0105] It is understandable that, in the final stage of injection, after the injection pump stops supplying liquid, the residual pressure in the system gradually decreases as the fluid is drained / permeates, and this decrease process conforms to the exponential change law of pressure release in a closed system in fluid mechanics. Therefore, exponential decay fitting is adopted. The specific steps are as follows: real-time acquisition of pressure data from the moment the injection pump stops until the system pressure drops to the safe threshold during the final stage.

[0106] The real-time collected pressure values ​​are continuously added to the constructed historical pressure sequence to form a complete pressure-time dataset for the final stage. According to fluid mechanics theory, the pressure decay of a closed pipeline system follows a first-order exponential decay model when there is no external pressure input. The core formula is: .

[0107] in For a moment System pressure, The initial pressure at the instant the injection pump stops. The time after the injection pump stops. is the pressure decay time constant.

[0108] In the above formula, The exponential factor representing pressure decay over time. The decrease due to the increase reflects the rate of decrease in system pressure. The pressure decay time constant is the value of which indicates that the pressure decays more slowly and the system has a stronger ability to maintain pressure. Conversely, the smaller the value, the faster the pressure decays.

[0109] The entire formula describes the system pressure from its initial value when there is no external pressure input. A dynamic process that decays exponentially to a stable state.

[0110] Since the exponential function is a nonlinear function, its parameters need to be solved through data linearization or nonlinear least squares method. In this embodiment, the nonlinear least squares method is used. Specifically, the optimization objective is to minimize the sum of squared residuals between the pressure predicted by the exponential function and the actual collected pressure. The pressure-time dataset of the final stage and the initial parameters of the exponential function are input, and the parameters are adjusted iteratively. The value of is taken until the sum of squared residuals converges to a minimum. If the residual is less than 0.001 MPa², then is extracted. The value of is used as the instantaneous pressure decay time constant.

[0111] Based on the current pressure value and the instantaneous decay time constant, the above exponential decay function is substituted to obtain the remaining time from the current time until the pressure drops to zero. Then, the remaining time is summed with the current time to obtain the predicted time when the pressure reaches zero.

[0112] Calculate the time deviation between the predicted zero pressure point and the preset injection cutoff point.

[0113] If the time deviation remains below the process tolerance for a preset judgment period, the time when the pressure drops to zero is determined to be consistent with the preset injection cutoff time; otherwise, they are determined to be inconsistent.

[0114] The aforementioned process tolerance is determined based on the response delay characteristics of the liquid injection system. Specifically, by collecting full-process data on multiple batches of lithium battery liquid injection processes, the actual time deviation between the pressure drop to zero and the preset liquid injection cutoff time is recorded. The standard deviation of these deviation data is calculated, and a value that can cover the range of most normal process fluctuations is selected as the process tolerance.

[0115] The preset judgment time is determined by monitoring the fluctuation pattern of time deviation during the liquid injection process of multiple lithium batteries, and statistically analyzing the period from the occurrence of fluctuation to the stabilization of time deviation. The preset judgment time must cover this period to avoid misjudgment due to the judgment time being too short.

[0116] In actual production, the pressure decay process is affected by factors such as equipment precision and electrolyte viscosity, resulting in unavoidable minor fluctuations. The preset judgment time and process tolerance settings are intended to allow for the existence of such normal process fluctuations. As long as the time deviation remains within the tolerance range for a reasonable duration, the time is judged to be consistent, avoiding misjudgment of inconsistency due to minor fluctuations, thereby reducing unnecessary process intervention and improving production efficiency.

[0117] In a preferred embodiment of the present invention, the specific method for adjusting the injection parameters to control the pressure decay rate is as follows: real-time calculation of the time deviation between the predicted pressure zero point and the preset injection cutoff time.

[0118] Based on the time deviation, the absolute value of the instantaneous pressure decay rate at the current moment, and the difference between the current moment and the preset injection cutoff moment, the gain parameter of the proportional-integral-derivative controller is dynamically adjusted, and after each adjustment, it is determined whether the time deviation has converged. If it has not converged, the process switches to the next gain adjustment stage.

[0119] The dynamic adjustment process uses fuzzy rules to fuzzify the three input variables: time deviation, decay rate, and remaining time. The adjustment amount of each gain is obtained through reasoning from the fuzzy rule base, and then output to the PID controller after defuzzification.

[0120] The fuzzy rule base stores fuzzy sets of input variables, fuzzy sets of output variables, membership functions, and fuzzy mapping rules.

[0121] The fuzzy set of input variables includes the absolute value of time deviation, the absolute value of pressure decay rate, and the difference between the current time and the preset injection cutoff time.

[0122] The output variable fuzzy set includes proportional gain adjustment parameters, integral gain adjustment parameters, and derivative gain adjustment parameters.

[0123] The membership function is used to convert between precise input and output values ​​and fuzzy semantic values.

[0124] The fuzzy mapping rule defines the output gain adjustment amount to be taken under different input conditions.

[0125] The establishment of the fuzzy rule base specifically includes: first, fuzzifying the input variables, defining their fuzzy sets and membership functions, dividing the absolute value of the time deviation and the absolute value of the pressure decay rate into three fuzzy levels: small, medium, and large, and dividing the difference between the current time and the preset injection cutoff time into three fuzzy levels: short, medium, and long.

[0126] The membership function adopts a triangular membership function. This method forms a triangular membership distribution by setting a vertex with a membership degree of 1 and two adjacent endpoints with a membership degree of 0 for each fuzzy set within a defined universe of discourse.

[0127] Next, a fuzzy set of output variables is defined, dividing the proportional gain adjustment parameter into negative large, negative small, zero, positive small, and positive large; the integral gain adjustment parameter into negative large, zero, and positive large; and the differential gain adjustment parameter into negative small, zero, positive small, and positive large.

[0128] Then, fuzzy mapping rules are established. Based on the correspondence between the input and output fuzzy sets, the output gain adjustment amount corresponding to each input condition is set in the form of IF-THEN rules. For example, if the time deviation is large, the pressure decay rate is small, and the remaining time is long, then the proportional gain adjustment parameter is given a positive value.

[0129] Finally, the defined input and output fuzzy sets, membership functions, and all fuzzy mapping rules are systematically stored to generate the fuzzy rule library.

[0130] The time deviation is input into the dynamically adjusted proportional-integral-derivative controller, which outputs the adjustment amount of the pressure relief valve opening.

[0131] The above adjustment operation is repeated until the time deviation remains below the process tolerance for a preset time period.

[0132] The specific logic of the above repeated adjustment operation is as follows: First, the proportional gain is dynamically adjusted, and after each adjustment cycle, it is determined whether the time deviation is continuously lower than the process tolerance within the preset judgment time. If the condition is met, the adjustment is exited; otherwise, the integral gain adjustment stage is entered, and the above judgment process is repeated. If the condition is still not met, the derivative gain adjustment is further started until the time deviation meets the requirements or the number of repetitions reaches the maximum number of adjustment times.

[0133] This invention clearly divides the injection process into three core stages: start-up, constant pressure, and finish-up. Targeted monitoring, verification, and control strategies are designed for each stage. During the start-up stage, the synchronous upward trend of pressure and flow rate and the preset correlation pattern are verified. During the constant pressure injection stage, the dynamic coupling relationship between pressure and flow rate is continuously analyzed, effectively identifying blockages and leaks and performing removal operations. In the finish-up stage, exponential decay fitting and PID dynamic adjustment ensure that the pressure zero point matches the injection cutoff time.

[0134] Understandably, the accuracy of lithium battery electrolyte injection directly affects battery performance and safety. Time deviation reflects the degree of deviation between the zero pressure point and the preset cutoff time, the absolute value of the instantaneous pressure decay rate reflects the speed of pressure change, and the difference between the current time and the preset cutoff time defines the adjustment time window. Based on these three parameters, the PID gain can be dynamically adjusted to precisely control the opening of the pressure relief valve, thereby strictly controlling the electrolyte injection cutoff time and ensuring that the accuracy of the electrolyte injection meets the process standards.

[0135] S4. Monitor the electrolyte level of the battery after electrolyte injection. For batteries whose electrolyte level is below the standard, perform corresponding adjustment operations to bring the electrolyte level up to the set standard.

[0136] The deviation in the injection volume is mainly caused by the equipment precision and electrolyte viscosity. In order to accurately control the electrolyte content, the adjustment operation needs to be carried out in a targeted manner according to the direction and degree of the injection volume deviation. The specific adjustment operation is as follows: if the actual injection volume is less than the standard injection volume, it is determined that the injection volume has not met the standard, and an electrolyte replenishment operation is performed.

[0137] If the actual injected volume is greater than the standard injected volume, it is determined that the injected volume exceeds the standard, and the excess electrolyte is extracted.

[0138] The electrolyte replenishment or extraction operation is performed according to the absolute deviation between the actual injected volume and the standard injected volume.

[0139] After the electrolyte injection is completed, the system detects whether the electrolyte level meets the standard in real time and automatically performs operations to add electrolyte or extract excess electrolyte, which significantly improves the pass rate of battery cell products.

[0140] See Figure 4As shown, a second aspect of the present invention provides an automatic control system for photovoltaic power generation, comprising: a liquid injection start-up control module, a constant pressure liquid injection monitoring module, a liquid injection end-of-pipe adjustment module, and a liquid injection volume verification module.

[0141] The injection start control module is connected to the constant pressure injection monitoring module, and the injection end adjustment module is connected to the injection volume verification module.

[0142] The injection start control module controls the injection pressure to increase stepwise to the target value and maintain it during the injection start phase according to the preset strategy. It verifies whether the injection pressure and flow rate maintain a synchronous upward trend during the pressure increase and whether they conform to the preset correlation mode during the pressure maintenance. If both are verified, the constant pressure injection phase is entered; otherwise, the injection is judged to be abnormal and a rejection operation is performed.

[0143] The constant pressure injection monitoring module continuously analyzes the preset correlation pattern between injection pressure and instantaneous flow rate. If the two deviate from this pattern, it determines whether there is a risk of blockage or leakage based on the deviation. If the risk is found, it performs a rejection operation; otherwise, it enters the injection completion stage.

[0144] The injection termination adjustment module predicts whether the time when the pressure drops to zero is consistent with the preset injection termination time based on the pressure decay characteristics of the injection process. If the prediction is inconsistent, the pressure decay rate is controlled by adjusting the injection parameters.

[0145] The electrolyte injection volume verification module monitors the electrolyte injection volume of batteries that have completed the injection process. For batteries whose electrolyte injection volume does not meet the standard, the module performs corresponding adjustment operations to bring the electrolyte injection volume up to the set standard.

[0146] The above description is merely an example and illustration of the structure of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the structure of the invention or exceed the scope defined by the present invention, they should all fall within the protection scope of the present invention.

Claims

1. A method for monitoring the safety of lithium battery production, characterized in that, include: During the injection start-up phase, the injection pressure is controlled to increase stepwise to the target value according to the preset strategy and then maintained. It is verified whether the injection pressure and flow rate maintain a synchronous upward trend during the pressure increase process and whether they conform to the preset correlation mode during the pressure maintenance process. If both are verified, the constant pressure injection phase is entered; otherwise, the injection is judged to be abnormal and a rejection operation is performed. During the constant pressure injection stage, the preset correlation pattern between injection pressure and instantaneous flow rate is continuously analyzed. If the two deviate from this pattern, the risk of blockage or leakage is determined based on the deviation. If the risk is determined to be present, a rejection operation is performed; otherwise, the injection closure stage is entered. During the final stage of injection, based on the pressure decay characteristics of the injection process, it is predicted whether the time when the pressure drops to zero is consistent with the preset injection cutoff time. If the prediction is inconsistent, the pressure decay rate is controlled by adjusting the injection parameters. Monitor the electrolyte level of batteries that have been filled with electrolyte. For batteries whose electrolyte level is below the standard, perform corresponding adjustment operations to bring the electrolyte level up to the set standard.

2. The lithium battery production safety monitoring method according to claim 1, characterized in that: The aforementioned synchronous upward trend specifically refers to: The injection flow rate was collected in real time during the pressure increase process and its rate of change relative to the previous sampling time was quantified. Extract the pressure growth rate at each sampling time as specified by the preset strategy; Calculate the correlation coefficients between the rate of increase in injection pressure and the rate of change in injection flow rate at each sampling time; Verify that the injection pressure and flow rate maintain a synchronous upward trend during the pressure increase process, based on any of the following conditions: I. The correlation coefficients at all sampling points during the pressure increase process are within the acceptable correlation coefficient range defined by historical samples; II. The number of sampling times outside the interval is less than the proportion of the total sampling times, and the maximum absolute deviation in the correlation coefficient outside the interval is less than or equal to the proportion of the interval span.

3. The method for monitoring the safety of lithium battery production according to claim 2, characterized in that: The specific conditions for meeting the preset association pattern are as follows: The injection flow rate was collected in real time during the pressure maintenance process to construct a measured flow rate sequence; Extract the target pressure value at the same sampling time from the preset strategy, and calculate the expected flow sequence through the preset pressure-flow mapping function calibrated by historical samples; Connect the discrete points in the expected flow sequence and the measured flow sequence in chronological order to construct the expected flow curve and the measured flow curve, respectively. Calculate the integral value of the absolute deviation between the expected flow rate curve and the measured flow rate curve during the pressure maintenance process. If the integral value of the absolute deviation is less than or equal to the preset allowable deviation value, it is determined that the preset correlation mode is met.

4. The lithium battery production safety monitoring method according to claim 3, characterized in that: The method for establishing the qualified correlation coefficient interval and the pressure-flow mapping function includes: Collect historical normal lithium battery production batch data, and screen production batches that have a high degree of matching with the current production conditions as historical samples. The production conditions include electrolyte viscosity, injection pipeline size and battery internal pore structure parameters. Extract the correlation coefficients between the pressure growth rate and the flow change rate during the growth process in the startup phase from historical samples, construct a set of correlation coefficients, divide the historical samples into preset percentiles, and define the intervals between the preset percentiles as qualified correlation coefficient intervals. For the pressure maintenance process, data pairs consisting of injection pressure and injection flow rate at the same sampling time are extracted, and the pressure-flow rate mapping function is obtained through statistical regression fitting.

5. The method for monitoring the safety of lithium battery production according to claim 1, characterized in that: The specific process for determining abnormal injection and performing a rejection operation is as follows: Extract the data sequence of injection pressure and injection flow rate during the anomaly detection period as the sequence to be detected; Calculate the dynamic time-normalized distance between the sequence to be judged and each typical abnormal sequence in the preset abnormal database; If the dynamic time warp distance is less than or equal to the preset distance threshold, the anomaly type is determined to be consistent with the corresponding typical case, and the frequency of occurrence of this type of anomaly and the associated injection equipment parameters are recorded. If the dynamic time warp distance is greater than the preset distance threshold, the abnormal data sequence is marked as a new type of abnormality and, after manual confirmation, is included in the preset abnormality database.

6. The method for monitoring the safety of lithium battery production according to claim 5, characterized in that: The specific time period for anomaly detection is: During the pressure increase process, the sampling time when the correlation coefficient first exceeds the qualified correlation coefficient range will be taken as the starting point of the anomaly judgment period, and the end time of the pressure increase process will be taken as the ending point of the anomaly judgment period. During the pressure maintenance process, the absolute deviation integral value is continuously calculated according to a preset cycle. The starting point of the cycle in which the integral value first exceeds the preset allowable deviation value is taken as the starting point of the abnormal judgment period, and the end time of the pressure maintenance process is taken as the end point of the abnormal judgment period.

7. The method for monitoring the safety of lithium battery production according to claim 1, characterized in that: The specific process for determining whether a blockage or leakage exists based on the aforementioned deviation is as follows: Based on the injection pressure and injection flow rate sequence, the instantaneous flow resistance is calculated and an instantaneous flow resistance sequence is constructed; A first fitting slope is obtained by linearly fitting the instantaneous flow resistance sequence. The first and second derivatives of the instantaneous injection flow rate sequence are calculated to form the decay rate sequence and decay acceleration sequence, respectively. A second fitting slope is obtained by linearly fitting the absolute values ​​of the decay rate sequence. If the first fitting slope is positive and the second fitting slope is negative, and the average flow attenuation acceleration is greater than zero, then a leak is determined to exist. If the first fitting slope is positive, the second fitting slope is positive, and the average flow attenuation acceleration is less than zero, then a blockage is determined to exist.

8. The method for monitoring the safety of lithium battery production according to claim 1, characterized in that: The specific process for determining whether the predicted pressure drop to zero coincides with the preset injection cutoff time is as follows: The injection pressure value at the end of the injection phase is continuously added to the constructed historical pressure sequence. Based on its decay trend, exponential decay fitting is performed to obtain the instantaneous pressure decay time constant. Based on the current pressure value and the instantaneous decay time constant, the remaining time required for the pressure to decay to zero is predicted, thus obtaining the predicted zero pressure time. Calculate the time deviation between the predicted pressure zero point and the preset injection cutoff time; If the time deviation remains below the process tolerance for a preset judgment period, the time when the pressure drops to zero is determined to be consistent with the preset injection cutoff time; otherwise, they are determined to be inconsistent.

9. A method for monitoring the safety of lithium battery production according to claim 8, characterized in that: The specific method for adjusting the injection parameters to control the pressure decay rate is as follows: Real-time calculation of the time deviation between the predicted zero pressure point and the preset injection cutoff point; Based on the time deviation, the absolute value of the instantaneous pressure decay rate at the current moment, and the difference between the current moment and the preset injection cutoff moment, the gain parameter of the proportional-integral-derivative controller is dynamically adjusted, and after each adjustment, it is determined whether the time deviation has converged. If it has not converged, the process switches to the next gain adjustment stage. The time deviation is input into the dynamically adjusted proportional-integral-derivative controller, which outputs the adjustment amount of the pressure relief valve opening. The above adjustment operation is repeated until the time deviation remains below the process tolerance for a preset time period or the number of repetitions reaches the maximum number of adjustment times.

10. A lithium battery production safety monitoring system, characterized in that: include: The injection start control module controls the injection pressure to increase stepwise to the target value and maintain it during the injection start phase according to the preset strategy. It verifies whether the injection pressure and flow rate maintain a synchronous upward trend during the pressure increase and whether they conform to the preset correlation mode during the pressure maintenance. If both are verified, the constant pressure injection phase is entered; otherwise, the injection is judged to be abnormal and a rejection operation is performed. The constant pressure injection monitoring module continuously analyzes the preset correlation pattern between injection pressure and instantaneous flow rate. If the two deviate from this pattern, it determines whether there is a risk of blockage or leakage based on the deviation. If it is determined that there is, it performs a rejection operation; otherwise, it enters the injection completion stage. The injection termination adjustment module predicts whether the time when the pressure drops to zero is consistent with the preset injection termination time based on the pressure decay characteristics of the injection process. If the prediction is inconsistent, the pressure decay rate is controlled by adjusting the injection parameters. The electrolyte injection volume verification module monitors the electrolyte injection volume of batteries that have completed the injection process. For batteries whose electrolyte injection volume does not meet the standard, the module performs corresponding adjustment operations to bring the electrolyte injection volume up to the set standard.

Images