A lithium battery production safety monitoring method and system

Abstract

The present application belongs to the technical field of safety monitoring, and specifically relates to a lithium battery production safety monitoring method and system, which verifies whether the injection pressure and flow rate keep synchronous rising in the pressure growth process in the injection starting stage and whether the pressure maintenance stage conforms to the preset correlation mode, determines an abnormality and removes the abnormality if any of the two conditions fails, enters the constant pressure injection stage after passing the verification, continuously analyzes the preset correlation mode between the injection pressure and the instantaneous flow rate, determines whether there is a risk of blockage or leakage if a deviation is monitored, removes the abnormality if the risk exists, otherwise, enters the finishing stage, predicts whether the time when the pressure drops to zero is consistent with the preset injection cutoff time, dynamically adjusts the injection parameters if the two times are inconsistent, and monitors the injection amount after the injection is completed, and performs a supplement or extraction operation on the battery that does not meet the standard, so that the static and one-sided analysis problems of the existing monitoring method are effectively solved through dynamic relationship monitoring, whole-process data analysis and self-adaptive parameter adjustment.

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) The present application detects whether the liquid level meets the standard in real time after the liquid injection is completed, and automatically performs the operation of adding electrolyte or extracting excess electrolyte, through the closed-loop adjustment of the injection amount and the early rejection of abnormality, the product qualification rate of the battery cell is significantly improved. BRIEF DESCRIPTION OF DRAWINGS

[0019] The present application will be further described with the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of the following drawings.

[0020] Figure 1 The present application is a method flowchart.

[0021] Figure 2 The present application is a logic determination diagram for determining whether there is a blockage or leakage.

[0022] Figure 3 The present application is a method flowchart for determining whether the preset correlation mode is met.

[0023] Figure 4 The present application is a module connection diagram. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0025] Referring to Figure 1 The first aspect of the present application provides a lithium battery production safety monitoring method, comprising: S1, in the liquid injection starting stage, the injection pressure is controlled to increase to the target value according to the preset strategy, and then maintained, it is verified whether the injection pressure and flow rate keep synchronous rising trend in the pressure increasing process, and whether the pressure maintaining process conforms to the preset correlation mode, if both are verified, then enter the constant pressure injection stage, otherwise, determine that the liquid injection is abnormal and execute the rejection operation.

[0026] It should be noted that verifying whether the injection pressure and flow rate keep synchronous rising trend in the pressure increasing process and whether the pressure maintaining process conforms to the preset correlation mode is the prerequisite for normal start of the injection process.

[0027] Considering that the electrolyte needs to be gradually filled into the internal gap of the lithium battery in the initial stage of electrolyte injection, in this process, the injection pressure drives the flow of electrolyte, and the corresponding injection flow rate increases with the increase of pressure. When the pressure reaches the target value and enters the maintenance process, the injection dynamics state should be from transient to quasi-steady state, at this time, the pressure and flow rate in the system should follow the preset correlation mode marked by historical normal production data.

[0028] The above-mentioned synchronous rising trend is specifically that the injection flow rate value is collected in real time during the pressure increasing process and quantified as the change rate relative to the previous sampling time.

[0029] The pressure growth rate at each sampling time marked by the preset strategy is extracted.

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

[0031] It should be further noted that the above-mentioned preset strategy is to establish a pressure-flow-time correspondence relationship through step pressure infiltration characteristic test according to process requirements such as battery model and electrolyte viscosity, for example, for a specific battery model, the pressure step growth series is set to 5 levels, the pressure target value of each level is 0.1 MPa, 0.2 MPa, 0.3 MPa, 0.4 MPa, 0.5 MPa, the pressure climbing rate of each level is 0.05 MPa / s, and the residence time at each pressure level is 10 s. The pressure-flow-time correspondence relationship is established by collecting normal injection data through multiple experiments.

[0032] The strategy includes but is not limited to the number of pressure step growth, the pressure target value of each level, the pressure climbing rate of each level, and the residence time at each pressure level, according to which the pressure at each sampling time 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 the mass flow meter arranged on the main pipeline between the injection pump outlet and the battery injection port, and then the difference comparison-time normalization is used to quantify the flow rate change rate relative to the previous sampling time.

[0034] Specifically: a unified time interval is set as the pressure data sampling time according to the preset strategy, and the injection flow rate is collected at the same time, for the collected flow rate data, the flow rate values of adjacent two times are compared in time sequence, the difference between the latter flow rate and the former flow rate is calculated, the adjacent flow rate difference is compared with the set adjacent two sampling time intervals, and the flow rate change rate is obtained.

[0035] The correlation coefficient of the injection pressure growth rate and the injection flow rate change rate at each sampling time is calculated.

[0036] It can be understood that the injection pressure and the flow rate are synchronously increased, and the correlation coefficient can convert the correlation between the pressure growth rate and the flow rate change rate into a quantitative value, directly reflecting the consistency of the change trend of the two.

[0037] The correlation coefficient specifically quantifies the linear correlation degree between the pressure growth rate sequence and the flow rate change rate sequence by Pearson correlation coefficient, and the Pearson correlation coefficient is prior art, which will not be described here.

[0038] According to any one of the following conditions, the injection pressure and the flow rate are verified to maintain a synchronous rising trend during the pressure growth process: I. The correlation coefficients of all sampling time points during the pressure growth process are in the qualified correlation coefficient interval marked by the historical samples.

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

[0040] It can be understood that when the lithium battery injection is in the starting stage and under normal working conditions, the cooperative change of the injection pressure and the flow rate will follow the stable law verified by historical normal production, specifically I, however, external factors may interfere during the production process, for example, temperature changes will cause fluctuations in the viscosity of the electrolyte, and viscosity changes will affect the balance of inertial force and viscous force of the fluid, causing the pressure response to lag or lead the flow rate change, resulting in the correlation coefficient of the pressure growth rate and the flow rate change rate at some sampling time points, which may temporarily exceed the qualified correlation coefficient interval, but the overall is still in a stable rising trend.

[0041] The total sampling time proportion value refers to the product of the total sampling time and the first preset proportion, and the interval span proportion value refers to the product of the interval span and the second preset proportion. The determination of the two preset proportions is derived from statistical analysis of a large amount of historical normal production batch data, and the determination logic of the second preset proportion is the same as that of the first preset proportion. Taking the first preset proportion as an example, the specific determination process is as follows: collect historical batch data judged as normal under the same production conditions, calculate the proportion of the number of sampling time points exceeding the preset qualified interval in the total sampling time points in the measured flow sequence during the entire injection pressure maintenance process for each historical normal batch, and statistically analyze the distribution characteristics of all historical batch data, and take the mode as the first preset proportion.

[0042] The above-mentioned qualified correlation coefficient interval and the establishment method of the following pressure-flow mapping function have the following common points: by collecting historical normal lithium battery production batch data, selecting production batches with high matching degree with the current production conditions as historical samples, the production conditions include electrolyte viscosity, injection pipe 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, abnormal values are removed by a box plot method, a principle of covering 90% of normal intervals in historical normal samples by a qualified correlation coefficient interval is adopted, two preset percentiles are set as upper and lower boundaries of the interval, for example, 5% percentile is set as the lower limit and 95% percentile is set as the upper limit, and a value range between the 5% percentile and the 95% percentile is defined as the qualified correlation coefficient interval.

[0053] It should be further explained that the implementer can customize the preset percentile according to the production requirements for the liquid injection stability.

[0054] Referring to FIG. 1, Figure 3 In a preferred embodiment of the present application, the preset correlation mode meets the condition, which is specifically as follows: a liquid injection flow value is collected in real time during a pressure maintaining process, and a measured flow sequence is constructed.

[0055] A target pressure value at the same sampling time in the preset strategy is extracted, and an expected flow sequence is calculated by using a preset pressure-flow mapping function calibrated by historical samples.

[0056] The construction of the above-mentioned pressure-flow mapping function is specifically performed in a statistical regression fitting manner, and for example, a linear regression fitting manner is used to construct a linear regression function: .

[0057] Wherein, is an expected flow, is a liquid injection pressure, is a slope, is an intercept.

[0058] The least square method can be used to solve the coefficients a and the intercept b, and the statistical library of Python is directly used to calculate the average of the liquid injection pressure, the average of the liquid injection flow, the covariance of the liquid injection pressure and flow, and the pressure variance.

[0059] The slope is a ratio of the covariance and the force variance, and the intercept is a difference between the average of the liquid injection flow and the product of the slope and the average of the liquid injection pressure.

[0060] It should be further explained that if the relationship between the pressure and the flow shows a rapid growth trend, a nonlinear function should be considered for mapping at this time, for example, an exponential function can be used, the exponential function uses a nonlinear least square fitting algorithm to directly regress the original data, and the undetermined parameters in the function are solved.

[0061] The liquid injection pressure at each sampling time is collected, and the corresponding liquid injection flow is calculated by using the preset pressure-flow mapping function, and an expected flow sequence is constructed.

[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 time when the correlation coefficient first exceeds the interval is the critical node for the liquid injection process to deviate from the normal collaborative state to the abnormal state during the pressure growth process in the lithium battery liquid injection starting stage, the first time when the correlation coefficient exceeds the interval is taken as the starting point, and the natural end point of the pressure growth process is taken as the end point, so that the whole process of abnormal start, abnormal development and abnormal end can be completely covered, and the loss of data in the initial stage of abnormality caused by late starting point can be avoided, and the initial parameter change characteristics of the fault cannot be traced back.

[0073] For the pressure maintenance process, the absolute deviation integral value is continuously calculated according to the preset period, and the starting point corresponding to the period when the integral value first exceeds the preset allowable deviation value is taken as the starting point of the abnormality determination period, and the natural end point of the pressure maintenance process is taken as the end point of the abnormality determination period.

[0074] Considering that the flow fluctuation in the pressure maintenance stage may be intermittent or gradual, the absolute deviation integral value is continuously calculated according to the preset period, and the implicit deviation is captured by period accumulation rather than single-point sampling, so that the integral value can reflect the cumulative effect of the deviation in the period.

[0075] For example, the deviation gradually increases in the first 5 seconds of the period, and the integral value exceeds the standard at 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, the starting point corresponding to the period when the integral value first exceeds the preset allowable deviation value is taken as the starting point of the abnormality determination period, so that the deviation change data in the early stage of the period can be avoided, and the natural end point of the pressure maintenance process is taken as the end point of the abnormality determination period, so that the to-be-determined data sequence is ensured to be in a process environment with constant pressure.

[0076] It should be noted that the implementer can set the cycle period corresponding to the preset period according to the liquid pump speed fine adjustment or the electrolyte supplement operation cycle, so as to quickly help locate the abnormality cause. For example, if the electrolyte supplement is performed at the starting point of a certain period, and the integral value of the period first exceeds the standard, it can be initially investigated whether the deviation is caused by the viscosity fluctuation of the electrolyte after the supplement.

[0077] The dynamic time warping distance of the to-be-determined sequence and each typical abnormal sequence in the preset abnormal database is calculated.

[0078] It should be noted that the preset abnormal database is constructed by the following steps: first, the data sequences of the liquid injection pressure and the liquid injection flow in the historical lithium battery liquid injection abnormal process 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, the collected data sequences are preprocessed, including data cleaning, standardization and time alignment, so as to eliminate noise and dimension influence and ensure the consistency of the sequences.

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

[0081] Finally, database generation is performed, and the mapping relationship between the typical abnormal category and the corresponding typical abnormal sequence is stored to form the preset abnormal database.

[0082] If the dynamic time warping distance is less than or equal to the preset distance threshold, it is determined that the abnormal type is consistent with the corresponding typical case, and the occurrence frequency of the type of abnormality and the associated liquid injection equipment parameters are recorded.

[0083] The calculation logic of the above dynamic time warping distance is as follows: an accumulated distance matrix of the to-be-judged sequence and the typical abnormal sequence is constructed, and the optimal bending path is found through the accumulated distance matrix, which can minimize the accumulated distance between the two sequences after warping. The minimum accumulated distance is the dynamic time warping distance.

[0084] The preset distance threshold can be obtained through experiments, specifically: collect historical abnormal data, calculate the dynamic time warping distance of each type of sequence in the preset abnormal database, and classify the distance calculated with the correct type of sequence into the same distance set, and the distance calculated with the wrong type of sequence into the different distance set.

[0085] The values in the same distance set and the different distance set are arranged in ascending order respectively, and the maximum value of the same distance set and the minimum value of the different distance set are taken, and the interval formed by the two values is the safety interval.

[0086] Take the value at one quarter of the safety interval as the initial threshold, and then use the initial threshold to judge a batch of verification data with known abnormal labels. If the omission 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 misjudgment 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 principle, repeated iteration adjustment is performed until the misjudgment rate and the misjudgment rate meet the process requirements. The threshold at this time is the preset distance threshold.

[0087] If the dynamic time warping distance is greater than the preset distance threshold, the abnormal data sequence is marked as a new type of abnormality, and after manual confirmation, the feature data is included in the preset abnormal database.

[0088] Referring to Figure 2 As shown in FIG. 2, S2, in the constant pressure liquid injection stage, a preset correlation mode between the liquid injection pressure and the instantaneous flow rate is continuously analyzed. If it is monitored that the two deviate from this mode, it is determined whether there is a risk of blockage or leakage based on the deviation performance. If it is determined that there is, the rejection operation is performed, otherwise, the liquid injection is entered into the end stage.

[0089] In view of the constant pressure liquid injection stage, the liquid injection process follows the basic law of fluid mechanics, and in this stage, whether the liquid injection is normal will be directly reflected in the dynamic change of flow resistance and the attenuation characteristics of flow rate. Therefore, the specific process of the deviation performance judgment of whether there is a blockage or leakage is: based on the injection pressure and injection flow rate sequence, the instantaneous flow resistance is calculated, which is the ratio of injection pressure and injection flow rate at the same time, and forms an instantaneous flow resistance sequence.

[0090] Linear fitting is performed on the instantaneous flow resistance sequence to obtain a first fitting slope.

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

[0092] It should be noted that the first derivative reflects the instantaneous change rate of flow rate with time, i.e. the attenuation speed. If the derivative is negative, it represents a decrease in flow rate, which is consistent with the flow rate change characteristics when there is a blockage or leakage. The first derivatives of the instantaneous injection flow rate sequence at the beginning and end are calculated using forward difference method and backward difference method, and the central difference method is used to calculate the first derivatives of other points. All the first derivatives at different times are arranged in chronological order, i.e. to form an attenuation speed sequence. The calculation here is a prior art and will not be described in detail.

[0093] The second derivative reflects the change rate of the attenuation speed with time, i.e. the attenuation acceleration. If the second derivative is positive, it represents a decrease in the attenuation speed, i.e. the flow rate decreases slowly when there is a leakage. If it is negative, it represents an increase in the attenuation speed, i.e. the flow rate decreases rapidly when there is a blockage. Then, based on the attenuation speed sequence, the second derivative is calculated using the same logic as the first derivative, and the attenuation acceleration sequence is constructed.

[0094] The absolute value of the attenuation speed sequence is linearly fitted to obtain a second fitting slope, and the calculation logic of the second fitting slope is consistent with that of the first fitting slope.

[0095] If the first fitting slope is positive and the second fitting slope is negative, and the average value of the flow rate attenuation acceleration is greater than zero, it is determined that there is a leakage.

[0096] If the first fitting slope is positive and the second fitting slope is positive, and the average value of the flow rate attenuation acceleration is less than zero, it is determined that there is a blockage.

[0097] It is considered that whether it is a leakage or a blockage fault, it will cause the effective through section of the electrolyte to decrease, and there will be a phenomenon of flow rate decrease and flow resistance increase. Therefore, the positive first fitting slope is a common prerequisite for determining the existence of a fault.

[0098] On the basis of confirming the existence of the fault, the dynamic characteristics of the flow decay are analyzed, that is, the leakage and the blockage are distinguished. Specifically, if the leakage occurs, since the gap of the leakage point is usually of a fixed size, as the liquid injection proceeds, the liquid injection flow state gradually tends to be stable, the rate of flow decline gradually slows down, that is, the absolute value of the flow decay rate decreases with time, and therefore the second fitting slope is negative, which is one of the key characteristics that distinguish the leakage fault from the blockage fault.

[0099] Secondly, when the leakage occurs, the flow decay rate gradually slows down, and the change amount of the decay rate is positive, for example, the decay rate changes from -0.5 mL / s² to -0.3 mL / s², the decay acceleration is +0.2 mL / s³, and the average value of the flow decay acceleration in the whole period is still greater than zero. This index quantitatively verifies the trend that the flow decay is getting slower and slower from the mathematical level, and further points to the leakage fault.

[0100] In summary, if the above conditions are met, it can be determined that there is a leakage.

[0101] If the blockage occurs, for example, if there are impurities in the electrolyte, as the electrolyte continues to be injected, the impurities may further accumulate, causing the cross section of the injection channel to continuously decrease, the flow resistance of the electrolyte to continuously increase, the absolute value of the flow decay rate to increase with time, and finally the second fitting slope to be positive, and the average value of the flow decay acceleration in the whole period to be less than zero. Therefore, if the above conditions are met, it can be determined that there is a blockage.

[0102] S3. In the end stage of the liquid injection, whether the time when the pressure drops to zero is consistent with the preset injection cutoff time is predicted according to the pressure decay characteristics of the liquid injection process, and if the prediction is inconsistent, the pressure decay rate is adjusted by adjusting the liquid injection parameters.

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

[0104] The historical pressure sequence is an initial data sequence formed by continuously storing the pressure values of the liquid injection pipeline or the battery injection port in the early stage of the end stage of the lithium battery liquid injection in real time based on the process rhythm matching of the liquid injection stages by the preset pressure acquisition device, and the acquisition frequency is matched.

[0105] It can be understood that the liquid injection ending stage meets the liquid injection pump stopping to supply liquid, the residual pressure in the system gradually decays with the fluid emptying / penetration, and the decay process conforms to the exponential change law of pressure release of a closed system in fluid mechanics, so an exponential decay fitting is adopted, and the specific steps are as follows: real-time acquisition of pressure data from the moment when the liquid injection pump stops to the moment when the system pressure drops to a safety threshold in the ending stage.

[0106] The real-time acquired pressure value is continuously added to the historical pressure sequence constructed to form a complete ending stage pressure-time data set. According to the theory of fluid mechanics, the pressure decay of a closed pipeline system under no external pressure input follows a first-order exponential decay model, and its core formula is: .

[0107] Among them, is the system pressure at time , is the initial pressure at the moment when the liquid injection pump stops, is the time after the liquid injection pump stops, is the pressure decay time constant.

[0108] In the above formula, represents the exponential factor of pressure decay, which decreases with the increase of time , reflecting the decay rate of system pressure, is the pressure decay time constant, and the greater the value, the slower the pressure decay, and the stronger the ability of the system to maintain pressure, and vice versa.

[0109] The entire formula describes the dynamic process of the system pressure decaying from the initial value to a stable state in an exponential form under no external pressure input.

[0110] Since the exponential function is a nonlinear function, it needs to be solved by data linearization conversion or nonlinear least squares method. In this embodiment, the nonlinear least squares method is adopted, and the specific steps are as follows: taking the residual sum of squares of the predicted pressure of the exponential function and the actual collected pressure as the optimization target, inputting the ending stage pressure-time data set and the initial parameters of the exponential function, and adjusting the value of through iteration until the residual sum of squares converges to a minimum value, such as less than 0.001 MPa², and the value of is extracted as the instantaneous pressure decay time constant.

[0111] Based on the pressure value at the current time and the instantaneous decay time constant, the remaining time from the current time to the pressure dropping to zero is obtained by substituting the above exponential decay function, and then the remaining time and the current time are summed to obtain the predicted pressure zero point time.

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

[0113] If the time deviation continues to be lower than the process tolerance within the preset determination duration, it is determined that the time when the pressure drops to zero is consistent with the preset liquid injection cutoff time, otherwise it is determined that it is not consistent.

[0114] The process tolerance is determined according to the response delay characteristics of the liquid injection system. Specifically, by collecting full-process data of multiple batches of lithium battery liquid injection processes, recording the actual time deviation between the time when the pressure drops to zero and the preset liquid injection cutoff time, calculating the standard deviation of these deviation data, and selecting a numerical range that can cover most normal process fluctuations as the process tolerance.

[0115] The preset determination duration is determined by monitoring the fluctuation rule of the time deviation in multiple sets of lithium battery liquid injection processes, and counting the period from the appearance of the fluctuation to the stabilization of the time deviation. The preset determination duration needs to cover this period to avoid misjudgment due to too short determination duration.

[0116] In actual production, the pressure decay process will be affected by factors such as equipment precision and electrolyte viscosity, and there will be inevitable small fluctuations. The settings of the preset determination duration and the process tolerance are to allow the existence of such normal process fluctuations. As long as the time deviation continues to be within the tolerance range within a reasonable duration, it is determined that the time is consistent, avoiding misjudgment due to small fluctuations, thereby reducing unnecessary process intervention and improving production efficiency.

[0117] In a preferred embodiment of the present application, the specific method of adjusting the liquid injection parameters to control the pressure decay rate is to calculate the time deviation between the predicted pressure zero point time and the preset liquid injection cutoff time in real time.

[0118] According to the time deviation, the absolute value of the instantaneous pressure decay rate at the current time, and the difference between the current time and the preset liquid injection cutoff time, the gain parameters of the proportional-integral-derivative controller are dynamically adjusted, and after each adjustment, it is judged whether the time deviation converges. If it does not converge, switch to the next gain adjustment stage.

[0119] The dynamic adjustment process uses fuzzy rules. After the three input variables of time deviation, decay rate, and remaining time are fuzzified, the adjustment amount of each gain is obtained through the fuzzy rule base reasoning, and then de-fuzzified and output to the PID controller.

[0120] The fuzzy rule base stores the input variable fuzzy set, output variable fuzzy set, membership function, and fuzzy mapping rule.

[0121] The input variable fuzzy set includes the absolute value of the time deviation, the absolute value of the pressure decay rate, and the difference between the current time and the preset liquid 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 repeatedly performing the adjustment operation is: first, dynamically adjusting the proportional gain, and after each adjustment period, judging whether the time deviation is continuously lower than the process tolerance within the preset judgment duration, if the condition is met, the adjustment is exited, otherwise, entering the integral gain adjustment stage, repeating the above judgment process, if it still does not meet the condition, further starting the differential gain adjustment until the time deviation meets the requirement or the number of repetitions reaches the maximum adjustment number.

[0133] The present application clearly divides the liquid injection process into three core stages of start-up, constant pressure and end, and designs specific monitoring, verification and control strategies for each stage. By verifying the synchronous rising trend of pressure and flow and the preset correlation mode in the start-up stage, and continuously analyzing the pressure-flow dynamic coupling relationship in the constant pressure injection stage, the blockage and leakage can be effectively identified and the rejection operation is performed. In the end stage, through exponential decay fitting and PID dynamic adjustment, the pressure zero point and the injection cutoff time are matched.

[0134] It can be understood that the accuracy of the lithium battery injection amount directly affects the battery performance and safety, the time deviation reflects the deviation degree of the pressure zero point time 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 determines the time window of the adjustment. Based on the three parameters, the PID gain is dynamically adjusted to accurately control the pressure relief valve opening degree, so as to strictly control the injection cutoff time and ensure the accuracy of the injection amount to meet the process standard.

[0135] S4, the battery completing the injection is subjected to injection amount monitoring, and the battery with unqualified injection amount is subjected to corresponding adjustment operation to make the injection amount reach the set standard.

[0136] The injection amount deviation is mainly caused by equipment accuracy and electrolyte viscosity factors. In order to accurately control the electrolyte content, the adjustment operation needs to be treated according to the direction and degree of the injection amount deviation. The corresponding adjustment operation is specifically: if the actual injection amount is less than the standard injection amount, it is determined that the injection amount is unqualified, and the electrolyte supplement operation is performed.

[0137] If the actual injection amount is greater than the standard injection amount, it is determined that the injection amount exceeds the standard, and the excess electrolyte extraction operation is performed.

[0138] The electrolyte supplement or extraction operation is performed according to the absolute deviation between the actual injection amount and the standard injection amount.

[0139] After the injection is completed, the system detects whether the liquid level meets the standard in real time, and automatically performs the operation of supplementing electrolyte or extracting excess electrolyte, which significantly improves the product qualification rate of the battery cell.

[0140] Referring to Figure 4As shown, the second aspect of the present application provides a photovoltaic power automatic control 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.

[0141] 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.

[0142] The liquid injection start control module, in the liquid injection start stage, controls the liquid injection pressure to increase to a target value according to a preset strategy, maintains the target value, verifies whether the liquid injection pressure and flow rate keep a synchronous rising trend in the pressure increasing process, and whether they conform to a preset correlation mode in the pressure maintaining process, if both are verified, enters the constant pressure liquid injection stage, otherwise, determines that the liquid injection is abnormal, and executes the rejection operation.

[0143] The constant pressure liquid injection monitoring module continuously analyzes the preset correlation mode between the liquid injection pressure and the instantaneous flow rate, if it is monitored that the two deviate from the mode, determines whether there is a risk of blockage or leakage based on the deviation performance, if it is determined that there is, executes the rejection operation, otherwise, enters the liquid injection end stage.

[0144] The liquid injection end adjustment module, according to the pressure decay characteristics of the liquid injection process, predicts whether the time when the pressure drops to zero is consistent with the preset liquid injection cutoff time, if the prediction is inconsistent, adjusts the pressure decay speed by adjusting the liquid injection parameters.

[0145] The liquid injection amount verification module monitors the liquid injection amount of the completed liquid injection battery, and executes corresponding adjustment operation on the battery whose liquid injection amount does not meet the standard to make the liquid injection amount reach the set standard.

[0146] The above content is only an example and description of the structure of the present application, and those skilled in the art can make various modifications or supplements or adopt similar ways to replace the described specific embodiments, as long as they do not deviate from the structure of the present application or exceed the scope defined by the present application, and should belong to the protection scope of the present application.

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. The specific conditions for meeting the preset association mode 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 curve and the measured flow 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 it conforms to the preset correlation mode. During the constant pressure injection stage, continuously analyze the preset correlation mode between the injection pressure and the instantaneous flow. If it is detected that the two deviate from this mode, determine whether there is a risk of blockage or leakage based on the deviation performance. If it is determined that there is, perform the rejection operation; otherwise, enter the injection termination stage. 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 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.

4. The lithium battery production safety monitoring method 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.

5. The lithium battery production safety monitoring method according to claim 4, 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 end point of the anomaly judgment period. During the pressure maintenance process, the integral value of the absolute deviation 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.

6. 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.

7. 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, the determination is inconsistent.

8. The method for monitoring the safety of lithium battery production according to claim 7, 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.

9. 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