A temperature regulation and performance optimization method in a dry electrode preparation process

Abstract

The application provides a temperature regulation and performance optimization method in a dry electrode preparation process, S1, collecting real-time temperature of dry electrode preparation to generate a temperature gradient map; S2, identifying a binder morphology transition risk point based on the temperature gradient map, if the temperature gradient exceeds a temperature gradient threshold value, filtering and denoising, and constructing a refined temperature distribution model; S3, extracting a fiber transition key interval, clustering and grouping regional temperature data to obtain a morphology transition prediction sequence, calculating a connection parameter combined with a network structure formation index, and determining an electrode internal uniformity score; S4, if the electrode internal uniformity score is lower than a qualified threshold value, selecting a high interference interval from the morphology transition prediction sequence, adjusting a heat source input through a feedback control algorithm, and obtaining an optimized temperature regulation scheme; S5, updating equipment parameters according to the optimized temperature regulation scheme, extracting battery performance indicators and comparing initial values, determining a production efficiency improvement value, and obtaining a final dry electrode preparation output.

Description

Technical Field

[0001] This invention relates to the field of dry electrode technology, and in particular to a method for temperature control and performance optimization in the dry electrode preparation process. Background Technology

[0002] Dry electrode technology, a crucial direction in battery manufacturing, focuses on achieving uniform mixing and structural optimization of electrode materials through solvent-free processes. This is invaluable for improving battery performance, reducing production costs, and achieving environmental goals. With the rapid development of the new energy industry, dry processes have gradually become a focus of industry attention because they avoid the use of solvents in traditional wet processes, reducing environmental pollution and lowering energy consumption.

[0003] However, research and application in this field still face many challenges, and breakthroughs in key technological bottlenecks are urgently needed. Currently, dry electrode fabrication methods have significant shortcomings in practical operation, mainly due to insufficient precise control over process conditions. Many existing technologies often struggle to balance the various physical changes during material mixing, resulting in poor structural stability of the final electrode material, affecting battery cycle life and energy density. Especially in complex production environments, fluctuations in process parameters directly interfere with the formation of the material's internal microstructure. This problem cannot be solved simply by upgrading equipment or adjusting parameters; rather, it requires a deeper understanding of the material's behavioral changes during processing.

[0004] Against this backdrop, the core technical challenge in dry electrode fabrication lies in the impact of temperature changes on material morphology. Temperature, a critical factor in the processing, undergoes uncontrollable changes due to the frictional heat generated during high-intensity mixing. This heat accumulation further interferes with the transformation of the binder material from granular to fibrous form. If the temperature distribution is uneven or improperly controlled, the binder material cannot form a uniform network structure, leading to poor internal connectivity of the electrode and consequently affecting conductivity and mechanical strength. For example, in actual production, excessively high temperatures in some areas may cause premature softening or even degradation of the binder material, while insufficient temperatures in other areas may prevent sufficient deformation, ultimately resulting in uneven electrode material distribution and impacting battery performance.

[0005] Therefore, how to precisely control the temperature distribution in a solvent-free environment and achieve the orderly transformation of the binder material morphology through temperature management has become a key issue that urgently needs to be addressed in dry electrode fabrication technology. Solving this problem not only concerns the optimization of the electrode material's microstructure but also directly affects the overall performance and production efficiency of battery products. Summary of the Invention

[0006] To address the technical problems mentioned in the background section, the present invention provides a method for temperature control and performance optimization in the dry electrode fabrication process, the method comprising: S1: Collect the real-time temperature of dry electrode preparation, extract distribution characteristics by combining thermal imaging, and generate a temperature gradient map; S2, based on the temperature gradient map, identify the risk points of binder morphological transformation. If the temperature gradient exceeds the temperature gradient threshold, filter and denoise the data to construct a refined temperature distribution model. S3, extract the key intervals of fiber transformation, obtain the morphological transformation prediction sequence by clustering the temperature data of the grouped regions, and calculate the connection parameters by combining the network structure formation index to determine the uniformity score inside the electrode. S4. If the uniformity score inside the electrode is lower than the qualified threshold, a high-interference interval is selected from the morphological transformation prediction sequence, and the heat source input is adjusted through the feedback control algorithm to obtain an optimized temperature control scheme. S5 updates equipment parameters based on the optimized temperature control scheme, evaluates the improvement trend of conductivity and mechanical strength, obtains the optimized electrode model, extracts battery performance indicators and compares them with the initial values, determines the production efficiency improvement value, and obtains the final dry electrode preparation output.

[0007] Furthermore, step S1 includes: arranging an infrared temperature sensor array at the roll gap of the roller press to collect real-time temperature data at a sampling frequency of 100 milliseconds.

[0008] Furthermore, step S1 also includes: using thermal imaging technology to perform a comprehensive scan of the material crushing area, mapping discrete sensor data to a two-dimensional spatial coordinate system, using the Sobel operator to extract the temperature change rate between adjacent pixels, calculating the temperature gradient value of each region, and constructing a temperature gradient map.

[0009] Furthermore, step S2 includes: Step S21: By scanning the temperature gradient map, preliminary marking is made for possible interference points. Step S22: Use a data filtering algorithm to process outliers that exceed the temperature gradient threshold, remove noise, and obtain a denoised temperature distribution dataset. Step S23: For the denoised temperature distribution dataset, analyze the changing trend of the temperature gradient, detect gradient changes in key areas, and determine potential locations of morphological transformation. Step S24: Further analyze the potential locations of morphological transformation, combine the physical properties of material bonding to obtain the corresponding distribution characteristics, extract the correlation between temperature gradient and material bonding, and construct a refined temperature distribution model. Step S25: If the detected gradient change exceeds the gradient change threshold, the abnormal area is reconstructed to determine the core area of ​​the material morphology transformation and complete the accurate location of the interference point.

[0010] Furthermore, step S22 includes: using a Gaussian filtering algorithm to process outlier data that exceeds the temperature gradient threshold, setting an appropriate smoothing window size, smoothing local abrupt pixel changes, removing environmental noise interference, and outputting a denoised temperature distribution dataset.

[0011] Furthermore, step S23 includes: calculating the temperature difference between adjacent data points in the denoised temperature distribution dataset; when a nonlinear fluctuation in the temperature gradient of a continuous region is detected between 0.8 degrees per millimeter and 1.5 degrees per millimeter, it is identified as a potential location point for morphological transformation.

[0012] Furthermore, step S23 also includes: converting the absolute temperature value and gradient variance of the potential location points of morphological transformation into multidimensional feature vectors.

[0013] Furthermore, step S25 includes: if the refined temperature distribution model detects that the gradient change in a certain local area exceeds the gradient change threshold, then the Kriging space interpolation algorithm is used to perform distance-weighted calculation using the normal temperature data around the local area, and the data of the local area is reconstructed to obtain the final temperature distribution optimization result.

[0014] Furthermore, step S3 includes: obtaining temperature distribution data from the refining temperature distribution model, performing preliminary sorting of temperature values ​​for different regions, applying preset interval division rules to divide the overall temperature range into segments of 5.0 degrees Celsius each, and extracting key intervals that are highly correlated with the transformation of binder fibers.

[0015] Furthermore, step S3 also includes: applying the K-means clustering algorithm to group the regional temperature data, using the temperature value and uniformity variance as the algorithm input features, setting the number of clusters to 3, dividing the electrode region into high-conversion region, medium-conversion region and low-conversion region, and generating a regional classification set based on temperature distribution, i.e., morphological transformation prediction sequence.

[0016] The technical solution provided by this invention has the following beneficial effects: This invention discloses a method for temperature control and performance optimization in the dry electrode fabrication process. Addressing the problem of uneven temperature distribution during fabrication leading to uneven material morphology transformation and consequently affecting electrode internal connectivity and battery performance, this invention acquires real-time temperature data from production equipment sensors and generates temperature gradient maps using thermal imaging technology. This accurately identifies interference points in the morphology transformation of the bonding material. When the temperature gradient exceeds a threshold, a data filtering algorithm refines the distribution model, extracts key fiber transformation intervals, and uses a clustering algorithm to predict the morphology transformation sequence. Connectivity parameters are calculated to evaluate electrode uniformity scores. When the score falls below a threshold, a feedback control algorithm optimizes the heat source input and updates equipment parameters, ultimately improving conductivity and mechanical strength. This invention, through a closed-loop mechanism of temperature control and structural optimization, significantly improves production efficiency and battery performance indicators, achieving high-quality output from dry electrode fabrication. Attached Figure Description

[0017] Figure 1 This is a flowchart of a method for temperature control and performance optimization in the dry electrode fabrication process according to the present invention. Detailed Implementation

[0018] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0019] like Figure 1 As shown in this embodiment, a method for temperature control and performance optimization in the dry electrode fabrication process may specifically include: S1 collects the real-time temperature of dry electrode preparation, extracts distribution characteristics by combining thermal imaging, and generates a temperature gradient map.

[0020] In one embodiment, real-time temperature data during the dry electrode fabrication process is collected by sensors in the production equipment and stored as an initial temperature dataset. Thermal imaging technology is used to perform a comprehensive scan of the material area, extracting spatial distribution information from the initial temperature dataset to generate a temperature distribution image of the material area. Temperature gradient values ​​for each area are calculated, and a temperature gradient map is constructed. If any region in the temperature gradient map has a gradient value exceeding a threshold, in-depth analysis is performed on the temperature data of that region to identify the distribution locations of anomalies. Based on the distribution locations of the anomalies, temperature change records for the corresponding time period are extracted from the initial temperature dataset to determine if there are any persistent anomalies. Time series analysis is performed on the temperature change records of persistent anomalies, and a support vector machine algorithm is used to classify the anomaly trends. A data file for targeted adjustment strategies is generated and output to the production equipment control module.

[0021] In one possible implementation, the dry electrode fabrication process is extremely sensitive to temperature, and temperature fluctuations directly affect the degree of fiberization of the binder. An array of infrared temperature sensors is deployed at the roll gap of the roller press to collect real-time temperature data at a sampling frequency of 100 milliseconds, forming an initial temperature dataset. Thermal imaging technology is then used to comprehensively scan the material's rolling area, mapping the discrete sensor data onto a two-dimensional coordinate system to generate a temperature distribution image of the material region. This multi-source data fusion approach effectively eliminates blind spots from single-point temperature measurements, comprehensively reflecting the heat distribution on the electrode surface and providing a high-resolution data foundation for subsequent precise temperature control.

[0022] For example, for the generated temperature distribution image, the Sobel operator is used to extract the rate of temperature change between adjacent pixels, calculate the temperature gradient value of each region, and construct a temperature gradient map. A gradient value threshold of 2.5 degrees Celsius per centimeter is set. When the gradient value of a region in the map reaches 3.2 degrees Celsius per centimeter, the region is determined to exceed the preset gradient value threshold. Subsequently, the center coordinates of the abnormal region are extracted as the location of the abnormal point, and the temperature change records for the five minutes before and after this coordinate point are retrieved from the initial temperature dataset. This time series is smoothed using a sliding window method. If the temperature continuously deviates from the baseline value for 30 consecutive seconds, it is confirmed as a persistent anomaly. This process can accurately filter transient thermal interference and precisely locate the actual process defect area.

[0023] It should be noted that after acquiring records of persistently abnormal temperature changes, these records are used as input feature vectors and imported into a pre-trained support vector machine (SVM) algorithm model. This model uses a radial basis function kernel to map the input features to a high-dimensional space to find the optimal hyperplane, thereby classifying abnormal trends into specific anomaly types such as localized overheating of the rollers, material agglomeration and frictional heating, or cooling equipment failure. If the determination is material agglomeration and frictional heating, then, based on the area and location of the abnormal region in the temperature gradient map, a targeted adjustment strategy data file is generated to reduce the local feed rate and fine-tune the roller pressure. This file is directly output to the production equipment control module for closed-loop regulation. This machine learning-based intelligent diagnosis and feedback mechanism significantly shortens the anomaly response time and substantially improves the yield and consistency of dry-process electrodes.

[0024] S2, based on the temperature gradient map, identify the risk points of binder morphological transformation. If the temperature gradient exceeds the temperature gradient threshold, filter and denoise to construct a refined temperature distribution model.

[0025] Optionally, this step also includes: Step S21: By scanning the temperature gradient map, preliminary marking is made for possible interference points.

[0026] Step S22: Use a data filtering algorithm to process outliers that exceed the temperature gradient threshold, remove noise, and obtain a denoised temperature distribution dataset.

[0027] Step S23: For the denoised temperature distribution dataset, analyze the changing trend of the temperature gradient, detect gradient changes in key areas, and determine potential locations of morphological transformation.

[0028] Step S24: Further analyze the potential locations of morphological transformation, combine the physical properties of material bonding to obtain the corresponding distribution characteristics, extract the correlation between temperature gradient and material bonding, and construct a refined temperature distribution model.

[0029] Step S25: If the detected gradient change exceeds the gradient change threshold, the abnormal area is reconstructed to determine the core area of ​​the material morphology transformation and complete the accurate location of the interference point.

[0030] Specifically, after acquiring the raw temperature data, initial annotations were made for instantaneous abnormally high temperatures caused by powder splashing or reflections from the metal surface of the rolling equipment during dry electrode fabrication. A local temperature fluctuation threshold of 15.5 degrees Celsius was set, and coordinate points where temperature jumps exceeded this range within adjacent time windows were marked as potential interference points. Subsequently, a Gaussian filtering algorithm was used to process the abnormal points exceeding the temperature gradient threshold. By setting an appropriate smoothing window size, locally abrupt pixel changes were smoothed, environmental noise interference was eliminated, and a denoised temperature distribution dataset was output.

[0031] For example, for the denoised dataset, the spatial variation trend of the temperature gradient is analyzed. In the dry electrode process, binders such as polytetrafluoroethylene (PTFE) undergo a key morphological transformation—fibrillation—when subjected to shear force. This physical process is accompanied by specific microscopic thermal changes. The temperature difference between adjacent data points in the denoised temperature distribution dataset is calculated. When a nonlinear fluctuation in the temperature gradient of a continuous region is detected between 0.8 degrees per millimeter and 1.5 degrees per millimeter, it is identified as a potential location of the morphological transformation. Combining the bonding and thermal physical properties of PTFE in the 120-140 degree Celsius range, the absolute temperature value, gradient variance, and other quantitative indicators of the potential location of the morphological transformation are transformed into multidimensional feature vectors to construct a preliminary morphological transformation distribution map.

[0032] In one possible implementation, the correlation features between temperature gradient and material bonding state are extracted. Using the aforementioned multi-dimensional feature vector as input, a random forest regression algorithm is employed for model training to construct a refined temperature distribution model. This model can output the probability distribution of bonding uniformity in each region, thereby identifying the distribution pattern of interference points caused by adhesive agglomeration. If the refined temperature distribution model detects that the gradient change in a local area exceeds the 2.0 degree / millimeter gradient change threshold, a Kriging space interpolation algorithm is used. Distance-weighted calculations are performed using normal temperature data from the surrounding area to reconstruct the data for that local area, obtaining the final optimized temperature distribution result. By comparing the data difference matrix before and after reconstruction, the boundary coordinates of the core region of the material's fibrillation morphology transformation are finally determined, completing the precise location of interference points such as agglomeration.

[0033] S3, extract the key intervals of fiber transformation, obtain the morphological transformation prediction sequence by clustering the temperature data of the grouped regions, calculate the connection parameters by combining the network structure formation index, and determine the uniformity score inside the electrode.

[0034] Specifically, the key intervals for fiber transformation are extracted from the refining temperature distribution model to determine the uniformity of the material particles' transformation into fibrous structures. Clustering algorithms are used to group regional temperature data to obtain morphological transformation prediction sequences, acquire network structure formation indicators, and calculate connection parameters through simulation algorithms to determine the uniformity score inside the electrode.

[0035] In one embodiment, temperature distribution data is obtained from a refining temperature distribution model. Temperature values ​​in different regions are initially organized, and key regions related to fiber transformation are extracted using preset interval division rules, yielding temperature range data for these key regions. Based on the temperature range data of the key regions, the fiber transformation of material particles under different temperature conditions is analyzed. Statistical tools are used to quantify the uniformity of the transformation process, determining the uniformity distribution of material particle transformation into fibrous form. For the uniformity distribution results, combined with regional temperature data, a clustering algorithm is applied to group temperature values, generating a region classification set based on temperature distribution, resulting in a morphological transformation prediction sequence. Based on the morphological transformation prediction sequence, the formation trend of the network structure is analyzed, core indicators related to structure formation are extracted, and connection parameters between regions are calculated using simulation algorithms to determine the connectivity evaluation data of the network structure. For the connectivity evaluation data, combined with the uniformity distribution results, the structural consistency within the electrode is comprehensively analyzed. Preset scoring rules are used to score the uniformity of each region, yielding the uniformity score result within the electrode. Based on the uniformity score of the electrode interior, comprehensive evaluation data related to fiber conversion and network structure are generated. If the score is lower than the preset score threshold, the temperature distribution data is analyzed a second time to extract potential non-uniform regions and determine reference data for improvement directions.

[0036] In one possible implementation, temperature distribution data of the electrode manufacturing process is obtained from a refining temperature distribution model, and a preset interval division rule is applied to the temperature values ​​of different regions. The overall temperature range is divided into 5.0-degree Celsius increments, and key intervals highly correlated with the fiber transformation of the binder are extracted, for example, the temperature range of 85.0-115.0 degrees Celsius is extracted as the key interval. Based on the temperature range data of this key interval, the fiber transformation of the active material particles and the binder under different temperature conditions is analyzed. The uniformity of the transformation process is quantified using analysis of variance (ANOVA), with the variance of the fiber length distribution generated in each sampling area used as a measure. The smaller the variance, the more uniform the transformation, thus determining the uniformity distribution of the material particle to fibrous transformation.

[0037] For example, based on the above uniformity distribution results and the corresponding regional temperature data, the K-means clustering algorithm is applied to group the regional temperature data. Using the regional temperature data and uniformity variance as input features, and setting the number of clusters to 3, the electrode region is divided into high-conversion, medium-conversion, and low-conversion zones, generating a region classification set based on temperature distribution, i.e., a morphological transformation prediction sequence. Based on this morphological transformation prediction sequence, the formation trend of the binder fiber network structure is analyzed, and the fiber cross-point density is extracted as a core indicator related to structure formation. Using a Monte Carlo simulation algorithm, the cross-point density and average fiber length of each region are input, and the connectivity probability between adjacent regions is calculated as a connection parameter. For example, regions with a connectivity probability greater than 0.85 are considered valid connections, thus determining the connectivity evaluation data of the network structure.

[0038] It should be noted that, based on the acquired connectivity assessment data and the uniformity distribution results, a comprehensive analysis of the structural consistency within the electrode is performed. A preset weighted scoring rule is used to score the uniformity of each region, with the uniformity distribution results assigned a weight of 0.6 and the connectivity assessment data assigned a weight of 0.4, resulting in a uniformity score for the electrode's internal structure. Based on this score, comprehensive assessment data is generated. If a region's score is detected to be below a preset threshold of 80.0, a secondary analysis of the temperature distribution data for that region is performed. Potential non-uniform regions with scores below 60.0 are extracted and used as reference data to determine the direction of subsequent temperature field adjustments and improvements.

[0039] S4. If the uniformity score inside the electrode is lower than the qualified threshold, a high-interference interval is selected from the morphological transformation prediction sequence, and the heat source input is adjusted through a feedback control algorithm to obtain an optimized temperature control scheme.

[0040] In one embodiment, uniformity score data is collected from inside the electrode. A preliminary assessment is performed by comparing the score to a passing threshold to determine if the standard requirements are met, thus obtaining an initial assessment result. If the initial assessment result shows that the uniformity score is below the passing threshold, relevant data is extracted from the morphological transformation prediction sequence to determine the specific location and range of the high-interference interval. For the data in the high-interference interval, a feedback control algorithm is used to dynamically calculate the heat source input parameters to obtain an optimized heat source configuration scheme. Based on the optimized heat source configuration scheme, the operating parameters of the temperature control system are adjusted to generate a preliminary temperature control command set. The temperature control command set drives the heat source equipment to operate, and real-time temperature distribution data inside the electrode is collected to determine whether the expected uniformity improvement effect has been achieved. If the temperature distribution data still does not meet the uniformity score requirements, a secondary fine-tuning of the heat source input is performed based on the feedback control algorithm to obtain the final adjustment scheme. The operating status of the temperature control system is updated using the final adjustment scheme, and the uniformity score inside the electrode is continuously monitored to determine the stability of the optimization result.

[0041] In one possible implementation, the uniformity score data collected inside the electrode is compared with a preset pass threshold.

[0042] Specifically, the passing threshold for uniformity scoring is set at 85 points. When the initial score is 76 points, it is determined that the internal structure of this batch of material does not meet the standard requirements during the refining process. At this point, based on the previously generated morphological transformation prediction sequence, the high-interference intervals that cause the low score are extracted.

[0043] For example, by analyzing the temperature gradient changes in the prediction sequence, the region with coordinates X = 120 to 150 and Y = 45 to 80 to the right of the electrode center was located. The temperature fluctuation range in this region reached 8.5℃, which deviated from the stable temperature environment required for normal fiber conversion, and was thus identified as a high-interference region.

[0044] In one possible implementation, a proportional-integral-derivative (PID) control algorithm is used to dynamically calculate the heat source input parameters for the high-interference interval data extracted above.

[0045] Specifically, the deviation between the real-time temperature in the high-interference range and the target conversion temperature is used as the algorithm input. The proportional element responds to the temperature difference, the integral element eliminates steady-state error, and the derivative element predicts the temperature change trend, thereby calculating the optimized heat source configuration scheme.

[0046] For example, the algorithm output indicates that the power of the induction heating coil in that local area needs to be increased by 150W and the heating pulse frequency adjusted to 50Hz. Based on this optimization scheme, the operating parameters of the temperature control system are adjusted to generate a preliminary temperature control instruction set containing specific power and frequency values.

[0047] In one possible implementation, the heat source device is driven to perform heating operations by the aforementioned temperature control instruction set, and temperature distribution data inside the electrode is collected in real time using a temperature sensor.

[0048] Specifically, if the real-time data shows that the temperature fluctuation in the area has dropped to 3.2℃, but the corresponding uniformity score has only risen to 82 points, which still does not meet the threshold requirement of 85 points, then the secondary fine-tuning mechanism will be triggered.

[0049] For example, a secondary fine-tuning of the heat source input is performed based on a feedback control algorithm, reducing the heating power by 20W and extending the heat preservation time by 5 seconds to obtain the final adjustment scheme. After updating the operating status using this final scheme, the uniformity score inside the electrode is continuously monitored until the score stabilizes above 88 points, confirming that the uniformity of the transformation of material particles into fibrous structures inside the electrode has reached a stable state.

[0050] S5 updates equipment parameters based on the optimized temperature control scheme, evaluates the improvement trend of conductivity and mechanical strength, obtains the optimized electrode model, extracts battery performance indicators and compares them with the initial values, determines the production efficiency improvement value, and obtains the final dry electrode preparation output.

[0051] In one embodiment, a set of equipment parameter configuration schemes is generated by adjusting the temperature control scheme. The impact data of each configuration on the production environment is recorded to obtain a preliminary parameter adjustment set. From the preliminary parameter adjustment set, test data on conductivity and mechanical strength are extracted, and their changing trends are analyzed to construct a correlation mapping between performance changes and parameter configurations, identifying key influencing factors. For these key influencing factors, the structural design of the electrode model is optimized, generating optimized electrode model data and obtaining the model's performance results under different parameters. From the optimized electrode model data, relevant battery performance indicators are extracted, organized into a standardized dataset, and the stability of the performance improvement is assessed. If the stability of the performance improvement reaches a stability threshold, a verification algorithm is used to compare the initial records with the optimized data to calculate the increase in production efficiency and obtain the specific numerical value of efficiency improvement. Based on the specific numerical value of efficiency improvement, the final parameters of the dry electrode preparation process are adjusted, generating data records of the preparation results and completing the entire optimization process.

[0052] In one possible implementation, a configuration scheme containing the operating parameters of the roller press is generated based on the previous temperature control scheme. The initial linear pressure of the roller press is set to 15 MPa, the belt speed is set to 5 meters per minute, and the dust concentration and local ambient temperature data in the production environment under this configuration are recorded simultaneously to obtain a preliminary set of parameter adjustments.

[0053] Specifically, from the finished products corresponding to this parameter adjustment set, the resistivity of the electrode is extracted as conductivity data, and the tensile strength of the electrode is extracted as mechanical strength data. Using the collected 2.5 ohm-cm resistivity and 1.2 MPa tensile strength as inputs, the Pearson correlation coefficient algorithm is used to analyze the changing trends of the above properties and rolling parameters. By calculating the correlation coefficient matrix, a correlation mapping between performance changes and parameter configuration is constructed, determining that linear pressure is a key influencing factor affecting the formation of the electrode micronetwork.

[0054] For example, the structural design of the three-dimensional pore network model of the dry electrode is optimized, focusing on the key influencing factor of linear pressure. The fiber distribution density of the polytetrafluoroethylene binder is adjusted in the model to generate the optimized electrode model data. This model data is then input into a finite element analysis algorithm to simulate the stress distribution under different linear pressures ranging from 16 MPa to 18 MPa, obtaining the porosity performance of the model under the corresponding parameters.

[0055] In one possible implementation, the first-discharge specific capacity index of the simulated battery is extracted from the electrode model data after the above-mentioned structural optimization, and organized into a standardized dataset. An analysis of variance (ANOVA) method is used to process this dataset and calculate the variance of the specific capacity data. If the calculated variance value is lower than a variance threshold set at 0.05, the stability of the performance improvement is considered to have met the standard.

[0056] Specifically, once stability is achieved, a dynamic time warping algorithm is used to compare the initial recorded single-roll electrode production time sequence with the optimized time sequence. By calculating the shortest matching path between the two time sequences, the production efficiency improvement is determined to be 12.5%. Based on this specific efficiency improvement, the final parameters of the dry electrode fabrication process are fixed and adjusted, updating the actual production conveyor speed to 5.6 meters per minute, and generating a fabrication result data record containing all equipment settings for the entire process, thus completing the entire parameter optimization process.

[0057] The above are only some preferred embodiments of the present invention, but the present invention is not limited thereto, and many improvements and modifications can be made. Any improvements and modifications made based on the basic principles of the present invention should be considered to fall within the protection scope of the present invention.

Claims

1. A method for temperature control and performance optimization in the dry electrode fabrication process, characterized in that, The method includes: S1: Collect the real-time temperature of dry electrode preparation, extract distribution characteristics by combining thermal imaging, and generate a temperature gradient map; S2, based on the temperature gradient map, identify the risk points of binder morphological transformation. If the temperature gradient exceeds the temperature gradient threshold, filter and denoise the data to construct a refined temperature distribution model. S3, extract the key intervals of fiber transformation, obtain the morphological transformation prediction sequence by clustering the temperature data of the grouped regions, and calculate the connection parameters by combining the network structure formation index to determine the uniformity score inside the electrode. S4. If the uniformity score inside the electrode is lower than the qualified threshold, a high-interference interval is selected from the morphological transformation prediction sequence, and the heat source input is adjusted through the feedback control algorithm to obtain an optimized temperature control scheme. S5 updates equipment parameters based on the optimized temperature control scheme, evaluates the improvement trend of conductivity and mechanical strength, obtains the optimized electrode model, extracts battery performance indicators and compares them with the initial values, determines the production efficiency improvement value, and obtains the final dry electrode preparation output.

2. The method according to claim 1, characterized in that, Step S1 includes: arranging an infrared temperature sensor array at the roll gap of the roller press to collect real-time temperature data at a sampling frequency of 100 milliseconds.

3. The method according to claim 2, characterized in that, Step S1 further includes: using thermal imaging technology to perform a comprehensive scan of the material crushing area, mapping discrete sensor data to a two-dimensional spatial coordinate system, using the Sobel operator to extract the temperature change rate between adjacent pixels, calculating the temperature gradient value of each region, and constructing a temperature gradient map.

4. The method according to claim 1, characterized in that, Step S2 includes: Step S21: By scanning the temperature gradient map, preliminary marking is made for possible interference points. Step S22: Use a data filtering algorithm to process outliers that exceed the temperature gradient threshold, remove noise, and obtain a denoised temperature distribution dataset. Step S23: For the denoised temperature distribution dataset, analyze the changing trend of the temperature gradient, detect gradient changes in key areas, and determine potential locations of morphological transformation. Step S24: Further analyze the potential locations of morphological transformation, combine the physical properties of material bonding to obtain the corresponding distribution characteristics, extract the correlation between temperature gradient and material bonding, and construct a refined temperature distribution model. Step S25: If the detected gradient change exceeds the gradient change threshold, the abnormal area is reconstructed to determine the core area of ​​the material morphology transformation and complete the accurate location of the interference point.

5. The method according to claim 4, characterized in that, Step S22 includes: using a Gaussian filtering algorithm to process outlier data that exceeds the temperature gradient threshold, setting an appropriate smoothing window size, smoothing local abrupt pixel changes, removing environmental noise interference, and outputting a denoised temperature distribution dataset.

6. The method according to claim 5, characterized in that, Step S23 includes: calculating the temperature difference between adjacent data points in the denoised temperature distribution dataset; when a nonlinear fluctuation in the temperature gradient of a continuous region is detected between 0.8 degrees per millimeter and 1.5 degrees per millimeter, it is identified as a potential location point for morphological transformation.

7. The method according to claim 6, characterized in that, Step S23 further includes: converting the absolute temperature value and gradient variance of the potential location points of morphological transformation into multidimensional feature vectors.

8. The method according to claim 7, characterized in that, Step S25 includes: if the refined temperature distribution model detects that the gradient change in a certain local area exceeds the gradient change threshold, then the Kriging space interpolation algorithm is used to perform distance weighted calculation using the normal temperature data around the local area, and the data of the local area is reconstructed to obtain the final temperature distribution optimization result.

9. The method according to claim 1, characterized in that, Step S3 includes: obtaining temperature distribution data from the refining temperature distribution model, performing preliminary sorting of temperature values ​​for different regions, applying preset interval division rules to divide the overall temperature range into segments of 5.0 degrees Celsius each, and extracting key intervals that are highly correlated with the transformation of binder fibers.

10. The method according to claim 9, characterized in that, Step S3 further includes: applying the K-means clustering algorithm to group the regional temperature data, using the temperature value and uniformity variance as the algorithm input features, setting the number of clusters to 3, dividing the electrode region into high conversion region, medium conversion region and low conversion region, and generating a regional classification set based on temperature distribution, i.e., morphological transformation prediction sequence.

Images