Multi-stage safety control and energy optimization method for lithium battery pack

Abstract

The invention relates to the technical field of lithium battery pack management, and discloses a multi-stage safety control and energy optimization method for a lithium battery pack, which comprises the following steps of: acquiring a voltage fluctuation value and a temperature fluctuation value of a single battery in the operation of the lithium battery pack, extracting a characteristic period of the single battery through time domain transformation, and further constructing voltage fluctuation degree and temperature fluctuation intensity. And determining the safety risk level at the current moment according to the correlation between the voltage fluctuation degree and the temperature fluctuation intensity in each characteristic period and the average level. And analyzing a voltage drop trend and a temperature rise trend at adjacent sampling moments, constructing an energy loss trend degree, obtaining an energy optimization demand degree in combination with a safety risk level so as to adjust feedback regulation and control strength of an energy controller, and performing real-time adjustment on the regulation and control strength by using a self-adaptive control model. The method realizes multi-stage evaluation of the safety state of the lithium battery pack and energy optimization management, and is suitable for lithium battery pack management in the fields of electric vehicles, energy storage systems and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of lithium battery pack management, in particular to a multi-level safety control and energy optimization method for lithium battery packs. BACKGROUND

[0002] In terms of safety control, the prior art has many deficiencies. Traditional safety monitoring often only focuses on the absolute values of the voltage and temperature of single batteries, while ignoring their fluctuation characteristics and time-domain variation rules. For example, some methods only set fixed voltage and temperature thresholds to determine the safety state, which cannot capture potential risk signals such as frequent voltage fluctuations and the degree of temperature fluctuations, resulting in insufficient early warning capabilities for early safety hazards. Existing safety risk assessments are mostly based on a single parameter or a simple combination of parameters, failing to fully consider the correlation between voltage fluctuations and temperature fluctuations, making it difficult to accurately assess the overall safety risk level of lithium battery packs. In addition, when safety risks occur, existing control methods lack dynamic adjustment mechanisms and cannot timely optimize the control effort according to the changes in risk levels, affecting the effectiveness of safety control.

[0003] In terms of energy optimization, the prior art also has obvious defects. Many energy optimization methods fail to fully consider the energy loss trend during the operation of lithium battery packs, and only make simple adjustments based on the current energy state, resulting in low energy utilization. For example, during the process of single battery voltage drop and temperature rise, existing methods cannot accurately analyze the comprehensive impact of voltage drop trend and temperature rise trend on energy loss, making it difficult to construct an accurate energy loss trend index. At the same time, the calculation of energy optimization demand often ignores the influence of safety risk level, causing energy optimization and safety control to be disconnected, and failing to achieve optimal energy utilization under the premise of ensuring safety. In addition, the adjustment of the control effort of the energy controller lacks adaptive ability and cannot perform real-time feedback adjustment according to the changes in energy optimization demand, resulting in low control precision and further affecting the energy optimization effect. SUMMARY

[0004] The purpose of the present application is to provide a multi-level safety control and energy optimization method for lithium battery packs to solve the problems raised in the background.

[0005] To achieve the above purpose, the present application provides the following technical solution: a multi-level safety control and energy optimization method for lithium battery packs, the method comprising: obtaining voltage fluctuation values and temperature fluctuation values of single batteries during the operation of the lithium battery pack; performing time-domain transformation on the voltage fluctuation values and the temperature fluctuation values to extract characteristic periods of the voltage fluctuation values and the temperature fluctuation values, respectively; By analyzing the extent of frequent change of voltage fluctuation of single battery in lithium battery pack and the extent of temperature fluctuation value changing with time in each characteristic period, voltage fluctuation degree and temperature fluctuation intensity are constructed respectively; According to the correlation between voltage fluctuation degree and temperature fluctuation intensity of each characteristic period before the current time, and the average level of voltage fluctuation degree and the average level of temperature fluctuation intensity, the safety risk level of lithium battery pack at the current time is obtained; The voltage drop trend and temperature rise trend of single battery at the current time and a plurality of adjacent sampling time points are analyzed, the energy loss trend degree of lithium battery pack at the current time is constructed, and the energy optimization demand degree of lithium battery pack at the current time is obtained in combination with the safety risk level of lithium battery pack, the feedback control strength in the energy controller at the current time is obtained by using the change difference of energy optimization demand degree and the regulation and control strength in the energy controller at the current time and the preset regulation and control strength adjustment amount. According to the feedback control strength and the actual control strength in the energy controller at the current time, the regulation and control strength of the energy controller is controlled and adjusted by using the adaptive control model.

[0006] Preferably, the extraction of the characteristic period of the voltage fluctuation value and the temperature fluctuation value further comprises: The voltage fluctuation values and the temperature fluctuation values of all historical sampling time points within a preset time length before the current time are respectively arranged in time order to form a voltage fluctuation vector and a temperature fluctuation vector at the current time; The voltage fluctuation vector at the current time is subjected to time domain transformation to obtain the amplitude change of all time points in the time domain, and the reciprocal of the maximum amplitude change corresponding to the time interval is taken as the voltage fluctuation characteristic period in a short time before the current time. Correspondingly, for the temperature fluctuation vector, the temperature fluctuation characteristic period in a short time before the current time is obtained by using fast Fourier transform.

[0007] Preferably, the construction of the voltage fluctuation degree further comprises: The voltage fluctuation values in each voltage fluctuation characteristic period before the current time are arranged in ascending order of time to form a voltage period vector of each voltage fluctuation characteristic period, and correspondingly, for the temperature fluctuation values in each temperature fluctuation characteristic period, a temperature period vector of each temperature fluctuation characteristic period is obtained. For each voltage period vector, a first-order difference sequence of the voltage period vector is obtained, the number of all non-zero elements in the first-order difference sequence is counted and recorded as the change frequency, and the mean value of the absolute values of all non-zero elements in the first-order difference sequence is calculated and recorded as the average change amplitude, and the product of the change frequency and the average change amplitude is taken as the voltage fluctuation degree in each voltage fluctuation characteristic period.

[0008] Preferably, the construction of the temperature fluctuation intensity further comprises: For each temperature cycle vector, the mean of the absolute values of all elements in the first-order difference sequence of the temperature cycle vector is calculated as the temperature fluctuation intensity of each temperature fluctuation characteristic cycle.

[0009] Preferably, the acquisition process of the safety risk level of the lithium battery pack at the current moment is: The voltage fluctuation degrees of all voltage fluctuation characteristic cycles are arranged in ascending order of time to form a voltage fluctuation degree vector at the current moment, and the temperature fluctuation intensities of all temperature fluctuation characteristic cycles are arranged in ascending order of time to form a temperature fluctuation intensity vector at the current moment. The correlation between the voltage fluctuation degree vector and the temperature fluctuation intensity vector is calculated, and the element mean in the voltage fluctuation degree vector and the element mean in the temperature fluctuation intensity vector are combined to obtain the safety risk level of the lithium battery pack at the current moment.

[0010] Preferably, the construction of the energy loss trend degree of the lithium battery pack at the current moment includes: The plurality of adjacent sampling moments closest in time to the current moment are all taken as adjacent sampling moments of the current moment, and the voltage fluctuation values and temperature fluctuation values of the plurality of adjacent sampling moments are used to obtain the voltage drop trend amount and the temperature rise trend amount of the current moment. The product of the voltage drop trend amount and the temperature rise trend amount at the current moment is determined as the energy loss trend degree of the lithium battery pack at the current moment.

[0011] Preferably, the voltage drop trend amount and the temperature rise trend amount at the current moment further include: The voltage fluctuation values and temperature fluctuation values of the plurality of adjacent sampling moments are arranged in ascending order of time respectively to form a voltage trend vector and a temperature trend vector at the current moment, and the first-order difference vectors of the voltage trend vector and the temperature trend vector at the current moment are calculated respectively. The sum of the absolute values of all negative values in the first-order difference vector of the voltage trend vector is taken as the voltage drop trend amount at the current moment, and the sum of all positive values in the first-order difference vector of the temperature trend vector is taken as the temperature rise trend amount at the current moment.

[0012] Preferably, the calculation process of the energy optimization demand degree of the lithium battery pack at the current moment is: The natural constant is taken as the base number to calculate the negative exponential function of the product of the safety risk level and the energy loss trend degree, and the result of taking 1 minus the exponential function is taken as the energy optimization demand degree of the lithium battery pack at the current moment.

[0013] Preferably, the calculation process of the feedback control strength in the energy controller at the current moment is: The feedback regulation strength in the energy controller at the current time is equal to the regulation strength in the energy controller at the current time, plus the product of the energy optimization demand at the current time and the energy optimization demand at the previous sampling time, and the product of the preset regulation strength adjustment amount, wherein the regulation strength adjustment amount is not more than 10% of the regulation strength in the energy controller at the current time.

[0014] Preferably, the process of controlling and adjusting the regulation strength of the energy controller by using the adaptive control model further comprises: Obtaining the difference between the feedback regulation strength and the actual regulation strength at the current time as a control error amount; According to the change trend of the control error amount, adjusting the proportional coefficient, integral coefficient and differential coefficient in the adaptive control model, and the adjusted coefficients are used to real-time correct the regulation strength of the energy controller.

[0015] Compared with the prior art, the present application has the following advantages: In the aspect of safety control, by obtaining the voltage fluctuation value and temperature fluctuation value of the single battery and performing time domain transformation to extract the characteristic period, the internal law of voltage and temperature fluctuation can be deeply mined. The voltage fluctuation degree and temperature fluctuation intensity are constructed to comprehensively evaluate the running state of the lithium battery pack from two dimensions of the frequent change degree of voltage fluctuation and the intensity degree of temperature fluctuation change with time. Compared with the traditional method of only focusing on the absolute value, the potential safety hazards can be more sensitively captured. According to the correlation between the voltage fluctuation degree and the temperature fluctuation intensity and the average level of the two, the safety risk level is obtained, which comprehensively considers the mutual influence of multiple parameters, making the safety risk assessment more accurate and comprehensive, which helps to discover safety risks in advance and take corresponding measures to improve the safety of the lithium battery pack.

[0016] In the aspect of energy optimization, the energy loss trend degree is constructed by analyzing the voltage drop trend and temperature rise trend of the single battery at adjacent sampling times, which can accurately reflect the energy loss of the lithium battery pack. The energy optimization demand degree is obtained by combining the safety risk level, which realizes the organic combination of safety control and energy optimization, and ensures energy optimization under the premise of safety guarantee. The feedback regulation strength is adjusted by using the change difference of the energy optimization demand degree and the preset regulation strength adjustment amount, and the regulation strength of the energy controller is real-time corrected by the adaptive control model, so that the regulation strength can be dynamically adjusted according to the actual situation, improving the accuracy and flexibility of the regulation, thereby effectively reducing the energy loss and improving the energy utilization rate.

[0017] The method for extracting the characteristic period in time domain, such as obtaining the reciprocal of the time interval corresponding to the maximum amplitude change of the voltage fluctuation vector as the voltage fluctuation characteristic period by performing time domain transformation on the voltage fluctuation vector, and obtaining the temperature fluctuation characteristic period by using fast Fourier transformation on the temperature fluctuation vector, can accurately capture the periodic characteristics of voltage and temperature fluctuations, and provide reliable basis for subsequent fluctuation degree and severity calculation. When constructing the voltage fluctuation degree, the product of the change frequency and the average change amplitude of the first-order difference sequence of the voltage period vector can comprehensively reflect the frequency and amplitude of the voltage fluctuation; when constructing the temperature fluctuation severity, the mean value of the absolute value of the first-order difference sequence elements of the temperature period vector can accurately reflect the severity of the temperature fluctuation. The construction of these indexes provides a scientific quantitative basis for the evaluation of the safety risk level.

[0018] In the process of obtaining the safety risk level, the correlation between the voltage fluctuation degree vector and the temperature fluctuation severity vector is combined with the mean values of the two, which fully considers the correlation and overall level between the parameters, so that the determination of the safety risk level is more scientific and reasonable. The construction of the energy loss trend degree comprehensively considers the influence of voltage and temperature changes on energy loss by calculating the product of the voltage drop trend and the temperature rise trend, which can accurately reflect the trend of energy loss. The calculation of the energy optimization demand degree adopts a negative exponential function form with a natural constant as the base, which can sensitively adjust the energy optimization demand degree according to the changes of the safety risk level and the energy loss trend degree, and provide accurate basis for the calculation of the feedback control strength.

[0019] The calculation of the feedback control strength considers the difference between the current time and the previous sampling time of the energy optimization demand degree, which can adjust the control strength in time according to the change of the energy optimization demand degree, and the preset control strength adjustment amount does not exceed 10% of the current control strength, which ensures the stability and safety of the control process. The adaptive control model adjusts the proportional coefficient, integral coefficient and differential coefficient according to the control error amount, realizes the real-time correction of the energy controller control strength, and improves the response speed and control accuracy of the control system. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 The working principle diagram of the multi-stage safety control and energy optimization method of the lithium battery pack described in the application; Figure 2 The design diagram for extracting the characteristic period of voltage / temperature fluctuation value; Figure 3 The flowchart for constructing the energy loss trend degree; Figure 4 The flowchart for calculating the voltage drop trend and the temperature rise trend. DETAILED DESCRIPTION

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see Figures 1-4 The multi-level safety control and energy optimization method for lithium battery packs involved in this invention has the following specific implementation steps: This method acquires the voltage and temperature fluctuation values ​​of individual cells during the operation of a lithium battery pack. In practical applications, voltage and temperature sensors installed within the lithium battery pack can collect the voltage and temperature data of each individual cell in real time, and then calculate the voltage and temperature fluctuation values.

[0023] The characteristic periods of voltage and temperature fluctuations are extracted by performing time-domain transformations on the voltage and temperature fluctuation values, respectively. Specifically, the voltage and temperature fluctuation values ​​of all historical sampling times within a preset time period before the current time are obtained and arranged into vectors according to their chronological order, serving as the voltage and temperature fluctuation vectors for the current time. The voltage fluctuation vector at the current time is then transformed in the time domain to obtain the amplitude changes at all time points. The reciprocal of the time interval corresponding to the largest amplitude change is taken as the characteristic period of voltage fluctuations in the short period before the current time. Correspondingly, for the temperature fluctuation vector, the characteristic period of temperature fluctuations in the short period before the current time is obtained using a fast Fourier transform.

[0024] By analyzing the frequency of voltage fluctuations and the severity of temperature fluctuations in individual cells of a lithium battery pack over time within each characteristic period, voltage fluctuation degree and temperature fluctuation severity are constructed respectively. Specifically, when constructing voltage fluctuation degree, the voltage fluctuation values ​​within each voltage fluctuation characteristic period before the current time are arranged in ascending order of time to form a voltage period vector for each voltage fluctuation characteristic period. For each voltage period vector, a first-order difference sequence is obtained, and the number of all non-zero elements in the first-order difference sequence is counted, recorded as the frequency of change. The mean of the absolute values ​​of all non-zero elements in the first-order difference sequence is calculated, recorded as the average amplitude of change. The product of the frequency of change and the average amplitude of change is taken as the voltage fluctuation degree within each voltage fluctuation characteristic period. Similarly, when constructing temperature fluctuation severity, for each temperature period vector, the mean of the absolute values ​​of all elements in the first-order difference sequence of the temperature period vector is calculated, which is taken as the temperature fluctuation severity within each temperature fluctuation characteristic period.

[0025] Based on the correlation between voltage fluctuation and temperature fluctuation severity in previous characteristic periods, as well as the average levels of voltage fluctuation and temperature fluctuation severity, the safety risk level of the lithium battery pack at the current moment is obtained. Specifically, the voltage fluctuation of all voltage fluctuation characteristic periods is arranged in ascending order to form the voltage fluctuation vector at the current moment; the temperature fluctuation severity of all temperature fluctuation characteristic periods is arranged in ascending order to form the temperature fluctuation severity vector at the current moment. The correlation between the voltage fluctuation vector and the temperature fluctuation severity vector is calculated. Combining the mean values ​​of the elements in the voltage fluctuation vector and the temperature fluctuation severity vector, the safety risk level of the lithium battery pack at the current moment is obtained.

[0026] This method analyzes the voltage decrease trend and temperature increase trend of individual cells at multiple adjacent sampling times to construct the energy loss trend of the lithium battery pack at the current moment. Combined with the safety risk level of the lithium battery pack, the energy optimization demand of the lithium battery pack at the current moment is obtained. By utilizing the differences in the energy optimization demand and the current and preset adjustment amounts of the control strength within the energy controller, the feedback control strength within the energy controller at the current moment is obtained. Specifically, when constructing the energy loss trend, the multiple moments with the closest time interval to the current moment are all considered as adjacent sampling times. Based on the changes in the voltage and temperature fluctuations of individual cells at multiple adjacent sampling times, the voltage decrease trend and temperature increase trend at the current moment are obtained. Specifically, the voltage and temperature fluctuation values ​​of individual cells at multiple adjacent sampling times are arranged in ascending order of time to form the voltage trend vector and temperature trend vector at the current moment. The first-order difference vectors of the voltage and temperature trend vectors at the current moment are then calculated. The sum of the absolute values ​​of all negative values ​​within the first-order difference vector of the voltage trend vector is taken as the voltage decrease trend at the current moment, and the sum of all positive values ​​within the first-order difference vector of the temperature trend vector is taken as the temperature increase trend at the current moment. The product of the voltage decrease trend and the temperature increase trend is determined as the energy loss trend of the lithium battery pack at the current moment. When calculating the energy optimization demand, a negative exponential function of the product of the safety risk level and the energy loss trend is calculated with the natural constant as the base. The result of subtracting this exponential function from 1 is taken as the energy optimization demand of the lithium battery pack at the current moment. When calculating the feedback control strength, the feedback control strength in the energy controller at the current moment is equal to the control strength in the energy controller at the current moment, plus the difference between the energy optimization demand at the current moment and the energy optimization demand at the previous sampling moment, multiplied by the preset control strength adjustment amount, wherein the control strength adjustment amount does not exceed 10% of the control strength in the energy controller at the current moment.

[0027] Based on the current feedback control intensity and the actual control intensity within the energy controller, an adaptive control model is used to adjust the control intensity of the energy controller. Specifically, the difference between the current feedback control intensity and the actual control intensity is obtained as the control error; based on the changing trend of the control error, the proportional coefficient, integral coefficient, and derivative coefficient in the adaptive control model are adjusted, and the adjusted coefficients are used to correct the control intensity of the energy controller in real time.

[0028] Example 1: In the actual operation of lithium battery packs, safety control and energy optimization are crucial, and accurately extracting the characteristic periods of voltage and temperature fluctuations is one of the fundamental steps in the entire method. The following details the specific implementation method for characteristic period extraction.

[0029] When acquiring historical sampling data within a preset time period prior to the current moment, it is necessary to clarify the basis for setting the preset time period and the specific operation. The selection of the preset time period needs to comprehensively consider factors such as the application scenario, operating characteristics, and data acquisition frequency of the lithium battery pack. For example, in some scenarios with high real-time requirements, the preset time period may be set to a shorter time, such as 5 minutes or 10 minutes, to more promptly reflect recent changes in the operating status of the lithium battery pack; while in other scenarios requiring long-term trend analysis, the preset time period may be extended to 30 minutes or even longer. In actual operation, the voltage and temperature of each individual battery cell are collected in real time through voltage and temperature sensors configured within the lithium battery pack, according to a set sampling frequency (such as once per second or once every few seconds). This collected data is stored in the corresponding data storage unit for subsequent processing.

[0030] The voltage and temperature fluctuation values ​​from historical sampling times are arranged into vectors according to their chronological order. For voltage fluctuation values, starting from the first sampling time within a preset time period, the voltage fluctuation values ​​at each time point are arranged sequentially to form an ordered voltage fluctuation vector. Similarly, the temperature fluctuation values ​​are arranged in the same chronological order to obtain a temperature fluctuation vector. It is important to note that the chronological order of the data must be strictly accurate, as subsequent analyses are based on time series data. If the chronological order is disordered, it will lead to inaccurate feature period extraction results, thus affecting the effectiveness of the entire method.

[0031] After obtaining the voltage fluctuation vector, a time-domain transformation is performed to extract the characteristic period of the voltage fluctuation. The purpose of the time-domain transformation is to convert the time-series data of voltage fluctuations into a form that is easier to analyze for characteristic periods. Various methods can be used for the time-domain transformation. For example, one can analyze the waveform changes of the voltage fluctuations to observe their fluctuations on the time axis, or calculate the amplitude changes of the voltage fluctuation values ​​at different time points. During the time-domain transformation, it is necessary to obtain the amplitude change information at all time points in the time domain. This can be done by comparing each element in the voltage fluctuation vector with its neighboring elements to calculate the amplitude change at adjacent time points, thereby obtaining the amplitude change over the entire time series.

[0032] After obtaining the amplitude changes at all time points, it is necessary to find the time interval corresponding to the largest amplitude change. The largest amplitude change usually indicates a significant change in voltage fluctuation within that time period, and the corresponding time interval may reflect an important characteristic period of voltage fluctuation. For example, suppose the amplitude change of voltage fluctuation is the largest within a certain time period, then the length of that time period is a key time interval. Then, the reciprocal of this time interval is taken as the characteristic period of voltage fluctuation in the short period before the current moment. For example, if the time interval corresponding to the largest amplitude change is 2 minutes, then the characteristic period of voltage fluctuation is 1 / 2 minute. -1 That is, 0.5 minutes -1 .

[0033] For the temperature fluctuation vector, the Fast Fourier Transform (FFT) is used to obtain the characteristic period of the temperature fluctuation. The Fast Fourier Transform is an efficient signal processing algorithm that can convert time-domain signals into frequency-domain signals, thus facilitating the analysis of the signal's frequency components. When applying the Fast Fourier Transform, the temperature fluctuation vector is first preprocessed, such as by normalization, to ensure the accuracy and stability of the transformation. Then, the Fast Fourier Transform algorithm is used to calculate the temperature fluctuation vector, obtaining its frequency domain representation. In the frequency domain, each frequency component corresponds to a periodic component of the temperature fluctuation. By analyzing the peak positions in the frequency domain, the main characteristic period of the temperature fluctuation can be determined. For example, the period corresponding to the frequency component with the largest amplitude in the frequency domain is the characteristic period of the temperature fluctuation.

[0034] In practical applications, noise interference that may occur during data acquisition also needs to be considered. To improve the accuracy of feature period extraction, the voltage fluctuation vector and temperature fluctuation vector can be filtered to remove noise signals before performing time-domain transformation and fast Fourier transform. Commonly used filtering methods include mean filtering, median filtering, or Gaussian filtering, and the specific method can be selected according to the characteristics of the noise.

[0035] Furthermore, adjusting the preset duration and optimizing the sampling frequency are also aspects that need attention. As the operating state of the lithium battery pack changes, the preset duration may need to be dynamically adjusted to adapt to different operating stages. For example, during the start-up phase of the lithium battery pack, the preset duration can be set shorter to quickly capture voltage and temperature changes during startup; while during the stable operation phase, the preset duration can be appropriately extended to analyze long-term operating trends. Simultaneously, the sampling frequency also affects the accuracy of feature period extraction. A higher sampling frequency can provide more detailed time-series data, but it also increases the workload of data processing and the consumption of computing resources. Therefore, a trade-off needs to be struck between accuracy and efficiency.

[0036] Throughout the feature period extraction process, real-time data processing is also crucial. Since the lithium battery pack operates in real time, the collected data needs to be processed promptly to ensure that the feature period reflects the current operating status. This requires the data processing system to have sufficient computing power and real-time processing capabilities to complete vector construction, transformation, and feature period calculation within a short timeframe.

[0037] In this embodiment, the extraction of the characteristic periods of voltage and temperature fluctuation values ​​is achieved by reasonably setting the preset duration, accurately collecting historical sampling data, constructing a time series vector, employing appropriate transformation methods (time-domain transformation and fast Fourier transform), and considering various influencing factors in practical applications (such as noise and real-time performance). This process provides crucial foundational data for subsequent steps such as constructing voltage and temperature fluctuation severity, assessing safety risk levels, and optimizing energy loss, ensuring the effectiveness and accuracy of the entire lithium battery pack's multi-level safety control and energy optimization methods.

[0038] Example 2: In the safety control and energy optimization system of lithium battery packs, the construction of voltage fluctuation is a key step in assessing the stability of battery operation. Its core lies in quantifying the frequency and amplitude characteristics of voltage fluctuations to provide a precise basis for subsequent safety risk level determination. The specific construction process of voltage fluctuation in Example 2 is described in detail below.

[0039] The voltage fluctuation values ​​within each voltage fluctuation characteristic period prior to the current moment need to be time-series organized. Here, "each voltage fluctuation characteristic period" originates from the characteristic period results extracted in Example 1; each period corresponds to a time interval with a specific fluctuation pattern. In actual operation, the system will filter the voltage fluctuation values ​​within each period from historical voltage data based on the start and end times of the characteristic period, and arrange them strictly in ascending order of time. For example, if there are three characteristic periods before the current moment: period 1 corresponds to 0-10 minutes, period 2 to 10-25 minutes, and period 3 to 25-40 minutes, then the voltage fluctuation values ​​within these three time periods need to be extracted separately to form three independent voltage period vectors. The order of elements in each vector must be strictly consistent with the sampling time to ensure the timing accuracy of subsequent differential calculations.

[0040] After constructing the voltage period vector, a first-order difference operation needs to be performed on each vector. The essence of the first-order difference is to calculate the change in voltage fluctuation values ​​at adjacent time points; its mathematical logic is as follows: for a vector... Its first-order difference sequence In engineering implementation, this operation can be implemented programmatically by iterating through vector elements and calculating the differences between adjacent values. It's important to note that difference operations have strict requirements on the data sampling frequency. If the sampling interval is uneven, the data must be resampled to ensure the physical meaning of the difference result. For example, when the sampling interval is 1 minute, the difference result represents the voltage fluctuation per minute; if the sampling interval fluctuates, the difference result will deviate from the actual fluctuation rate, affecting subsequent analysis.

[0041] After completing the first-order difference sequence calculation, a two-dimensional analysis is required: frequency of change statistics and average change amplitude calculation. The statistical logic for frequency of change involves counting the number of non-zero elements in the difference sequence, which physically represents the number of effective changes in voltage fluctuation within that characteristic period. Here, "effective change" is defined as a difference in voltage fluctuation values ​​between adjacent sampling points, i.e., the difference result is not zero. For example, if there are 20 non-zero elements in the difference sequence of a certain characteristic period, it indicates that the voltage fluctuation has undergone 20 effective changes within that period. It is important to note that if the sampling frequency is too high, noise signals may be included in the frequency of change. Therefore, threshold filtering can be applied to the difference sequence before statistics to remove small changes smaller than a preset noise threshold, avoiding incorrect statistics.

[0042] The average amplitude of voltage fluctuation is calculated by taking the average of the absolute values ​​of all non-zero elements in the difference sequence. The specific steps are: first, iterate through the difference sequence, summing the absolute values ​​of all non-zero elements, and then divide by the number of non-zero elements (i.e., the frequency of change). For example, if the sum of the absolute values ​​of the non-zero elements in a difference sequence is 150mV and the frequency of change is 20 times, then the average amplitude of change is 150 / 20 = 7.5mV / time. This indicator reflects the average amplitude of voltage fluctuations and can distinguish between "high-frequency small-amplitude fluctuations" and "low-frequency large-amplitude fluctuations." In practical applications, the unit of the average amplitude of change must be consistent with the unit of the voltage fluctuation value, and attention must be paid to numerical precision during the calculation process to avoid deviations in results due to floating-point operation errors.

[0043] Voltage fluctuation is obtained by multiplying the frequency of change by the average amplitude of change. The physical meaning of this calculation lies in coupling and quantifying the "frequency characteristic" and "amplitude characteristic" of the fluctuation to form a comprehensive evaluation index. For example, if the frequency of change in a certain characteristic period is 20 times, and the average amplitude of change is 7.5mV / time, then the voltage fluctuation is 20 × 7.5 = 150 mV. In practical systems, the numerical range of voltage fluctuation needs to be normalized according to parameters such as battery type and operating voltage range to facilitate horizontal comparison under different operating conditions. For example, for a lithium battery pack with a nominal voltage of 3.7V, the fluctuation value can be divided by 3.7V for normalization to obtain a dimensionless fluctuation index.

[0044] In engineering implementation, the following practical issues also need to be considered: First, data storage and processing efficiency. When the number of characteristic periods is large and the sampling frequency is high, the dimension of the voltage period vector will increase significantly. Efficient data structures (such as arrays or matrices) need to be used to store the vector, and parallel computing technology should be used to accelerate the differential operation and statistical process. Second, handling of abnormal data. If there are sudden changes in voltage data within a certain characteristic period (such as jump values ​​caused by sensor failure), it needs to be corrected using methods such as sliding window filtering or neighborhood interpolation to avoid interference from outliers in the fluctuation calculation. Third, real-time control. The calculation of voltage fluctuation needs to be updated synchronously with the battery operating status. Therefore, a reasonable calculation cycle needs to be set, such as updating the fluctuation index every 5 minutes, to ensure that the evaluation results can reflect the current state of the battery in a timely manner.

[0045] Furthermore, the threshold setting for voltage fluctuation needs to be combined with battery safety specifications and historical operating data. For example, by analyzing a large amount of fluctuation data under normal operating conditions, a normal distribution model can be established, and fluctuations exceeding three standard deviations can be considered abnormal, triggering a safety warning mechanism. At the same time, different types of lithium batteries (such as lithium iron phosphate and ternary lithium) have different normal fluctuation ranges due to differences in electrochemical characteristics, requiring the establishment of differentiated evaluation standards for battery types.

[0046] Example 3: In the safe operation and management of lithium battery packs, accurately assessing their safety risk level is a core element in achieving multi-level safety control, and constructing the temperature fluctuation severity and obtaining the safety risk level are key steps in this process. The following will detail the process of constructing the temperature fluctuation severity and obtaining the current safety risk level of the lithium battery pack.

[0047] When constructing the temperature fluctuation intensity, it is necessary to process the temperature fluctuation values ​​within each temperature fluctuation characteristic period. These temperature fluctuation characteristic periods are consistent with those extracted in Example 1, with each period corresponding to a specific time period of temperature change during the operation of the lithium battery pack. In actual operation, the system extracts the corresponding temperature fluctuation values ​​from historical temperature data based on the time range of each temperature fluctuation characteristic period and arranges them in ascending order of time, thereby forming a temperature period vector for each temperature fluctuation characteristic period. For example, if there are two temperature fluctuation characteristic periods before the current moment, with the first period being 0 to 15 minutes and the second period being 15 to 35 minutes, the system will extract the temperature fluctuation values ​​within these two time periods respectively, forming two temperature period vectors. The elements in each vector are strictly arranged in chronological order of sampling time to ensure the accuracy of subsequent analysis.

[0048] For each temperature period vector, the mean of the absolute values ​​of all elements in its first-order difference sequence needs to be calculated. This mean is used as the intensity of the temperature fluctuation in that characteristic period. The first-order difference sequence is calculated by sequentially calculating the difference between any two adjacent elements in the temperature period vector, resulting in a new sequence. For example, for a temperature period vector... Its first-order difference sequence is After obtaining the first-order difference sequence, it is necessary to calculate the mean of the absolute values ​​of all elements in the sequence. Specifically, this involves taking the absolute value of each element in the first-order difference sequence, summing these absolute values, and then dividing by the number of elements in the sequence. The result is the temperature fluctuation intensity. This indicator can intuitively reflect the severity of temperature fluctuation over time within a specific temperature fluctuation period. The greater the temperature fluctuation intensity, the more drastic the temperature change of the lithium battery pack within this period, which may indicate that the lithium battery pack's operating state is not stable enough and poses certain safety hazards.

[0049] After constructing the temperature fluctuation severity, the next step is to obtain the safety risk level of the lithium battery pack at the current moment. This requires arranging the voltage fluctuations of all voltage fluctuation characteristic periods in ascending order of time to form the voltage fluctuation vector for the current moment. The voltage fluctuations here are derived from the voltage fluctuations of each voltage fluctuation characteristic period calculated in Example 2. Similarly, the temperature fluctuation severity of all temperature fluctuation characteristic periods is arranged in ascending order of time to form the temperature fluctuation severity vector for the current moment. The construction of these two vectors must strictly adhere to the chronological order to ensure that each element in the vector corresponds to the correct time period, thereby guaranteeing the accuracy of subsequent correlation calculations.

[0050] The correlation between the voltage fluctuation vector and the temperature fluctuation severity vector is calculated. Various methods can be used to calculate the correlation, such as the Pearson correlation coefficient method. This method measures the degree of linear correlation between the two vectors, and its value typically ranges from -1 to 1. A correlation coefficient of 1 indicates a perfect positive correlation; a correlation coefficient of -1 indicates a perfect negative correlation; and a correlation coefficient of 0 indicates no linear correlation. In practical applications, calculating the correlation coefficient between these two vectors reveals the degree of association between voltage fluctuation and temperature fluctuation severity. For example, a high correlation coefficient indicates a strong correlation between voltage and temperature fluctuations, potentially meaning that voltage and temperature changes in the lithium battery pack interact during operation, requiring closer monitoring of its safety status.

[0051] In addition to calculating the correlation, it is also necessary to consider the mean values ​​of elements within the voltage fluctuation vector and the mean values ​​of elements within the temperature fluctuation severity vector. These two means reflect the average levels of voltage fluctuation and temperature fluctuation severity, respectively. For example, a higher mean value in the voltage fluctuation vector indicates a greater average degree of voltage fluctuation over the voltage fluctuation characteristic periods prior to the current moment; similarly, a higher mean value in the temperature fluctuation severity vector indicates a greater average degree of temperature fluctuation. These two mean indicators can reflect the average state of voltage and temperature fluctuations in the lithium battery pack as a whole, and are important bases for assessing the safety risk level.

[0052] By considering the correlation between the voltage fluctuation vector and the temperature fluctuation intensity vector, as well as the mean of these two vectors, a specific algorithm or model is used to determine the current safety risk level of the lithium battery pack. In practical applications, safety risk levels can be categorized into different levels, such as low risk, medium risk, high risk, and extremely high risk. For example, a high risk level can be defined as a high correlation with both high means, and a low risk level as a low correlation with both low means. The specific classification criteria need to be determined based on factors such as the type of lithium battery pack, the usage scenario, and historical operating data to ensure that the safety risk level assessment accurately reflects the actual safety status of the lithium battery pack.

[0053] Throughout the process, attention must be paid to the accuracy and real-time nature of the data. For example, when extracting temperature fluctuation values ​​and voltage fluctuations, it is essential to ensure that data acquisition and processing are error-free; when constructing vectors and calculating correlation and mean values, the correctness of the algorithm and the accuracy of the calculation must be guaranteed. Furthermore, to improve the reliability of the safety risk level assessment, more influencing factors can be considered, such as the charge / discharge state of the lithium battery pack and the consistency between individual cells. However, in this embodiment, the assessment is primarily based on the severity of voltage and temperature fluctuations, their correlation, and mean values.

[0054] Furthermore, special circumstances in practical applications need to be considered. For example, when anomalies are found in the data within a certain temperature fluctuation characteristic period or voltage fluctuation characteristic period, these anomalies need to be processed, such as being removed or corrected, to avoid their impact on the safety risk level assessment results. Simultaneously, as the lithium battery pack's operating time increases, historical data needs to be continuously updated to ensure that the assessment results reflect the current actual operating status of the lithium battery pack.

[0055] Example 4: In the energy management system of lithium battery packs, the construction of energy loss trend and the calculation of energy optimization requirements are key links connecting safety status assessment and actual control actions. The core lies in dynamically capturing the changing trends of voltage and temperature, quantifying the system's energy loss risk, and transforming it into optimization requirements. The implementation method of Example 4 is described in detail below with specific examples.

[0056] Taking a certain type of ternary lithium battery pack as an example, assuming it contains 16 individual cells and operates in an on-board energy storage scenario, when the system sampling time t=10:30:00, it is necessary to construct the energy loss trend at the current moment. First, determine the adjacent sampling times. Here, the five sampling points with the closest time interval are selected, namely t-1=10:29:50, t-2=10:29:40, t-3=10:29:30, t-4=10:29:20, and t-5=10:29:10. The voltage and temperature data at these five moments form the basis of the analysis.

[0057] When constructing the voltage drop trend vector, the system first extracts the individual cell voltage fluctuation values ​​at these five time points. Assume the voltage fluctuation values ​​(unit: mV) at each time point are as follows: t-5 120, t-4 115, t-3 110, t-2 105, t-1 100. These values ​​are then arranged in ascending order of time to form the voltage trend vector. Then, its first-order difference vector, i.e. the difference between adjacent time points, is calculated: 115-120=-5, 110-115=-5, 105-110=-5, 100-105=-5, thus obtaining the difference vector. Since all elements in the difference vector are negative, the sum of their absolute values ​​is 5 + 5 + 5 + 5 = 20. This value represents the voltage decrease trend at the current moment, in mV. This indicates that in the last 50 seconds, the individual cell voltage has been decreasing by 5mV every 10 seconds, showing a clear overall downward trend.

[0058] The temperature trend vector is constructed based on the temperature fluctuation values ​​at adjacent sampling times within the same batch. Assume the temperature fluctuation values ​​(unit: °C) at each time point are as follows: t-5: 25, t-4: 26, t-3: 28, t-2: 30, t-1: 32. Arrange these values ​​in ascending order of time to form the temperature trend vector. Calculate its first-order difference vector: 26-25=1, 28-26=2, 30-28=2, 32-30=2, to obtain the difference vector. All elements in this difference vector are positive, and their sum is 1+2+2+2=7, representing the current temperature increase trend in °C. This indicates that in the last 50 seconds, the temperature of a single cell has increased at a rate of approximately 1.4 °C per 10 seconds, showing a significant upward temperature trend.

[0059] Multiplying the voltage drop trend (20) by the temperature rise trend (7) yields an energy loss trend score of 20 × 7 = 140. This energy loss trend score is a comprehensive indicator. Its physical meaning is that a continuous voltage drop indicates battery capacity loss, while a continuous temperature rise reflects increased internal impedance or exacerbated side reactions. The product of these two factors amplifies the risk of energy loss. For example, even when the voltage drop trend is slow but the temperature rise is rapid, the product result will still be high, suggesting the need to pay attention to abnormal internal energy loss.

[0060] When calculating the energy optimization demand, it is assumed that the safety risk level obtained through Example 3 at the current moment is medium risk, defined as a value of 2 (assuming that the safety risk level is divided into 1-4 levels, with 1 being low risk and 4 being extremely high risk). Using the natural constant e as the base, the product of the safety risk level and the energy loss trend is calculated to be 2 × 140 = 280, and the negative exponential function value is e. -280 Since this value approaches 0, subtracting the exponential function from 1 yields approximately 1, meaning the energy optimization demand is close to 1. This indicates that under the medium-risk safety level, combined with a significant energy loss trend, the system's need for energy optimization is extremely urgent, requiring immediate activation of the control mechanism.

[0061] In another example, if the safety risk level is low (value 1) and the energy loss trend degree is 50, then the product is 50, and the negative exponential function value is e. -50 Subtracting this value from 1 yields approximately 0.9999997, indicating that the energy optimization demand level is close to 1 but slightly lower than the former. This demonstrates that even with a low safety risk level, energy optimization is still necessary if the energy loss trend is significant, reflecting the coupled consideration of dual factors by this indicator.

[0062] In practical engineering applications, the selection of the number of adjacent sampling times needs to balance computational efficiency and trend accuracy. If too few times are selected (e.g., 2 times), the trend analysis may be affected by random fluctuations; if too many times are selected (e.g., 20 times), it will increase the computational load and may include distant historical data, failing to reflect the latest trend. It is usually determined based on the sampling frequency; for example, when sampling once per second, selecting 5-10 adjacent times is appropriate.

[0063] The construction of voltage and temperature trend vectors requires attention to data synchronization. In the example above, both voltage and temperature data are from the same batch of five sampling times, ensuring a consistent time base for trend analysis. If temperature data is missing at a certain time, the system must supplement it using methods such as linear interpolation or previous value hold, to avoid trend deviations caused by data asynchrony.

[0064] Furthermore, trend calculations are highly sensitive to outliers. For example, if the voltage fluctuation suddenly drops to zero due to a sensor malfunction, it can cause an abnormally negative value in the difference vector, resulting in a sharp increase in the voltage drop trend. Therefore, in practical systems, outlier detection mechanisms must be incorporated, such as using the 3σ principle to remove data points that significantly deviate from the mean, or employing sliding window filtering to smooth the data, to ensure the reliability of trend calculations.

[0065] The energy optimization demand index ranges from 0 to 1, and its non-linear calculation method (negative exponential function) highlights the characteristics of high risk. When the product of the safety risk level and the energy loss trend is small, the demand index increases slowly; when the product exceeds a certain threshold, the demand index rapidly approaches 1. This non-linear relationship aligns with the engineering logic of "the higher the risk, the higher the priority for handling." For example, when the product increases from 100 to 200, the demand index increases from approximately 0.9995 to 0.999999, a significant increase that can promptly trigger high-intensity regulation.

[0066] In automotive applications, this mechanism can be integrated with the cooling system of the Battery Management System (BMS). When the energy optimization demand level approaches 1, the BMS automatically increases the cooling fan speed and adjusts the charging and discharging current to suppress the temperature rise and slow down the voltage drop, thereby reducing energy loss. In energy storage power station scenarios, it may be combined with balancing circuits or load scheduling strategies to achieve energy redistribution and optimization.

[0067] Example 5: In the energy optimization control of lithium battery packs, the calculation of feedback regulation intensity and the adjustment of the adaptive control model are key aspects of achieving precise energy management. The core of this approach lies in dynamically adjusting the regulation intensity to respond to changes in energy optimization needs and making real-time corrections based on the actual regulation effect. The implementation method of Example 5 is described in detail below with specific examples.

[0068] Taking a lithium battery pack in an energy storage system as an example, assume that the energy controller's regulation intensity needs to be adjusted in real time during a certain operating phase. Let the current time be t, the previous sampling time be t-1, and the preset regulation intensity adjustment amount be 5% of the current regulation intensity (with a limit not exceeding 10%).

[0069] First, calculate the feedback control strength at the current moment. Assume that at time t-1, the energy controller's control strength is 80 (this is a normalized value for ease of calculation and explanation), the energy optimization demand at time t-1 is 0.6, and the energy optimization demand at the current moment t is 0.8. Therefore, the difference in energy optimization demand between the current moment and the previous moment is 0.8 - 0.6 = 0.2. The preset control strength adjustment is 5% of the current control strength (i.e., the control strength of 80 at time t-1), which calculates to an adjustment of 80 × 5% = 4. According to the calculation method for feedback control strength, the feedback control strength at the current moment t is equal to the control strength of 80 at time t-1, plus the product of the demand difference of 0.2 and the preset adjustment of 4, i.e., 80 + 0.2 × 4 = 80 + 0.8 = 80.8. At this point, the feedback control intensity is 80.8, which is an increase compared to the control intensity at the previous moment. This is because the current energy optimization demand has increased compared to the previous moment, and the control intensity needs to be increased to meet the energy optimization demand.

[0070] In another scenario, if the control intensity at time t-1 remains 80, the energy optimization demand at time t-1 is 0.8, and the energy optimization demand at the current time t is 0.5, the difference in demand is 0.5 - 0.8 = -0.3. The preset control intensity adjustment remains 5% of 80, i.e., 4. Therefore, the feedback control intensity is 80 + (-0.3) × 4 = 80 - 1.2 = 78.8. In this case, the feedback control intensity decreases, indicating that due to the reduced energy optimization demand, the control intensity also decreases accordingly to avoid over-control.

[0071] In practical applications, the value of the preset control intensity adjustment needs to comprehensively consider the characteristics of the lithium battery pack and the operating scenario. For example, for lithium battery packs with larger capacity, the adjustment amount of the control intensity may be relatively large, but it must adhere to the limitation of not exceeding 10% of the current control intensity to ensure the stability and safety of the control process. If the adjustment amount is too large, it may cause large fluctuations in the output of the energy controller, affecting the normal operation of the lithium battery pack; if the adjustment amount is too small, it may not be able to respond to changes in energy optimization requirements in a timely manner, resulting in poor energy optimization effects.

[0072] Next, the adaptive control model is used to adjust the regulation intensity of the energy controller. Assuming the feedback regulation intensity at time t is 80.8, while the actual regulation intensity is 80, then the control error is 80.8 - 80 = 0.8. At this point, the proportional, integral, and derivative coefficients in the adaptive control model need to be adjusted according to the changing trend of the control error.

[0073] For example, if the control error gradually increases in the first few sampling times, it indicates that the deviation between the actual control strength and the feedback control strength is widening. In this case, the proportional gain needs to be increased to accelerate the response to the error, allowing the actual control strength to approach the feedback control strength more quickly. Conversely, if the error changes slowly, the integral gain may need to be adjusted appropriately to eliminate accumulated error and ensure long-term control accuracy. The derivative gain is used to predict the trend of error change. If the error changes rapidly, increasing the derivative gain allows for earlier adjustments to the control strength, improving system stability.

[0074] Adjusting the coefficients requires specific algorithms or rules. For example, it can be set that when the control error exceeds a certain threshold, the proportional coefficient is increased by a certain percentage; when the rate of change of the error exceeds a certain range, the derivative coefficient is adjusted accordingly. These adjustment rules need to be set based on the actual operating data and experience of the lithium battery pack to ensure that the adaptive control model can effectively correct the control intensity in real time.

[0075] Assume that after adjustment, the proportional gain increases from 0.5 to 0.6, the integral gain is adjusted from 0.2 to 0.25, and the derivative gain is adjusted from 0.1 to 0.15. The adjusted coefficients are used to correct the energy controller's regulation intensity in real time. The corrected regulation intensity will comprehensively consider the current error, the cumulative error, and the rate of error change, thereby more accurately tracking and responding to the regulation intensity.

[0076] In actual operation, the control error may change continuously. The adaptive control model needs to continuously monitor the trend of error change and adjust the coefficients accordingly. For example, when the error gradually decreases and approaches 0, the proportional coefficient can be appropriately reduced to avoid excessive adjustment of the control strength, which could lead to system oscillation. The integral coefficient can also be adjusted according to the accumulation of error to maintain the stability of control. The derivative coefficient can be reduced as the rate of error change decreases to reduce sensitivity to noise.

[0077] Furthermore, various disturbances during the operation of the lithium battery pack must be considered, such as changes in ambient temperature and sudden load fluctuations. These factors can lead to significant errors between the actual control force and the feedback control force. The adaptive control model needs to be able to quickly adapt to these disturbances and maintain the accuracy of control by adjusting the coefficients in a timely manner.

[0078] For example, when the external environmental temperature suddenly rises, the temperature rising trend of the lithium battery pack intensifies, and the energy loss trend degree may increase, resulting in an increase in the energy optimization demand degree, and the feedback regulation strength will also increase correspondingly. At this time, the actual regulation strength may not be able to keep up with the feedback regulation strength in time due to the influence of temperature change, resulting in a large control error amount. The adaptive control model needs to quickly detect this error change trend, increase the proportional coefficient and the differential coefficient to accelerate the adjustment speed of the regulation strength and adapt to the influence brought by the temperature change.

[0079] In the energy storage system, the adjustment of the adaptive control model also needs to cooperate with other subsystems. For example, when the regulation strength of the energy controller is adjusted, it may be necessary to simultaneously adjust the charge and discharge strategies of the battery, the operating state of the cooling system, etc. to achieve overall energy optimization and safety control.

[0080] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0081] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A multi-stage safety control and energy optimization method for lithium battery packs, characterized in that, The method comprises the following steps: obtaining voltage fluctuation values and temperature fluctuation values of single batteries in a lithium battery pack during operation; performing time domain transformation on the voltage fluctuation values and the temperature fluctuation values to extract characteristic periods of the voltage fluctuation values and the temperature fluctuation values; constructing voltage fluctuation degrees and temperature fluctuation intensities by analyzing the degree of frequent voltage fluctuation and the degree of temperature fluctuation intensity of the single batteries in the lithium battery pack in each characteristic period; obtaining a safety risk level of the lithium battery pack at the current time according to the correlation between the voltage fluctuation degrees and the temperature fluctuation intensities of each characteristic period before the current time, and the average level of the voltage fluctuation degrees and the average level of the temperature fluctuation intensities; constructing an energy loss trend degree of the lithium battery pack at the current time by analyzing voltage drop trends and temperature rise trends of the single batteries at a plurality of adjacent sampling times of the current time, and obtaining an energy optimization demand degree of the lithium battery pack at the current time in combination with the safety risk level of the lithium battery pack, and obtaining a feedback control intensity in the energy controller at the current time by using a change difference of the energy optimization demand degree and a preset control intensity adjustment amount of the control intensity in the energy controller at the current time; controlling and adjusting the control intensity of the energy controller by using an adaptive control model according to the feedback control intensity and an actual control intensity in the energy controller at the current time.

2. The multi-level safety control and energy optimization method for a lithium battery pack as described in claim 1, characterized in that, The extraction of the characteristic periods of the voltage fluctuation values and the temperature fluctuation values further comprises: obtaining voltage fluctuation values and temperature fluctuation values of all historical sampling times within a preset time length before the current time to form a voltage fluctuation vector and a temperature fluctuation vector in time sequence as the voltage fluctuation vector and the temperature fluctuation vector at the current time; performing time domain transformation on the voltage fluctuation vector at the current time to obtain amplitude changes at all time points in the time domain, and taking the reciprocal of the maximum amplitude change corresponding to the time interval as the voltage fluctuation characteristic period in a short time before the current time; correspondingly, for the temperature fluctuation vector, the temperature fluctuation characteristic period in a short time before the current time is obtained by using fast Fourier transformation.

3. The method of claim 1, wherein the method further comprises: determining if the battery is in a state of overcharge, overdischarge, or overcurrent; and if the battery is in a state of overcharge, overdischarge, or overcurrent, then reducing the power output of the battery. The construction of the voltage fluctuation degree further comprises: arranging the voltage fluctuation values in each voltage fluctuation characteristic period in ascending order of time to form a voltage period vector of each voltage fluctuation characteristic period, and correspondingly, for the temperature fluctuation values in each temperature fluctuation characteristic period, a temperature period vector of each temperature fluctuation characteristic period is obtained; for each voltage period vector, a first-order difference sequence of the voltage period vector is obtained, the number of all non-zero elements in the first-order difference sequence is counted and recorded as a change frequency, and the mean value of the absolute values of all non-zero elements in the first-order difference sequence is calculated and recorded as an average change amplitude, and the product of the change frequency and the average change amplitude is taken as the voltage fluctuation degree in each voltage fluctuation characteristic period.

4. The multi-level safety control and energy optimization method for a lithium battery pack as described in claim 3, characterized in that, The construction of the temperature fluctuation intensity further comprises: for each temperature period vector, the mean value of the absolute values of all elements in the first-order difference sequence of the temperature period vector is calculated as the temperature fluctuation intensity of each temperature fluctuation characteristic period.

5. The method of claim 4, wherein the method further comprises: determining if the battery is in a state of overcharge; and if the battery is in a state of overcharge, then reducing the charging current to the battery to a predetermined value. 5 The obtaining process of the safety risk level of the lithium battery pack at the current time is: The voltage fluctuation degrees of all voltage fluctuation characteristic periods are arranged in ascending order of time to form a voltage fluctuation degree vector of the current time, the temperature fluctuation intensities of all temperature fluctuation characteristic periods are arranged in ascending order of time to form a temperature fluctuation intensity vector of the current time, the correlation between the voltage fluctuation degree vector and the temperature fluctuation intensity vector is calculated, and the mean of the elements in the voltage fluctuation degree vector and the mean of the elements in the temperature fluctuation intensity vector are combined to obtain the safety risk level of the lithium battery pack at the current time.

6. The method of claim 1, wherein the method further comprises: determining if the battery is in a state of overcharge, overdischarge, or overcurrent; and if the battery is in a state of overcharge, overdischarge, or overcurrent, then reducing the power output of the battery. The construction of the energy loss trend degree of the lithium battery pack at the current time comprises: The single battery voltage fluctuation values and the temperature fluctuation values of the plurality of adjacent sampling times closest to the current time are obtained, and the voltage drop trend amount and the temperature rise trend amount at the current time are obtained according to the changes of the single battery voltage fluctuation values and the temperature fluctuation values of the plurality of adjacent sampling times. The product of the voltage drop trend amount and the temperature rise trend amount at the current time is determined as the energy loss trend degree of the lithium battery pack at the current time.

7. The multi-level safety control and energy optimization method for a lithium battery pack as described in claim 6, characterized in that, The voltage drop trend amount and the temperature rise trend amount at the current time further comprise: The single battery voltage fluctuation values and the temperature fluctuation values of the plurality of adjacent sampling times are arranged in ascending order of time to form a voltage trend vector and a temperature trend vector at the current time, and the first-order difference vectors of the voltage trend vector and the temperature trend vector at the current time are calculated respectively. The sum of the absolute values of all negative values in the first-order difference vector of the voltage trend vector is taken as the voltage drop trend amount at the current time, and the sum of all positive values in the first-order difference vector of the temperature trend vector is taken as the temperature rise trend amount at the current time.

8. The multi-level safety control and energy optimization method for a lithium battery pack as described in claim 1, characterized in that, The calculation process of the energy optimization demand degree of the lithium battery pack at the current time comprises: The natural constant is taken as the base number to calculate the negative exponential function of the product of the safety risk level and the energy loss trend degree, and the result of taking 1 minus the exponential function is taken as the energy optimization demand degree of the lithium battery pack at the current time.

9. The method for multi-stage safety control and energy optimization of a lithium battery pack of claim 1, wherein, The calculation process of the feedback control strength in the energy controller at the current time comprises: The feedback control strength in the energy controller at the current time is equal to the control strength in the energy controller at the current time, plus the difference between the energy optimization demand degree at the current time and the energy optimization demand degree at the previous sampling time, multiplied by the preset control strength adjustment amount, wherein the control strength adjustment amount does not exceed 10% of the control strength in the energy controller at the current time.

10. A multi-level safety control and energy optimization method for a lithium battery pack as described in claim 1, characterized in that, The process of controlling and adjusting the control strength of the energy controller by using the adaptive control model further comprises: The difference between the feedback control strength at the current time and the actual control strength is taken as the control error amount; According to the change trend of the control error amount, the proportional coefficient, the integral coefficient and the differential coefficient in the adaptive control model are adjusted, and the control strength of the energy controller is corrected in real time by using the adjusted coefficients.

Images