Battery state monitoring and management system based on Internet of Things

Abstract

The invention relates to the technical field of battery management, in particular to a battery state monitoring and management system based on the Internet of Things, which comprises a signal acquisition module, a voltage evaluation module, a temperature analysis module, a temperature evaluation module and a state adjustment module. According to the method, a comprehensive health index capable of reflecting the dynamic performance and the long-term aging trend of the battery at the same time is constructed by calculating the change rate of the state of charge and the attenuation rate of the internal resistance, and thermal imbalance and abnormal regions in the battery pack are accurately identified by combining the difference comparison between the temperature change rate of each local region and the global average rate. The overall health degree and the local thermal state are further comprehensively evaluated through a probability model, a more accurate state judgment result is output, finally, differentiated temperature control and charging and discharging adjustment instructions are generated according to the difference degree of local thermal response, accurate intervention on a problem area and normal management on a health area are achieved, and the reliability of the system is improved. And the depth of battery state monitoring and the fineness of management are obviously improved.

Description

Technical Field

[0001] This invention relates to the field of battery management technology, and in particular to a battery status monitoring and management system based on the Internet of Things. Background Technology

[0002] The field of battery management technology involves multiple aspects such as battery monitoring, management and maintenance, mainly including battery status monitoring, battery charge and discharge control, battery protection and battery health assessment.

[0003] The battery status monitoring and management system refers to a system that monitors and manages the battery status in real time in application scenarios such as two-wheeled and three-wheeled electric vehicles and robots. The system mainly collects data such as battery voltage, current, and temperature to analyze the battery's health status and perform operations such as charge and discharge management, equalization charging, and overcharge protection.

[0004] Current battery status monitoring technologies primarily rely on macroscopic data such as overall voltage and temperature for assessment. This approach lacks a refined understanding of the internal state of the battery pack and cannot effectively identify performance inconsistencies caused by aging or differences in operating conditions between different areas. Consequently, potential risks in certain areas are difficult to detect in advance. For example, abnormal temperature rises or accelerated degradation in individual cells may be masked by averaged data. As a result, management strategies can only be implemented in a uniform and extensive manner, failing to specifically alleviate the pressure on specific areas. Over time, this will exacerbate the imbalance of the internal state of the battery pack, affecting its overall performance consistency and shortening its effective lifespan. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing an Internet of Things-based battery status monitoring and management system.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A battery status monitoring and management system based on the Internet of Things includes: The signal acquisition module collects temperature and voltage signals from inside the robot's battery pack, converts the temperature signal into a temperature value, calculates the rate of change of state of charge based on the change of voltage signal over time, and forms multi-source synchronous data. The voltage assessment module reads the battery internal resistance increment over time and the initial internal resistance, calculates the battery health degradation rate, compares the state of charge change rate in the multi-source synchronous data with the battery health degradation rate, and obtains the battery health status assessment index. The temperature analysis module divides the battery pack into local areas based on the location distribution of the temperature sensors inside the battery pack. It extracts the temperature value of each local area from the multi-source synchronous data, calculates the rate of temperature change over time, compares it with the average rate of temperature change of the entire battery pack, and records the local thermal response difference data. The temperature assessment module comprehensively assesses the overall health and thermal state of the battery pack based on the battery health status assessment indicators and the local thermal response difference data, and generates a health assessment result. The status adjustment module optimizes the temperature control and charging / discharging commands for each local area of ​​the battery pack based on the health assessment results, and generates battery management results.

[0007] As a further aspect of the present invention, the multi-source synchronous data specifically includes temperature signal, state of charge change rate, and voltage signal; the battery health status assessment indicators include health degradation rate, state of charge change rate, and internal resistance increment; the local thermal response difference data includes local area temperature change rate, global average temperature change rate, and temperature change delay; the health assessment results include thermal response difference assessment and overall health status assessment; and the battery management results include temperature control command, charge / discharge command, and health management adjustment command.

[0008] As a further aspect of the present invention, the signal acquisition module includes: The network data acquisition submodule collects temperature and voltage signals inside the robot's battery pack through the Internet of Things sensor network, calls an analog-to-digital converter to convert the temperature signals into temperature values, and integrates the temperature values ​​and voltage signals based on a synchronization timestamp to establish a network sensor signal set. The battery state rate calculation submodule calls the voltage signal from the network sensor signal set, calculates the voltage difference between two adjacent time points, calculates the voltage change over time based on the voltage difference and the corresponding time interval, and calculates the state of charge change rate based on the change and the known capacity of the battery. The multi-dimensional data generation submodule takes the temperature value in the network sensor signal set and calls the rate of change of state of charge. It aligns and associates the two according to the synchronization timestamp, and synchronizes the aligned data through the Internet of Things network to generate multi-source synchronized data.

[0009] As a further aspect of the present invention, the voltage evaluation module includes: The internal resistance data reading submodule reads the battery internal resistance increment over time and the initial internal resistance to form a set of battery internal resistance parameters. The health degradation rate calculation submodule calculates the ratio of the battery internal resistance increment over time to the initial internal resistance in the battery internal resistance parameter set, which is used as the battery health degradation rate. The comprehensive health status assessment submodule calculates the ratio of the rate of change of state of charge in the multi-source synchronous data to the rate of battery health degradation, which serves as the battery health status assessment index.

[0010] As a further aspect of the present invention, the temperature analysis module includes: The region division and data extraction submodule divides the battery pack into local regions based on the location distribution of the temperature sensors inside the battery pack, extracts the temperature value of each local region from the multi-source synchronous data, and classifies and integrates the extracted temperature values ​​according to the divided local regions to establish a regional temperature dataset. The temperature rate calculation submodule calculates the temperature change rate of each local area based on the temperature value of each local area in the regional temperature dataset over time, calculates the global average temperature change rate of the battery pack based on the temperature values ​​of all local areas, and integrates the local rate and the global average rate to generate a temperature change rate set. The thermal response difference comparison submodule calls the temperature change rate set, compares the temperature change rate of each local area with the average temperature change rate of the battery pack, and records the difference value generated by the comparison to obtain local thermal response difference data.

[0011] As a further aspect of the present invention, the temperature evaluation module includes: The data integration input submodule associates the battery health status assessment index, which reflects the overall health status of the battery pack, with the local thermal response difference data, which reflects the local area of ​​the battery pack, to establish a comprehensive assessment input set. The state probability calculation submodule inputs the comprehensive evaluation input set into the Bayesian discriminant model, calculates the posterior probability of the battery pack belonging to multiple health states based on the preset prior probabilities and conditional probabilities within the Bayesian discriminant model, and obtains the health state membership probability. The comprehensive state discrimination submodule selects the state with the highest probability from the posterior probabilities of various health states based on the health state membership probability, and determines the overall health state and thermal state of the battery pack to generate a health assessment result.

[0012] As a further aspect of the present invention, the process of setting the prior probability and the conditional probability includes: Acquire historical operating data of multiple battery groups throughout their complete life cycle. The historical operating data includes battery health status assessment indicators at all time points, local thermal response difference data, and corresponding real health status labels. The frequency of each true health status label in historical operating data is counted, and the prior probability of the battery pack belonging to multiple health statuses is calculated based on the ratio of the frequency to the total data. Historical operating data is categorized based on the actual health status label. Within each category, the numerical range of battery health status assessment indicators and local thermal response difference data is discretized into segments. The frequency of data occurrence within each segment is counted, and a probability lookup table representing the distribution of input data under a specified health status is established. This probability lookup table is then used as a conditional probability.

[0013] As a further aspect of the present invention, the state adjustment module includes: The problem area identification submodule determines whether the overall health status and thermal status of the battery pack are in an abnormal range based on the health assessment results. If the determination is yes, it calls the local thermal response difference data, analyzes and identifies local areas where the difference data exceeds the normal range, and establishes a list of abnormal state areas. The regional instruction optimization submodule optimizes the temperature control instruction parameters and charge / discharge instruction parameters for each local region in the list of abnormal states, and combines the optimized two types of instructions to generate a health adjustment instruction set. The management strategy generation submodule compiles the temperature control instructions and charge / discharge instructions for each local area of ​​the battery pack into a unified management strategy based on the health adjustment instruction set, and generates battery management results.

[0014] As a further aspect of the present invention, the process of optimizing the temperature control command parameters and the charge / discharge command parameters respectively includes: For any local region in the list of abnormal states, obtain the local thermal response difference data corresponding to that local region; The local thermal response difference data is input into a preset mapping function. The mapping function maps the value of the local thermal response difference data to an adjustment coefficient. The adjustment coefficient is positively correlated with the value of the local thermal response difference data. The optimization of the temperature control command parameters specifically involves applying the adjustment coefficient to the reference cooling power value of a local area to generate a new cooling system power setting value. The optimization of the charge and discharge command parameters specifically involves applying adjustment coefficients to the upper limit of the reference charging current and the upper limit of the reference discharging current in the local area to generate new upper limit values ​​for the charging current and the discharging current.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, a comprehensive health index that can simultaneously reflect the dynamic performance and long-term aging trend of a battery is constructed by calculating the rate of change of state of charge and the rate of decay of internal resistance. By combining the difference between the temperature change rate of each local area and the global average rate, thermal imbalance and abnormal areas inside the battery pack are accurately identified. Furthermore, a probabilistic model is used to comprehensively evaluate the overall health and local thermal state, outputting more accurate state judgment results. Finally, based on the degree of difference in local thermal response, differentiated temperature control and charge / discharge adjustment commands are generated, realizing precise intervention in problem areas and normal management of healthy areas, significantly improving the depth of battery state monitoring and the precision of management. Attached Figure Description

[0016] Figure 1 This is a system flowchart of the present invention; Figure 2 This is a flowchart of the signal acquisition module of the present invention; Figure 3 This is a flowchart of the voltage evaluation module of the present invention; Figure 4 This is a flowchart of the temperature analysis module of the present invention; Figure 5 This is a flowchart of the temperature evaluation module of the present invention; Figure 6 This is a flowchart of the state adjustment module of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0018] Please see Figure 1 The IoT-based battery status monitoring and management system includes: The signal acquisition module collects temperature and voltage signals from inside the robot's battery pack, converts the temperature signal into a temperature value, calculates the rate of change of state of charge based on the change of voltage signal over time, and forms multi-source synchronous data. The voltage assessment module reads the battery internal resistance increment over time and the initial internal resistance, calculates the battery health degradation rate, compares the state of charge change rate in the multi-source synchronous data with the battery health degradation rate, and obtains the battery health status assessment index. The temperature analysis module divides the battery pack into local areas based on the location distribution of temperature sensors inside the battery pack. It extracts the temperature value of each local area from multi-source synchronous data, calculates the rate of temperature change over time, compares it with the average rate of temperature change of the entire battery pack, and records the local thermal response difference data. The temperature assessment module comprehensively assesses the overall health and thermal status of the battery pack based on battery health status assessment indicators and local thermal response difference data, and generates health assessment results. The status adjustment module optimizes the temperature control and charging / discharging commands for each local area of ​​the battery pack based on the health assessment results, and generates battery management results.

[0019] The multi-source synchronous data specifically includes temperature signals, rate of change of state of charge, and voltage signals. Battery health status assessment indicators include health degradation rate, rate of change of state of charge, and internal resistance increment. Local thermal response difference data includes local area temperature change rate, global average temperature change rate, and temperature change delay. Health assessment results include thermal response difference assessment and overall health status assessment. Battery management results include temperature control commands, charge / discharge commands, and health management adjustment commands.

[0020] Please see Figure 2 The signal acquisition module includes: The network data acquisition submodule collects temperature and voltage signals inside the robot's battery pack through the Internet of Things sensor network, calls an analog-to-digital converter to convert the temperature signals into temperature values, and integrates the temperature values ​​and voltage signals based on a synchronization timestamp to establish a network sensor signal set. First, data is collected using multiple IoT sensors deployed inside the robot's battery pack. Specifically, several thermistor-type temperature and voltage sensors are evenly distributed or positioned according to key hot spots along the predetermined locations of the cell array within the battery pack housing. For example, in a battery pack consisting of 100 cells, a combined temperature and voltage sensor node is placed every 10 cells along its length. At a specific time point, such as the 10th second after startup, the resistance value of the temperature sensor located near the positive terminal of the battery pack ("location one") changes due to the current ambient temperature. This resistance signal is transmitted to an analog-to-digital converter (ADC). The ADC, based on its internal preset reference voltage and voltage divider circuit, converts the received analog resistance signal into a digital quantity, such as "10110101". This digital quantity is then interpreted as the actual temperature value, such as 25.5 degrees Celsius, according to the sensor's factory-calibrated "digital quantity-temperature" lookup table. At exactly the same time point, the voltage sensor paired with the "Location 1" temperature sensor directly measures the voltage across the battery cell at its location and transmits the measured voltage signal, for example, 3.75 volts, to an analog-to-digital converter for digital processing. The IoT gateway module adds a high-precision synchronization timestamp to these two signal sources, for example, "November 3, 2025, 10:00:00:00." This process is executed synchronously on all sensor nodes within the battery pack, thus precisely aligning the collected temperature and voltage signals at all locations in time. Finally, all "temperature-voltage" data pairs with the same timestamp are combined to form a complete record. Multiple records with consecutive times constitute a network sensor signal set, such as a data stream containing consecutive data points like "timestamp 1, temperature 1, voltage 1," "timestamp 2, temperature 2, voltage 2," etc.

[0021] The battery state rate calculation submodule calls the voltage signal from the network sensor signal set, calculates the voltage difference between two adjacent time points, calculates the voltage change over time based on the voltage difference and the corresponding time interval, and calculates the rate of change of state of charge based on the change and the known capacity of the battery. First, extract the voltage signals from two consecutive time points in the dataset. For example, retrieve the voltage signal at timestamp "10:00:00:00:00", with a value of 3.75 volts, and the voltage signal at the next timestamp "10:00:01:000", with a value of 3.74 volts. Then, calculate the voltage difference between these two adjacent time points: 3.75 volts minus 3.74 volts, resulting in a difference of 0.01 volts. Simultaneously, calculate the precise time interval between the two signals, i.e., 1 second between "10:00:01:000" and "10:00:00:00". Next, calculate the voltage change over time based on the voltage difference and the corresponding time interval: 0.01 volts divided by 1 second, resulting in a voltage change of 0.01 volts per second. Finally, calculate the rate of change of state of charge (SOC) based on the obtained voltage change and the known battery capacity. This step is based on a pre-calibrated mapping relationship between the voltage drop rate and the state of charge (SOC) depletion rate of a specific battery model under specific operating conditions. This mapping relationship is established by applying a standard discharge current to the same battery model in different SOC ranges under laboratory conditions and recording the voltage change rate and the actual SOC depletion rate. For example, calibration results show that when the voltage change is 0.01 volts per second, for a battery with a known capacity of 50 amp-hours, its SOC depletion is 0.01%. Therefore, the final calculated SOC change rate is -0.01% per second.

[0022] The multi-dimensional data generation submodule targets the temperature value in the network sensor signal set, calls the rate of change of state of charge, aligns and associates the two according to the synchronization timestamp, and synchronizes the aligned data through the Internet of Things network to generate multi-source synchronized data. First, the temperature value with the same synchronization timestamp as the voltage signal is extracted from the network sensor signal set. For example, for the timestamp "10:00:01:000", the extracted temperature value is 25.8 degrees Celsius. Simultaneously, the state-of-charge (POC) rate of change, calculated from the battery state-of-charge rate (SOC) algorithm for the same timestamp, is retrieved; its value is -0.01% per second. Based on the shared synchronization timestamp "10:00:01:000", the data from these two different sources are aligned and correlated to form a multi-dimensional data record containing temperature, POC rate of change, and timestamp. Specifically, this is represented by a data structure, such as {timestamp: "10:00:01:000", temperature: 25.8 degrees Celsius, POC rate of change: -0.01% / second}. This process is repeated for each time point in the network sensor signal set, generating a series of multi-dimensional data records arranged chronologically. Finally, these integrated and aligned multidimensional data records are synchronously uploaded from the battery pack's processing unit to a cloud server or upper-level management platform via an IoT network, thereby generating multi-source synchronized data for subsequent analysis.

[0023] Please see Figure 3 The voltage evaluation module includes: The internal resistance data reading submodule reads the battery internal resistance increment over time and the initial internal resistance to form a set of battery internal resistance parameters. First, the battery's internal resistance increment over time and initial internal resistance are read from the non-volatile memory of the battery management unit. The initial internal resistance is a baseline value measured and recorded using standard testing methods (e.g., 1kHz AC internal resistance testing) when the battery leaves the factory or is first put into use; for example, the measured and recorded initial internal resistance is 20.0 milliohms. The battery's internal resistance increment over time refers to the difference between the current internal resistance value measured using the same testing method in the most recent internal resistance testing cycle and the initial internal resistance value. For example, if the most recently measured internal resistance is 22.5 milliohms, then the internal resistance increment is 22.5 milliohms minus 20.0 milliohms, which is 2.5 milliohms. Next, these two data points are standardized. The standardization is based on a reference dataset obtained through a complete life cycle aging test on the same batch of batteries. In the experiment, the maximum internal resistance increment was recorded as the battery progressed from a brand-new state to the end of its life (e.g., capacity decay to 80% of the nominal value). Assuming that experimental measurements show that the typical internal resistance increment at the end of the lifespan of this battery model is 20.0 milliohms, a set of battery internal resistance parameters is formed.

[0024] The healthy degradation rate calculation submodule calculates the ratio of the battery internal resistance increment over time to the initial internal resistance in the battery internal resistance parameter set, which is used as the battery healthy degradation rate. First, the incremental value of the battery's internal resistance over time and the original value of the initial internal resistance are extracted from the battery internal resistance parameter set. The extracted incremental internal resistance is 2.5 milliohms, and the initial internal resistance is 20.0 milliohms. Then, the ratio of these two values ​​is directly calculated. Specifically, the incremental internal resistance of 2.5 milliohms is used as the numerator, and the initial internal resistance of 20.0 milliohms is used as the denominator, followed by a division operation. The result is 2.5 divided by 20.0, which gives 0.125. This calculated ratio is directly defined as the battery's health degradation rate. This value reflects the degree of deterioration of the battery's internal resistance compared to its initial state. A higher ratio indicates a more severe degradation in the battery's health. This calculation result will be used for comprehensive evaluation.

[0025] The comprehensive health status assessment submodule calculates the ratio of the rate of change of state of charge in multi-source synchronous data to the rate of battery health degradation, which serves as an indicator for battery health status assessment. The rate of change of state of charge (SOC) is extracted from multi-source synchronous data, with a value of -0.01% per second, and the battery health degradation rate, calculated from the health degradation rate, is received, with a value of 0.125. The final battery health status assessment metric is calculated using the following formula: ,in, This represents the final calculated battery health status assessment index, which is a dimensionless value. The index value is inversely proportional to the denominator, which incorporates both the battery's dynamic performance and aging level. Therefore, a larger denominator indicates a worse overall battery condition, resulting in a smaller final index value. It is a dimensionless quantity characterizing the dynamic properties of a battery, calculated from the absolute value of the rate of change of state of charge and a reference time period. In this embodiment, the rate of change of state of charge is -0.01% per second, and its absolute value is 0.0001 / second. The reference time period is set based on the average time of a single complete operation, statistically derived from a large amount of historical task data in a battery application scenario (warehouse robot), and is set to 900 seconds. Therefore, The calculation is as follows: Here It is the rate of change of the state of charge. This is a reference time period. The battery health degradation rate is calculated by the health degradation rate calculation step. It is a dimensionless quantity, and its value is 0.125 in this example. and These are weighting factors corresponding to the dynamic characteristic parameters and the health degradation rate, respectively, both dimensionless. The basis for setting these two weights is to establish a predictive model by performing regression analysis on historical battery lifecycle data, using the battery's actual remaining lifespan as the target variable and the dynamic characteristic parameters and health degradation rate as independent variables. These two weighting factors are the coefficients of the independent variables in the model, obtained after normalization (the sum of the two is 1), and their values ​​reflect the contribution of each factor to the battery's health status. Regression analysis results show that the health degradation rate, which better reflects the degree of long-term irreversible aging, is more important; therefore, it is set as the weighting factor. It is 0.4. The value is 0.6. First, calculate the first term in the denominator: Next, calculate the second term in the denominator: Then, add the two terms together to get the value of the denominator: Finally, calculate the evaluation metrics: This indicator is a comprehensive dimensionless value that will be used in subsequent probability calculations.

[0026] Please see Figure 4 The temperature analysis module includes: The region division and data extraction submodule divides the battery pack into local regions based on the location distribution of the temperature sensors inside the battery pack, extracts the temperature value of each local region from multi-source synchronous data, and classifies and integrates the extracted temperature values ​​according to the divided local regions to establish a regional temperature dataset. First, based on the physical location distribution of the temperature sensors inside the battery pack, the battery pack is logically divided into local regions. For example, a long, narrow battery pack can be divided into three regions along its length: "Local Region 1" covers the first third of the cells near the positive terminal, "Local Region 2" covers the middle third, and "Local Region 3" covers the last third. Each region contains several temperature sensors. After division, the temperature value at each time point is extracted from the multi-source synchronous data stream from upstream. Specifically, each record in the data stream is parsed; for example, at timestamp "10:00:01:000", the data record contains the temperature readings of all sensors. These temperature values ​​are then categorized according to a preset correspondence between sensor IDs and regions. For example, the readings of sensors T1, T2, and T3 (26.1°C, 26.3°C, and 26.2°C) are categorized into "Local Region 1"; the readings of sensors T4, T5, and T6 (27.0°C, 27.2°C, and 27.1°C) are categorized into "Local Region 2"; and the readings of sensors T7, T8, and T9 (25.9°C, 26.0°C, and 25.8°C) are categorized into "Local Region 3". Next, the temperature values ​​within each region are aggregated, and their average value is calculated to represent the overall temperature of that region. For example, the average temperature for "Local Region 1" is calculated as (26.1 + 26.3 + 26.2) / 3 = 26.2°C. After performing the same operation on all regions, a time-series regional temperature dataset is established, where each time point corresponds to a set of representative temperature values ​​for each local region.

[0027] The temperature rate calculation submodule calculates the temperature change rate of each local area based on the temperature value of each local area in the regional temperature dataset over time, calculates the global average temperature change rate of the battery pack based on the temperature values ​​of all local areas, and integrates the local rate and the global average rate to generate a temperature change rate set. First, the rate of temperature change is calculated for each local region. This requires obtaining the region's temperature values ​​at two consecutive time points. For example, for "Local Region 1," the average temperature at "10:00:01:000" is 26.2 degrees Celsius, and at the next time point "10:00:02:000," the average temperature changes to 26.3 degrees Celsius. The rate of temperature change is calculated by subtracting the temperature of the previous time point from the temperature of the later time point, and then dividing by the time interval. Specifically, (26.3 - 26.2) degrees Celsius / 1 second, resulting in a temperature change rate of 0.1 degrees Celsius per second for "Local Region 1." Similarly, the same calculation is performed for "Local Region 2" (e.g., changing from 27.1 degrees Celsius to 27.4 degrees Celsius at a rate of 0.3 degrees Celsius per second) and "Local Region 3" (e.g., changing from 25.9 degrees Celsius to 26.0 degrees Celsius at a rate of 0.1 degrees Celsius per second). Simultaneously, the average rate of temperature change for the entire battery pack is also calculated. To calculate the global average temperature, the temperature values ​​from all sensors at a given moment are first summed and averaged. For example, at "10:00:01:000", the total average temperature from the nine sensors is 26.5 degrees Celsius; at "10:00:02:000", the total average temperature changes to 26.7 degrees Celsius. The global average temperature change rate is calculated as (26.7 - 26.5) degrees Celsius / second, yielding 0.2 degrees Celsius per second. Finally, the temperature change rates of all local areas ([0.1, 0.3, 0.1] degrees Celsius / second) are integrated with the calculated global average temperature change rate (0.2 degrees Celsius / second) to generate a structured set of temperature change rates.

[0028] The thermal response difference comparison submodule calls the temperature change rate set, compares the temperature change rate of each local area with the average temperature change rate of the battery pack, and records the difference value generated by the comparison to obtain local thermal response difference data. The temperature change rate of each local region is compared with the global average temperature change rate of the battery pack. Specifically, the local rate is subtracted from the global average rate. For "Local Region 1," the comparison process is: 0.1 degrees Celsius / second (local rate) minus 0.2 degrees Celsius / second (global average rate), resulting in a difference of -0.1 degrees Celsius / second. For "Local Region 2," the comparison process is: 0.3 degrees Celsius / second (local rate) minus 0.2 degrees Celsius / second (global average rate), resulting in a difference of +0.1 degrees Celsius / second. For "Local Region 3," the comparison process is: 0.1 degrees Celsius / second (local rate) minus 0.2 degrees Celsius / second (global average rate), resulting in a difference of -0.1 degrees Celsius / second. This subtraction operation is performed on all defined local regions. The difference value after each comparison is recorded. Finally, the difference values ​​of all local regions ([-0.1, +0.1, -0.1] degrees Celsius / second) are collected together to form a data array. This set is the local thermal response difference data, which quantifies the degree of deviation of different regions inside the battery pack from the overall average thermal behavior. Positive values ​​indicate that the region heats up faster than the average level, and negative values ​​indicate that the region heats up slower than the average level.

[0029] Please see Figure 5 The temperature assessment module includes: The data integration input submodule associates the battery health status assessment indicators that reflect the overall health status of the battery pack with the data reflecting the differences in local thermal response of each local area of ​​the battery pack to establish a comprehensive assessment input set; First, the battery health status assessment metric, with a value of 9.009, is retrieved from the voltage assessment step. Simultaneously, local thermal response difference data, an array containing three values ​​[-0.1, +0.1, -0.1] degrees Celsius per second, is also retrieved from the temperature analysis step. Next, these two different data dimensions are correlated to create a unified data structure that encapsulates the battery health status assessment metric and the local thermal response difference data as a whole. For example, a record is generated: {Overall metric: 9.009, Local difference: [-0.1, +0.1, -0.1]}. This combined data structure constitutes the comprehensive assessment input set.

[0030] The state probability calculation submodule takes the comprehensive evaluation input set and inputs it into the Bayesian discriminant model. Based on the prior probability and conditional probability preset in the Bayesian discriminant model, it calculates the posterior probability of the battery pack belonging to multiple health states and obtains the health state membership probability. The process of setting prior probabilities and conditional probabilities includes: Acquire historical operating data of multiple battery groups throughout their complete life cycle. The historical operating data includes battery health status assessment indicators at all time points, local thermal response difference data, and corresponding real health status labels. The frequency of each true health status label in historical operating data is counted, and the prior probability of the battery pack belonging to multiple health statuses is calculated based on the ratio of the frequency to the total data. Based on the real health status label, the historical operating data is classified. In each category of data, the numerical range of the battery health status assessment index and the local thermal response difference data is discretized into segments. The frequency of data occurrence in each segment is counted, and a probability lookup table representing the distribution of input data under a specified health status is established. The probability lookup table is used as the conditional probability. By embedding an improved Bayesian discriminant model, structured as a classifier, it applies Bayes' theorem to calculate the probability of a sample belonging to each predefined category and introduces an exponential penalty term to quantify the negative impact of thermal inhomogeneity. For an input sample, the model outputs its posterior probability of belonging to each health state. The posterior probability of belonging to a state is calculated as follows: ,in, Representing the posterior probability, it is a dimensionless numerical value that indicates the probability obtained from the input data. Under these conditions, the battery pack is in a healthy state. The probability of. It is a comprehensive evaluation input set, containing an overall index of 9.009 and local variance data [-0.1, +0.1, -0.1]. Representing the A health state. In this embodiment, 3 represents, respectively (healthy), (Sub-health) and (Fault). Representing the prior probability, this is a dimensionless value. This value is determined based on statistical analysis of a large amount of historical battery operating data, calculating the frequency of various health states. Based on the statistical results, a setting is established. , , . Representing conditional probability, it is a dimensionless value that indicates the state a battery is in given a given condition. Under the premise of observing the current input data The probability. This value is obtained from a pre-defined probability density lookup library. This lookup library is created by categorizing historical data by health status and then applying the input data to each category. Kernel density estimation is performed to establish a multidimensional probability density function for each health state. When new input data arrives, its corresponding probability density value is obtained by querying these functions. Based on the current input data, the database is consulted to obtain: , , . This is an index of thermal non-uniformity, measured in degrees Celsius per second. It is obtained by calculating the standard deviation of local thermal response differences [-0.1, +0.1, -0.1]. Its mean is approximately (-0.1 + 0.1 - 0.1) / 3 ≈ -0.033. The standard deviation is calculated as follows: Celsius per second. This is the thermal penalty coefficient, measured in seconds per degree Celsius, used to quantify the negative impact of thermal non-uniformity on battery health. The coefficient is determined by establishing a functional relationship (e.g., an exponential model) between thermal non-uniformity and battery capacity decay rate through accelerated aging experiments with different temperature gradients applied to multiple battery groups. The thermal penalty coefficient is the key parameter in this model characterizing the sensitivity of the decay rate to thermal non-uniformity; the greater the impact of thermal non-uniformity on battery life, the larger the coefficient value. It was determined through fitting experimental data. Set to 0.5 seconds per degree Celsius. It is the base of the natural logarithm. It is the index of the summation symbol, traversing all health states (from 1 to ...). ), This represents the total number of categories of the preset battery health state. First, calculate the common penalty factor: Next, the numerator, i.e., the unnormalized probability value for each state, is calculated: for (healthy): .for (Sub-health): .for (Fault): Then, calculate the denominator, which is the sum of all unnormalized probabilities: Finally, calculate the posterior probability of each state: , , This set of probability values ​​is the probability of belonging to a health status.

[0031] The comprehensive state discrimination submodule selects the state with the highest probability from the posterior probabilities of various health states based on the health state membership probability, and determines the overall health state and thermal state of the battery pack to generate a health assessment result. Based on the health status membership probabilities output by the state probability calculation, the final discrimination is performed. The received set of posterior probability values ​​shows that the battery pack belongs to a "healthy" state (0.904), a "sub-healthy" state (0.090), and a "faulty" state (0.006). The discrimination process involves selecting the state corresponding to the highest probability value from this set of posterior probability values. In this example, 0.904 is the highest probability value, corresponding to the "healthy" state. Therefore, the overall health status of the battery pack is determined to be "healthy," and since the local thermal response difference data did not trigger a specific threshold, its thermal status is determined to be "normal." Finally, this two-part judgment conclusion, "Overall health status: healthy; thermal status: normal," is combined to generate the final health assessment result for this evaluation.

[0032] Please see Figure 6 The status adjustment module includes: The problem area identification submodule determines whether the overall health and thermal status of the battery pack are in an abnormal range based on the health assessment results. If the determination is yes, it calls the local thermal response difference data, analyzes and identifies the local areas where the difference data exceeds the normal range, and establishes a list of abnormal state areas. First, the received health assessment result is evaluated. This result indicates the overall health and thermal state of the battery pack, for example, "Overall health: sub-healthy; Thermal state: localized overheating." An internally defined standard for abnormal states is a preset list containing all state combinations other than "healthy" and "normal." The received assessment result is matched against this list of abnormal states. In this example, both "sub-healthy" and "localized overheating" match the abnormal states in the list, therefore the determination is that the battery pack is in an abnormal range. After the determination is established, the local thermal response difference data that led to this assessment result is retrieved, for example, data of [-0.05, +0.28, -0.02] degrees Celsius per second. Subsequently, the local areas exceeding the normal range in the difference data are analyzed and identified. Here, "normal range" is a preset threshold range. The threshold range is set based on statistical analysis of local thermal response differences in a large number of "healthy" batteries operating under various conditions. The mean and standard deviation of this dataset are calculated, and the mean plus or minus three times the standard deviation (i.e., the 99.7% confidence interval) is taken as the normal fluctuation range, for example, [-0.15, +0.15] degrees Celsius per second. The difference data for each local region is compared to this range. -0.05 is within the range, +0.28 exceeds the upper limit, and -0.02 is within the range. Therefore, "Local Region Two" with a difference value of +0.28 is identified as an anomaly. Finally, the identifiers of all identified anomaly regions (e.g., "Local Region Two") are compiled to create a list of anomalous regions.

[0033] The regional instruction optimization submodule optimizes the temperature control instruction parameters and charge / discharge instruction parameters for each local region in the list of abnormal states, and combines the optimized two types of instructions to generate a health adjustment instruction set. The process of optimizing the temperature control command parameters and the charge / discharge command parameters includes: For any local region in the list of abnormal states, obtain the local thermal response difference data corresponding to that local region; The local thermal response difference data is input into a preset mapping function. The mapping function maps the value of the local thermal response difference data to an adjustment coefficient. The adjustment coefficient is positively correlated with the value of the local thermal response difference data. The optimization of the temperature control command parameters specifically involves applying the adjustment coefficient to the reference cooling power value of a local area to generate a new cooling system power setting value. The optimization of the charge and discharge command parameters specifically involves applying the adjustment coefficient to the upper limit of the reference charging current and the upper limit of the reference discharging current in the local area to generate new upper limit values ​​of the charging current and the discharging current. For each local region in the list of abnormal states, parameter optimization of temperature control and charge / discharge commands is performed. Taking "Local Region Two" in the list as an example, firstly, the local thermal response difference data that caused it to be identified as abnormal, i.e., +0.28 degrees Celsius / second, is obtained. Next, this difference data value is input into a preset mapping function. This mapping function maps the input difference value to an adjustment coefficient, and its establishment process is based on experiments: on a test platform, different degrees of heat accumulation are artificially created in specific areas of the battery pack, the corresponding thermal response difference values ​​are recorded, and simultaneously, the additional cooling power required to restore thermal balance or the required reduction in current ratio is determined. Finally, a mathematical relationship between the difference value and the adjustment coefficient is established through polynomial regression fitting. This relationship is set to be positively correlated, that is, the larger the difference, the stronger the adjustment. For example, the function specifies that when the difference value is +0.28 degrees Celsius / second, the corresponding temperature control adjustment coefficient is 1.4, and the charge / discharge adjustment coefficient is 0.8. The optimization of the temperature control command parameters involves obtaining the baseline cooling power value for "Local Region Two," calculated based on the battery thermal model under standard operating conditions, for example, 20 watts. An adjustment factor of 1.4 is applied to this baseline value to generate a new cooling power setting value; the calculation process is 20 watts multiplied by 1.4, resulting in 28 watts. The optimization of the charge / discharge command parameters involves obtaining the baseline charging current limit and baseline discharging current limit for the cell group corresponding to "Local Region Two." These two baseline limits are specified by the cell manufacturer's safety datasheet, for example, 5 amps and 15 amps respectively. An adjustment factor of 0.8 is applied to these two baseline limits to generate new current limit values. The new charging current limit is 5 amps multiplied by 0.8, resulting in 4 amps; the new discharging current limit is 15 amps multiplied by 0.8, resulting in 12 amps. Finally, the optimized temperature control command parameters (28 watts) and charge / discharge command parameters (4 amps, 12 amps) are combined to generate a health adjustment command set for "Local Region Two."

[0034] The management strategy generation submodule compiles the temperature control instructions and charge / discharge instructions for each local area of ​​the battery pack into a unified management strategy based on the health adjustment instruction set, and generates battery management results. After generating a set of health adjustment instructions based on the list of abnormal state areas, a unified management strategy is compiled. First, all instructions are organized: for "Local Area Two," the temperature control instruction is set to 28 watts of cooling power, and the charge / discharge instructions are set to a maximum charging current of 4 amps and a maximum discharging current of 12 amps. For other areas not appearing in the abnormal state area list, such as "Local Area One" and "Local Area Three," their control instructions maintain their respective baseline values, for example, a cooling power of 5 watts, a maximum charging current of 5 amps, and a maximum discharging current of 15 amps. Next, these temperature control instructions and charge / discharge instructions, which are tailored to each local area of ​​the battery pack and contain differentiated settings, are compiled according to the communication protocol and data format that the battery management unit's main controller can parse. The compilation process includes packaging information such as area identifiers, instruction types, and instruction values ​​into a structured data frame. This unified management strategy ensures coordinated control of the entire battery pack, enabling targeted and powerful intervention in problematic areas while ensuring that the operating efficiency of other normal areas is not unnecessarily affected. Finally, this assembled data packet, containing the latest instructions for all regions, is sent to the underlying execution unit as a battery management result to directly control the corresponding cooling and power units.

[0035] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A battery status monitoring and management system based on the Internet of Things, characterized in that, The system includes: The signal acquisition module collects temperature and voltage signals from inside the robot's battery pack, converts the temperature signal into a temperature value, calculates the rate of change of state of charge based on the change of voltage signal over time, and forms multi-source synchronous data. The voltage assessment module reads the battery internal resistance increment over time and the initial internal resistance, calculates the battery health degradation rate, compares the state of charge change rate in the multi-source synchronous data with the battery health degradation rate, and obtains the battery health status assessment index. The temperature analysis module divides the battery pack into local areas based on the location distribution of the temperature sensors inside the battery pack. It extracts the temperature value of each local area from the multi-source synchronous data, calculates the rate of temperature change over time, compares it with the average rate of temperature change of the entire battery pack, and records the local thermal response difference data. The temperature assessment module comprehensively assesses the overall health and thermal state of the battery pack based on the battery health status assessment indicators and the local thermal response difference data, and generates a health assessment result. The status adjustment module optimizes the temperature control and charging / discharging commands for each local area of ​​the battery pack based on the health assessment results, and generates battery management results.

2. The battery status monitoring and management system based on the Internet of Things according to claim 1, characterized in that, The multi-source synchronized data specifically includes temperature signals, rate of change of state of charge, and voltage signals. The battery health status assessment indicators include health degradation rate, rate of change of state of charge, and internal resistance increment. The local thermal response difference data includes local area temperature change rate, global average temperature change rate, and temperature change delay. The health assessment results include thermal response difference assessment and overall health status assessment. The battery management results include temperature control commands, charge / discharge commands, and health management adjustment commands.

3. The battery status monitoring and management system based on the Internet of Things according to claim 1, characterized in that, The signal acquisition module includes: The network data acquisition submodule collects temperature and voltage signals inside the robot's battery pack through the Internet of Things sensor network, calls an analog-to-digital converter to convert the temperature signals into temperature values, and integrates the temperature values ​​and voltage signals based on a synchronization timestamp to establish a network sensor signal set. The battery state rate calculation submodule calls the voltage signal from the network sensor signal set, calculates the voltage difference between two adjacent time points, calculates the voltage change over time based on the voltage difference and the corresponding time interval, and calculates the state of charge change rate based on the change and the known capacity of the battery. The multi-dimensional data generation submodule takes the temperature value in the network sensor signal set and calls the rate of change of state of charge. It aligns and associates the two according to the synchronization timestamp, and synchronizes the aligned data through the Internet of Things network to generate multi-source synchronized data.

4. The battery status monitoring and management system based on the Internet of Things according to claim 3, characterized in that, The voltage evaluation module includes: The internal resistance data reading submodule reads the battery internal resistance increment over time and the initial internal resistance to form a set of battery internal resistance parameters. The health degradation rate calculation submodule calculates the ratio of the battery internal resistance increment over time to the initial internal resistance in the battery internal resistance parameter set, which is used as the battery health degradation rate. The comprehensive health status assessment submodule calculates the ratio of the rate of change of state of charge in the multi-source synchronous data to the rate of battery health degradation, which serves as the battery health status assessment index.

5. The battery status monitoring and management system based on the Internet of Things according to claim 4, characterized in that, The temperature analysis module includes: The region division and data extraction submodule divides the battery pack into local regions based on the location distribution of the temperature sensors inside the battery pack, extracts the temperature value of each local region from the multi-source synchronous data, and classifies and integrates the extracted temperature values ​​according to the divided local regions to establish a regional temperature dataset. The temperature rate calculation submodule calculates the temperature change rate of each local area based on the temperature value of each local area in the regional temperature dataset over time, calculates the global average temperature change rate of the battery pack based on the temperature values ​​of all local areas, and integrates the local rate and the global average rate to generate a temperature change rate set. The thermal response difference comparison submodule calls the temperature change rate set, compares the temperature change rate of each local area with the average temperature change rate of the battery pack, and records the difference value generated by the comparison to obtain local thermal response difference data.

6. The battery status monitoring and management system based on the Internet of Things according to claim 5, characterized in that, The temperature assessment module includes: The data integration input submodule associates the battery health status assessment index, which reflects the overall health status of the battery pack, with the local thermal response difference data, which reflects the local area of ​​the battery pack, to establish a comprehensive assessment input set. The state probability calculation submodule inputs the comprehensive evaluation input set into the Bayesian discriminant model, calculates the posterior probability of the battery pack belonging to multiple health states based on the preset prior probabilities and conditional probabilities within the Bayesian discriminant model, and obtains the health state membership probability. The comprehensive state discrimination submodule selects the state with the highest probability from the posterior probabilities of various health states based on the health state membership probability, and determines the overall health state and thermal state of the battery pack to generate a health assessment result.

7. The battery status monitoring and management system based on the Internet of Things according to claim 6, characterized in that, The process of setting the prior probability and the conditional probability includes: Acquire historical operating data of multiple battery groups throughout their complete life cycle. The historical operating data includes battery health status assessment indicators at all time points, local thermal response difference data, and corresponding real health status labels. The frequency of each true health status label in historical operating data is counted, and the prior probability of the battery pack belonging to multiple health statuses is calculated based on the ratio of the frequency to the total data. Historical operating data is categorized based on the actual health status label. Within each category, the numerical range of battery health status assessment indicators and local thermal response difference data is discretized into segments. The frequency of data occurrence within each segment is counted, and a probability lookup table representing the distribution of input data under a specified health status is established. This probability lookup table is then used as a conditional probability.

8. The battery status monitoring and management system based on the Internet of Things according to claim 6, characterized in that, To calculate the posterior probability of a battery pack belonging to multiple health states, the formula is: ； in, This indicates that the input data has been obtained. Under these conditions, the battery pack is in a healthy state. The probability, It is a comprehensive evaluation of the input set. Representing the A kind of health state, Represents prior probability. Representing conditional probability, it is determined that the battery is in a certain state. Under the premise of observing the current input data The probability, It is an index of thermal non-uniformity. It is the thermal penalty coefficient. It is the base of the natural logarithm. It is the index of the summation symbol, traversing from 1 to... health status, This represents the total number of preset battery health status categories.

9. The battery status monitoring and management system based on the Internet of Things according to claim 6, characterized in that, The status adjustment module includes: The problem area identification submodule determines whether the overall health status and thermal status of the battery pack are in an abnormal range based on the health assessment results. If the determination is yes, it calls the local thermal response difference data, analyzes and identifies local areas where the difference data exceeds the normal range, and establishes a list of abnormal state areas. The regional instruction optimization submodule optimizes the temperature control instruction parameters and charge / discharge instruction parameters for each local region in the list of abnormal states, and combines the optimized two types of instructions to generate a health adjustment instruction set. The management strategy generation submodule compiles the temperature control instructions and charge / discharge instructions for each local area of ​​the battery pack into a unified management strategy based on the health adjustment instruction set, and generates battery management results.

10. The battery status monitoring and management system based on the Internet of Things according to claim 9, characterized in that, The process of optimizing the temperature control command parameters and the charge / discharge command parameters respectively includes: For any local region in the list of abnormal states, obtain the local thermal response difference data corresponding to that local region; The local thermal response difference data is input into a preset mapping function. The mapping function maps the value of the local thermal response difference data to an adjustment coefficient. The adjustment coefficient is positively correlated with the value of the local thermal response difference data. The optimization of the temperature control command parameters specifically involves applying the adjustment coefficient to the reference cooling power value of a local area to generate a new cooling system power setting value. The optimization of the charge and discharge command parameters specifically involves applying adjustment coefficients to the upper limit of the reference charging current and the upper limit of the reference discharging current in the local area to generate new upper limit values ​​for the charging current and the discharging current.

Images