Device and method for batch detection of remaining capacity of disposable batteries in high-altitude environment

Abstract

The application provides a disposable battery residual capacity batch detection device and method in high-altitude environment, and belongs to the lithium hyposulfuryl chloride battery detection technical field.The application comprises a detection device main body and an upper computer, and the device main body comprises a power supply, temperature monitoring, control, indication, switching, battery test and parameter acquisition unit;the power supply unit comprises a power supply interface, a power supply switch, a switching power supply and a voltage reduction circuit;the temperature monitoring unit comprises a temperature measurement module and a thermocouple, the control unit comprises a controller MCU;the indication unit comprises a loudspeaker, a switching indicator lamp and a three-color indicator lamp, the switching unit comprises a de-passivation switching circuit, a multi-channel switching circuit and a pulse switching circuit;the battery test unit comprises a battery socket and a resistance load;and the parameter acquisition unit comprises a pulse test and reference voltage acquisition module.The application solves the problem that the disposable battery cannot be accurately detected due to the continuous passivation phenomenon, and can realize the accurate prediction of the disposable battery residual capacity.

Description

Technical Field

[0001] This invention relates to a device and method for batch testing of the remaining capacity of disposable batteries in high-altitude environments, belonging to the field of lithium thionyl chloride battery testing technology. Background Technology

[0002] As the core power source for maintaining billing timing, the accurate detection of the remaining capacity of the clock battery in smart energy meters directly affects the clock's timekeeping accuracy and the reliability of power grid metering. However, high-energy-density lithium primary batteries suffer from a non-linear deviation between the open-circuit voltage (OCV) and the true capacity (typical error >12%) due to the dynamic thickening of the electrode passivation film over service time. Furthermore, mixed load conditions (μA static + mA pulse) cause voltage response hysteresis. Traditional deep discharge methods destroy the battery's maintenance-free characteristics, and electrochemical impedance spectroscopy (EIS) equipment is expensive and difficult to deploy on a large scale. Existing technical solutions cannot simultaneously achieve the goals of rapid detection, high accuracy, and non-destructive evaluation, resulting in a large number of clock failures stemming from capacity decay and persistently high maintenance costs. Summary of the Invention

[0003] The purpose of this invention is to solve the problems existing in the prior art, and to provide a device and method for batch testing of the remaining capacity of disposable batteries in high-altitude environments.

[0004] The objective of this invention is achieved through the following technical solution:

[0005] A batch testing device for the remaining capacity of disposable batteries in high-altitude environments, the device comprising: a testing device body and a host computer; the testing device body comprising: a power supply unit, a temperature monitoring unit, a control unit, an indicator unit, a switching unit, a battery testing unit, and a parameter acquisition unit;

[0006] The signal input / output terminals of the host computer are respectively connected to the signal input / output terminals of the temperature monitoring unit, the control unit, and the parameter acquisition unit; the power supply unit is electrically connected to the temperature monitoring unit, the control unit, the switching unit, and the parameter acquisition unit; the signal output terminal of the control unit is respectively connected to the signal input terminals of the indicator unit and the switching unit; the signal output terminal of the switching unit is respectively connected to the signal input terminals of the control unit and the battery testing unit; and the signal output terminal of the battery testing unit is connected to the signal input terminal of the parameter acquisition unit.

[0007] The host computer is used to store the data collected by the temperature monitoring unit and the parameter acquisition unit, and to send instructions to the temperature monitoring unit, the control unit and the parameter acquisition unit to perform voltage-temperature correlation identification and remaining capacity prediction.

[0008] The power supply unit is used to supply power to the temperature monitoring unit, control unit, switching unit and parameter acquisition unit;

[0009] The temperature monitoring unit is used to convert the collected temperature signals into electrical signals and transmit them to the host computer.

[0010] The control unit is used to control the switching unit to perform passivation film elimination and pulse load testing;

[0011] The indicator unit is used to alarm when a test fault occurs, and transmits signals to the control unit to indicate the working status of the switching unit and the operating status of the device;

[0012] The switching unit is used to connect a resistive load to both ends of the battery to eliminate the passivation film, and then select different batteries to connect different pulse loads to both ends of the battery for pulse load testing.

[0013] The battery test unit is used to enable quick plugging and unplugging for battery testing;

[0014] The parameter acquisition unit is used to acquire the voltage across the battery terminals and the reference voltage during pulse load testing.

[0015] Preferably, the power supply unit includes: a power switch and a switching power supply, wherein the power switch is connected to the switching power supply; and the switching power supply is electrically connected to the switching unit, the control unit, the parameter acquisition unit and the temperature monitoring unit respectively.

[0016] The temperature monitoring unit includes: a temperature measuring module and a thermocouple, with the temperature measuring module connected to the thermocouple; the thermocouple is attached to the switching power supply, the parameter acquisition unit, and the battery testing unit respectively;

[0017] The control unit is a controller MCU, and the controller MCU is connected to the switching unit and the indication unit;

[0018] The indicator unit includes a speaker, a switching indicator light, and a three-color indicator light, which are respectively connected to the control unit.

[0019] The battery testing unit includes: a battery socket and a resistive load, the battery socket is connected to the resistive load, the resistive load is connected to a connector, and the connector is connected to a switching unit and a parameter acquisition unit.

[0020] Preferably, the parameter acquisition unit is a pulse test and reference voltage acquisition module, which is connected to the battery socket and the reference voltage chip, respectively.

[0021] Preferably, the switching unit includes: a depassivation switching circuit, a multi-channel switching circuit, and a pulse switching circuit; the depassivation switching circuit is connected to the battery socket and the resistive load, respectively; the multi-channel switching circuit is connected to the battery socket and the pulse test and reference voltage acquisition module; and the pulse switching circuit is connected to the pulse test and reference voltage acquisition module.

[0022] A detection method for batch testing of remaining capacity of disposable batteries in high-altitude environments includes the following steps:

[0023] Step S101, Reference Voltage Acquisition Stage: Install the battery test unit on the connector of the testing device and place the battery to be tested into the battery socket; after the host computer sends the reference voltage acquisition command, the pulse test and reference voltage acquisition module acquires the reference voltage VOUT, sends it to the host computer and stores it, and loads the dynamic threshold for the host computer later; the reference voltage VOUT is used on the one hand to carry out reference voltage compensation in combination with high-altitude environmental parameters, and on the other hand, it is synchronously pre-stored in the temperature-ΔV mapping table;

[0024] Step S102, Temperature Acquisition Stage: The host computer sends a temperature acquisition command, the temperature measurement module sends the acquired temperature data to the host computer and completes storage for temperature compensation; the host computer transmits the acquired temperature data to the control unit in real time.

[0025] Step S103, Depassivation Stage: After the host computer sends the depassivation command, the depassivation switching circuit connects the battery under test to the two ends of the corresponding resistive load and forms a separate closed loop; the passivation film structure is efficiently broken through the depassivation current.

[0026] Step S104, Settling Stage: The host computer sends a settling command, and the depassivation switching circuit removes the resistive load at both ends of the battery under test and sets the battery for a predetermined time.

[0027] Step S105, Pulse Load Test and Temperature Acquisition Stage: After the host computer sends the pulse load test command, the multi-channel switching circuit connects the battery under test to the pulse test and reference voltage acquisition module and starts the voltage acquisition at both ends of the battery; after the acquisition starts, the pulse switching circuit connects the selected pulse load to both ends of the battery under test and then removes it. After the pulse load is removed, the multi-channel switching circuit removes the battery under test from the pulse test and reference voltage acquisition module, stops the voltage acquisition, sends the acquired pulse load test voltage data to the host computer and completes the storage, and the host computer marks the abnormality.

[0028] Step S106, Feature Value Extraction Stage: After receiving the voltage data and anomaly markers collected in step S105, the host computer first performs data repair, then extracts the feature values ​​from the repaired data, and after reference voltage compensation, uses them as input to the LSTM-BP network hybrid model.

[0029] Step S107, Remaining Capacity Prediction Stage: The host computer inputs the feature values ​​extracted in step S106 into the LSTM-BP network hybrid model, outputs the final predicted value and calculates the loss, and performs the final lifetime prediction. After the prediction is completed, the host computer stores the test data in the log database.

[0030] Step S108, Cyclic Testing Phase: Repeat steps S105 to S107 until the test battery with the serial number configured by the user in the host computer program is completed. The host computer will collect log data in real time and perform interference source tracing and model optimization.

[0031] Preferably, the formula for loading the dynamic threshold in step S101 is as follows:

[0032] (1)

[0033] (2)

[0034] in: This is the upper limit of the dynamic threshold. This is the lower limit of the dynamic threshold. This is the battery's nominal voltage. This is the reference voltage deviation coefficient.

[0035] Preferably, the specific steps for the host computer to perform anomaly marking in step S105 are as follows:

[0036] Step S1051: The host computer compares the sampled reference voltage VOUT. and If the value exceeds this range, the marking amplitude is abnormal.

[0037] Step S1052: Calculate the slopes k1 and k2 and the rate of change of the voltage at three consecutive points. ,like Furthermore, the temperature fluctuation is greater than ±2℃, and the marking slope is abnormal;

[0038] Step S1053: The real-time comparison pulse test and reference voltage acquisition module acquires the reference voltage VOUT and the voltage VREF of the reference voltage chip. If the absolute value of ΔV is >5mV at room temperature or >8mV at temperature ≤-20℃, it is marked as an abnormal drift.

[0039] Preferably, the data repair performed in step S106 includes the following situations:

[0040] (1) Minor anomalies: For single-point amplitude anomalies, i.e., ΔV of 3-5mV or slope anomalies with temperature fluctuations of 1-2℃, the baseline is first corrected using the following formula.

[0041] (3)

[0042] in, This is the corrected voltage value after compensation with the reference voltage. This represents the original voltage value acquired at the i-th sampling time.

[0043] Then correct to the 25℃ equivalent value using the following formula;

[0044] (4)

[0045] in, For actual measured temperature, This is the battery voltage value equivalent to a reference temperature of 25°C after temperature compensation.

[0046] (2) Serious abnormality: For two or more consecutive abnormalities or ΔV>8mV, control the multi-channel switching unit to switch to the backup channel, increase the sampling frequency to 5kHz and retest, and take the average value after comparing the two ΔV deviations <3mV;

[0047] (3) High-frequency drift: If the ΔV drift is greater than 3 times / hour when the temperature is ≤20℃, call the temperature-ΔV mapping table and update ΔV.

[0048] Preferably, the LSTM-BP network hybrid model in step S107 includes: data loading, normalization processing, dataset partitioning, LSTM layer, fully connected layer and BP layer, and prediction output;

[0049] Load data, with input columns represented as vectors. , in, Let t represent the remaining capacity of the battery at time t. These are the n-dimensional pulse test feature values ​​corresponding to time t;

[0050] Normalize the input data:

[0051] (5)

[0052] (6)

[0053] in, , Let Y represent the minimum and maximum values ​​of the time series array Y, respectively. , These represent the minimum and maximum values ​​of the parameter array in the nth dimension, respectively.

[0054] The dataset is divided into three parts: 70% as the training set, 15% as the validation set, and 15% as the test set.

[0055] The processed data sequence is input into the LSTM layer. The multi-layer LSTM network can extract the long-term and short-term dependencies in the time series and improve the model's learning ability and robustness.

[0056] The features extracted from the LSTM layer are then used as input to the BP layer through a fully connected layer. Each input node of the fully connected layer is connected to each output node. The final predicted value is then output through the BP layer, and the loss is calculated to perform the final lifetime prediction. After the prediction is completed, the host computer stores the ΔV value, temperature, anomaly type, repair method, and anomaly feature value drift of this test in the anomaly-compensation-effect log library.

[0057] Preferably, the specific steps for interference tracing and model optimization in step S108 are as follows:

[0058] Step S1081, Interference Source Tracing: If ΔV drift > 5 times / day, trigger hardware fault prompt; if ΔV is generally > 8mV when temperature ≤ -25℃, update temperature-ΔV mapping table;

[0059] Step S1082, Model Optimization: When there are ≥500 abnormal samples in the log library, optimize the correction coefficient of the voltage-temperature correlation identification model through linear regression and update the control unit algorithm simultaneously;

[0060] Among them, the voltage-temperature correlation identification model is a parameter mapping-anomaly determination model based on the battery's electrochemical characteristics and temperature dependence. Its core function is to quantify the impact of temperature fluctuations on battery voltage characteristics, providing a basis for slope anomaly identification and temperature compensation. Its input includes the real-time voltage sequence after reference voltage compensation. i=1,2,...,n The data includes the voltage value at the i-th sampling moment after compensation with reference voltage, the synchronously collected ambient temperature T, covering the high altitude and low temperature range of -30℃ to 10℃, the slope characteristics of the voltage sequence, and the historical correlation data of temperature-voltage slope-anomaly markers in the log database.

[0061] The slope characteristic of the voltage sequence is a slope at three consecutive points:

[0062] (7)

[0063] in, , Let be the voltage slopes from point i-1 to point i and from point i to point i+1, respectively, and Δt be the sampling time interval; and be the rate of change of the slope. , The rate of change of the voltage slope between two adjacent segments; the output is the slope anomaly judgment result and the temperature correction coefficient k(T), where k(T) is the voltage correction coefficient corresponding to the current temperature T;

[0064] The voltage-temperature correlation identification model achieves the binding of temperature and voltage characteristics through a two-layer correlation: the first layer is the dynamic correlation between temperature and voltage slope, based on Fick's law, which fits the slope change rate threshold under normal temperature fluctuations, i.e., the reaction rate of temperature influence. The formula is as follows:

[0065] (8)

[0066] in, The slope change rate threshold under normal temperature fluctuations is defined as follows: a and b are initial fitting coefficients, determined through 50 sets of calibration samples; ΔT = |T - T0|, where ΔT is the temperature fluctuation value, T is the current measured temperature, and T0 is the temperature at the previous moment; when the measured... At that time, the slope abnormality was determined to be caused by temperature interference;

[0067] The second layer involves temperature and voltage compensation. A preset reference correction formula is used to uniformly correct voltage values ​​at different temperatures to a reference temperature of 25℃. The formula is as follows:

[0068] (9)

[0069] in, To correct the battery equivalent voltage value to a reference temperature of 25°C, The voltage value is after compensation by the reference voltage, k(T) is the temperature correction coefficient, the initial value is set to 0.002, and T is the current measured ambient temperature;

[0070] When the log database accumulates ≥500 abnormal samples, the linear regression equation is solved using the least squares method, with the goal of minimizing the error between the compensated voltage and the true voltage. The regression equation is:

[0071] (10)

[0072] in, This is the actual battery voltage value measured using high-precision calibration equipment. The voltage value is the value after compensation with a reference voltage, k(T) is the temperature correction factor, and T is the measured ambient temperature. This is the regression error term;

[0073] Ultimately, k(T) is dynamically converged to 0.0018 to 0.0022 within the range of -30℃ to 10℃; the optimized temperature correction coefficient k(T) is synchronously updated to the control unit for subsequent temperature compensation calculations, significantly improving the accuracy of temperature compensation.

[0074] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0075] This invention provides replacement prompts for aging channel hardware through interference tracing and model optimization, and also supplements low-temperature calibration points; furthermore, it improves temperature compensation accuracy with increasing testing volume. This invention addresses the problem of misjudgment caused by other disposable battery remaining capacity detection technologies not incorporating ΔV calibration thresholds through interference tracing-based active waveform anomaly identification; it also solves the problem of directly discarding abnormal data due to ignoring the correlation between temperature and slope; and it resolves the issue of ΔV calibration at room temperature not adapting to low-temperature drift at high altitudes.

[0076] This invention reduces the mean absolute error (MAE) of remaining capacity estimation under different high-altitude and low-temperature environments through temperature compensation function. It eliminates the need for large-scale sample tests at each temperature point and enables high-precision capacity prediction of batch batteries with only a few calibration tests.

[0077] This invention, based on the synergistic analysis of dynamic voltage response and static environmental parameters (temperature / reference voltage) under pulsed load excitation, achieves high-precision capacity prediction by dynamically stripping away passivation film interference. This drives the transformation of operation and maintenance models towards predictive maintenance, significantly improving the reliability and economy of power grid systems. This invention solves the problem of inaccurate remaining capacity detection in primary batteries due to continuous passivation, enabling accurate and rapid prediction of the remaining capacity of primary batteries. Attached Figure Description

[0078] Figure 1 This is a top view schematic diagram of a batch testing device for the remaining capacity of disposable batteries in high-altitude environments.

[0079] Figure 2 This is a schematic diagram of the internal structure of a batch testing device for the remaining capacity of disposable batteries in high-altitude environments.

[0080] Figure 3 This is a rear view schematic diagram of a batch detection device for the remaining capacity of disposable batteries in high-altitude environments.

[0081] Figure 4 This is a front view schematic diagram of a batch testing device for the remaining capacity of disposable batteries in high-altitude environments.

[0082] Figure 5 This is a schematic diagram of the module connection for a batch testing device for the remaining capacity of disposable batteries in high-altitude environments.

[0083] Figure 6 An automated testing flowchart for batch testing of the remaining capacity of disposable batteries in high-altitude environments.

[0084] Figure 7 This is a voltage waveform acquired by the pulse test and reference voltage acquisition module.

[0085] Figure 8 This is a schematic diagram of the LSTM-BP network hybrid model.

[0086] In the attached diagram, the following labels represent different components: 1 is the battery socket, 2 is resistive load one, 3 is resistive load two, 4 is connector one, 5 is connector two, 6 is the switching power supply, 7 is the temperature measurement module, 8 is the speaker, 9 is the passivation switching circuit one, 10 is the multi-channel switching circuit one, 11 is the indicator unit circuit, 12 is the controller MCU, 13 is the multi-channel switching circuit two, 14 is the passivation switching circuit two, 15 is the pulse switching circuit, 16 is the pulse test and reference voltage acquisition module, 17 is the power interface, 18 is the temperature measurement module communication interface, 19 is the pulse test and reference voltage acquisition module communication interface, 20 is the controller MCU communication interface, 21 is the three-color indicator light, and 22 is the power switch. Detailed Implementation

[0087] The present invention will be further described in detail below with reference to the accompanying drawings: This embodiment is implemented under the premise of the technical solution of the present invention, and detailed implementation methods are given, but the protection scope of the present invention is not limited to the following embodiments.

[0088] like Figures 1 to 5 As shown in this embodiment, a batch testing device for the remaining capacity of disposable batteries in a high-altitude environment is disclosed. The device includes a main body and a host computer. The main body includes: a power supply unit, a temperature monitoring unit, a control unit, an indicator unit, a switching unit, a battery testing unit, and a parameter acquisition unit, wherein:

[0089] The host computer is used to store the data collected by the temperature monitoring unit and the parameter acquisition unit, and to send instructions to the temperature monitoring unit, the control unit and the parameter acquisition unit to perform voltage-temperature correlation identification and remaining capacity prediction. After the data collected by the temperature monitoring unit and the parameter acquisition unit are sent to the host computer, the voltage-temperature correlation identification and remaining capacity prediction are both completed by the host computer.

[0090] The power supply unit includes: a power interface 17, a power switch 22, a switching power supply 6, and a step-down circuit; the power switch 22 is connected to the switching power supply 6 and controls the input switching of the switching power supply 6; the switching power supply 6 is connected to a switching unit, a control unit, a parameter acquisition unit, and a temperature monitoring unit; the power interface 17 can be connected to 220V AC mains power and is connected to the power switch 22, which controls the input switching of the switching power supply 6; the switching power supply 6 can convert 220V AC mains power to 12V, and its output is divided into two paths: the first path is connected to the step-down circuit as the input of the step-down circuit; the second path is connected to the switching unit; the step-down circuit can convert 12V to 5V, and its output is divided into four paths: the first path is used to power the MCU 12 and is connected to the control unit; the second path is connected to the parameter acquisition unit; the third path is connected to the temperature measurement module 7 of the temperature monitoring unit; and the fourth path is connected to the three-color indicator light 21 of the indicator unit.

[0091] The temperature monitoring unit collects temperature data for temperature and compensation. The temperature monitoring unit includes a temperature measurement module 7 and a thermocouple. The temperature measurement module 7 is connected to the thermocouple and the temperature measurement module communication interface 18 on the outer surface of the rear end of the main body of the detection device. It can collect the temperature at four different locations. The thermocouple is attached to the switching power supply 6, the pulse test and reference voltage acquisition module 16 and the battery test unit respectively to collect temperature data.

[0092] The control unit is a controller MCU12, which is connected to the controller MCU communication interface 20 on the rear outer surface of the depassivation unit, the switching unit, the indication unit, and the main body of the detection device.

[0093] The indicator unit includes a speaker 8, a switching indicator light, a three-color indicator light 21, and an indicator unit circuit 11. The speaker 8, the switching indicator light, and the three-color indicator light 21 are connected to the control unit via the indicator unit circuit 11. The speaker 8 can be used to alarm when a test fault occurs and is connected to the controller MCU 12. The switching indicator light indicates the working status of the de-passivation switching circuit 9, the de-passivation switching circuit 14, the multi-channel switching circuit 10, the multi-channel switching circuit 13, and the pulse switching circuit 15 in the switching unit and is connected to the controller MCU 12. The de-passivation switching circuit 9, the multi-channel switching circuit 10, the controller MCU 12, the multi-channel switching circuit 13, the de-passivation switching circuit 14, and the pulse switching circuit 15 are equipped with switching indicator lights on the front panel of the detection device. The three-color indicator light 21 is used to observe the device's operation, stop, and error status, indicating the device's operating status and is connected to the controller MCU 12.

[0094] The switching unit includes a depassivation switching circuit 1-9, a depassivation switching circuit 2-14, a multi-channel switching circuit 1-10, a multi-channel switching circuit 2-13, and a pulse switching circuit 15. The depassivation switching circuit 1-9 and the depassivation switching circuit 2-14 are respectively connected to the battery socket 1 of the battery testing unit and the resistive load 1-2 and the resistive load 2-3. The resistive load 1-2, the resistive load 2-3, and the battery socket 1 of the battery testing unit are connected through relays in the depassivation switching circuit 1-9 and the depassivation switching circuit 2-14. The multi-channel switching circuit 1-10 and the multi-channel switching circuit 2-13 are connected to the battery socket 1 of the battery testing unit and the pulse test and reference voltage acquisition module 16 of the parameter acquisition unit. The pulse switching circuit 15 is connected to the pulse test and reference voltage acquisition module 16 of the parameter acquisition unit.

[0095] The depassivation switching circuit 19 and the depassivation switching circuit 214 are connected to the battery socket 1 and the resistive load of the battery test unit. When the depassivation command is sent, the corresponding resistive load 12 and resistive load 23 can be connected to the battery socket 1 to form a closed loop to eliminate the passivation film.

[0096] The multi-channel switching circuit 10 and the multi-channel switching circuit 13 are connected to the battery socket 1 of the battery test unit and the pulse test and reference voltage acquisition module 16 of the parameter acquisition unit. When a channel switching command is sent, the corresponding battery socket 1 and the pulse test and reference voltage acquisition module 16 of the parameter acquisition unit can form a closed loop. The pulse test and reference voltage acquisition module 16 of the parameter acquisition unit can directly acquire the voltage at both ends of the battery socket 1.

[0097] The pulse switching circuit 15 and the pulse test and reference voltage acquisition module 16 of the parameter acquisition unit are connected. After sending the channel switching command, when sending the pulse switching command, the pulse load and the pulse test and reference voltage acquisition module 16 of the parameter acquisition unit can form a closed loop, that is, the pulse load and the battery socket 1 form a closed loop to perform pulse testing.

[0098] The battery testing unit includes a battery socket 1, a resistive load 1 2 and a resistive load 2 3. The resistive load 1 2 and the resistive load 2 3 are respectively connected to the connector 1 4 and the connector 2 5 at the upper end of the testing device. The connector 1 4 and the connector 2 5 are connected to the switching unit and the parameter acquisition unit.

[0099] The parameter acquisition unit is a pulse test and reference voltage acquisition module 16. The pulse test and reference voltage acquisition module 16 can acquire the voltage at both ends of the battery and the reference voltage during pulse load testing. The pulse test and reference voltage acquisition module 16 is connected to the battery socket 1, multi-channel switching circuit one 10, multi-channel switching circuit two 13, pulse switching circuit 15, reference voltage chip and pulse test and reference voltage acquisition module communication interface 19 provided on the rear outer surface of the main body of the detection device of the battery test unit.

[0100] The front outer surface of the main body of the detection device is provided with a switching indicator light, a three-color indicator light 21 and a power switch 22. The rear outer surface of the main body of the detection device is provided with a power interface 17, a controller MCU communication interface 20, a pulse test and reference voltage acquisition module communication interface 19 and a temperature measurement module communication interface 18. The upper end of the detection device is provided with connector 4 and connector 5 to connect to the battery test unit, which can realize quick plugging and unplugging.

[0101] like Figures 6 to 8 As shown, a batch testing method for the remaining capacity of disposable batteries in high-altitude environments, taking Saft batteries as an example, includes the following automatic testing process: reference voltage acquisition stage, temperature acquisition stage, passivation removal stage, resting stage, pulse load testing stage, feature value extraction stage, remaining capacity prediction stage, and cycle testing stage. The specific process is as follows:

[0102] Step S101, Reference Voltage Acquisition Stage: Install the quick-pluggable battery test unit on the connector at the top of the device body, and place the Saft battery in the battery holder with a resistive load of 180Ω; after the host computer sends the reference voltage acquisition command, the pulse test and reference voltage acquisition module 16 acquires the reference voltage VOUT, sends it to the host computer and stores it—this VOUT is used on the one hand to carry out reference voltage compensation in combination with high-altitude environmental parameters, and on the other hand, it is synchronously pre-stored in the "temperature-ΔV mapping table" and used for the host computer to load dynamic thresholds later.

[0103] (1)

[0104] (2)

[0105] in: This is the upper limit of the dynamic threshold. This is the lower limit of the dynamic threshold. This is the battery's nominal voltage. The reference voltage deviation coefficient;

[0106] Step S102, Temperature Acquisition Stage: The host computer sends a temperature acquisition command, and the temperature measurement module 7 sends the acquired temperature data to the host computer and stores it for temperature compensation. The acquisition range of the temperature acquisition module covers the typical low temperature range at high altitudes (-30℃ to 10℃). The acquired temperature data is transmitted to the control unit in real time and is synchronously cached with the pulse test voltage data.

[0107] Step S103, Depassivation Stage: After the host computer sends the depassivation command, the depassivation switching circuit connects the test battery with the serial number and predetermined time configured by the user in its program to the two ends of the corresponding resistive load and forms a separate closed loop; through the depassivation current that neither damages the internal electrochemical mechanism of the battery nor consumes too much effective capacity, the double-layer passivation film structure with LiCl as the main component is efficiently broken, wherein the predetermined depassivation time is t101;

[0108] Step S104, Settling Stage: The host computer sends a settling command, and the depassivation switching circuit removes the deresistivity load at both ends of the battery, wherein the settling time is predetermined t102.

[0109] Step S105, Pulse Load Test and Temperature Acquisition Stage: After the host computer sends the pulse load test command, the multi-channel switching circuit connects the battery under test (which the user has configured with a serial number and predetermined time in their program) to the pulse test and reference voltage acquisition module 16 and starts acquiring the voltage across the battery. One second after acquisition begins, the pulse switching circuit 15 connects the selected pulse load to both ends of the battery, holds it for a predetermined time, and then removes it (the predetermined time for pulse load connection is t103). After the pulse load is removed, the multi-channel switching circuit removes the battery under test from the pulse test and reference voltage acquisition module 16. Three seconds later, the pulse test and reference voltage acquisition module 16 stops voltage acquisition, and the acquired voltage data is sent to the host computer and stored (the predetermined pulse test time is 5 seconds, and the pulse load is 36Ω). The host computer compares the sampled voltage... and If the value exceeds the limit, an abnormal amplitude is marked; calculate the slope of three consecutive points. , And the slope change rate γ, if γ>50% and temperature fluctuation>±2℃, mark the slope as abnormal; compare VOUT and VREF in real time, if the absolute value of ΔV>5mV (room temperature) or>8mV (≤-20℃), mark the drift as abnormal;

[0110] Step S106, Feature Value Extraction Stage: After receiving voltage data and anomaly markers, the host computer first performs repair:

[0111] (1) Minor anomalies: For single-point amplitude anomalies (ΔV of 3-5mV) or slope anomalies of temperature fluctuations of 1-2℃, first check the following:

[0112] (3)

[0113] in, This is the corrected voltage value after compensation with the reference voltage. This represents the original voltage value acquired at the i-th sampling time.

[0114] Correct the benchmark, then use

[0115] (4)

[0116] in, For actual measured temperature, This is the battery voltage value equivalent to a reference temperature of 25°C after temperature compensation.

[0117] Corrected to the 25℃ equivalent value;

[0118] (2) Serious abnormality: For more than 2 consecutive abnormalities or ΔV>8mV, control the multi-channel switching unit to switch to the backup channel, increase the sampling frequency to 5kHz and retest, and take the average value after comparing the two ΔV deviations of <3mV;

[0119] (3) High-frequency drift: If ΔV drift > 3 times / hour at ≤20℃, call the 'temperature-ΔV mapping table' to update ΔV;

[0120] The host computer extracts the slope of the feature value AD from the repaired data. The slope of DE The slope of CD DC time constant The voltage values ​​of C+D, D+2 (the second point to the right of D), D+3 (the third point to the right of D), and D+4 (the fourth point to the right of D), after being compensated by the reference voltage, are used as the inputs to the LSTM-BP network hybrid model.

[0121] Step S107, Remaining Capacity Prediction Stage: The host computer inputs the feature values ​​described in step S106 into the LSTM-BP network hybrid model, which consists of data loading, normalization processing, dataset partitioning, LSTM layer, fully connected layer, BP layer and prediction output module.

[0122] Load data, with input columns represented as vectors. , in, Let t represent the remaining capacity of the battery at time t. These are the n-dimensional pulse test feature values ​​corresponding to time t;

[0123] Normalize the input data:

[0124] (5)

[0125] (6)

[0126] in, , Let Y represent the minimum and maximum values ​​of the time series array Y, respectively. , These represent the minimum and maximum values ​​of the parameter array in the nth dimension, respectively.

[0127] The dataset is divided into three parts: 70% as the training set, 15% as the validation set, and 15% as the test set.

[0128] The processed data sequence is input into the LSTM layer. The multi-layer LSTM network can extract the long-term and short-term dependencies in the time series and improve the model's learning ability and robustness.

[0129] The features extracted from the LSTM layer are then used as input to the BP layer through a fully connected layer. Each input node of the fully connected layer is connected to each output node. The BP layer then outputs the final predicted value and calculates the loss to perform the final lifetime prediction. The prediction result is output through the prediction output module. After the prediction is completed, the host computer stores the ΔV value, temperature, anomaly type, repair method, and anomaly feature value drift of this test in the 'anomaly-compensation-effect' log library.

[0130] Step S108, Cyclic Testing Phase: Repeat steps S105 to S107 until the test of the battery with the serial number configured by the user in the host computer program is completed. During the repetition of S105-S107, the host computer will collect and log data in real time.

[0131] (1) Interference source tracing: If ΔV drift > 5 times / day, trigger hardware fault prompt; if ΔV is generally > 8mV when the temperature is ≤ -25℃, update the 'temperature-ΔV mapping table';

[0132] (2) Model optimization: When there are ≥500 abnormal samples in the log library, the correction coefficient of the temperature-voltage correlation model is optimized by linear regression (dynamically adjusted by 0.002), and the control unit algorithm is updated synchronously;

[0133] Example 1:

[0134] In this specific embodiment, the batch testing device for the remaining capacity of disposable batteries in high-altitude environments can perform batch testing on up to 32 batteries and has a reference voltage compensation function. The reference voltage compensation addresses the voltage acquisition deviation caused by differences in the hardware of different ADC chips (such as inconsistent gain or zero-point offset) in other disposable battery remaining capacity testing technologies. First, the voltage VREF of a reference voltage chip with a clear and stable output value is used as the "standard reference". The pulse test and reference voltage acquisition module 16 acquires this reference voltage VOUT. By comparing the acquired value with the actual value of the reference voltage, the deviation correction coefficient ΔV of each module is calculated.

[0135] (7)

[0136] like Figure 7 As shown, the voltage points A, B, C, D, and E are acquired by the pulse test and reference voltage acquisition module 16. Subsequently, during the actual voltage acquisition for the battery pulse test, the original acquired characteristic values ​​are reverse-corrected.

[0137] (8)

[0138] (9)

[0139] (10)

[0140] (11)

[0141] This offsets the systematic errors caused by hardware differences, unifies different acquisition results to the real voltage reference, and ensures the consistency and accuracy of the acquired data.

[0142] The batch testing device for the remaining capacity of primary batteries in high-altitude environments addresses the impact of temperature on the electrochemical characteristics of batteries under typical low-temperature conditions (-30℃ to 10℃) at high altitudes. It constructs a temperature compensation function through "parameter characteristic analysis - linear correction modeling - equivalent value conversion." The core logic is as follows:

[0143] According to Fick's law, increased temperature accelerates the internal reaction and diffusion processes of the battery, causing individual voltage parameters (such as points D and E) to change significantly with temperature (points D and E increase at high temperatures and decrease at low temperatures). If used directly for capacity prediction, this can easily lead to errors. However, experiments have shown that the combined voltage parameters extracted by pulse testing are significantly less affected by temperature interference than individual voltage parameters because the temperature effects between parameters cancel each other out.

[0144] Based on this, temperature compensation is achieved through the following steps:

[0145] Experimental calibration: At different low temperatures at high altitudes (such as -30℃, -20℃, 0℃) and at a reference temperature of 25℃, pulse tests were conducted on a batch of batteries to obtain the correspondence between core parameters such as C+D, D+2, D+3, and D+4 and the remaining capacity at each temperature, and to clarify the deviation pattern of parameters relative to 25℃ at different temperatures.

[0146] Correlation Model Construction: Using 25℃ as the baseline, a linear correlation model of "temperature-parameter correction coefficient" is established—based on experimental data, specific linear correction coefficients are set for different temperature points, and the specific correction formula is as follows:

[0147] (12)

[0148] Where U is the measured parameter value, U′ is the equivalent value corrected to 25℃, and k(T) is the correction coefficient corresponding to temperature T (dynamically optimized through linear regression, ranging from 0.92 to 1.15).

[0149] Equivalent value conversion: The parameters measured at various temperatures at high altitudes are multiplied by the corresponding correction coefficients through the above correlation model to convert them into equivalent parameter values ​​at a reference temperature of 25℃, thus eliminating the interference of temperature on the correspondence between "parameter-remaining capacity".

[0150] This temperature compensation function can reduce the mean absolute error (MAE) of remaining capacity estimation in different high-altitude and low-temperature environments from less than 9% without correction to about 5%. It eliminates the need for large-scale sample tests for each temperature point and enables high-precision capacity prediction of batch batteries with only a few calibration tests.

[0151] The channel switching functions of passivation switching circuits, multi-channel switching circuits, and pulse switching circuits are all implemented by small relays or MOSFETs.

[0152] The passivation switching circuit, multi-channel switching circuit, and pulse switching circuit allow users to configure their channel switching functions via parameters in the host computer program. The depassivation switching circuit is used to simultaneously depassivate one or more of up to 32 batteries, with a depassivation current range of 0mA to 150mA. The multi-channel switching circuit is used to select different channels for voltage acquisition. The pulse switching circuit allows users to configure eight pulse loads in the host computer program and connect the selected pulse loads to the two ends of the selected batteries for testing, with a pulse load current range of 0mA to 150mA.

[0153] The aforementioned automated testing process addresses the issues that other disposable battery remaining capacity detection technologies may encounter due to differences in the initial passivation degree of the battery (e.g., batteries stored for a longer time have thicker passivation films) and activity differences at high altitudes and low temperatures—fixed parameters may lead to incomplete passivation removal (large testing errors) or excessive capacity consumption (misjudgment of remaining capacity). In step S103, the predetermined passivation removal time t101, the predetermined resting time t102 in step S104, and the predetermined pulse test time t103 in step S105 are configured by the user in the host computer program. Specifically, t101 ranges from 100s to 1800s, with a preferred passivation removal time of 900s; t102 ranges from 200s to 2000s, with a preferred resting time of 300s; and t103 ranges from 0.5s to 10s, with a preferred pulse test time of 1s.

[0154] To address the insufficient robustness of other single-use battery remaining capacity detection technologies that only compensate for reference voltage or temperature interference and lack proactive anomaly handling, this paper employs an "interference source identification - dual-parameter collaborative repair - closed-loop optimization" mechanism to accurately address the combined effects of reference voltage hardware deviation and temperature fluctuations.

[0155] (1) Active identification of waveform anomalies for interference tracing

[0156] Because the pulse test and the reference voltage acquisition module 16 synchronously acquire the VREF and battery pulse voltage waveform output by the reference voltage chip, and the temperature acquisition module acquires the battery and ADC chip temperature in real time (covering the high altitude and low temperature range of -30℃ to 10℃), the three data are transmitted to the control unit; the control unit constructs a "voltage-temperature" correlation identification model to accurately locate the waveform abnormalities caused by reference voltage deviation and temperature fluctuation;

[0157] The voltage-temperature correlation identification model is a parameter mapping-anomaly detection model based on the battery's electrochemical characteristics and temperature dependence. Its core function is to quantify the impact of temperature fluctuations on battery voltage characteristics, providing a basis for slope anomaly identification and temperature compensation. Its input includes a real-time voltage sequence compensated for a reference voltage. i=1,2,...,n The data includes the voltage value at the i-th sampling moment after compensation with reference voltage, the synchronously collected ambient temperature T, covering the high altitude and low temperature range of -30℃ to 10℃, the slope characteristics of the voltage sequence, and the historical correlation data of temperature-voltage slope-anomaly markers in the log database.

[0158] The slope characteristic of the voltage sequence is a slope at three consecutive points:

[0159] (13)

[0160] in, , Let be the voltage slopes from point i-1 to point i and from point i to point i+1, respectively, and Δt be the sampling time interval; and be the rate of change of the slope. , The output is the rate of change of the slope between two adjacent voltage segments; the output is the slope anomaly determination result (e.g., " "When the voltage fluctuation is greater than 50% and the temperature fluctuation is greater than ±2℃, it is considered abnormal" and temperature correction factor k(T) (k(T) is the voltage correction factor corresponding to the current temperature T).

[0161] The model achieves the binding of temperature and voltage characteristics through two layers of correlation: the first layer is a dynamic correlation between temperature and voltage slope, based on Fick's law, which fits the slope change rate threshold under normal temperature fluctuations, i.e., the temperature affects the reaction rate. The formula is as follows:

[0162] (14)

[0163] in, The slope change rate threshold under normal temperature fluctuations, a and b are the initial fitting coefficients, determined through 50 sets of calibration samples; ΔT=|T-T0|, where ΔT is the temperature fluctuation value, T is the current measured temperature, and T0 is the temperature at the previous moment;

[0164] When measured When the slope anomaly is determined to be caused by temperature interference, the second layer is a compensation correlation between temperature and voltage values. A preset reference correction formula is used to uniformly correct voltage values ​​at different temperatures to a reference temperature of 25℃. The formula is as follows:

[0165] (15)

[0166] in, To correct the battery equivalent voltage value to a reference temperature of 25°C, The voltage value is the voltage value after compensation by the reference voltage, k(T) is the temperature correction coefficient, the initial value is set to 0.002, and T is the current measured ambient temperature.

[0167] When the log database accumulates ≥500 abnormal samples, with the goal of "minimizing the error between the compensated voltage and the true voltage", the linear regression equation is solved using the least squares method. The regression equation is:

[0168] (16)

[0169] in, This is the actual battery voltage value measured using high-precision calibration equipment. ε is the voltage value after compensation by the reference voltage, k(T) is the temperature correction coefficient to be optimized, T is the measured ambient temperature, and ε is the regression error term.

[0170] Ultimately, k(T) is dynamically converged to 0.0018 to 0.0022 within the range of -30℃ to 10℃; the optimized temperature correction coefficient k(T) is synchronously updated to the control unit for subsequent temperature compensation calculations, significantly improving the accuracy of temperature compensation.

[0171] Amplitude anomaly identification: Based on the reference voltage compensation coefficient ΔV, set the upper and lower limits of the dynamic threshold Thdyn:

[0172] (17)

[0173] (18)

[0174] If the voltage at the sampling point exceeds the threshold, it is determined to be an amplitude abnormality caused by hardware deviation of the reference voltage (such as ADC gain drift), which solves the problem of easy misjudgment due to the fact that other single-use battery remaining capacity detection technologies do not incorporate the ΔV calibration threshold;

[0175] Slope anomaly identification: Calculate the voltage slope at three consecutive points:

[0176] (19)

[0177] and rate of change:

[0178] (20)

[0179] like If the battery capacity is >50% and the synchronous temperature fluctuation is >±2℃, it is determined to be an abnormal slope caused by a sudden change in battery activity due to temperature fluctuation. This solves the problem that other technologies for detecting the remaining capacity of primary batteries ignore the correlation between temperature and slope and directly discard abnormal data.

[0180] Reference voltage drift identification: Compare the collected VREF with the true value of VOUT. If the absolute value of ΔV is >5mV (at room temperature) or >8mV (≤-20℃), it is determined to be an abnormal zero-point drift of the reference voltage due to low temperature. This solves the problem that other single-use battery remaining capacity detection technologies only calibrate ΔV at room temperature and are not adapted to low-temperature drift at high altitudes.

[0181] (2) Dual-parameter synergistic hierarchical repair

[0182] The control unit, based on the type of anomaly, integrates reference voltage compensation and temperature compensation to achieve precise repair.

[0183] Minor Anomalies: ΔV-Temperature Co-correction: When a single point amplitude anomaly (ΔV deviation 3-5mV) or a slight slope anomaly caused by temperature fluctuation of 1-2℃ is detected, the correction is first performed by... Correct the reference deviation, and then correct the voltage to the equivalent value at 25℃ based on the temperature-voltage correlation model. The feature value extraction error is ≤2% (the conventional single ΔV compensation error is ≥7%).

[0184] Severe anomaly: Retest + Dynamic ΔV calibration: When the ΔV deviation > 8mV or the temperature fluctuation > 3℃ causes more than 2 consecutive anomalies, the control pulse test and reference voltage acquisition module 16 pauses the test and re-acquires VREF to calibrate ΔV (calibration frequency increased to 1 time / 30s). At the same time, the sampling frequency is increased from 1kHz to 5kHz for retesting. When the ΔV deviation of the two retests is < 3mV, the average value is taken as the valid data.

[0185] High-frequency drift: Temperature-adaptive ΔV update: When the ΔV drift frequency is >3 times / hour in an environment of ≤-20℃, the "temperature-ΔV mapping table" is automatically activated to dynamically update the ΔV compensation coefficient and avoid the cumulative error caused by continuous drift at low temperatures.

[0186] (3) Closed-loop optimization iteration

[0187] The host computer establishes an "anomaly-compensation-effect" log database, storing the ΔV value, temperature, repair method, and characteristic value error for each anomaly; data analysis is performed by the host computer to achieve the following:

[0188] Interference source tracing: If the ΔV drift frequency is >5 times / day, it is determined that the channel hardware is aging and replacement is prompted; if the ΔV drift is generally >8mV at ≤-25℃, update the "Temperature-ΔV Mapping Table" to supplement the low temperature calibration points;

[0189] Model optimization: For every 500 abnormal samples accumulated, the correction coefficient of the temperature-voltage correlation model is optimized by linear regression (dynamically adjusted from 0.002 to 0.0018~0.0022) so that the temperature compensation accuracy increases with the increase of the number of tests.

[0190] The above description is merely a preferred embodiment of the present invention. These specific embodiments are different implementations based on the overall concept of the present invention, and the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A detection method for a batch detection device for the remaining capacity of disposable batteries in high-altitude environments, characterized in that, Includes the following steps: Step S101, Reference Voltage Acquisition Stage: Install the battery test unit on the connector of the detection device and put the battery to be tested into the battery socket (1); after the host computer sends the reference voltage acquisition command, the pulse test and reference voltage acquisition module (16) acquires the reference voltage VOUT, sends it to the host computer and stores it, and loads the dynamic threshold for the host computer later; the reference voltage VOUT is used on the one hand to carry out reference voltage compensation in combination with high altitude environmental parameters, and on the other hand, it is synchronously pre-stored in the temperature-ΔV mapping table; Step S102, Temperature Acquisition Stage: The host computer sends a temperature acquisition command, and the temperature measurement module (7) sends the acquired temperature data to the host computer and completes storage for temperature compensation; the host computer transmits the acquired temperature data to the control unit in real time. Step S103, Depassivation Stage: After the host computer sends the depassivation command, the depassivation switching circuit connects the battery under test to the two ends of the corresponding resistive load and forms a separate closed loop; the passivation film structure is efficiently broken through the depassivation current. Step S104, Settling Stage: The host computer sends a settling command, and the depassivation switching circuit removes the resistive load at both ends of the battery under test and sets the battery for a predetermined time. Step S105, Pulse Load Test and Temperature Acquisition Stage: After the host computer sends the pulse load test command, the multi-channel switching circuit connects the battery under test to the pulse test and reference voltage acquisition module 16 and starts the acquisition of the voltage at both ends of the battery. After the acquisition begins, the pulse switching circuit (15) connects the selected pulse load to both ends of the battery under test and then removes it. After the pulse load is removed, the multi-channel switching circuit removes the battery under test from the pulse test and reference voltage acquisition module 16, stops voltage acquisition, sends the acquired pulse load test voltage data to the host computer and completes storage, and the host computer marks the abnormality. Step S106, Feature Value Extraction Stage: After receiving the voltage data and anomaly markers collected in step S105, the host computer first performs data repair, then extracts the feature values ​​from the repaired data, and after reference voltage compensation, uses them as input to the LSTM-BP network hybrid model. Step S107, Remaining Capacity Prediction Stage: The host computer inputs the feature values ​​extracted in step S106 into the LSTM-BP network hybrid model, outputs the final predicted value and calculates the loss, and performs the final lifetime prediction. After the prediction is completed, the host computer stores the test data in the log database. Step S108, Cyclic Testing Phase: Repeat steps S105 to S107 until the test battery with the serial number configured by the user in the host computer program is completed. The host computer will collect log data in real time and perform interference source tracing and model optimization. The specific steps for interference tracing and model optimization described in step S108 are as follows: Step S1081, Interference Source Tracing: If ΔV drift > 5 times / day, trigger hardware fault prompt; if ΔV is generally > 8mV when temperature ≤ -25℃, update temperature-ΔV mapping table; Step S1082, Model Optimization: When there are ≥500 abnormal samples in the log library, optimize the correction coefficient of the voltage-temperature correlation identification model through linear regression and update the control unit algorithm simultaneously; Among them, the voltage-temperature correlation identification model is a parameter mapping-anomaly determination model based on the battery's electrochemical characteristics and temperature dependence. Its core function is to quantify the impact of temperature fluctuations on battery voltage characteristics, providing a basis for slope anomaly identification and temperature compensation. Its input includes the real-time voltage sequence after reference voltage compensation. i=1,2,...,n The data includes the voltage value at the i-th sampling moment after compensation with reference voltage, the synchronously collected ambient temperature T, covering the high altitude and low temperature range of -30℃ to 10℃, the slope characteristics of the voltage sequence, and the historical correlation data of temperature-voltage slope-anomaly markers in the log database. The slope characteristic of the voltage sequence is a slope at three consecutive points: in, , Let be the voltage slopes from point i-1 to point i and from point i to point i+1, respectively, and Δt be the sampling time interval; and be the rate of change of the slope. , The rate of change of the voltage slope between two adjacent segments; the output is the slope anomaly judgment result and the temperature correction coefficient k(T), where k(T) is the voltage correction coefficient corresponding to the current temperature T; The voltage-temperature correlation identification model achieves the binding of temperature and voltage characteristics through a two-layer correlation: the first layer is the dynamic correlation between temperature and voltage slope, based on Fick's law, which fits the slope change rate threshold under normal temperature fluctuations, i.e., the reaction rate of temperature influence. The formula is as follows: in, The slope change rate threshold under normal temperature fluctuations is defined as follows: a and b are initial fitting coefficients, determined through 50 sets of calibration samples; ΔT = |T - T0|, where ΔT is the temperature fluctuation value, T is the current measured temperature, and T0 is the temperature at the previous moment; when the measured... At that time, the slope abnormality was determined to be caused by temperature interference; The second layer involves temperature and voltage compensation. A preset reference correction formula is used to uniformly correct voltage values ​​at different temperatures to a reference temperature of 25℃. The formula is as follows: in, To correct the battery equivalent voltage value to a reference temperature of 25°C, The voltage value is after compensation by the reference voltage, k(T) is the temperature correction coefficient, the initial value is set to 0.002, and T is the current measured ambient temperature; When the log database accumulates ≥500 abnormal samples, the linear regression equation is solved using the least squares method, with the goal of minimizing the error between the compensated voltage and the true voltage. The regression equation is: in, This is the actual battery voltage value measured using high-precision calibration equipment. ε is the voltage value after compensation by the reference voltage, k(T) is the temperature correction coefficient, T is the measured ambient temperature, and ε is the regression error term. Ultimately, k(T) is dynamically converged to 0.0018 to 0.0022 within the range of -30℃ to 10℃; the optimized temperature correction coefficient k(T) is synchronously updated to the control unit for subsequent temperature compensation calculations, significantly improving the accuracy of temperature compensation.

2. The detection method according to claim 1, characterized in that, The formula for loading the dynamic threshold in step S101 is as follows: in: This is the upper limit of the dynamic threshold. This is the lower limit of the dynamic threshold. This is the battery's nominal voltage. This is the reference voltage deviation coefficient.

3. The detection method according to claim 2, characterized in that, The specific steps for the host computer to perform anomaly marking in step S105 are as follows: Step S1051: The host computer compares the sampled reference voltage VOUT. and If the value exceeds this range, the marking amplitude is abnormal. Step S1052: Calculate the voltage slopes k1, k2 and the rate of change γ at three consecutive points. If γ > 50% and the temperature fluctuation > ±2℃, mark the slope as abnormal. Step S1053: Real-time comparison pulse test and reference voltage acquisition module (16) acquires reference voltage VOUT and reference voltage chip voltage VREF. If the absolute value of ΔV is >5mV at room temperature or >8mV at temperature ≤-20℃, mark it as an abnormal drift.

4. The detection method according to claim 2, characterized in that, The data repair described in step S106 includes the following situations: (1) Minor anomalies: For single-point amplitude anomalies, i.e., ΔV of 3-5mV or slope anomalies with temperature fluctuations of 1-2℃, the baseline is first corrected using the following formula. in, This is the corrected voltage value after compensation with the reference voltage. This represents the original voltage value acquired at the i-th sampling time. Then correct to the 25℃ equivalent value using the following formula; in, For actual measured temperature, This is the battery voltage value equivalent to a reference temperature of 25°C after temperature compensation. (2) Serious abnormality: For two or more consecutive abnormalities or ΔV>8mV, control the multi-channel switching unit to switch to the backup channel, increase the sampling frequency to 5kHz and retest, and take the average value after comparing the two ΔV deviations <3mV; (3) High-frequency drift: If the ΔV drift is greater than 3 times / hour when the temperature is ≤20℃, call the temperature-ΔV mapping table and update ΔV.

5. The detection method according to claim 1, characterized in that, The LSTM-BP hybrid network model described in step S107 includes: data loading, normalization processing, dataset partitioning, LSTM layer, fully connected layer and BP layer, and prediction output; Load data, with input columns represented as vectors. , in, Let t represent the remaining capacity of the battery at time t. These are the n-dimensional pulse test feature values ​​corresponding to time t; Normalize the input data: in, , Let Y represent the minimum and maximum values ​​of the time series array Y, respectively. , These represent the minimum and maximum values ​​of the parameter array in the nth dimension, respectively. The dataset is divided into three parts: 70% as the training set, 15% as the validation set, and 15% as the test set. The processed data sequence is input into the LSTM layer. The multi-layer LSTM network can extract the long-term and short-term dependencies in the time series and improve the model's learning ability and robustness. The features extracted from the LSTM layer are then used as input to the BP layer through a fully connected layer. Each input node of the fully connected layer is connected to each output node. The final predicted value is then output through the BP layer, and the loss is calculated to perform the final lifetime prediction. After the prediction is completed, the host computer stores the ΔV value, temperature, anomaly type, repair method, and anomaly feature value drift of this test in the anomaly-compensation-effect log library.

6. A batch detection device for the remaining capacity of disposable batteries in a high-altitude environment based on the detection method of any one of claims 1-5, characterized in that, The device includes: a detection device body and a host computer; the detection device body includes: a power supply unit, a temperature monitoring unit, a control unit, an indicator unit, a switching unit, a battery testing unit, and a parameter acquisition unit; The signal input / output terminals of the host computer are respectively connected to the signal input / output terminals of the temperature monitoring unit, the control unit, and the parameter acquisition unit; the power supply unit is electrically connected to the temperature monitoring unit, the control unit, the switching unit, and the parameter acquisition unit; the signal output terminal of the control unit is respectively connected to the signal input terminals of the indicator unit and the switching unit; the signal output terminal of the switching unit is respectively connected to the signal input terminals of the control unit and the battery testing unit; and the signal output terminal of the battery testing unit is connected to the signal input terminal of the parameter acquisition unit. The host computer is used to store the data collected by the temperature monitoring unit and the parameter acquisition unit, and to send instructions to the temperature monitoring unit, the control unit and the parameter acquisition unit to perform voltage-temperature correlation identification and remaining capacity prediction. The power supply unit is used to supply power to the temperature monitoring unit, control unit, switching unit and parameter acquisition unit; The temperature monitoring unit is used to convert the collected temperature signals into electrical signals and transmit them to the host computer. The control unit is used to control the switching unit to perform passivation film elimination and pulse load testing; The indicator unit is used to alarm when a test fault occurs, and transmits signals to the control unit to indicate the working status of the switching unit and the operating status of the device; The switching unit is used to connect a resistive load to both ends of the battery to eliminate the passivation film, and then select different batteries to connect different pulse loads to both ends of the battery for pulse load testing. The battery test unit is used to enable quick plugging and unplugging for battery testing; The parameter acquisition unit is used to acquire the voltage across the battery terminals and the reference voltage during pulse load testing.

7. The batch detection device for remaining capacity of disposable batteries in high-altitude environments according to claim 6, characterized in that, The power supply unit includes: a power switch (22) and a switching power supply (6), the power switch (22) and the switching power supply (6) are connected; the switching power supply (6) is electrically connected to the switching unit, the control unit, the parameter acquisition unit and the temperature monitoring unit respectively; The temperature monitoring unit includes: a temperature measuring module (7) and a thermocouple, the temperature measuring module (7) being connected to the thermocouple; the thermocouple is attached to the switching power supply (6), the parameter acquisition unit and the battery testing unit respectively; The control unit is a controller MCU (12), which is connected to the switching unit and the indicator unit; The indicator unit includes: a speaker (8), a switching indicator light and a three-color indicator light (21), and the speaker (8), the switching indicator light and the three-color indicator light (21) are respectively connected to the control unit; The battery testing unit includes a battery socket (1) and a resistive load. The battery socket (1) is connected to the resistive load, the resistive load is connected to a connector, and the connector is connected to a switching unit and a parameter acquisition unit.

8. The batch detection device for the remaining capacity of disposable batteries in high-altitude environments according to claim 7, characterized in that, The parameter acquisition unit is a pulse test and reference voltage acquisition module (16), which is connected to the battery socket (1) and the reference voltage chip respectively.

9. The batch detection device for remaining capacity of disposable batteries in high-altitude environments according to claim 8, characterized in that, The switching unit includes: a depassivation switching circuit, a multi-channel switching circuit, and a pulse switching circuit (15). The depassivation switching circuit is connected to the battery socket (1) and the resistive load, respectively. The multi-channel switching circuit is connected to the battery socket (1) and the pulse test and reference voltage acquisition module (16). The pulse switching circuit (15) is connected to the pulse test and reference voltage acquisition module (16).

Images