Multi-fault diagnosis device for lithium battery system

Abstract

The invention provides a multi-fault diagnosis device for a lithium battery system, which is characterized by comprising a data acquisition module, the voltage estimation module is used for obtaining the estimated voltage of each battery according to the total voltage, the terminal voltage and the SOC value; the fault value calculation module is used for calculating a first fault value according to the estimated voltage and the measured voltage of each battery; the battery envelope relation generation module is used for calculating a second fault value corresponding to each battery according to the measurement voltage of each battery and the total voltage; the first fault judgment module is used for judging whether the first fault value is larger than a first fault threshold value or not, if yes, the lithium battery system has a fault, and if not, the lithium battery system operates normally; and the second fault judgment module is used for obtaining a fault diagnosis result according to the second fault value and the second fault judgment condition. In a word, the method can accurately identify the fault of the lithium battery system and perform early warning judgment.

Description

Technical Field

[0001] The invention belongs to the technical field of lithium-ion battery management, and in particular relates to a multi-fault diagnosis device for a lithium battery system. Background Art

[0002] Lithium-ion battery system (LBS) is an important energy storage and power component and has been widely used in various fields, especially in energy, automobile and aerospace. In recent years, spontaneous combustion and fire incidents of LBS have become increasingly frequent, which has attracted people's attention to its safety issues. The main reason for this phenomenon is the abuse of LBS caused by wrong decisions of the battery management system. Wrong decisions may induce abuse behaviors such as overcharging and over-discharging of LBS, and the abuse behaviors will intensify over time, eventually causing thermal runaway and thermal diffusion of LBS. However, the wrong decisions of the battery management system are mainly due to information distortion caused by internal faults of LBS. At the same time, due to external environmental factors and component aging, sensor failure and internal short circuit of the battery are difficult to avoid. Therefore, it is crucial to detect faults in the early stages of failure and provide early warning.

[0003] At present, fault diagnosis technologies for LBS are mainly divided into model-based frameworks and data-driven frameworks. The main focus of current fault diagnosis technologies is to detect and identify faults. However, it is basically assumed that only one fault can occur in the same time period. In addition, there are very few analysis technologies for fault severity, and existing quantitative analysis technologies are all for single fault scenarios. Related fault warning technologies are still in their infancy, and existing technologies mostly rely on voltage information, and warning analysis is relatively simple.

[0004] Therefore, existing fault warning technologies still have problems such as simple warning analysis and inability to provide multiple fault warnings. Summary of the invention

[0005] The present invention is made to solve the above-mentioned problems, and its purpose is to provide a multi-fault diagnosis device for a lithium battery system.

[0006] The present invention provides a multi-fault diagnosis device for a lithium battery system, which is used for performing fault diagnosis on a battery pack of a lithium battery system, and has the following characteristics: a data acquisition module, which is used for real-time acquisition of the total voltage of the battery pack, as well as the terminal voltage, SOC value and measured voltage of each battery in the battery pack; a voltage estimation module, which is used for obtaining the estimated voltage of each battery according to the total voltage, the terminal voltage and the SOC value; a fault value calculation module, which is used for calculating a first fault value according to the estimated voltage and the measured voltage of each battery; a battery envelope relationship generation module, which is used for calculating a second fault value corresponding to each battery according to the measured voltage and the total voltage of each battery; a first fault judgment module, which stores a preset first fault threshold, which is used for judging whether the first fault value is greater than the first fault threshold, if so, the lithium battery system has a fault, if not, the lithium battery system operates normally; a second fault judgment module, which stores a preset second fault judgment condition, which is used for obtaining a fault diagnosis result according to the second fault value and the second fault judgment condition, wherein the fault diagnosis result includes a total voltage sensor fault, a battery voltage sensor fault, a battery current sensor fault and a battery micro-short circuit fault.

[0007] In the lithium battery system multi-fault diagnosis device provided by the present invention, it can also have the following characteristics: wherein the voltage estimation module includes a long short-term memory network, the long short-term memory network generates a corresponding output vector according to an input vector, the input vector is composed of the total voltage and the terminal voltage and SOC value of the battery, the output vector is the estimated voltage of the battery, and the input vector corresponding to the i-th battery at the k-th time step is The expression is: Where U T,k-α is the total voltage of the battery pack corresponding to the k-αth time step, U i,k-α is the terminal voltage of the ith battery at the k-αth time step, SOC i,k-α is the SOC value corresponding to the i-th battery at the k-αth time step, and the estimated voltage corresponding to the i-th battery at the kth time step The expression is: Where W 1:n is the weight matrix of the LSTM network, and f(·) is the activation function of the LSTM network.

[0008] The multi-fault diagnosis device for lithium battery system provided by the present invention may also have the following features: wherein the fault value calculation module calculates the first fault value corresponding to the battery in the cycle according to all estimated voltages and measured voltages in a cycle, and the first fault value d corresponding to the i-th battery E (U e ,U m ) is: Where n is the total number of batteries in the battery pack, b is the length of the time window, is the estimated voltage of the ith battery at the kth time step, is the measured voltage of the i-th battery at the k-th time step.

[0009] The multi-fault diagnosis device for lithium battery system provided by the present invention may also have the following features: wherein the second fault value of the i-th battery The calculation expression is:

[0010] τ l =[(l-1)b+1,lb], l = [(γ-1)h+1,γh], where h is the time window length, ξ is a positive constant, and U T is the total voltage of the battery pack at the kth time step, is the measured voltage of the ith battery at the kth time step, and b is the length of the time window.

[0011] The lithium battery system multi-fault diagnosis device provided by the present invention may also have the following features: wherein the second fault judgment condition includes: When , the current sensor of the i-th battery has a fault in the time period γ; and When , there is a total voltage sensor fault in the time period γ; and When , the voltage sensor of the i-th battery has a fault in the time period γ; and When , the i-th battery has a micro-short circuit fault in the time period γ.

[0012] The lithium battery system multi-fault diagnosis device provided by the present invention may also have such a feature, and also includes: a current sensor fault time calculation module, a current sensor fault current calculation module, a voltage sensor fault time calculation module, a voltage sensor fault voltage calculation module, a micro-short circuit fault time calculation module and a micro-short circuit current calculation module, wherein the data acquisition module is also used to collect the terminal current, fault current and battery capacity of the battery in real time, the current sensor fault time calculation module is used to calculate the start time point and end time point of the current sensor fault according to the terminal current and the fault current, and the current sensor fault current calculation module is used to calculate the start time point and the end time point of the current sensor fault according to the current sensor fault The start time point and end time point, terminal voltage and battery capacity are used to calculate the average fault current of the current sensor fault. The voltage sensor fault time calculation module is used to obtain the start time point and end time point of the total voltage sensor fault or the battery voltage sensor fault according to the second fault value; the voltage sensor fault voltage calculation module is used to calculate the average fault voltage of the total voltage sensor fault or the battery voltage sensor fault according to the terminal voltage and the total voltage. The micro-short circuit fault time calculation module is used to obtain the occurrence time of the micro-short circuit fault according to the second fault value. The micro-short circuit current calculation module is used to calculate the micro-short circuit resistance according to the measured current of the current sensor.

[0013] The lithium battery system multi-fault diagnosis device provided by the present invention may also have the following features: wherein, the process in which the current sensor fault time calculation module calculates the end time point of the current sensor fault includes the following steps: step S1, taking the time step of the current sensor fault as the starting point, and calculating the corresponding SOC change rate every first interval time according to the starting point; step S2, judging whether the difference between the SOC change rate corresponding to the current first interval time and the SOC change rate corresponding to the previous first interval time is greater than a first preset value, and if so, taking the current first interval time as the inflection point area; step S3, judging whether the working mode of the battery is the discharge mode and whether the working condition has changed, and if so, Then calculate the SOC change rate corresponding to the next first interval time, and execute step S2, if not, execute step S4; step S4, calculate the corresponding SOC change rate every second interval time in the inflection point area; step S5, determine whether the difference between the SOC change rate corresponding to the current second interval time and the SOC change rate corresponding to the previous second interval time is greater than the second preset value, if so, execute step S6; step S6, determine whether the working mode of the battery is the discharge mode and whether the working condition has changed, if so, calculate the SOC change rate corresponding to the next second interval time, and execute step S5, if not, take the current second interval time as the end time point, and the calculation expression of the SOC change rate is: In the formula is the SOC change rate of the i-th battery, I r is the terminal current of the ith battery, If is the fault current of the ith battery, C i,cap is the nominal capacity of the ith battery, and the calculation expression of the average fault current is: Where ΔC h In discharge mode, the voltage V u Change to voltage V l The capacity changes or charging mode is normally from the voltage V l Change to voltage V u Capacity change, ΔC m In the discharge mode, the current sensor fails and the voltage V u Change to voltage V l The capacity changes or the current sensor fails in the charging mode when the voltage V l Change to voltage V u Capacity change, T end is the end time point of the current sensor fault, The voltage V is reached in discharge mode or charge mode l The corresponding time, T start is the starting time point of the current sensor failure, The voltage V is reached in discharge mode or charge mode u The corresponding time.

[0014] The lithium battery system multi-fault diagnosis device provided by the present invention may also have the following features: wherein, in the voltage sensor fault time calculation module, the time point when the total voltage sensor fault is detected is taken as the starting point, and the search is performed backward from the starting point and the first The time step is taken as the end time point of the total voltage sensor failure, and the first The time step of is taken as the starting time point of the total voltage sensor failure, and the time point when the voltage sensor failure of the i-th battery is detected is taken as the starting point, and the search is performed backward from the starting point and the first searched The time step is taken as the end time point of the voltage sensor failure of the i-th battery, and the search is forward from the starting point and the first searched The time step is taken as the starting time point of the voltage sensor failure of the i-th battery. In the voltage sensor failure voltage calculation module, the average fault voltage of the total voltage sensor failure is The expression is: Where n is the total number of voltage sensors, p is the total number of normal voltage sensors, is the measured voltage of the voltage sensor corresponding to the pth battery at time step k, is the measured voltage of the total voltage sensor at time step k, b is the time window length, and the average fault voltage of the voltage sensor corresponding to the i-th battery The expression is: In the formula is the measured voltage of the voltage sensor corresponding to the i-th battery at time step k.

[0016] The multi-fault diagnosis device for lithium battery system provided by the present invention may also have the following features: wherein, in the micro-short circuit fault time calculation module, the time point when the micro-short circuit fault is detected is taken as the starting point, and the search is carried out from the starting point forward and the first searched The time step is taken as the occurrence time of the total voltage sensor failure. In the micro-short-circuit current calculation module, the micro-short-circuit resistance corresponding to the i-th battery The calculation expression is: Where t1 is the selected time step, t0 is the time step when the micro short circuit fault is first detected, C i,cap is ΔC in the time interval [t0, t1] obtained by using the unscented Kalman filter r Estimated value, ΔSOC i,e is the estimated value of ΔSOC in the time interval [t0, t1] obtained by using the unscented Kalman filter, I m (t) is the measured current of the current sensor, U SOC＝20％ is the voltage of the ith battery when SOC = 20%, U SOC＝90％ is the voltage of the ith battery when SOC=90%.

[0017] In the lithium battery system multi-fault diagnosis device provided by the present invention, it can also have such a feature, and also includes: an early warning classification module, wherein the early warning classification module includes a single fault classification unit and a multiple fault classification unit, the single fault classification unit is used to calculate the corresponding early warning classification result according to the fault type, duration and amplitude, and the multiple fault classification unit is used to calculate the corresponding early warning classification result according to the sensor failure rate, battery failure rate and the cumulative occurrence time of multiple faults.

[0018] Functions and Effects of the Invention

[0019] According to the multi-fault diagnosis device for lithium battery system involved in the present invention, on the one hand, accurate fault judgment results can be generated through the voltage estimation module, fault value calculation module, battery envelope relationship generation module, first fault judgment module and second fault judgment module; on the other hand, the duration of each fault and the early warning classification result of the lithium battery system under the fault can be obtained through each fault calculation module and early warning classification module. Therefore, the multi-fault diagnosis device for lithium battery system of the present invention can accurately identify the faults of the lithium battery system and make early warning judgments. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a block diagram of a multi-fault diagnosis device for a lithium battery system according to an embodiment of the present invention;

[0021] Figure 2 is a schematic diagram of a lithium battery system fault diagnosis process in an embodiment of the present invention;

[0022] Figure 3 It is a schematic diagram of a flow chart of a current sensor fault time calculation module calculating an end time point of a current sensor fault in an embodiment of the present invention;

[0023] Figure 4 is a membership function of input variables and output variables in a single fault scenario in an embodiment of the present invention;

[0024] Figure 5 is a membership function of input variables and output variables in a multi-fault scenario in an embodiment of the present invention;

[0025] Figure 6 is a schematic diagram of an estimated voltage curve and an estimated error curve in an embodiment of the present invention;

[0026] Figure 7 is a schematic diagram of a second fault value in an embodiment of the present invention;

[0027] Figure 8 is a schematic diagram of search results of a current sensor fault duration in an embodiment of the present invention;

[0028] Fig. 9 is a schematic diagram of average fault voltage of a voltage sensor and a total voltage sensor fault in an embodiment of the present invention;

[0029] Fig.10 Schematic diagram of a micro short circuit fault in an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the following embodiments and accompanying drawings specifically illustrate the multi-fault diagnosis device for a lithium battery system of the present invention.

[0031] This embodiment provides a lithium battery system multi-fault diagnosis device for performing fault diagnosis on a battery pack of a lithium battery system.

[0032] Figure 1 4 is a block diagram of a multi-fault diagnosis device for a lithium battery system in an embodiment of the present invention.

[0033] like Figure 1 As shown, the lithium battery system multi-fault diagnosis device 100 includes a data acquisition module 11, a voltage estimation module 12, a fault value calculation module 13, a battery envelope relationship generation module 14, a first fault judgment module 15, a second fault judgment module 16, a current sensor fault time calculation module 17, a current sensor fault current calculation module 18, a voltage sensor fault time calculation module 19, a voltage sensor fault voltage calculation module 20, a micro-short circuit fault time calculation module 21, a micro-short circuit current calculation module 22, an early warning classification module 23 and a control module 24 that controls the above modules.

[0034] The data acquisition module 11 is used to collect the total voltage of the battery pack, as well as the terminal voltage, SOC value, measurement voltage, terminal current, fault current and battery capacity of each battery in the battery pack in real time.

[0035] The voltage estimation module 12 is used to obtain the estimated voltage of each battery according to the total voltage, the terminal voltage and the SOC value.

[0036] The voltage estimation module 12 includes a long short-term memory network, which generates a corresponding output vector according to an input vector. The input vector is composed of a total voltage, a terminal voltage of the battery, and a SOC value. The output vector is an estimated voltage of the battery.

[0037] The input vector corresponding to the i-th battery at the k-th time step The expression is:

[0038]

[0039] Where U T,k-α is the total voltage of the battery pack corresponding to the k-αth time step, U i,k-α is the terminal voltage of the ith battery at the k-αth time step, SOC i,k-α is the SOC value corresponding to the i-th battery at the k-αth time step,

[0040] The estimated voltage of the i-th battery at the k-th time step The expression is:

[0041]

[0042] Where W 1:nis the weight matrix of the LSTM network, and f(·) is the activation function of the LSTM network.

[0043] The fault value calculation module 13 is used to calculate a first fault value according to the estimated voltage and the measured voltage of each battery.

[0044] The fault value calculation module 13 calculates the first fault value corresponding to the battery in a cycle according to all estimated voltages and measured voltages in the cycle.

[0045] The first fault value d corresponding to the i-th battery E (U e ,U m ) is:

[0046]

[0047] Where n is the total number of batteries in the battery pack, b is the length of the time window, is the estimated voltage of the ith battery at the kth time step, is the measured voltage of the i-th battery at the k-th time step.

[0048] The battery envelope relationship generation module 14 is used to calculate the second fault value corresponding to each battery according to the measured voltage of each battery and the total voltage.

[0049] Among them, the second fault value of the i-th battery The calculation expression is:

[0050]

[0051] τ l =[(l-1)b+1,lb],

[0052] l=[(γ-1)h+1,γh],

[0053] Where h is 10, ξ is a positive constant, and U T is the total voltage of the battery pack at the kth time step, is the measured voltage of the ith battery at the kth time step, b is the length of the time window, and in this embodiment, b=100.

[0054] The first fault judgment module 15 stores a preset first fault threshold value, which is used to judge whether the first fault value is greater than the first fault threshold value. If so, there is a fault in the lithium battery system; if not, the lithium battery system operates normally.

[0055] The second fault judgment module 16 stores a preset second fault judgment condition, and is used to obtain a fault diagnosis result according to the second fault value and the second fault judgment condition. The fault diagnosis result includes a total voltage sensor fault, a battery voltage sensor fault, a battery current sensor fault, and a battery micro-short circuit fault.

[0056] The second fault judgment condition includes:

[0057] When , the current sensor of the i-th battery has a fault in the time period γ;

[0058] and When , there is a total voltage sensor fault in the time period γ;

[0059] and When , the voltage sensor of the i-th battery has a fault in the time period γ; and When , the i-th battery has a micro-short circuit fault in the time period γ.

[0060] The following describes the process of performing lithium battery system fault diagnosis using the lithium battery system multi-fault diagnosis device 100 in conjunction with the accompanying drawings.

[0061] Figure 2 It is a schematic diagram of the flow of lithium battery system fault diagnosis in an embodiment of the present invention.

[0062] like Figure 2 As shown, the lithium battery system multi-fault diagnosis device 100 is used to perform lithium battery system fault diagnosis, including the following steps:

[0063] Step T1, using the data acquisition module 11 to collect the total voltage of the battery pack, as well as the terminal voltage, SOC value and measured voltage of each battery in the battery pack in real time.

[0064] Step T2, using the voltage estimation module 12 to obtain the estimated voltage of each battery according to the total voltage, the terminal voltage and the SOC value.

[0065] Step T3: The fault value calculation module 13 is used to calculate a first fault value according to the estimated voltage and the measured voltage of each battery.

[0066] Step T4: The battery envelope relationship generation module 14 is used to calculate the second fault value corresponding to each battery according to the measured voltage of each battery and the total voltage.

[0067] Step T5, using the first fault judgment module 15 to judge whether the first fault value is greater than the first fault threshold, if so, the lithium battery system has a fault, if not, the lithium battery system operates normally.

[0068] Step T6, using the second fault judgment module 16 to obtain a fault diagnosis result according to the second fault value and the second fault judgment condition.

[0069] In this embodiment, when the first fault judgment module 15 judges that the lithium battery system has a fault, the fault diagnosis result is further obtained through the second fault judgment module 16. Therefore, in other embodiments, according to actual needs, the fault judgment and diagnosis can be performed in the order of step T1, step T2, step T3, step T5, step T4, and step T6, that is, only when the first fault judgment module 15 judges that the lithium battery system has a fault, the second fault value required for calculation by the second fault judgment module 16 is generated by the battery envelope relationship generation module 14, thereby reducing the amount of model calculation.

[0070] The current sensor fault time calculation module 17 is used to calculate the start time point and the end time point of the current sensor fault according to the terminal current and the fault current.

[0071] Figure 3 It is a schematic diagram of a flow chart of a current sensor fault time calculation module in an embodiment of the present invention calculating an end time point of a current sensor fault.

[0072] like Figure 3 As shown, the process of the current sensor fault time calculation module 17 calculating the end time point of the current sensor fault includes the following steps:

[0073] Step S1, taking the time step of the current sensor failure as the starting point, and calculating the corresponding SOC change rate at every first interval time according to the starting point.

[0074] Among them, the calculation expression of SOC change rate is:

[0075]

[0076] In the formula is the SOC change rate of the i-th battery, I r is the terminal current of the ith battery, I f is the fault current of the ith battery, C i,cap is the nominal capacity of the ith battery.

[0077] Step S2, determine whether the difference between the SOC change rate corresponding to the current first interval time and the SOC change rate corresponding to the previous first interval time is greater than a first preset value. If so, take the current first interval time as the inflection point area and execute step S3. If not, take the next first interval time as the current first interval time and execute step S2.

[0078] Step S3, determining whether the working mode of the battery is the discharge mode and whether the working condition has changed, if so, calculating the SOC change rate corresponding to the next first interval time and executing step S2, if not, executing step S4.

[0079] Step S4, calculating the corresponding SOC change rate at every second interval time in the inflection point region.

[0080] Step S5, determine whether the difference between the SOC change rate corresponding to the current second interval time and the SOC change rate corresponding to the previous second interval time is greater than a second preset value, if so, execute step S6, if not, take the next second interval time as the current second interval time, and execute step S5.

[0081] Step S6, determining whether the working mode of the battery is the discharge mode and whether the working condition has changed, if so, calculating the SOC change rate corresponding to the next second interval time, and executing step S5, if not, taking the current second interval time as the end time point.

[0082] In this embodiment, the calculation method of the start time point of the current sensor fault is the same as the calculation method of the end time point of the current sensor fault. According to the above process, the search is adjusted to proceed forward from the starting point to obtain the corresponding start time point.

[0083] The current sensor fault current calculation module 18 is used to calculate the average fault current of the current sensor fault according to the start time point and end time point of the current sensor fault, the terminal voltage and the battery capacity.

[0084] Among them, the calculation expression of the average fault current is:

[0085]

[0086] Where ΔC h In discharge mode, the voltage V u Change to voltage V l The capacity changes or charging mode is normally from the voltage V l Change to voltage V u Capacity change, ΔC m In the discharge mode, the current sensor fails and the voltage V u Change to voltage V l The capacity changes or the current sensor fails in the charging mode when the voltage V l Change to voltage V u Capacity change, T end is the end time point of the current sensor fault, The voltage V is reached in discharge mode or charge mode l The corresponding time, T startis the starting time point of the current sensor failure, The voltage V is reached in discharge mode or charge mode u The corresponding time.

[0087] The voltage sensor fault time calculation module 19 is used to obtain the start time point and the end time point of the total voltage sensor fault or the battery voltage sensor fault according to the second fault value.

[0088] When a total voltage sensor fault is detected, the voltage sensor fault time calculation module 19 takes the time point when the total voltage sensor fault is detected as the starting point, searches backward from the starting point and calculates the first The time step is taken as the end time point of the total voltage sensor failure, and the first The time step is taken as the starting time point of the total voltage sensor failure.

[0089] When a voltage sensor failure is detected, the voltage sensor failure time calculation module 19 takes the time point when the voltage sensor failure of the i-th battery is detected as the starting point, searches backward from the starting point and calculates the first The time step is taken as the end time point of the voltage sensor failure of the i-th battery, and the search is forward from the starting point and the first searched The time step is taken as the starting time point of the voltage sensor failure of the i-th battery.

[0090] The voltage sensor fault voltage calculation module 20 is used to calculate the average fault voltage of the total voltage sensor fault or the battery voltage sensor fault according to the terminal voltage and the total voltage.

[0091] Among them, the average fault voltage of the total voltage sensor fault The expression is:

[0092]

[0093] Where n is the total number of voltage sensors, p is the total number of normal voltage sensors, is the measured voltage of the voltage sensor corresponding to the pth battery at time step k, is the measured voltage of the total voltage sensor at time step k, and b is the length of the time window.

[0094] The average fault voltage of the voltage sensor corresponding to the i-th battery The expression is:

[0095]

[0096] In the formula is the measured voltage of the voltage sensor corresponding to the i-th battery at time step k.

[0097] The micro-short circuit fault time calculation module 21 is used to obtain the occurrence time of the micro-short circuit fault according to the second fault value. Since the micro-short circuit fault of the battery is irreversible, it is only necessary to determine the occurrence time of the fault.

[0098] Among them, in the micro-short circuit fault time calculation module 21, the time point when the micro-short circuit fault is detected is taken as the starting point, and the search is carried out from the starting point forward and the first searched The time step is taken as the occurrence time of the total voltage sensor failure.

[0099] The micro short circuit current calculation module 22 is used to calculate the micro short circuit resistance according to the measured current of the current sensor.

[0100] Among them, in the micro short-circuit current calculation module 22, the micro short-circuit resistance corresponding to the i-th battery is The calculation expression is:

[0101]

[0102] Where t1 is the selected time step, t0 is the time step when the micro short circuit fault is first detected, C i,cap is ΔCr in the time interval [t0, t1] obtained by using the unscented Kalman filter 的 Estimated value, ΔSOC i,e is the estimated value of ΔSOC in the time interval [t0, t1] obtained by using the unscented Kalman filter, I m (t) is the measured current of the current sensor, U SOC＝20％ is the voltage of the ith battery when SOC = 20%, U SOC＝90％ is the voltage of the ith battery when SOC=90%. SOC＝20％ is 3.508V, U SOC＝90％ is 4.102V.

[0103] The early warning classification module 23 includes a single fault classification unit 231 and a multiple fault classification unit 232 .

[0104] The single fault classification unit 231 is used to calculate the corresponding warning classification result according to the fault type, duration and amplitude. In this embodiment, the amplitude is the amplitude of the corresponding fault, and the current sensor fault corresponds to the average fault current. The total voltage sensor failure corresponds to the average fault voltage The voltage sensor failure corresponds to the average fault voltage The micro short circuit fault corresponds to the average micro short circuit current

[0105] The single fault classification unit 231 is used for single fault scenarios. First, the magnitude and duration of the fault are converted into a fuzzy set with a domain of (0,1) using quantization factors α and β. The expression is:

[0106]

[0107] In the formula and |A| VF They represent the upper and lower limits of the fault amplitude when the voltage sensor or the total voltage sensor fails, and |A| CF They represent the upper and lower limits of the fault amplitude when the current sensor fails, and |A| MF They represent the upper and lower limits of the fault amplitude in case of micro short circuit fault, and T They represent the upper and lower bounds of the selected data analysis area respectively.

[0108] Secondly, the single fault classification unit 231 defines the fault amplitude and duration on the fuzzy set, and its expression is:

[0109] FA VF =α VF (A VF - |A| VF ),

[0110] FT VF =β T T VF ,

[0111] FA CF =α CF (A CF - |A| CF ),

[0112] FT CF =β T T CF ,

[0113] FA MF =α MF (A MF - |A| MF ),

[0114] FT MF =β T T MF ,

[0115] Where A VF , ACF and A MF are the fault amplitudes when the voltage sensor or total voltage sensor fails, the current sensor fails, and the micro short circuit fails, respectively. VF 、T CF and T MF They are the duration of voltage sensor or total voltage sensor fault, current sensor fault and micro short circuit fault respectively.

[0116] Finally, the single fault classification unit 231 sets the input variables to FA and FT, and the output variable is the warning level, i.e., the warning classification result. VF and FA CF It is divided into four fuzzy subsets: tiny (MS), small (S), medium (M) and large (L), FA MF , FT VF , FT CF and FT MF It is divided into three fuzzy subsets: S, M and L; the warning level is divided into four fuzzy subsets: L1, L2, L3 and L4. For current sensor fault, voltage sensor or total voltage sensor fault and battery micro short circuit fault, the fuzzy logic has 16, 16 and 9 fuzzy rules respectively.

[0117] The multiple fault grading unit 232 is used to calculate the corresponding warning grading result according to the sensor failure rate, the battery failure rate and the cumulative occurrence time of multiple faults.

[0118] The multi-fault classification unit 232 is used for multi-fault scenarios, and the input variable is the sensor failure rate FR S , Battery failure rate FR M , and the cumulative occurrence time of multiple faults FT S and FT M , whose expression is:

[0119]

[0120]

[0121] Where N1 is the number of sensors in the lithium battery system, N2 is the number of batteries in the lithium battery system, and N1 FS is the number of faulty sensors when the sensor fails, T FS is the fault duration of the sensor failure, N2 FM is the number of faulty batteries during micro short circuit failure, T FM It is the fault duration of micro short circuit fault.

[0122] The output variable obtained by the multi-fault classification unit 232 according to the input variables is the warning level, that is, the warning classification result. All four input variables are divided into three fuzzy subsets: S, M and L. The output variable is also divided into three fuzzy subsets: L3, L4 and L5. The fuzzy logic has a total of 81 fuzzy rules.

[0123] The control module 24 stores a control program for controlling the operation of each module.

[0124] The following is a further description of the multi-fault diagnosis device for a lithium battery system in conjunction with specific scenarios.

[0125] In this embodiment, the battery pack of the lithium battery system is composed of 5 single ternary lithium batteries, namely batteries 1 to 5. The initialization parameters are: the total number of time steps in the cycle b = 100, the first fault threshold is 0.35, ξ = 0.01, and the voltage range [V u ,V l ]=[4.0V,3.5V], the first interval time is 15%, and the second interval time is 2.5%.

[0126] Figure 4 It is the membership function of the input variable and the output variable in the single fault scenario in the embodiment of the present invention.

[0127] Figure 5 It is the membership function of the input variables and the output variables in the multi-fault scenario in the embodiment of the present invention.

[0128] In this embodiment, different faults and corresponding amplitudes and trigger times are set in the lithium battery system, as shown in the following table:

[0129]

[0130] In the above table, the first column is the fault type, the second column is the amplitude corresponding to each fault, and the third column is the trigger time, i.e., the duration, of each fault.

[0131] Figure 6 is a schematic diagram of an estimated voltage curve and an estimated error curve in an embodiment of the present invention.

[0132] like Figure 6 As shown, (a) is a schematic diagram of the curve of the estimated voltage of each battery generated by the voltage estimation module 12, the horizontal axis is the period T, the unit is hour h, and the vertical axis is the voltage estimation value, the unit is V, (b) is a schematic diagram of the curve of the estimation error of each battery, the horizontal axis is the period T, the unit is hour h, and the vertical axis is the estimation error, the unit is V.

[0133] Figure 7 is a schematic diagram of a second fault value in an embodiment of the present invention.

[0134] like Figure 7 As shown, from top to bottom are the Euclidean distance and the Γ generated by the battery envelope relationship generation module 14. 1,i , Γ 2,i , Γ 3,i , ΔΓ 1,i , ΔΓ 2,i and ΔΓ 3,i The corresponding curve diagrams, the horizontal axis is the period T, the vertical axis is the corresponding value, and each figure contains the data curve corresponding to each battery.

[0135] according to Figure 6 and Figure 7 , through the voltage estimation module 12 and the battery envelope relationship generation module 14, the corresponding fault type can be accurately identified after the fault is triggered.

[0136] Figure 8 It is a schematic diagram of search results of current sensor fault duration in an embodiment of the present invention.

[0137] like Figure 8 As shown, (a) is a schematic diagram of the current sensor fault time calculation module 17 performing forward search and backward search according to the first interval time, the horizontal axis is the period, and the total coordinate is the SOC change rate. (b) is a schematic diagram of the start time point of the search using the second interval time in the forward search, the horizontal axis is the period, and the total coordinate is the SOC change rate. (c) is a schematic diagram of the end time point of the search using the second interval time in the backward search, the horizontal axis is the period, and the total coordinate is the SOC change rate.

[0138] According to the current sensor fault time calculation module 17, the voltage sensor fault time calculation module 19 and the micro short circuit fault time calculation module 21, the duration of each fault is calculated as shown in the following table:

[0139]

[0140]

[0141] In the above table, the first column is each fault, and the second and third columns are the start time, i.e., the start time point, and the end time, i.e., the end time point, corresponding to each fault.

[0142] The average fault current calculated by the current sensor fault current calculation module 18 is shown in the following table:

[0143] cycle Battery 3 Battery 4 Battery 5 Cycle 2 19.9193A 19.9476A 19.9438A Cycle 3 19.7992A 19.8721A 19.8021A Cycle 5 19.5855A 19.6095A 19.5721A average value 19.7680A 19.8097A 19.7727A

[0144] The first column in the above table is for each cycle, and the second to fourth columns are the average fault currents of battery 3, battery 4 and battery 5 in the corresponding cycles respectively.

[0145] Fig. 9 Schematic diagram of average fault voltage of voltage sensor and total voltage sensor failure in an embodiment of the present invention.

[0146] like Fig. 9 As shown, (a) is a schematic diagram of the average fault voltage of the total voltage sensor fault generated by the voltage sensor fault voltage calculation module 20, the horizontal axis is the period T, and the vertical axis is the average fault voltage. (b) is a schematic diagram of the average fault voltage of the voltage sensor fault generated by the voltage sensor fault voltage calculation module 20, the horizontal axis is the period T, and the vertical axis is the average fault voltage.

[0147] Fig.10 Schematic diagram of a micro short circuit fault in an embodiment of the present invention.

[0148] like Fig.10 As shown, the micro-short-circuit current calculation module 22 calculates: (a) is a schematic diagram of the SOC and SOC estimation of battery 1, with the horizontal axis being the period T and the vertical axis being the SOC value. (b) is the ΔC of battery 1 e , ΔC Ah and ΔC r Schematic diagram, the horizontal axis is the charge and discharge cycle, the vertical axis is ΔC, the unit is Ah. (c) is the corresponding and Schematic diagram of the battery, with the horizontal axis representing the charge and discharge cycle and the vertical axis representing the corresponding data.

[0149] Finally, the warning classification module 23 calculates the warning levels of different fault types in different scenarios based on the above data, as shown in the following table:

[0150]

[0151]

[0152] The first column in the above table is the various scenarios, the second column is the various fault types, and the third column is the warning level corresponding to each fault type.

[0153] Functions and Effects of the Embodiments

[0154] According to the multi-fault diagnosis device for lithium battery system involved in this embodiment, on the one hand, accurate fault judgment results can be generated through the voltage estimation module, fault value calculation module, battery envelope relationship generation module, first fault judgment module and second fault judgment module; on the other hand, the duration of each fault and the early warning classification result of the lithium battery system under the fault can be obtained through each fault calculation module and early warning classification module. In short, this method can accurately identify the fault of the lithium battery system and make early warning judgments.

[0155] Those skilled in the art should understand that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A multi-fault diagnosis device for a lithium battery system, used for diagnosing faults of a battery pack of a lithium battery system, characterized in that: include: A data acquisition module, used for collecting the total voltage of the battery pack, and the terminal voltage, SOC value and measured voltage of each battery in the battery pack in real time; A voltage estimation module, used for obtaining an estimated voltage of each of the batteries according to the total voltage, the terminal voltage and the SOC value; a fault value calculation module, configured to calculate a first fault value according to the estimated voltage and the measured voltage of each of the batteries; A battery envelope relationship generating module, configured to calculate a second fault value corresponding to each of the batteries according to the measured voltage of each of the batteries and the total voltage; A first fault judgment module stores a preset first fault threshold, and is used to judge whether the first fault value is greater than the first fault threshold. If so, the lithium battery system has a fault; if not, the lithium battery system operates normally; a second fault judgment module storing a preset second fault judgment condition, and used to obtain a fault diagnosis result according to the second fault value and the second fault judgment condition; The fault diagnosis results include a total voltage sensor fault, a battery voltage sensor fault, a battery current sensor fault, and a battery micro-short circuit fault.

2. The multi-fault diagnosis device for lithium battery system according to claim 1, characterized in that: in, The voltage estimation module includes a long short-term memory network, The long short-term memory network generates a corresponding output vector according to the input vector, The input vector is composed of the total voltage and the terminal voltage and the SOC value of the battery, The output vector is the estimated voltage of the battery, The input vector corresponding to the i-th battery at the k-th time step The expression is: Where U T，k-α is the total voltage of the battery pack corresponding to the k-αth time step, U i，k-α is the terminal voltage of the ith battery at the k-αth time step, SOC i，k-α is the SOC value corresponding to the i-th battery at the k-αth time step, The estimated voltage of the i-th battery at the k-th time step The expression is: Where W1:n is the weight matrix of the LSTM network, and f(·) is the activation function of the LSTM network.

3. The multi-fault diagnosis device for lithium battery system according to claim 1, characterized in that: in, The fault value calculation module calculates a first fault value corresponding to the battery in a cycle according to all the estimated voltages and the measured voltages in the cycle, The first fault value d corresponding to the i-th battery E (U e , U m ) is: Where n is the total number of batteries in the battery pack, b is the length of the time window, is the estimated voltage of the ith battery at the kth time step, is the measured voltage of the ith battery at the kth time step.

4. The multi-fault diagnosis device for lithium battery system according to claim 1, characterized in that: in, The second fault value of the i-th battery The calculation expression is: τ l =[(l-1)b+1,lb], l=[(γ-1)h+1,γh], Where h is the time window length, ξ is a positive constant, and U T is the total voltage of the battery pack at the kth time step, is the measured voltage of the ith battery at the kth time step, and b is the length of the time window.

5. The multi-fault diagnosis device for lithium battery system according to claim 4, characterized in that: in, The second fault judgment condition includes: When , the current sensor of the i-th battery has a fault in the time period γ; and When , there is a total voltage sensor fault in the time period γ; and When , the voltage sensor of the i-th battery has a fault in the time period γ; and When , the i-th battery has a micro-short circuit fault in the time period γ.

6. The multi-fault diagnosis device for lithium battery system according to claim 1, characterized in that: Also includes: Current sensor fault time calculation module, current sensor fault current calculation module, voltage sensor fault time calculation module, voltage sensor fault voltage calculation module, micro short circuit fault time calculation module and micro short circuit current calculation module, The data acquisition module is also used to collect the terminal current, fault current and battery capacity of the battery in real time. The current sensor fault time calculation module is used to calculate the start time point and the end time point of the current sensor fault according to the terminal current and the fault current. The current sensor fault current calculation module is used to calculate the average fault current of the current sensor fault according to the starting time point and the ending time point of the current sensor fault, the terminal voltage and the battery capacity, The voltage sensor fault time calculation module is used to obtain the start time point and the end time point of the total voltage sensor fault or the battery voltage sensor fault according to the second fault value, The voltage sensor fault voltage calculation module is used to calculate the average fault voltage of the total voltage sensor fault or the battery voltage sensor fault according to the terminal voltage and the total voltage. The micro short circuit fault time calculation module is used to obtain the occurrence time of the micro short circuit fault according to the second fault value, The micro short circuit current calculation module is used to calculate the micro short circuit resistance according to the measured current of the current sensor.

7. The multi-fault diagnosis device for a lithium battery system according to claim 6, Features: The process of calculating the end time point of the current sensor fault by the current sensor fault time calculation module includes the following steps: Step S1, taking the time step of the current sensor failure as a starting point, and calculating the corresponding SOC change rate at every first interval time according to the starting point; Step S2, determining whether the difference between the SOC change rate corresponding to the current first interval time and the SOC change rate corresponding to the previous first interval time is greater than a first preset value, and if so, taking the current first interval time as an inflection point area; Step S3, determining whether the working mode of the battery is the discharge mode and whether the working condition has changed, if so, calculating the SOC change rate corresponding to the next first interval time and executing step S2, if not, executing step S4; Step S4, calculating the corresponding SOC change rate at every second interval in the inflection point area; Step S5, determining whether the difference between the SOC change rate corresponding to the current second interval time and the SOC change rate corresponding to the previous second interval time is greater than a second preset value, if so, executing step S6; Step S6, determining whether the working mode of the battery is the discharge mode and whether the working condition has changed, if so, calculating the SOC change rate corresponding to the next second interval time, and executing step S5, if not, taking the current second interval time as the end time point, The calculation expression of the SOC change rate is: In the formula is the SOC change rate of the i-th battery, I r is the terminal current of the ith battery, I f is the fault current of the ith battery, C i,cap is the nominal capacity of the ith battery, The calculation expression of the average fault current is: Where ΔC h In discharge mode, the voltage V u Change to voltage V l The capacity changes or charging mode is normally from the voltage V l Change to voltage V u Capacity change, ΔC m In the discharge mode, the current sensor fails and the voltage V u Change to voltage V l The capacity changes or the current sensor fails in the charging mode when the voltage V l Change to voltage V u Capacity change, T end is the end time point of the current sensor fault, The voltage V is reached in discharge mode or charge mode l The corresponding time, T start is the starting time point of the current sensor failure, The voltage V is reached in discharge mode or charge mode u The corresponding time.

8. The multi-fault diagnosis device for a lithium battery system according to claim 6, characterized in that: in, In the voltage sensor failure time calculation module, the time point when the total voltage sensor failure is detected is taken as the starting point, and the first The time step is taken as the end time point of the total voltage sensor fault, and the first searched The time step is taken as the starting time point of the total voltage sensor failure, The time point when the voltage sensor failure of the i-th battery is detected is taken as the starting point, and the first The time step is taken as the end time point of the voltage sensor failure of the i-th battery, and the search is forward from the starting point and the first searched The time step is taken as the starting time point of the voltage sensor failure of the i-th battery, In the voltage sensor fault voltage calculation module, the average fault voltage of the total voltage sensor fault is The expression is: Where n is the total number of voltage sensors, p is the total number of normal voltage sensors, is the measured voltage of the voltage sensor corresponding to the pth battery at time step k, is the measured voltage of the total voltage sensor at time step k, b is the length of the time window, The average fault voltage of the voltage sensor corresponding to the i-th battery The expression is: In the formula is the measured voltage of the voltage sensor corresponding to the i-th battery at time step k.

9. The multi-fault diagnosis device for lithium battery system according to claim 6, characterized in that: in, In the micro-short circuit fault time calculation module, the time point when the micro-short circuit fault is detected is taken as the starting point, and the first The time step is taken as the occurrence time of the total voltage sensor failure, In the micro short-circuit current calculation module, the micro short-circuit resistance corresponding to the i-th battery The calculation expression is: Where t1 is the selected time step, t0 is the time step when the micro short circuit fault is first detected, C i，cap is ΔC in the time interval [t0, t1] obtained by using the unscented Kalman filter r Estimated value, ΔSOC i，e is the estimated value of ΔSOC in the time interval [t0, t1] obtained by using the unscented Kalman filter, I m (t) is the measured current of the current sensor, U SOC＝20％ is the voltage of the ith battery when SOC = 20%, U SOC＝90％ is the voltage of the ith battery when SOC=90%.

10. The multi-fault diagnosis device for lithium battery system according to claim 6, characterized in that: Also includes: Early warning classification module, The warning classification module includes a single fault classification unit and a multiple fault classification unit. The single fault classification unit is used to calculate the corresponding warning classification result according to the fault type, duration and amplitude. The multiple fault grading unit is used to calculate the corresponding early warning grading result according to the sensor failure rate, the battery failure rate and the cumulative occurrence time of multiple faults.

Images