Battery voltage acquisition method and device based on Kalman filtering, electronic equipment and readable storage medium

Abstract

The invention relates to the technical field of battery voltage acquisition, in particular to a battery voltage acquisition method and device based on Kalman filtering, electronic equipment and a readable storage medium, and the method comprises the steps: obtaining measurement information at a plurality of historical acquisition moments and Kalman gains at the plurality of historical acquisition moments; and updating the process noise covariance and the observation noise covariance of the previous acquisition moment of the specified acquisition moment into the process noise covariance of the specified acquisition moment and the observation noise covariance of the specified acquisition moment according to the measurement information of the plurality of historical acquisition moments and the Kalman gains of the plurality of historical acquisition moments. According to the method, the process noise covariance and the observation noise covariance can be updated in real time, and the technical scheme of the invention has a strong tracking characteristic, so that the optimal voltage estimation value is closer to a true value, the filtering effect on periodic noise is further improved, and the accuracy of voltage acquisition is improved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of battery voltage collection, and in particular to a battery voltage collection method, device, electronic device and readable storage medium based on Kalman filtering. Background Art

[0002] Battery voltage is a key parameter for battery status management and a key indicator for measuring whether the battery has overcharged, over-discharged or other dangerous behaviors. Accurate battery voltage collection is of great significance for battery safety management and ensuring the normal operation of the battery system. However, the battery voltage collection process will be affected by the accuracy of the sampling chip, the interference of various modules in the circuit, the external environment, etc., which will reduce the accuracy of battery voltage collection.

[0003] Filtering is a technical means to reduce noise. The filtering methods for battery voltage acquisition mainly include hardware filtering and software filtering. Hardware filtering is to shield noise by adding additional filtering circuits, which has poor effect in suppressing system noise and increases hardware costs. Software filtering mainly includes average method and Kalman filtering method. The average method performs filtering by taking the average value of certain historical sampling points. The calculation is simple but the effect is poor. The Kalman filtering method mainly estimates and corrects the state at the current moment based on the estimated value at the previous acquisition moment and the measured value at the current moment. In the Kalman filtering process, the statistical characteristics of the system process noise and the system observation noise are not easy to determine, and they are also affected by changes in the system state and environment, which has a great impact on the final filtering effect. In the existing battery voltage sampling and filtering process based on the Kalman filtering algorithm, the covariance of the system process noise and the system observation noise is set to a fixed value based on experience, without considering the changes of the system noise with the system state and the external environment, and without considering the influence of historical data, resulting in poor voltage acquisition and tracking effect in actual use, and unsatisfactory filtering effect for periodic noise. Summary of the invention

[0004] In order to solve the problems in the related art, the embodiments of the present disclosure provide a battery voltage acquisition method, device, electronic device and readable storage medium based on Kalman filtering.

[0005] In a first aspect, an embodiment of the present disclosure provides a battery voltage acquisition method based on Kalman filtering, comprising:

[0006] Obtain the process noise covariance and observation noise covariance of the battery to be tested at the previous acquisition time of the specified acquisition time;

[0007] Establishing the basic state equation and basic observation equation of the battery to be tested;

[0008] Establishing a voltage a priori estimation equation, a voltage observation estimation equation, and a Kalman gain equation of the battery to be tested according to the basic state equation, the basic observation equation, the process noise covariance and the observation noise covariance at the previous acquisition moment of the specified acquisition moment;

[0009] Determine a voltage a priori estimate value, a voltage observation estimate value, and a Kalman gain at a specified acquisition time according to the voltage a priori estimate equation, the voltage observation estimate equation, and the Kalman gain equation;

[0010] Determine the optimal voltage estimate of the battery to be tested at the specified acquisition time according to the voltage a priori estimate at the specified acquisition time, the voltage observation estimate at the specified acquisition time, and the Kalman gain at the specified acquisition time;

[0011] Acquire measurement innovation information at multiple historical collection moments and Kalman gains at multiple historical collection moments, wherein the multiple historical collection moments include the designated collection moment and collection moments before the designated collection moment;

[0012] A forgetting factor and a fading factor are set, and a weighting coefficient of a collection moment before a specified collection moment is determined according to the forgetting factor; the fading factor is used to set the weights between the measurement innovations of the multiple historical collection moments; the calculation formula of the fading factor is as follows:

[0013] N i =2(m+1-i) / [m(m+1)], i=1,2,...,m;

[0014] Among them, N i is the i-th fading factor from the specified collection moment, and m is the number of moments included in the historical collection moment;

[0015] According to the weighting coefficient, the fading factor, the measurement innovations of the multiple historical collection moments, and the Kalman gains of the multiple historical collection moments, the process noise covariance and the observation noise covariance of the previous collection moment of the specified collection moment are updated to the process noise covariance of the specified collection moment and the observation noise covariance of the specified collection moment; the process noise covariance of the specified collection moment and the observation noise covariance of the specified collection moment are used to calculate the optimal voltage estimate at the next moment of the specified collection moment.

[0016] According to an embodiment of the present disclosure, the updating of the process noise covariance at the specified acquisition moment and the observation noise covariance at the specified acquisition moment by the process noise covariance and the observation noise covariance at the previous acquisition moment of the specified acquisition moment according to the weighting coefficient, the fading factor, the measurement innovations at the multiple historical acquisition moments, and the Kalman gains at the multiple historical acquisition moments includes:

[0017] Update the process noise covariance of the designated acquisition moment according to the process noise covariance of the previous acquisition moment, the weighting coefficient, the fading factor, the measurement innovations of the multiple historical acquisition moments, the Kalman gains of the multiple historical acquisition moments, and the product of the weighting coefficient and the Kalman gains of the multiple historical acquisition moments;

[0018] The observed noise covariance at the specified acquisition moment is updated according to the observed noise covariance at the previous acquisition moment, the weighting coefficient, the fading factor, the measurement innovations at the multiple historical acquisition moments, and the product of the weighting coefficient and the measurement innovations at the multiple historical acquisition moments.

[0019] According to an embodiment of the present disclosure, the weighting coefficient is calculated using the following formula:

[0020] d k-1 =(1-b) / (1-b k );

[0021] The process noise covariance is updated using the following formula:

[0022] Q k =(1-d k-1 )Q k-1 +d k-1 [Kg k ,Kg k-1 ,...,Kg k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T {[Kg k ,Kg k-1 ,...,Kg k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T} T ;

[0023] Among them, k is the specified collection time, k-1 is the collection time before the specified collection time, b is the forgetting factor, Q k is the process noise covariance at time k, Q k-1 is the process noise covariance at time k-1, d k-1 is the weighting coefficient at time k-1, Kg k ,Kg k-1 ,...,Kg k-m+1 is the Kalman gain corresponding to m historical acquisition moments, ek is the measurement information at time k, Kg k is the Kalman gain at time k, e k-m+1 is the measurement innovation at time k-m+1, T is the transposition operation, N1e k ,N2e k-1 ,...,N m e k-m+1 It is the product of the measurement innovation corresponding to m historical acquisition moments and the corresponding fading factor;

[0024] The observation noise covariance at time k is updated using the following formula:

[0025] R k =(1-d k-1 )R k-1 +d k-1 [N1e k ,N2e k-1 ,...,N m e k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T ;

[0026] Among them, R k is the observation noise covariance at time k, R k-1 is the observation noise covariance at time k-1.

[0027] According to an embodiment of the present disclosure, the voltage a priori estimation equation, voltage observation estimation equation, and Kalman gain equation of the battery to be tested are established according to the basic state equation, the basic observation equation, the process noise covariance and the observation noise covariance at the previous acquisition time of the specified acquisition time, including:

[0028] According to the basic state equation and the basic observation equation, respectively establishing a voltage a priori estimation equation and a voltage observation estimation equation of the battery to be tested;

[0029] Establishing an error covariance calculation equation according to the basic state equation and the process noise covariance at a previous acquisition moment of the specified acquisition moment;

[0030] The Kalman gain equation is established according to the error covariance calculation equation and the observation noise covariance at a collection time before the designated collection time.

[0031] According to an embodiment of the present disclosure, the basic state equation calculates the voltage state value at the specified acquisition time based on the voltage state value at the previous acquisition time, the system state parameter, and the system process noise at the specified acquisition time;

[0032] The basic observation equation calculates the voltage observation value at the specified acquisition time according to the voltage observation value at the specified acquisition time, the system observation parameter, and the system observation noise at the specified acquisition time;

[0033] The voltage a priori estimation equation calculates the a priori voltage estimation value at the specified acquisition time according to the voltage state value at the previous acquisition time and the system state parameters;

[0034] The voltage observation estimation equation calculates the voltage observation estimation value at the specified collection time according to the a priori voltage estimation value at the specified collection time, the system observation noise, and the voltage collection value at the specified collection time;

[0035] The error covariance calculation equation calculates the optimal estimated covariance at the specified acquisition time according to the covariance at the previous acquisition time, the system state parameter, and the process noise covariance at the previous acquisition time.

[0036] The Kalman gain equation calculates the Kalman gain at the specified acquisition time according to the optimal estimated covariance at the specified acquisition time, the system observation parameters, and the observation noise covariance at the previous acquisition time of the specified acquisition time.

[0037] According to an embodiment of the present disclosure, determining the optimal voltage estimate of the battery to be tested at the specified acquisition time according to the voltage a priori estimate at the specified acquisition time, the voltage observation estimate at the specified acquisition time, and the Kalman gain at the specified acquisition time includes:

[0038] Obtaining a voltage collection value of the battery to be tested at a specified collection time;

[0039] Calculate the measurement information at the specified collection time according to the voltage collection value at the specified collection time and the voltage observation estimate value at the specified collection time;

[0040] The optimal voltage estimate at the specified acquisition time is calculated according to the a priori voltage estimate at the specified acquisition time, the Kalman gain at the specified acquisition time, and the measurement innovation at the specified acquisition time.

[0041] According to an embodiment of the present disclosure, the basic state equation is shown in the following formula:

[0042] X k =AX k-1 +ω k ;

[0043] Among them, k-1 is the previous acquisition time of the specified acquisition time, X k is the voltage state value at time k, X k-1 is the voltage state value at time k-1, ωk is the process noise at time k, A is the system state parameter;

[0044] The basic observation equation is shown in the following formula:

[0045] Z k =CXk+υ k ;

[0046] Among them, Z k is the voltage observation value at time k, C is the system observation parameter, υ k is the observation noise at time k;

[0047] According to an embodiment of the present disclosure, the process noise and the observation noise are both Gaussian white noise;

[0048] The system state parameter A and the system observation parameter C are both single-element matrices with one element.

[0049] According to an embodiment of the present disclosure, the voltage a priori estimation equation is shown in the following formula:

[0050] X k '=AX k-1 ;

[0051] Among them, X k ' is the prior voltage estimate at time k;

[0052] The voltage observation estimation equation is shown in the following formula:

[0053] Z k '=CX k ';

[0054] Among them, Z k ' is the estimated value of voltage observation at time k;

[0055] The error covariance calculation equation is shown in the following formula:

[0056] P k '=AP k-1 A T +Q k-1

[0057] Among them, P k ' is the optimal estimated covariance at time k, P k-1 is the covariance at time k-1, Q k-1 is the process noise covariance at time k-1, T is the transposition operation;

[0058] The calculation method of calculating the measurement innovation at the specified collection time is as follows:

[0059] e k=Y k -Z k '=Y k -CX k ';

[0060] Among them, e k is the measurement information at time k, Y k is the voltage collection value at time k;

[0061] The Kalman gain equation is shown in the following formula:

[0062] Kg k =P k 'C T / (CP k 'C T +R k-1 );

[0063] Among them, Kg k is the Kalman gain at time k, R k-1 is the observation noise covariance at time k-1.

[0064] According to an embodiment of the present disclosure, the optimal voltage estimate at the specified acquisition time is calculated according to the a priori voltage estimate at the specified acquisition time, the Kalman gain at the specified acquisition time, and the measurement innovation at the specified acquisition time, and the calculation method is as follows:

[0065] X k + =X k '+Kg k e k ;

[0066] Among them, X k + is the optimal voltage estimate at time k.

[0067] According to an embodiment of the present disclosure, the method further includes:

[0068] The covariance at the specified acquisition time is updated according to the Kalman gain at the specified acquisition time, the optimal estimated covariance at the specified acquisition time, and the system observation parameters; the covariance at the specified acquisition time is used to calculate the optimal estimated error covariance at the next moment of the specified acquisition time.

[0069] According to an embodiment of the present disclosure, the covariance of the updated designated acquisition time is calculated as follows:

[0070] P k =(E-Kg k C)P k ';

[0071] Where E is the identity matrix.

[0072] In a second aspect, an embodiment of the present disclosure provides a battery voltage sampling device based on Kalman filtering, comprising:

[0073] A first acquisition module is configured to acquire the process noise covariance and the observation noise covariance of the battery to be tested at a previous acquisition time at a specified acquisition time;

[0074] A first establishing module is configured to establish a basic state equation and a basic observation equation of the battery to be tested;

[0075] A second establishing module is configured to establish a voltage a priori estimation equation, a voltage observation estimation equation, and a Kalman gain equation of the battery to be tested according to the basic state equation, the basic observation equation, the process noise covariance and the observation noise covariance at the previous acquisition moment of the specified acquisition moment;

[0076] A first determination module is configured to determine a voltage a priori estimation value, a voltage observation estimation value, and a Kalman gain at a specified acquisition time according to the voltage a priori estimation equation, the voltage observation estimation equation, and the Kalman gain equation;

[0077] A second determination module is configured to determine the optimal voltage estimate of the battery to be tested at the specified collection time according to the voltage a priori estimate at the specified collection time, the voltage observation estimate at the specified collection time, and the Kalman gain at the specified collection time;

[0078] A second acquisition module is configured to acquire measurement innovations at multiple historical acquisition moments and Kalman gains at multiple historical acquisition moments, wherein the multiple historical acquisition moments include the designated acquisition moment and the acquisition moment before the designated acquisition moment;

[0079] The setting module is configured to set a forgetting factor and a fading factor, and determine a weighting coefficient of a collection moment before a specified collection moment according to the forgetting factor; the fading factor is used to set the weights between the measurement innovations of the multiple historical collection moments; the calculation formula of the fading factor is as follows:

[0080] N i =2(m+1-i) / [m(m+1)], i=1,2,...,m;

[0081] Among them, N i is the i-th fading factor from the specified collection moment, and m is the number of moments included in the historical collection moment;

[0082] The first updating module is configured to update the process noise covariance and the observation noise covariance of the previous acquisition moment of the specified acquisition moment to the process noise covariance of the specified acquisition moment and the observation noise covariance of the specified acquisition moment according to the weighting coefficient, the fading factor, the measurement information of the multiple historical acquisition moments, and the Kalman gain of the multiple historical acquisition moments; the process noise covariance of the specified acquisition moment and the observation noise covariance of the specified acquisition moment are used to calculate the optimal voltage estimate value at the next moment of the specified acquisition moment.

[0083] According to an embodiment of the present disclosure, the first update module includes:

[0084] A first updating submodule is configured to update the process noise covariance of the specified acquisition moment according to the process noise covariance of the previous acquisition moment of the specified acquisition moment, the weighting coefficient, the fading factor, the measurement innovation of the multiple historical acquisition moments, the Kalman gains of the multiple historical acquisition moments, and the product of the weighting coefficient and the Kalman gains of the multiple historical acquisition moments;

[0085] The second updating submodule is configured to update the observed noise covariance at the specified acquisition moment according to the observed noise covariance at the previous acquisition moment, the weighting coefficient, the fading factor, the measurement updates at the multiple historical acquisition moments, and the product of the weighting coefficient and the measurement updates at the multiple historical acquisition moments.

[0086] According to an embodiment of the present disclosure, the weighting coefficient is calculated using the following formula:

[0087] d k-1 =(1-b) / (1-b k );

[0088] Among them, k is the specified collection time, k-1 is the collection time before the specified collection time, b is the forgetting factor, d k-1 is the weighting coefficient at time k-1.

[0089] According to an embodiment of the present disclosure, the process noise covariance is updated using the following formula:

[0090] Q k =(1-d k-1 )Q k-1 +d k-1 [Kg k ,Kg k-1 ,...,Kg k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T{[Kg k ,Kg k-1 ,...,Kg k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T} T ;

[0091] Among them, Q k is the process noise covariance at time k, Q k-1 is the process noise covariance at time k-1, Kg k ,Kg k-1 ,...,Kg k-m+1 is the Kalman gain corresponding to m historical acquisition moments, e k is the measurement information at time k, Kg k is the Kalman gain at time k, e k-m+1 is the measurement innovation at time k-m+1, T is the transposition operation, N1e k ,N2e k-1 ,...,N m e k-m+1 It is the product of the measurement innovation corresponding to m historical acquisition moments and the corresponding fading factor;

[0092] The observation noise covariance at time k is updated using the following formula:

[0093] R k =(1-d k-1 )R k-1 +d k-1 [N1e k ,N2e k-1 ,...,N m e k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T ;

[0094] Among them, R k is the observation noise covariance at time k, R k-1 is the observation noise covariance at time k-1.

[0095] According to an embodiment of the present disclosure, the second establishing module includes:

[0096] A first establishing submodule is configured to establish a voltage a priori estimation equation and a voltage observation estimation equation of the battery to be tested respectively according to the basic state equation and the basic observation equation;

[0097] A second establishing submodule is configured to establish an error covariance calculation equation according to the basic state equation and the process noise covariance at a previous acquisition moment of the designated acquisition moment;

[0098] The third establishing submodule is configured to establish the Kalman gain equation according to the error covariance calculation equation and the observation noise covariance at the previous acquisition moment of the designated acquisition moment.

[0099] According to an embodiment of the present disclosure, the basic state equation calculates the voltage state value at the specified acquisition time based on the voltage state value at the previous acquisition time, the system state parameter, and the system process noise at the specified acquisition time;

[0100] The basic observation equation calculates the voltage observation value at the specified acquisition time according to the voltage observation value at the specified acquisition time, the system observation parameter, and the system observation noise at the specified acquisition time;

[0101] The voltage a priori estimation equation calculates the a priori voltage estimation value at the specified acquisition time according to the voltage state value at the previous acquisition time and the system state parameters;

[0102] The voltage observation estimation equation calculates the voltage observation estimation value at the specified collection time according to the a priori voltage estimation value at the specified collection time, the system observation noise, and the voltage collection value at the specified collection time;

[0103] The error covariance calculation equation calculates the optimal estimated covariance at the specified acquisition time according to the covariance at the previous acquisition time, the system state parameter, and the process noise covariance at the previous acquisition time.

[0104] The Kalman gain equation calculates the Kalman gain at the specified acquisition time according to the optimal estimated covariance at the specified acquisition time, the system observation parameters, and the observation noise covariance at the previous acquisition time of the specified acquisition time.

[0105] According to an embodiment of the present disclosure, the second determining module includes:

[0106] An acquisition submodule, configured to acquire a voltage acquisition value of the battery to be tested at a specified acquisition time;

[0107] A first calculation submodule is configured to calculate the measurement information at the specified collection time according to the voltage collection value at the specified collection time and the voltage observation estimation value at the specified collection time;

[0108] The second calculation submodule is configured to calculate the optimal voltage estimate at the specified collection time according to the a priori voltage estimate at the specified collection time, the Kalman gain at the specified collection time, and the measurement innovation at the specified collection time.

[0109] According to an embodiment of the present disclosure, the basic state equation is shown in the following formula:

[0110] X k =AX k-1 +ω k ;

[0111] Among them, k is the specified collection time, k-1 is the collection time before the specified collection time, and X k is the voltage state value at time k, X k-1 is the voltage state value at time k-1, ω k is the process noise at time k, A is the system state parameter;

[0112] The basic observation equation is shown in the following formula:

[0113] Z k =CX k +υ k ;

[0114] Among them, Z k is the voltage observation value at time k, C is the system observation parameter, υ k is the observation noise at time k.

[0115] According to an embodiment of the present disclosure, the process noise and the observation noise are both Gaussian white noise;

[0116] The system state parameter A and the system observation parameter C are both single-element matrices with one element.

[0117] According to an embodiment of the present disclosure, the voltage a priori estimation equation is shown in the following formula:

[0118] X k '=AX k-1 ;

[0119] Among them, X k ' is the prior voltage estimate at time k;

[0120] The voltage observation estimation equation is shown in the following formula:

[0121] Z k '=CX k ';

[0122] Among them, Z k ' is the estimated value of voltage observation at time k;

[0123] The error covariance calculation equation is shown in the following formula:

[0124] P k '=AP k-1 A T +Q k-1

[0125] Among them, P k ' is the optimal estimated covariance at time k, P k-1 is the covariance at time k-1, Q k-1 is the process noise covariance at time k-1, T is the transposition operation;

[0126] The calculation method of calculating the measurement innovation at the specified collection time is as follows:

[0127] e k =Y k -Z k '=Y k -CX k ';

[0128] Among them, e k is the measurement information at time k, Y k is the voltage collection value at time k;

[0129] The Kalman gain equation is shown in the following formula:

[0130] Kg k =P k 'C T / (CP k 'C T +R k-1 );

[0131] Among them, Kg k is the Kalman gain at time k, R k-1 is the observation noise covariance at time k-1.

[0132] According to an embodiment of the present disclosure, the optimal voltage estimate at the specified acquisition time is calculated according to the a priori voltage estimate at the specified acquisition time, the Kalman gain at the specified acquisition time, and the measurement innovation at the specified acquisition time, and the calculation method is as follows:

[0133] X k + =X k '+Kg k e k ;

[0134] Among them, X k + is the optimal voltage estimate at time k.

[0135] According to an embodiment of the present disclosure, the device further includes:

[0136] The second updating module is configured to update the covariance at the specified acquisition time according to the Kalman gain at the specified acquisition time, the optimal estimated covariance at the specified acquisition time, and the system observation parameters; the covariance at the specified acquisition time is used to calculate the optimal estimated error covariance at the next moment of the specified acquisition time.

[0137] According to an embodiment of the present disclosure, the covariance of the updated designated acquisition time is calculated as follows:

[0138] P k =(E-Kg k C)P k ';

[0139] Where E is the identity matrix.

[0140] In a third aspect, an embodiment of the present disclosure provides an electronic device, comprising a memory and a processor, wherein the memory is used to store one or more computer instructions, and wherein the one or more computer instructions are executed by the processor to implement a method as described in any one of the first aspects.

[0141] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium on which computer instructions are stored. When the computer instructions are executed by a processor, the method described in the first aspect is implemented.

[0142] According to the technical solution provided by the embodiment of the present disclosure, a battery voltage acquisition method based on Kalman filtering is disclosed, including: obtaining the process noise covariance and observation noise covariance of the battery to be tested at the previous acquisition moment of the specified acquisition moment; establishing the basic state equation and basic observation equation of the battery to be tested; according to the basic state equation, the basic observation equation, and the process noise covariance and observation noise covariance of the previous acquisition moment of the specified acquisition moment, establishing a voltage a priori estimation equation, a voltage observation estimation equation, and a Kalman gain equation of the battery to be tested; according to the voltage a priori estimation equation, the voltage observation estimation equation, and the Kalman gain equation, determining the voltage a priori estimation value, the voltage observation estimation value, and the Kalman gain at the specified acquisition moment; according to the voltage a priori estimation equation, the voltage observation estimation equation, and the Kalman gain equation, determining the voltage a priori estimation value, the voltage observation estimation value, and the Kalman gain at the specified acquisition moment; according to the voltage a priori estimation value, the voltage observation estimation value, and the Kalman gain equation at the specified acquisition moment The invention discloses a method for calculating the optimal voltage estimate of the battery to be tested at the specified acquisition moment by using the voltage observation estimate at the specified acquisition moment and the Kalman gain at the specified acquisition moment; obtaining measurement information of multiple historical acquisition moments and the Kalman gains of multiple historical acquisition moments, wherein the multiple historical acquisition moments include the specified acquisition moment and the acquisition moments before the specified acquisition moment; updating the process noise covariance and the observation noise covariance of the acquisition moment before the specified acquisition moment to the process noise covariance of the specified acquisition moment and the observation noise covariance of the specified acquisition moment according to the measurement information of the multiple historical acquisition moments and the Kalman gains of the multiple historical acquisition moments; and using the process noise covariance of the specified acquisition moment and the observation noise covariance of the specified acquisition moment to calculate the optimal voltage estimate of the next moment of the specified acquisition moment.

[0143] The technical solution disclosed in the present invention can update the process noise covariance and the observation noise covariance in real time by acquiring the Kalman gain and measurement information of multiple historical acquisition moments and setting corresponding weight coefficients and fading factors. The technical solution disclosed in the present invention has a strong tracking characteristic, which makes the optimal voltage estimate value closer to the true value, further improves the filtering effect of periodic noise, and improves the accuracy of voltage acquisition.

[0144] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0145] Other features, objectives and advantages of the present disclosure will become more apparent through the following detailed description of non-limiting embodiments in conjunction with the accompanying drawings. In the accompanying drawings:

[0146] Figure 1 A flow chart of a battery voltage acquisition method based on Kalman filtering according to an embodiment of the present disclosure is shown.

[0147] Figure 2A flow chart of updating process noise covariance and observation noise covariance according to an embodiment of the present disclosure is shown.

[0148] Figure 3 A complete flow chart of a battery voltage acquisition method based on Kalman filtering according to an embodiment of the present disclosure is shown.

[0149] Figure 4 A schematic diagram showing filtering according to a battery voltage acquisition method according to an embodiment of the present disclosure, sampling based on an existing Kalman algorithm, and a voltage fluctuation range of sampling before filtering is shown.

[0150] Figure 5 A schematic diagram showing filtering according to a battery voltage acquisition method according to an embodiment of the present disclosure, sampling based on an existing Kalman algorithm, and comparing voltage fluctuation deviation values ​​of samples before filtering with true values ​​is shown.

[0151] Figure 6 A structural block diagram of a battery voltage sampling device based on Kalman filtering according to an embodiment of the present disclosure is shown.

[0152] Figure 7 A structural block diagram of updating process noise covariance and observation noise covariance in an embodiment of the present disclosure is shown.

[0153] Figure 8 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown.

[0154] Fig. 9 A schematic diagram showing the structure of a computer system suitable for implementing the method according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0155] Hereinafter, exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily implement them. In addition, for the sake of clarity, parts not related to the description of the exemplary embodiments are omitted in the accompanying drawings.

[0156] In the present disclosure, it should be understood that terms such as "include" or "have" are intended to indicate the presence of features, numbers, steps, behaviors, components, parts, or a combination thereof disclosed in the present specification, and are not intended to exclude the possibility that one or more other features, numbers, steps, behaviors, components, parts, or a combination thereof exist or are added.

[0157] It should also be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present disclosure may be combined with each other. The present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0158] Battery voltage is a key parameter for battery status management and a key indicator for measuring whether the battery has dangerous behaviors such as overcharge and over discharge. Accurate collection of battery voltage is of great significance for battery safety management and ensuring the normal operation of the battery system. However, the battery voltage sampling process will be affected by the accuracy of the sampling chip, the interference of various modules in the circuit, the external environment, etc., thereby reducing the accuracy of battery voltage sampling.

[0159] The filtering methods for battery voltage sampling mainly include hardware filtering and software filtering. Hardware filtering is to shield noise by adding additional filtering circuits, which has poor effect in suppressing system noise and increases hardware cost. Software filtering mainly includes average method and Kalman filtering method. The average method performs filtering by taking the average value of certain historical sampling points. The calculation is simple but the effect is poor. The Kalman filtering method mainly estimates and corrects the state at the current moment based on the estimated value at the last acquisition moment and the measured value at the current moment. In the Kalman filtering process, the statistical characteristics of the system process noise and the system observation noise are not easy to determine, and they are also affected by changes in the system state and environment, which has a great impact on the final filtering effect. In the existing battery voltage sampling filtering process based on the Kalman filtering algorithm, the covariance of the system process noise and the system observation noise is set to a fixed value based on experience, without considering the changes of the system noise with the system state and the external environment, and without considering the influence of historical data, resulting in poor voltage sampling tracking effect and unsatisfactory filtering effect in actual use.

[0160] Based on the above technical problems, the present invention discloses a battery voltage acquisition method based on Kalman filtering, comprising: obtaining the process noise covariance and observation noise covariance of the battery to be tested at the previous acquisition moment of the specified acquisition moment; establishing the basic state equation and basic observation equation of the battery to be tested; establishing a voltage a priori estimation equation, a voltage observation estimation equation, and a Kalman gain equation of the battery to be tested according to the basic state equation, the basic observation equation, and the process noise covariance and observation noise covariance at the previous acquisition moment of the specified acquisition moment; determining a voltage a priori estimation value, a voltage observation estimation value, and a Kalman gain at the specified acquisition moment according to the voltage a priori estimation equation, the voltage observation estimation equation, and the Kalman gain equation; determining a voltage a priori estimation value, a voltage observation estimation value, and a Kalman gain at the specified acquisition moment according to the voltage a priori estimation value at the specified acquisition moment, the voltage observation estimation value, and the Kalman gain equation; determining a voltage a priori estimation value at the specified acquisition moment, a voltage observation estimation value, and a Kalman gain at the specified acquisition moment according to the voltage a priori estimation value at the specified acquisition moment, the voltage observation estimation value, and the Kalman gain equation; determining a voltage a priori estimation value at the specified acquisition moment, the voltage observation estimation value, and the .... The voltage observation estimate value and the Kalman gain of the specified acquisition moment are used to determine the optimal voltage estimate value of the battery to be tested at the specified acquisition moment; measurement information of multiple historical acquisition moments and the Kalman gains of multiple historical acquisition moments are obtained, and the multiple historical acquisition moments include the specified acquisition moment and the acquisition moment before the specified acquisition moment; according to the measurement information of the multiple historical acquisition moments and the Kalman gains of the multiple historical acquisition moments, the process noise covariance and the observation noise covariance of the acquisition moment before the specified acquisition moment are updated to the process noise covariance of the specified acquisition moment and the observation noise covariance of the specified acquisition moment; the process noise covariance of the specified acquisition moment and the observation noise covariance of the specified acquisition moment are used to calculate the optimal voltage estimate value at the next moment of the specified acquisition moment.

[0161] The technical solution disclosed in the present invention can update the process noise covariance and the observation noise covariance in real time by acquiring the Kalman gain and measurement information of multiple historical acquisition moments and setting corresponding weight coefficients and fading factors. The technical solution disclosed in the present invention has a strong tracking characteristic, which makes the optimal voltage estimate value closer to the true value, further improves the filtering effect of periodic noise, and improves the accuracy of voltage acquisition.

[0162] Figure 1 A flow chart of a battery voltage acquisition method based on Kalman filtering according to an embodiment of the present disclosure is shown.

[0163] like Figure 1 As shown, the battery voltage acquisition method based on Kalman filtering includes the following steps S101-S108:

[0164] In step S101, the process noise covariance and the observation noise covariance of the battery to be tested at a sampling time before the specified sampling time are obtained.

[0165] According to an embodiment of the present disclosure, when the previous acquisition moment of the designated acquisition moment is the initial moment, the voltage state value is any value of the battery to be tested within the voltage variation range, and the process noise covariance and the observation noise covariance are related to the accuracy of the sensor. The values ​​of the two at the initial moment can be set to values ​​close to the sensor accuracy, and in the process of continuous iteration of the Kalman filter algorithm, they can be adjusted in real time according to the actual measurement state.

[0166] In step S102, a basic state equation and a basic observation equation of the battery to be tested are established.

[0167] According to the embodiments of the present disclosure, the sampling process of the battery voltage can be regarded as a discrete control system, and the filter calibration can be performed through the basic state equation and the basic observation equation of the Kalman filter equation. In the scheme of the present disclosure, the sampling value of the battery voltage is regarded as both the voltage state value of the system and the voltage observation value of the system.

[0168] According to an embodiment of the present disclosure, the basic state equation calculates the voltage state value at the specified acquisition time according to the voltage state value at the previous acquisition time, the system state parameter, and the system process noise at the specified acquisition time. The basic state equation is shown in the following formula:

[0169] X k =AX k-1 +ω k ;

[0170] Among them, k is the specified collection time, k-1 is the collection time before the specified collection time, and X k is the voltage state value at time k, X k-1 is the voltage state value at time k-1, ω k is the process noise at time k, and A is the system state parameter. The designated acquisition time k is the time when the voltage of the battery to be tested is sampled once, and k-1 is the previous acquisition time separated by a sampling interval from the designated acquisition time. The sampling interval can be set to a time length according to actual needs, such as 1 second, 1 minute, etc. In addition, since the voltage of the battery to be tested is a slow variable, the system state parameter A is a single-element matrix with an element of 1.

[0171] According to an embodiment of the present disclosure, the basic observation equation calculates the voltage observation value at the specified acquisition time according to the voltage observation value at the specified acquisition time, the system observation parameter, and the system observation noise at the specified acquisition time. The basic observation equation is shown in the following formula:

[0172] Z k =CX k +υ k ;

[0173] Among them, Z k is the voltage observation value at time k, C is the system observation parameter, υ k is the observation noise at time k. Since the voltage of the battery to be measured is a slow variable, the system observation parameter C is a single-element matrix with 1 as the element.

[0174] According to an embodiment of the present disclosure, the process noise ω k and observation noise υ k is Gaussian white noise.

[0175] In step S103, the voltage a priori estimation equation, voltage observation estimation equation and Kalman gain equation of the battery to be tested are established according to the basic state equation, the basic observation equation, the process noise covariance and the observation noise covariance of the previous acquisition moment of the specified acquisition moment.

[0176] According to an embodiment of the present disclosure, the step of establishing a priori voltage estimation equation, voltage observation estimation equation, and Kalman gain equation of the battery to be tested includes:

[0177] According to the basic state equation and the basic observation equation, a voltage a priori estimation equation and a voltage observation estimation equation of the battery to be tested are established respectively.

[0178] An error covariance calculation equation is established according to the basic state equation and the process noise covariance at a previous acquisition moment of the designated acquisition moment.

[0179] The Kalman gain equation is established according to the error covariance calculation equation.

[0180] According to the embodiment of the present disclosure, the voltage prior estimation equation is established based on the basic state equation to make a preliminary estimate of the voltage state value at the specified acquisition time. The voltage observation estimation equation is established based on the basic observation equation to make a preliminary estimate of the voltage observation value at the specified acquisition time.

[0181] According to an embodiment of the present disclosure, the voltage a priori estimation equation is shown in the following formula:

[0182] X k '=AX k-1 ;

[0183] Among them, X k ' is the a priori voltage estimate at time k. The a priori voltage estimate at time k is calculated by the voltage state value at time k-1 and the system state parameters. The voltage state value at time k-1 is calculated by the basic state equation. When k=1, X0 is the initial value of the voltage state. X0 can be set to any value within the voltage variation range of the battery to be tested.

[0184] According to an embodiment of the present disclosure, the voltage observation and estimation equation calculates the voltage observation and estimation value at a specified acquisition moment based on the prior voltage estimation value at the specified acquisition moment, the system observation noise, and the voltage acquisition value at the specified acquisition moment. The voltage observation and estimation equation is shown as the following formula:

[0185] Z k ' = CX k ';

[0186] Wherein, Z k ' is the voltage observation and estimation value at the k-th moment.

[0187] According to an embodiment of the present disclosure, the error covariance calculation equation calculates the optimal estimation covariance at a specified acquisition moment based on the covariance at the previous acquisition moment of the specified acquisition moment, the system state parameters, and the process noise covariance at the previous acquisition moment of the specified acquisition moment. The error covariance calculation equation is shown as the following formula:

[0188] P k ' = AP k-1 A T + Q k-1 ;

[0189] Wherein, P k ' is the optimal estimation covariance at the k-th moment, P k-1 is the covariance at the (k - 1)-th moment, Q k-1 is the process noise covariance at the (k - 1)-th moment, and T is the transpose operation. When k = 1, the range of the initial value P0 of the covariance is 0 < P0 ≤ 1. The initial value Q0 of the process noise covariance is related to the accuracy of the sensor and can be set to a value close to the sensor accuracy.

[0190] According to an embodiment of the present disclosure, the Kalman gain equation calculates the Kalman gain at a specified acquisition moment based on the optimal estimation covariance at the specified acquisition moment, the system observation parameters, and the observation noise covariance at the previous acquisition moment of the specified acquisition moment. The Kalman gain equation is shown as the following formula:

[0191] Kg k = P k 'C T / (CP k 'C T + R k-1 );

[0192] Wherein, Kg k is the Kalman gain at the k-th moment, R k-1 is the observation noise covariance at the (k - 1)-th moment, and the initial value R0 of the observation noise covariance is related to the accuracy of the sensor and can be set to a value close to the sensor accuracy.

[0193] In step S104, the voltage a priori estimation value, the voltage observation estimation value, and the Kalman gain at the specified acquisition time are determined according to the voltage a priori estimation equation, the voltage observation estimation equation, and the Kalman gain equation.

[0194] In step S105, the optimal voltage estimate of the battery to be tested at the specified collection time is determined according to the voltage a priori estimate at the specified collection time, the voltage observation estimate at the specified collection time, and the Kalman gain at the specified collection time.

[0195] According to an embodiment of the present disclosure, determining the optimal voltage estimation value of the battery to be tested at a specified acquisition time includes:

[0196] Obtain the voltage collection value of the battery to be tested at a specified collection time.

[0197] The measurement information at the specified collection time is calculated according to the voltage collection value at the specified collection time and the voltage observation estimation value at the specified collection time.

[0198] The optimal voltage estimate at the specified acquisition time is calculated according to the a priori voltage estimate at the specified acquisition time, the Kalman gain at the specified acquisition time, and the measurement innovation at the specified acquisition time.

[0199] According to an embodiment of the present disclosure, the optimal voltage estimate at the specified acquisition time is calculated according to the a priori voltage estimate at the specified acquisition time, the Kalman gain at the specified acquisition time, and the measurement innovation at the specified acquisition time, wherein the measurement innovation and the optimal voltage estimate are calculated as follows:

[0200] e k =Y k -Z k '=Y k -CX k ';

[0201] X k + =X k '+Kg k e k ;

[0202] Among them, e k is the measurement information at time k, Y k is the voltage collection value at time k, X k + is the optimal voltage estimate at time k.

[0203] In step S106, measurement information of multiple historical acquisition moments and Kalman gains of multiple historical acquisition moments are obtained, and the multiple historical acquisition moments include the designated acquisition moment and the acquisition moment before the designated acquisition moment. According to an embodiment of the present disclosure, after each acquisition time interval, the Kalman filter of the present disclosure will iterate once to generate new measurement information, Kalman gain and optimal voltage estimation value of the acquisition moment. As the number of Kalman filter iterations increases, measurement information and Kalman gain of multiple historical acquisition moments can be generated. The measurement information and Kalman gain of the historical acquisition moments can update and adjust the process noise covariance and the observation noise covariance in real time, enhance the tracking ability of the true value of the voltage and the filtering effect of the periodic noise during the battery voltage acquisition process, and improve the acquisition accuracy of the battery voltage.

[0204] In step S107, a forgetting factor and a fading factor are set, and a weighting coefficient of the previous acquisition moment of the specified acquisition moment is determined according to the forgetting factor; the fading factor is used to set the weights between the measurement innovations of the multiple historical acquisition moments. The calculation formula of the fading factor is as follows:

[0205] N i =2(m+1-i) / [m(m+1)], i=1,2,...,m;

[0206] Among them, N i is the i-th fading factor from the specified collection moment, and m is the number of moments included in the historical collection moments. The weighting coefficient can determine the weighted ratio between the process noise covariance of the previous collection moment of the specified collection moment and the measurement innovations of multiple historical collection moments and the Kalman gain of multiple historical collection moments when updating the process noise covariance of the specified collection moment. When updating the observation noise covariance of the specified collection moment, the weighting coefficient is used to determine the weighted ratio between the observation noise covariance of the previous collection moment of the specified collection moment and the measurement innovations of multiple historical collection moments. The fading factor is used to set the weights between the measurement innovations of the multiple historical collection moments and can prevent data saturation. In addition, the fading factor can make the weight of the measurement innovation close to the specified collection moment greater than that of the specified collection moment farther away from the specified collection moment, thereby reducing the interference of distant historical data on the final result.

[0207] In step S108, based on the weighting coefficient, the fading factor, the measurement innovations of the multiple historical collection moments, and the Kalman gains of the multiple historical collection moments, the process noise covariance and the observation noise covariance of the previous collection moment of the specified collection moment are updated to the process noise covariance of the specified collection moment and the observation noise covariance of the specified collection moment; the process noise covariance of the specified collection moment and the observation noise covariance of the specified collection moment are used to calculate the optimal voltage estimate of the next collection moment of the specified collection moment.

[0208] The inventors noticed that in the traditional battery voltage sampling and filtering algorithm based on the Kalman algorithm, the process noise covariance and the observation noise covariance are often set to fixed values ​​based on experience, without considering the influence of factors such as system state changes and external environment changes on the statistical characteristics of noise. In addition, the traditional Kalman filtering algorithm only uses the data of the previous collection time at the current collection time to predict the data at the current collection time, without considering the influence of historical data, resulting in poor tracking effect in the battery voltage collection process based on the Kalman filtering algorithm, and unsatisfactory filtering effect on periodic noise. After in-depth research, the inventors found that by obtaining multiple Kalman gains and measurement information at historical sampling times or multiple consecutive historical sampling times within a period of time, and updating the process noise covariance and observation noise covariance in real time, the filtering effect can be significantly improved.

[0209] Figure 2 FIG. 4 is a flowchart of updating the process noise covariance and the observation noise covariance according to an embodiment of the present disclosure. Figure 2 As shown, the process noise covariance and the observation noise covariance of the previous acquisition moment of the specified acquisition moment are updated according to the weighting coefficient, the fading factor, the measurement innovation of the multiple historical acquisition moments, and the Kalman gain of the multiple historical acquisition moments, including steps S1071-S1072:

[0210] In step S1071, the process noise covariance at the specified acquisition moment is updated according to the process noise covariance at the previous acquisition moment, the weighting coefficient, the fading factor, the measurement innovations at the multiple historical acquisition moments, the Kalman gains at the multiple historical acquisition moments, and the product of the weighting coefficient and the Kalman gains at the multiple historical acquisition moments.

[0211] According to an embodiment of the present disclosure, the weighting coefficient is calculated using the following formula:

[0212] d k-1 =(1-b) / (1-b k );

[0213] The process noise covariance is updated using the following formula:

[0214] Q k =(1-d k-1 )Q k-1 +d k-1 [Kg k ,Kg k-1 ,...,Kg k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T {[Kg k ,Kg k-1 ,...,Kg k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T} T ;

[0215] Among them, k is the specified collection time, k-1 is the collection time before the specified collection time, b is the forgetting factor, Q k is the process noise covariance at time k, Q k-1 is the process noise covariance at time k-1, d k-1 is the weighting coefficient at time k-1, Kg k is the Kalman gain at time k, Kg k ,Kg k-1 ,...,Kg k-m+1 is the Kalman gain corresponding to m historical acquisition moments, e k is the measurement information at time k, e k-1 is the measurement information at time k-1, e k-m+1 is the measurement innovation at time k-m+1, T is the transposition operation, N1e k ,N2e k-1 ,...,N m e k-m+1 It is the product of the measurement innovation corresponding to m historical acquisition moments and the corresponding fading factor. The forgetting factor b is used to calculate the weighted ratio of the process noise covariance at the previous acquisition moment to the historical data. The value range of the forgetting factor is generally 0.9. <b<1。

[0216] In step S1072, the observed noise covariance at the specified acquisition moment is updated according to the observed noise covariance at the previous acquisition moment, the weighting coefficient, the fading factor, the measurement updates at the multiple historical acquisition moments, and the product of the weighting coefficient and the measurement updates at the multiple historical acquisition moments.

[0217] According to an embodiment of the present disclosure, the observation noise covariance at time k is updated using the following formula:

[0218] R k =(1-d k-1 )R k-1 +d k-1 [N1e k ,N2e k-1 ,...,N m e k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T ;

[0219] Among them, R k is the observation noise covariance at time k, R k-1 is the observation noise covariance at time k-1.

[0220] In the actual calculation process of the Kalman filter algorithm, the system process noise and system observation noise have a great influence on the calculation results. Excessive deviation of the noise value may even cause filter divergence. However, the actual values ​​of the system process noise and system observation noise are not easy to determine, and as the battery runs, the battery will gradually decay and age, causing the battery system characteristics to change. Therefore, the system process noise and system observation noise will also change continuously. Traditional solutions do not take into account the impact and changes of historical data on the system process noise and measurement noise.

[0221] This solution optimizes the calculation of the system process noise covariance matrix and the observation noise covariance based on the measurement information, fading factors, and Kalman gains at multiple historical moments. It can calculate and update the system process noise and the system measurement noise in real time according to the system historical data and current data, and introduce fading factors to adjust the weights of the measurement information at different historical moments. This can prevent the omission of historical data information from causing inaccurate updates of the system process noise covariance and the system observation noise covariance, and can also prevent the problem of excessive weight of historical data at a longer distance from weakening the effect of historical data at a shorter distance, thereby further improving the calculation accuracy of the battery voltage acquisition method based on Kalman filtering of the present invention and its adaptability to actual operating conditions.

[0222] According to an embodiment of the present disclosure, the method further includes:

[0223] Updating the covariance at the specified acquisition time according to the Kalman gain at the specified acquisition time, the optimal estimation covariance at the specified acquisition time, and system observation parameters; the covariance at the specified acquisition time is used to calculate the optimal estimation error covariance at the next time of the specified acquisition time. When performing a calculation of the optimal voltage estimation value in the Kalman filtering algorithm of the present disclosure, the optimal estimation covariance P k ' at the specified sampling time will be obtained through calculation. After this calculation is completed, the covariance P k ' is used to update the covariance P k-1 to obtain the covariance P k at the specified acquisition time, completing the iteration of the covariance.

[0224] Specifically, the calculation method for updating the covariance at the specified acquisition time is as follows:

[0225] P k = (E - Kg k C)P k '; where E is the identity matrix.

[0226] Figure 3 FIG. shows a complete flowchart of a battery voltage acquisition method based on Kalman filtering according to an embodiment of the present disclosure. As Figure 3 shown:

[0227] First, initialize the Kalman filtering state equation, and respectively obtain the voltage state value X0, covariance P0, process noise covariance Q0, and observation noise covariance R0 at the initial time. Among them, the value of X0 is any value within the voltage change range of the battery to be measured, the value of P0 is 0 < P0 ≤ 1, and the values of Q0 and R0 are values close to the sensor accuracy.

[0228] Then, perform a priori estimation of the battery voltage at the specified acquisition time k. Based on the basic state equation and the basic observation equation, according to X k ' = AX k-1 calculate the a priori voltage estimation value at the specified acquisition time k;

[0229] Then, calculate the optimal estimation covariance at the specified sampling time k. According to P k ' = AP k-1 A T + Q k-1 calculate, where the process noise covariance Q k-1 is not a fixed value, but is iteratively updated with the specified acquisition time.

[0230] Then, according to e k = Yk -Z k '=Y k -CX k 'Calculate the measurement information at the specified acquisition time k, Y k is the voltage acquisition value at the specified acquisition time, and C is the system observation parameter, which is a single-element matrix with element 1.

[0231] According to Kg k =P k 'C T / (CP k 'C T +R k-1 ) calculates the Kalman gain at the specified acquisition time k, where the observation noise covariance R k-1 It is not a fixed value, but is updated iteratively at the specified collection time.

[0232] According to X k + =X k '+Kg k e k , perform state correction on the prior voltage estimate and calculate the optimal voltage estimate at the specified acquisition time.

[0233] According to P k =(E-Kg k C)P k ', the covariance at time k-1 in the above operation is updated to the covariance at time k, so as to be used in subsequent Kalman filter iterations.

[0234] According to d k-1 =(1-b) / (1-b k ) calculates the weighting coefficient,

[0235] According to N i =2(m+1-i) / [m(m+1)] calculates the fading factor,

[0236] according to

[0237] Q k =(1-d k-1 )Q k-1 +d k-1 [Kg k ,Kg k-1 ,...,Kg k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T {[Kg k ,Kg k-1 ,...,Kgk-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T} T ,

[0238] R k =(1-d k-1 )R k-1 +d k-1 [N1e k ,N2e k-1 ,...,N m e k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T , the process noise covariance and the observation noise covariance are updated to improve the filtering effect of the battery voltage-based Kalman filter algorithm on periodic noise.

[0239] Figure 4 A schematic diagram showing filtering according to a battery voltage acquisition method according to an embodiment of the present disclosure, sampling based on an existing Kalman algorithm, and a voltage fluctuation range of sampling before filtering is shown. Figure 5 A schematic diagram showing the voltage fluctuation deviation value of the battery voltage acquisition method according to the embodiment of the present disclosure, filtering, sampling based on the existing Kalman algorithm, and comparison between the sampling before filtering and the true value. Figure 4 As shown, the figure shows the true value of the battery voltage, the sampled value with noise before filtering, the estimated value after filtering by the existing Kalman algorithm, and the estimated value after filtering based on the method of the present invention. It can be seen that before filtering, the battery voltage acquisition value fluctuates more than the true value. After filtering with the existing Kalman algorithm, the degree of fluctuation is improved but the effect is not good. After filtering with the method of the present invention, the optimal voltage estimate is closer to the true value and the effect is better.

[0240] like Figure 5 As shown, in order to further quantitatively evaluate the filtering effect of the present invention, the sampled value with noise before filtering, the estimated value after filtering by the existing Kalman algorithm, and the estimated value after filtering based on the method of the present invention are compared with the real voltage value to obtain the corresponding deviation range. The deviation fluctuation range of the battery voltage sampling before filtering is about ±0.4%, the deviation fluctuation range of the estimated value after filtering by the existing Kalman algorithm is about ±0.2%, and the deviation fluctuation range of the voltage estimated value after filtering according to the method of the present invention is about ±0.077%. It can be seen that the method of the present invention can effectively remove periodic noise and significantly improve the accuracy of battery voltage sampling.

[0241] Figure 6 The structure block diagram of the battery voltage sampling device based on Kalman filtering according to an embodiment of the present disclosure is shown. The device can be implemented as part or all of the electronic device through software, hardware or a combination of both.

[0242] like Figure 6 As shown, the battery voltage sampling device 600 based on Kalman filtering includes a first acquisition module 610, a first establishment module 620, a second establishment module 630, a first determination module 640, a second determination module 650, a second acquisition module 660, a setting module 670, and a first update module 680.

[0243] The first acquisition module 610 is configured to acquire the process noise covariance and the observation noise covariance of the battery to be tested at a previous acquisition time before the specified acquisition time.

[0244] According to an embodiment of the present disclosure, when the previous acquisition moment of the designated acquisition moment is the initial moment, the voltage state value is any value of the battery to be tested within the voltage variation range, and the process noise covariance and the observation noise covariance are related to the accuracy of the sensor. The values ​​of the two at the initial moment can be set to values ​​close to the sensor accuracy, and can be adjusted in real time according to the state of the battery to be tested.

[0245] The first establishing module 620 is configured to establish the basic state equation and the basic observation equation of the battery to be tested.

[0246] According to the embodiments of the present disclosure, the sampling process of the battery voltage can be regarded as a discrete control system, and the filter calibration can be performed through the basic state equation and the basic observation equation of the Kalman filter equation. In the scheme of the present disclosure, the sampling value of the battery voltage is regarded as both the voltage state value of the system and the voltage observation value of the system.

[0247] According to an embodiment of the present disclosure, the basic state equation calculates the voltage state value at the specified acquisition time according to the voltage state value at the previous acquisition time, the system state parameter, and the system process noise at the specified acquisition time. The basic state equation is shown in the following formula:

[0248] X k =AX k-1 +ω k ;

[0249] Among them, k is the specified collection time, k-1 is the collection time before the specified collection time, and X k is the voltage state value at time k, X k-1 is the voltage state value at time k-1, ω kis the process noise at time k, and A is the system state parameter. The specified acquisition time is the time when the voltage of the battery to be tested is sampled once, k-1 is the previous acquisition time separated by a sampling interval from the specified acquisition time, and the sampling interval can be set to a time length according to actual needs, such as 1 second, 1 minute, etc. In addition, since the voltage of the battery to be tested is a slow variable, the system state parameter A is a single-element matrix with an element of 1.

[0250] According to an embodiment of the present disclosure, the basic observation equation calculates the voltage observation value at the specified acquisition time according to the voltage observation value at the specified acquisition time, the system observation parameter, and the system observation noise at the specified acquisition time. The basic observation equation is shown in the following formula:

[0251] Z k =CXk+υk;

[0252] Among them, Z k is the voltage observation value at time k, C is the system observation parameter, υ k is the observation noise at time k. Since the voltage of the battery to be measured is a slow variable, the system observation parameter C is a single-element matrix with 1 as the element.

[0253] According to an embodiment of the present disclosure, the process noise ω k and observation noise υ k is Gaussian white noise.

[0254] The second establishing module 630 is configured to establish the voltage a priori estimation equation, the voltage observation estimation equation, and the Kalman gain equation of the battery to be tested according to the basic state equation, the basic observation equation, the process noise covariance and the observation noise covariance at the previous acquisition moment of the specified acquisition moment.

[0255] According to an embodiment of the present disclosure, the second establishing module includes:

[0256] The first establishing submodule is configured to respectively establish a voltage a priori estimation equation and a voltage observation estimation equation of the battery to be tested according to the basic state equation and the basic observation equation.

[0257] The second establishing submodule is configured to establish an error covariance calculation equation according to the basic state equation and the process noise covariance at a previous acquisition moment of the designated acquisition moment.

[0258] The third establishing submodule is configured to establish the Kalman gain equation according to the error covariance calculation equation.

[0259] According to the embodiment of the present disclosure, the voltage prior estimation equation is established based on the basic state equation to make a preliminary estimate of the voltage state value at the specified acquisition time. The voltage observation estimation equation is established based on the basic observation equation to make a preliminary estimate of the voltage observation value at the specified acquisition time.

[0260] According to an embodiment of the present disclosure, the voltage a priori estimation equation is shown in the following formula:

[0261] X k '=AX k-1 ;

[0262] Among them, X k ' is the a priori voltage estimate at time k. The a priori voltage estimate at time k is calculated by the voltage state value at time k-1 and the system state parameters. The voltage state value at time k-1 is calculated by the basic state equation. When k=1, X0 is the initial value of the voltage state. X0 can be set to any value within the voltage variation range of the battery to be tested. The value range of the initial value of the voltage state X0 is related to the type of the battery to be tested.

[0263] According to an embodiment of the present disclosure, the voltage observation estimation equation calculates the voltage observation estimation value at the specified acquisition time according to the a priori voltage estimation value at the specified acquisition time, the system observation noise, and the voltage acquisition value at the specified acquisition time. The voltage observation estimation equation is shown in the following formula:

[0264] Z k '=CX k ';

[0265] Among them, Z k ' is the estimated value of voltage observation at time k.

[0266] According to an embodiment of the present disclosure, the error covariance calculation equation calculates the optimal estimated covariance at the specified acquisition time according to the covariance at the previous acquisition time, the system state parameter, and the process noise covariance at the previous acquisition time. The error covariance calculation equation is shown in the following formula:

[0267] P k '=AP k-1 A T +Q k-1

[0268] Among them, P k ' is the optimal estimated covariance at time k, P k-1 is the covariance at time k-1, Q k-1is the process noise covariance at time k-1, and T represents the transpose operation. When k = 1, the initial value P0 of the covariance ranges from 0 < P0 ≤ 1. The initial value Q0 of the process noise covariance is related to the accuracy of the sensor and can be set to a value close to the sensor accuracy.

[0269] According to an embodiment of the present disclosure, the Kalman gain equation calculates the Kalman gain at a specified acquisition time according to the optimal estimation covariance at the specified acquisition time, the system observation parameters, and the observation noise covariance at the previous acquisition time of the specified acquisition time. The Kalman gain equation is shown in the following formula:

[0270] Kg k = P k 'C T / (CP k 'C T + R k-1 );

[0271] where, Kg k is the Kalman gain at time k, R k-1 is the observation noise covariance at time k-1. The initial value R0 of the observation noise covariance is related to the accuracy of the sensor and can be set to a value close to the sensor accuracy.

[0272] The first determination module 640 is configured to determine the prior voltage estimate value, the observed voltage estimate value, and the Kalman gain at a specified acquisition time according to the voltage prior estimate equation, the voltage observation estimate equation, and the Kalman gain equation.

[0273] The second determination module 650 is configured to determine the optimal voltage estimate value of the battery under test at a specified acquisition time according to the prior voltage estimate value at the specified acquisition time, the observed voltage estimate value at the specified acquisition time, and the Kalman gain at the specified acquisition time.

[0274] According to an embodiment of the present disclosure, the second determination module includes:

[0275] An acquisition sub-module is configured to acquire the voltage acquisition value of the battery under test at a specified acquisition time.

[0276] A first calculation sub-module is configured to calculate the measurement innovation at a specified acquisition time according to the voltage acquisition value at the specified acquisition time and the observed voltage estimate value at the specified acquisition time;

[0277] A second calculation sub-module is configured to calculate the optimal voltage estimate value at a specified acquisition time according to the prior voltage estimate value at the specified acquisition time, the Kalman gain at the specified acquisition time, and the measurement innovation at the specified acquisition time.

[0278] According to an embodiment of the present disclosure, the optimal voltage estimate at the specified acquisition time is calculated according to the a priori voltage estimate at the specified acquisition time, the Kalman gain at the specified acquisition time, and the measurement innovation at the specified acquisition time, wherein the measurement innovation and the optimal voltage estimate are calculated as follows:

[0279] e k =Y k -Z k '=Y k -CX k ';

[0280] X k + =X k '+Kg k e k ;

[0281] Among them, e k is the measurement information at time k, Y k is the voltage collection value at time k, X k + is the optimal voltage estimate at time k.

[0282] The second acquisition module 660 is configured to acquire measurement information of multiple historical acquisition moments and Kalman gains of multiple historical acquisition moments, wherein the multiple historical acquisition moments include the designated acquisition moment and the acquisition moment before the designated acquisition moment. According to an embodiment of the present disclosure, after each acquisition time interval, the Kalman filter of the present disclosure will iterate once to generate new measurement information, Kalman gain and optimal voltage estimation value of the acquisition moment. As the number of Kalman filter iterations increases, measurement information and Kalman gain of multiple historical acquisition moments can be generated. The measurement information and Kalman gain of the historical acquisition moments can update and adjust the process noise covariance and the observation noise covariance in real time, enhance the tracking capability of the true value of the voltage and the filtering effect of the periodic noise during the battery voltage acquisition process, and improve the acquisition accuracy of the battery voltage.

[0283] The setting module 670 is configured to set a forgetting factor and a fading factor, and determine the weighting coefficient of the previous acquisition moment of the specified acquisition moment according to the forgetting factor; the fading factor is used to set the weight between the measurement innovations of the multiple historical acquisition moments. The calculation formula of the fading factor is as follows:

[0284] N i =2(m+1-i) / [m(m+1)], i=1,2,...,m;

[0285] Among them, N iis the i-th fading factor from the specified acquisition moment, and m is the number of moments included in the historical acquisition moments. The weighting coefficient can determine the weighted ratio between the process noise covariance of the previous acquisition moment of the specified acquisition moment and the measurement innovations of multiple historical acquisition moments and the Kalman gain of multiple historical acquisition moments when updating the process noise covariance of the specified acquisition moment. When updating the observation noise covariance of the specified acquisition moment, the weighting coefficient is used to determine the weighted ratio between the observation noise covariance of the previous acquisition moment of the specified acquisition moment and the measurement innovations of multiple historical acquisition moments. The fading factor is used to set the weights between the measurement innovations of the multiple historical acquisition moments and prevent data saturation. In addition, the fading factor can make the weight of the measurement innovation close to the specified acquisition moment greater than that of the specified acquisition moment farther away from the specified acquisition moment, thereby reducing the interference of distant historical data on the final result.

[0286] The first updating module 680 is configured to update the process noise covariance and the observation noise covariance of the previous acquisition moment of the specified acquisition moment to the process noise covariance of the specified acquisition moment and the observation noise covariance of the specified acquisition moment according to the weighting coefficient, the fading factor, the measurement information of the multiple historical acquisition moments, and the Kalman gain of the multiple historical acquisition moments; the process noise covariance of the specified acquisition moment and the observation noise covariance of the specified acquisition moment are used to calculate the optimal voltage estimate at the next moment of the specified acquisition moment.

[0287] The inventors noticed that in the traditional battery voltage sampling and filtering algorithm based on the Kalman algorithm, the process noise covariance and the observation noise covariance are often set to fixed values ​​based on experience, without considering the influence of factors such as system state changes and external environment changes on the statistical characteristics of noise. In addition, the traditional Kalman filtering algorithm uses the data of the previous collection time at the current collection time to predict the data at the current collection time, without considering the influence of historical data, resulting in poor tracking effect in the battery voltage collection process based on the Kalman filtering algorithm, and unsatisfactory filtering effect on periodic noise. After in-depth research, the inventors found that by obtaining multiple Kalman gains and measurement information at historical sampling times or multiple historical sampling times within a period of time, and updating the process noise covariance and observation noise covariance in real time, the filtering effect can be significantly improved.

[0288] Figure 7 FIG. 4 shows a structural block diagram of updating the process noise covariance and the observation noise covariance in an embodiment of the present disclosure. Figure 7 As shown, the first update module includes a first update submodule 681 and a second update submodule 682:

[0289] The first updating submodule 681 is configured to update the process noise covariance at the specified acquisition moment according to the process noise covariance at the previous acquisition moment of the specified acquisition moment, the weighting coefficient, the fading factor, the measurement innovations at the multiple historical acquisition moments, the Kalman gains at the multiple historical acquisition moments, and the product of the weighting coefficient and the Kalman gains at the multiple historical acquisition moments.

[0290] According to an embodiment of the present disclosure, the weighting coefficient is calculated using the following formula:

[0291] d k-1 =(1-b) / (1-b k );

[0292] The process noise covariance is updated using the following formula:

[0293] Q k =(1-d k-1 )Q k-1 +d k-1 [Kg k ,Kg k-1 ,...,Kg k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T {[Kg k ,Kg k-1 ,...,Kg k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T} T ;

[0294] Among them, k is the specified collection time, k-1 is the collection time before the specified collection time, b is the forgetting factor, Q k is the process noise covariance at time k, Q k-1 is the process noise covariance at time k-1, d k-1 is the weighting coefficient at time k-1, Kg k-1 is the Kalman gain at time k-1, Kg k ,Kg k-1 ,...,Kg k-m+1 is the Kalman gain corresponding to m historical acquisition moments, e k is the measurement information at time k, e k-1 is the measurement information at time k-1, e k-m+1 is the measurement innovation at time k-m+1, T is the transposition operation, N1ek ,N2e k-1 ,...,N m e k-m+1 It is the product of the measurement information corresponding to m historical collection moments and the corresponding fading factor.

[0295] The second updating submodule 682 is configured to update the observed noise covariance at the specified acquisition moment based on the observed noise covariance at the previous acquisition moment, the weighting coefficient, the fading factor, the measurement updates at the multiple historical acquisition moments, and the product of the weighting coefficient and the measurement updates at the multiple historical acquisition moments.

[0296] According to an embodiment of the present disclosure, the observation noise covariance is updated using the following formula:

[0297] R k =(1-d k-1 )R k-1 +d k-1 [N1e k ,N2e k-1 ,...,N m e k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T ;

[0298] Among them, R k is the observation noise covariance at time k, R k-1 is the observation noise covariance at time k-1.

[0299] According to an embodiment of the present disclosure, the device further includes:

[0300] The second updating module is configured to update the covariance at the specified acquisition time according to the Kalman gain at the specified acquisition time, the optimal estimated covariance at the specified acquisition time, and the system observation parameters; the covariance at the specified acquisition time is used to calculate the optimal estimated error covariance at the next moment of the specified acquisition time. When the Kalman filter algorithm of the present disclosure calculates the optimal voltage estimate value once, it will calculate the optimal estimated covariance P at the specified sampling time. k ', after this calculation is completed, use the optimal estimated covariance P k 'For covariance P k-1 Update to obtain the covariance P at the specified acquisition time k , completing the iteration of covariance.

[0301] Specifically, the calculation method of updating the covariance of the specified acquisition time is as follows:

[0302] P k =(E-Kg k C)P k '; Where E is the identity matrix.

[0303] The present disclosure also discloses an electronic device, Figure 8 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown.

[0304] like Figure 8 As shown, the electronic device includes a memory and a processor, wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the method according to an embodiment of the present disclosure.

[0305] A battery voltage acquisition method based on Kalman filtering, characterized by comprising:

[0306] Obtain the process noise covariance and observation noise covariance of the battery to be tested at the previous acquisition time of the specified acquisition time;

[0307] Establishing the basic state equation and basic observation equation of the battery to be tested;

[0308] Establishing a voltage a priori estimation equation, a voltage observation estimation equation, and a Kalman gain equation of the battery to be tested according to the basic state equation, the basic observation equation, the process noise covariance and the observation noise covariance at the previous acquisition moment of the specified acquisition moment;

[0309] Determine a voltage a priori estimate value, a voltage observation estimate value, and a Kalman gain at a specified acquisition time according to the voltage a priori estimate equation, the voltage observation estimate equation, and the Kalman gain equation;

[0310] Determine the optimal voltage estimate of the battery to be tested at the specified acquisition time according to the voltage a priori estimate at the specified acquisition time, the voltage observation estimate at the specified acquisition time, and the Kalman gain at the specified acquisition time;

[0311] Acquire measurement innovation information at multiple historical collection moments and Kalman gains at multiple historical collection moments, wherein the multiple historical collection moments include the designated collection moment and collection moments before the designated collection moment;

[0312] According to the measurement updates of the multiple historical collection moments and the Kalman gains of the multiple historical collection moments, the process noise covariance and the observation noise covariance of the previous collection moment of the specified collection moment are updated to the process noise covariance of the specified collection moment and the observation noise covariance of the specified collection moment; the process noise covariance of the specified collection moment and the observation noise covariance of the specified collection moment are used to calculate the optimal voltage estimate at the next moment of the specified collection moment.

[0313] Set a forgetting factor and a fading factor, and determine the weighting coefficient of the previous acquisition moment of the specified acquisition moment according to the forgetting factor; the fading factor is used to set the weights between the measurement innovations of the multiple historical acquisition moments. The calculation formula of the fading factor is as follows:

[0314] N i =2(m+1-i) / [m(m+1)], i=1,2,...,m;

[0315] Among them, N i is the ith fading factor from the specified collection moment, and m is the number of moments included in the historical collection moments.

[0316] The updating of the process noise covariance at the designated collection moment and the observation noise covariance at the designated collection moment by using the process noise covariance and the observation noise covariance at the previous collection moment of the designated collection moment according to the weighting coefficient, the fading factor, the measurement innovations at the multiple historical collection moments, and the Kalman gains at the multiple historical collection moments comprises:

[0317] Update the process noise covariance of the designated acquisition moment according to the process noise covariance of the previous acquisition moment, the weighting coefficient, the fading factor, the measurement innovations of the multiple historical acquisition moments, the Kalman gains of the multiple historical acquisition moments, and the product of the weighting coefficient and the Kalman gains of the multiple historical acquisition moments;

[0318] The observed noise covariance at the specified acquisition moment is updated according to the observed noise covariance at the previous acquisition moment, the weighting coefficient, the fading factor, the measurement innovations at the multiple historical acquisition moments, and the product of the weighting coefficient and the measurement innovations at the multiple historical acquisition moments.

[0319] The weighting coefficient is calculated using the following formula:

[0320] d k-1 =(1-b) / (1-b k );

[0321] The fading factor is set using the following formula:

[0322] Q k =(1-d k-1 )Q k-1 +d k-1 [Kg k ,Kg k-1 ,...,Kg k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T {[Kg k ,Kg k-1 ,...,Kg k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T} T ;

[0323] Among them, k is the specified collection time, k-1 is the collection time before the specified collection time, b is the forgetting factor, Q k is the process noise covariance at time k, Q k-1 is the process noise covariance at time k-1, d k-1 is the weighting coefficient at time k-1, Kg k ,Kg k-1 ,...,Kg k-m+1 is the Kalman gain corresponding to m historical acquisition moments, e k is the measurement information at time k, Kg k is the Kalman gain at time k, e k-m+1 is the measurement innovation at time k-m+1, T is the transposition operation, N1e k ,N2e k-1 ,...,N m e k-m+1 It is the product of the measurement innovation corresponding to m historical acquisition moments and the corresponding fading factor;

[0324] The observation noise covariance at time k is updated using the following formula:

[0325] R k =(1-d k-1 )R k-1 +d k-1 [N1e k ,N2e k-1 ,...,N m e k-m+1 ][N1e k ,N2e k-1 ,...,N m ek-m+1 ] T ;

[0326] Among them, R k is the observation noise covariance at time k, R k-1 is the observation noise covariance at time k-1.

[0327] The voltage a priori estimation equation, voltage observation estimation equation, and Kalman gain equation of the battery to be tested are established according to the basic state equation, the basic observation equation, the process noise covariance and the observation noise covariance of the previous acquisition time at the specified acquisition time, including:

[0328] According to the basic state equation and the basic observation equation, respectively establishing a voltage a priori estimation equation and a voltage observation estimation equation of the battery to be tested;

[0329] Establishing an error covariance calculation equation according to the basic state equation and the process noise covariance at a previous acquisition moment of the specified acquisition moment;

[0330] The Kalman gain equation is established according to the error covariance calculation equation and the observation noise covariance at a collection time before the designated collection time.

[0331] The basic state equation calculates the voltage state value at the specified acquisition time according to the voltage state value at the previous acquisition time, the system state parameter, and the system process noise at the specified acquisition time;

[0332] The basic observation equation calculates the voltage observation value at the specified acquisition time according to the voltage observation value at the specified acquisition time, the system observation parameter, and the system observation noise at the specified acquisition time;

[0333] The voltage a priori estimation equation calculates the a priori voltage estimation value at the specified acquisition time according to the voltage state value at the previous acquisition time and the system state parameters;

[0334] The voltage observation estimation equation calculates the voltage observation estimation value at the specified collection time according to the a priori voltage estimation value at the specified collection time, the system observation noise, and the voltage collection value at the specified collection time;

[0335] The error covariance calculation equation calculates the optimal estimated covariance at the specified acquisition time according to the covariance at the previous acquisition time, the system state parameter, and the process noise covariance at the previous acquisition time.

[0336] The Kalman gain equation calculates the Kalman gain at the specified acquisition time according to the optimal estimated covariance at the specified acquisition time, the system observation parameters, and the observation noise covariance at the previous acquisition time of the specified acquisition time.

[0337] Determining the optimal voltage estimate of the battery to be tested at the specified acquisition time according to the voltage a priori estimate at the specified acquisition time, the voltage observation estimate at the specified acquisition time, and the Kalman gain at the specified acquisition time, including:

[0338] Obtaining a voltage collection value of the battery to be tested at a specified collection time;

[0339] Calculate the measurement information at the specified collection time according to the voltage collection value at the specified collection time and the voltage observation estimate value at the specified collection time;

[0340] The optimal voltage estimate at the specified acquisition time is calculated according to the a priori voltage estimate at the specified acquisition time, the Kalman gain at the specified acquisition time, and the measurement innovation at the specified acquisition time.

[0341] The basic state equation is shown in the following formula:

[0342] X k =AX k-1 +ω k ;

[0343] Among them, k-1 is the previous acquisition time of the specified acquisition time, X k is the voltage state value at time k, X k-1 is the voltage state value at time k-1, ω k is the process noise at time k, A is the system state parameter;

[0344] The basic observation equation is shown in the following formula:

[0345] Z k =CXk+υ k ;

[0346] Among them, Z k is the voltage observation value at time k, C is the system observation parameter, υ k is the observation noise at time k;

[0347] The process noise and the observation noise are both Gaussian white noise;

[0348] The system state parameter A and the system observation parameter C are both single-element matrices with one element.

[0349] The voltage prior estimation equation is shown in the following formula:

[0350] X k '=AX k-1 ;

[0351] Among them, X k' is the prior voltage estimate at time k;

[0352] The voltage observation estimation equation is shown in the following formula:

[0353] Z k '=CX k ';

[0354] Among them, Z k ' is the estimated value of voltage observation at time k;

[0355] The error covariance calculation equation is shown in the following formula:

[0356] P k '=AP k-1 A T +Q k-1

[0357] Among them, P k ' is the optimal estimated covariance at time k, P k-1 is the covariance at time k-1, Q k-1 is the process noise covariance at time k-1, T is the transposition operation;

[0358] The calculation method of calculating the measurement innovation at the specified collection time is as follows:

[0359] e k =Y k -Z k '=Y k -CX k ';

[0360] Among them, e k is the measurement information at time k, Y k is the voltage collection value at time k;

[0361] The Kalman gain equation is shown in the following formula:

[0362] Kg k =P k 'C T / (CP k 'C T +R k-1 );

[0363] Among them, Kg k is the Kalman gain at time k, R k-1 is the observation noise covariance at time k-1.

[0364] The optimal voltage estimate at the specified acquisition time is calculated according to the a priori voltage estimate at the specified acquisition time, the Kalman gain at the specified acquisition time, and the measurement information at the specified acquisition time. The calculation method is as follows:

[0365] X k + =X k '+Kg k e k ;

[0366] Among them, X k + is the optimal voltage estimate at time k.

[0367] The method further comprises:

[0368] The covariance at the specified acquisition time is updated according to the Kalman gain at the specified acquisition time, the optimal estimated covariance at the specified acquisition time, and the system observation parameters; the covariance at the specified acquisition time is used to calculate the optimal estimated error covariance at the next moment of the specified acquisition time.

[0369] The calculation method of updating the covariance of the specified acquisition time is as follows:

[0370] P k =(E-Kg k C)P k ';

[0371] Where E is the identity matrix.

[0372] Fig. 9 A schematic diagram showing the structure of a computer system suitable for implementing the method according to an embodiment of the present disclosure is shown.

[0373] like Fig. 9 As shown, the computer system includes a processing unit, which can perform the various methods in the above-mentioned embodiments according to the program stored in the read-only memory (ROM) or the program loaded from the storage part into the random access memory (RAM). In the RAM, various programs and data required for the operation of the computer system are also stored. The processing unit, ROM and RAM are connected to each other through a bus. The input / output (I / O) interface is also connected to the bus.

[0374] The following components are connected to the I / O interface: an input part including a keyboard, a mouse, etc.; an output part including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage part including a hard disk, etc.; and a communication part including a network interface card such as a LAN card, a modem, etc. The communication part performs a communication process via a network such as the Internet. The drive is also connected to the I / O interface as needed. Removable media, such as disks, optical disks, magneto-optical disks, semiconductor memories, etc., are installed on the drive as needed so that the computer program read therefrom is installed into the storage part as needed. Among them, the processing unit can be implemented as a processing unit such as a CPU, a GPU, a TPU, an FPGA, an NPU, etc.

[0375] In particular, according to an embodiment of the present disclosure, the method described above can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program tangibly contained on a machine-readable medium, and the computer program includes a program code for executing the above method. In such an embodiment, the computer program can be downloaded and installed from a network through a communication part, and / or installed from a removable medium.

[0376] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0377] The units or modules involved in the embodiments described in the present disclosure may be implemented by software or programmable hardware. The units or modules described may also be set in a processor, and the names of these units or modules do not constitute limitations on the units or modules themselves in some cases.

[0378] As another aspect, the present disclosure further provides a computer-readable storage medium, which may be a computer-readable storage medium included in the electronic device or computer system in the above embodiment; or a computer-readable storage medium that exists independently and is not assembled into a device. The computer-readable storage medium stores one or more programs, and the programs are used by one or more processors to execute the method described in the present disclosure.

[0379] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present disclosure is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the inventive concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other.

Claims

1. A battery voltage acquisition method based on Kalman filtering, characterized in that: include: Obtain the process noise covariance and observation noise covariance of the battery to be tested at the previous acquisition time of the specified acquisition time; Establishing the basic state equation and basic observation equation of the battery to be tested; Establishing a voltage a priori estimation equation, a voltage observation estimation equation, and a Kalman gain equation of the battery to be tested according to the basic state equation, the basic observation equation, the process noise covariance and the observation noise covariance at the previous acquisition moment of the specified acquisition moment; Determine a voltage a priori estimate value, a voltage observation estimate value and a Kalman gain at a specified acquisition time according to the voltage a priori estimate equation, the voltage observation estimate equation and the Kalman gain equation; Determine the optimal voltage estimate of the battery to be tested at the specified acquisition time according to the voltage a priori estimate, the voltage observation estimate and the Kalman gain at the specified acquisition time; Acquire measurement innovation information at multiple historical collection moments and Kalman gains at multiple historical collection moments, wherein the multiple historical collection moments include the designated collection moment and collection moments before the designated collection moment; A forgetting factor and a fading factor are set, and a weighting coefficient of a collection moment before a specified collection moment is determined according to the forgetting factor; the fading factor is used to set the weights between the measurement innovations of the multiple historical collection moments; the calculation formula of the fading factor is as follows: N i =2(m+1-i) / [m(m+1)],i=1,2,...,m; Among them, N i is the i-th fading factor from the specified collection moment, and m is the number of moments included in the historical collection moment; According to the weighting coefficient, the fading factor, the measurement information of the multiple historical collection moments, and the Kalman gains of the multiple historical collection moments, the process noise covariance and the observation noise covariance of the previous collection moment of the specified collection moment are updated to the process noise covariance of the specified collection moment and the observation noise covariance of the specified collection moment; the process noise covariance of the specified collection moment and the observation noise covariance of the specified collection moment are used to calculate the optimal voltage estimate value of the next collection moment of the specified collection moment.

2. The method according to claim 1, characterized in that The updating of the process noise covariance at the designated collection moment and the observation noise covariance at the designated collection moment by using the process noise covariance and the observation noise covariance at the previous collection moment of the designated collection moment according to the weighting coefficient, the fading factor, the measurement innovations at the multiple historical collection moments, and the Kalman gains at the multiple historical collection moments comprises: Update the process noise covariance of the designated acquisition moment according to the process noise covariance of the previous acquisition moment, the weighting coefficient, the fading factor, the measurement innovations of the multiple historical acquisition moments, the Kalman gains of the multiple historical acquisition moments, and the product of the weighting coefficient and the Kalman gains of the multiple historical acquisition moments; The observed noise covariance at the specified acquisition moment is updated according to the observed noise covariance at the previous acquisition moment, the weighting coefficient, the fading factor, the measurement innovations at the multiple historical acquisition moments, and the product of the weighting coefficient and the measurement innovations at the multiple historical acquisition moments.

3. The method according to claim 2, characterized in that: The weighting coefficient is calculated using the following formula: d k-1 =(1-b) / (1-b k ); Among them, k is the specified collection time, k-1 is the collection time before the specified collection time, b is the forgetting factor, d k-1 is the weighting coefficient at time k-1.

4. The method according to claim 3, characterized in that: The process noise covariance is updated using the following formula: Q k (1-d k-1 )Q k-1 +d k-1 [Kg k ,Kg k-1 ,...,Kg k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T {[Kg k ,Kg k-1 ,...,Kg k- m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T } T ; Among them, Q k is the process noise covariance at time k, Q k-1 is the process noise covariance at time k-1, Kg k ,Kg k-1 ,...,Kg k-m+1 is the Kalman gain corresponding to m historical acquisition moments, e k is the measurement information at time k, Kg k is the Kalman gain at time k, e k-m+1 is the measurement innovation at time k-m+1, T is the transposition operation, N1e k ,N2e k-1 ,...,N m e k-m+1 It is the product of the measurement innovation corresponding to m historical acquisition moments and the corresponding fading factor; The observation noise covariance at time k is updated using the following formula: R k =(1-d k-1 )R k-1 +d k-1 [N1e k ,N2e k-1 ,...,N m yes k-m+1 ][N1e k ,N2e k-1 ,...,N m yes k-m+1 ] T ; Among them, R k is the observation noise covariance at time k, R k-1 is the observation noise covariance at time k-1.

5. The method according to claim 1, characterized in that The voltage a priori estimation equation, voltage observation estimation equation, and Kalman gain equation of the battery to be tested are established according to the basic state equation, the basic observation equation, the process noise covariance and the observation noise covariance of the previous acquisition time at the specified acquisition time, including: According to the basic state equation and the basic observation equation, respectively establishing a voltage a priori estimation equation and a voltage observation estimation equation of the battery to be tested; Establishing an error covariance calculation equation according to the basic state equation and the process noise covariance at a previous acquisition moment of the specified acquisition moment; The Kalman gain equation is established according to the error covariance calculation equation and the observation noise covariance at a collection time before the designated collection time.

6. The method according to claim 5, characterized in that: The basic state equation calculates the voltage state value at the specified acquisition time according to the voltage state value at the previous acquisition time, the system state parameter, and the system process noise at the specified acquisition time; The basic observation equation calculates the voltage observation value at the specified acquisition time according to the voltage observation value at the specified acquisition time, the system observation parameter, and the system observation noise at the specified acquisition time; The voltage a priori estimation equation calculates the a priori voltage estimation value at the specified acquisition time according to the voltage state value at the previous acquisition time and the system state parameters; The voltage observation estimation equation calculates the voltage observation estimation value at the specified collection time according to the a priori voltage estimation value at the specified collection time, the system observation noise, and the voltage collection value at the specified collection time; The error covariance calculation equation calculates the optimal estimated covariance at the specified acquisition time according to the covariance at the previous acquisition time, the system state parameter, and the process noise covariance at the previous acquisition time. The Kalman gain equation calculates the Kalman gain at the specified acquisition time according to the optimal estimated covariance at the specified acquisition time, the system observation parameters, and the observation noise covariance at the previous acquisition time of the specified acquisition time.

7. The method according to claim 6, characterized in that Determining the optimal voltage estimate of the battery to be tested at the specified acquisition time according to the voltage a priori estimate at the specified acquisition time, the voltage observation estimate at the specified acquisition time, and the Kalman gain at the specified acquisition time, including: Obtaining a voltage collection value of the battery to be tested at a specified collection time; Calculate the measurement information at the specified collection time according to the voltage collection value at the specified collection time and the voltage observation estimate value at the specified collection time; The optimal voltage estimate at the specified acquisition time is calculated according to the a priori voltage estimate at the specified acquisition time, the Kalman gain at the specified acquisition time, and the measurement innovation at the specified acquisition time.

8. The method according to claim 7, characterized in that: The basic state equation is shown in the following formula: X k =AX k-1 +ω k ; Among them, k is the specified collection time, k-1 is the collection time before the specified collection time, and X k is the voltage state value at time k, X k-1 is the voltage state value at time k-1, ω k is the process noise at time k, A is the system state parameter; The basic observation equation is shown in the following formula: WITH k =CX k +υ k ; Among them, Z k is the voltage observation value at time k, C is the system observation parameter, υ k is the observation noise at time k.

9. The method according to claim 8, characterized in that The process noise and the observation noise are both Gaussian white noise; The system state parameter A and the system observation parameter C are both single-element matrices with one element.

10. The method according to claim 8, characterized in that: The voltage prior estimation equation is shown in the following formula: X k '=AX k-1 ; Among them, X k ' is the prior voltage estimate at time k; The voltage observation estimation equation is shown in the following formula: From k '=CX k '; Among them, Z k ' is the estimated value of voltage observation at time k; The error covariance calculation equation is shown in the following formula: P k '=AP k-1 From T +Q k-1 ; Among them, P k ' is the optimal estimated covariance at time k, P k-1 is the covariance at time k-1, Q k-1 is the process noise covariance at time k-1, T is the transposition operation; The calculation method of calculating the measurement innovation at the specified collection time is as follows: and k =And k -Z k '=And k -CX k '; Among them, e k is the measurement information at time k, Y k is the voltage collection value at time k; The Kalman gain equation is shown in the following formula: Kg k =P k 'C T / (CP k 'C T +R k-1 ); Among them, Kg k is the Kalman gain at time k, R k-1 is the observation noise covariance at time k-1.

11. The method according to claim 10, characterized in that The optimal voltage estimate at the specified acquisition time is calculated according to the a priori voltage estimate at the specified acquisition time, the Kalman gain at the specified acquisition time, and the measurement information at the specified acquisition time. The calculation method is as follows: X k + =X k '+Kg k e k ; Among them, X k + is the optimal voltage estimate at time k.

12. The method according to claim 10, characterized in that The method further comprises: The covariance at the specified acquisition time is updated according to the Kalman gain at the specified acquisition time, the optimal estimated covariance at the specified acquisition time, and the system observation parameters; the covariance at the specified acquisition time is used to calculate the optimal estimated error covariance at the next moment of the specified acquisition time.

13. The method according to claim 12, characterized in that The calculation method of updating the covariance of the specified acquisition time is as follows: P.S k (E-Kg k C)P k 'R Where E is the identity matrix.

14. A battery voltage sampling device based on Kalman filtering, characterized in that: include: A first acquisition module is configured to acquire the process noise covariance and the observation noise covariance of the battery to be tested at a previous acquisition time at a specified acquisition time; A first establishing module is configured to establish a basic state equation and a basic observation equation of the battery to be tested; A second establishing module is configured to establish a voltage a priori estimation equation, a voltage observation estimation equation, and a Kalman gain equation of the battery to be tested according to the basic state equation, the basic observation equation, the process noise covariance and the observation noise covariance at the previous acquisition moment of the specified acquisition moment; A first determination module is configured to determine a voltage a priori estimation value, a voltage observation estimation value, and a Kalman gain at a specified acquisition time according to the voltage a priori estimation equation, the voltage observation estimation equation, and the Kalman gain equation; A second determination module is configured to determine the optimal voltage estimate of the battery to be tested at the specified collection time according to the voltage a priori estimate at the specified collection time, the voltage observation estimate at the specified collection time, and the Kalman gain at the specified collection time; A second acquisition module is configured to acquire measurement innovations at multiple historical acquisition moments and Kalman gains at multiple historical acquisition moments, wherein the multiple historical acquisition moments include the designated acquisition moment and the acquisition moment before the designated acquisition moment; The setting module is configured to set a forgetting factor and a fading factor, and determine a weighting coefficient of a collection moment before a specified collection moment according to the forgetting factor; the fading factor is used to set the weights between the measurement innovations of the multiple historical collection moments; the calculation formula of the fading factor is as follows: N i =2(m+1-i) / [m(m+1)],i=1,2,...,m; Among them, N i is the i-th fading factor from the specified collection moment, and m is the number of moments included in the historical collection moment; The first updating module is configured to update the process noise covariance and the observation noise covariance of the previous acquisition moment of the specified acquisition moment to the process noise covariance of the specified acquisition moment and the observation noise covariance of the specified acquisition moment according to the weighting coefficient, the fading factor, the measurement information of the multiple historical acquisition moments, and the Kalman gain of the multiple historical acquisition moments; the process noise covariance of the specified acquisition moment and the observation noise covariance of the specified acquisition moment are used to calculate the optimal voltage estimate value at the next moment of the specified acquisition moment.

15. The device according to claim 14, characterized in that The first update module includes: A first updating submodule is configured to update the process noise covariance of the specified acquisition moment according to the process noise covariance of the previous acquisition moment of the specified acquisition moment, the weighting coefficient, the fading factor, the measurement innovation of the multiple historical acquisition moments, the Kalman gains of the multiple historical acquisition moments, and the product of the weighting coefficient and the Kalman gains of the multiple historical acquisition moments; The second updating submodule is configured to update the observed noise covariance at the specified acquisition moment according to the observed noise covariance at the previous acquisition moment, the weighting coefficient, the fading factor, the measurement updates at the multiple historical acquisition moments, and the product of the weighting coefficient and the measurement updates at the multiple historical acquisition moments.

16. The device according to claim 15, characterized in that: The weighting coefficient is calculated using the following formula: d k-1 =(1-b) / (1-b k ); Among them, k is the specified collection time, k-1 is the collection time before the specified collection time, b is the forgetting factor, d k-1 is the weighting coefficient at time k-1.

17. The device according to claim 16, characterized in that: The process noise covariance is updated using the following formula: Q k (1-d k-1 )Q k-1 +d k-1 [Kg k ,Kg k-1 ,...,Kg k-m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T {[Kg k ,Kg k-1 ,...,Kg k- m+1 ][N1e k ,N2e k-1 ,...,N m e k-m+1 ] T } T ; Among them, Q k is the process noise covariance at time k, Q k-1 is the process noise covariance at time k-1, Kg k ,Kg k-1 ,...,Kg k-m+1 is the Kalman gain corresponding to m historical acquisition moments, e k is the measurement information at time k, Kg k is the Kalman gain at time k, e k-m+1 is the measurement innovation at time k-m+1, T is the transposition operation, N1e k ,N2e k-1 ,...,N m e k-m+1 It is the product of the measurement innovation corresponding to m historical acquisition moments and the corresponding fading factor; The observation noise covariance at time k is updated using the following formula: R k =(1-d k-1 )R k-1 +d k-1 [N1e k ,N2e k-1 ,...,N m yes k-m+1 ][N1e k ,N2e k-1 ,...,N m yes k-m+1 ] T ; Among them, R k is the observation noise covariance at time k, R k-1 is the observation noise covariance at time k-1.

18. The device according to claim 14, characterized in that The second establishing module comprises: A first establishing submodule is configured to establish a voltage a priori estimation equation and a voltage observation estimation equation of the battery to be tested respectively according to the basic state equation and the basic observation equation; A second establishing submodule is configured to establish an error covariance calculation equation according to the basic state equation and the process noise covariance at a previous acquisition moment of the designated acquisition moment; The third establishing submodule is configured to establish the Kalman gain equation according to the error covariance calculation equation and the observation noise covariance at the previous acquisition moment of the designated acquisition moment.

19. The device according to claim 18, characterized in that: The basic state equation calculates the voltage state value at the specified acquisition time according to the voltage state value at the previous acquisition time, the system state parameter, and the system process noise at the specified acquisition time; The basic observation equation calculates the voltage observation value at the specified acquisition time according to the voltage observation value at the specified acquisition time, the system observation parameter, and the system observation noise at the specified acquisition time; The voltage a priori estimation equation calculates the a priori voltage estimation value at the specified acquisition time according to the voltage state value at the previous acquisition time and the system state parameters; The voltage observation estimation equation calculates the voltage observation estimation value at the specified collection time according to the a priori voltage estimation value at the specified collection time, the system observation noise, and the voltage collection value at the specified collection time; The error covariance calculation equation calculates the optimal estimated covariance at the specified acquisition time according to the covariance at the previous acquisition time, the system state parameter, and the process noise covariance at the previous acquisition time. The Kalman gain equation calculates the Kalman gain at the specified acquisition time according to the optimal estimated covariance at the specified acquisition time, the system observation parameters, and the observation noise covariance at the previous acquisition time of the specified acquisition time.

20. The device according to claim 19, characterized in that The second determining module includes: An acquisition submodule, configured to acquire a voltage acquisition value of the battery to be tested at a specified acquisition time; A first calculation submodule is configured to calculate the measurement information at the specified collection time according to the voltage collection value at the specified collection time and the voltage observation estimation value at the specified collection time; The second calculation submodule is configured to calculate the optimal voltage estimate at the specified collection time according to the a priori voltage estimate at the specified collection time, the Kalman gain at the specified collection time, and the measurement innovation at the specified collection time.

21. The device according to claim 20, characterized in that: The basic state equation is shown in the following formula: X k =AX k-1 +ω k ; Among them, k is the specified collection time, k-1 is the collection time before the specified collection time, and X k is the voltage state value at time k, X k-1 is the voltage state value at time k-1, ω k is the process noise at time k, A is the system state parameter; The basic observation equation is shown in the following formula: WITH k =CX k +υ k ; Among them, Z k is the voltage observation value at time k, C is the system observation parameter, υ k is the observation noise at time k.

22. The device according to claim 21, characterized in that: The process noise and the observation noise are both Gaussian white noise; The system state parameter A and the system observation parameter C are both single-element matrices with one element.

23. The device according to claim 21, characterized in that: The voltage prior estimation equation is shown in the following formula: X k '=AX k-1 ; Among them, X k ' is the prior voltage estimate at time k; The voltage observation estimation equation is shown in the following formula: From k '=CX k '; Among them, Z k ' is the estimated value of voltage observation at time k; The error covariance calculation equation is shown in the following formula: P k '=AP k-1 From T +Q k-1 Among them, P k ' is the optimal estimated covariance at time k, P k-1 is the covariance at time k-1, Q k-1 is the process noise covariance at time k-1, T is the transposition operation; The calculation method of calculating the measurement innovation at the specified collection time is as follows: and k =And k -Z k '=And k -CX k '; Among them, e k is the measurement information at time k, Y k is the voltage collection value at time k; The Kalman gain equation is shown in the following formula: Kg k =P k 'C T / (CP k 'C T +R k-1 ); Among them, Kg k is the Kalman gain at time k, R k-1 is the observation noise covariance at time k-1.

24. The device according to claim 23, characterized in that The optimal voltage estimate at the specified acquisition time is calculated according to the a priori voltage estimate at the specified acquisition time, the Kalman gain at the specified acquisition time, and the measurement information at the specified acquisition time. The calculation method is as follows: X k + =X k '+Kg k e k ; Among them, X k + is the optimal voltage estimate at time k.

25. The device according to claim 23, characterized in that The device also includes: The second updating module is configured to update the covariance at the specified acquisition time according to the Kalman gain at the specified acquisition time, the optimal estimated covariance at the specified acquisition time, and the system observation parameters; the covariance at the specified acquisition time is used to calculate the optimal estimated error covariance at the next moment of the specified acquisition time.

26. The method according to claim 25, characterized in that The calculation method of updating the covariance of the specified acquisition time is as follows: P.S k (E-Kg k C)P k 'R Where E is the identity matrix.

27. An electronic device, characterized in that: The method comprises a memory and a processor; wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the method steps described in any one of claims 1-13.

28. A computer-readable storage medium having computer instructions stored thereon, characterized in that: When the computer instructions are executed by a processor, the method steps described in any one of claims 1 to 13 are implemented.

Images