Electricity calculation method, system, medium and electronic equipment based on LMS self-adaptation

Abstract

The present invention provides a method, system, medium and electronic device for calculating power based on LMS adaptation, which relates to the technical field of lithium-ion batteries, including: using EMS to obtain a third resistance value in real time, and determining the initial health state of the battery at the current moment according to the preset first resistance value, the second resistance value and the third resistance value at the current moment; obtaining the initial health state at consecutive preset moments within a preset time period as the original signal, performing adaptive filtering, and obtaining the target health state at each preset moment; determining the target power that the system can discharge within the preset time period according to the target health state at each preset moment, the total power when the battery is fully charged, the discharge efficiency of the system, the depth of battery discharge and the state of charge at each preset moment. The present invention solves the problems of certain hysteresis and poor reliability when calculating the battery power in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium-ion batteries, and in particular to a method, system, medium and electronic equipment for calculating electric quantity based on LMS self-adaptation. Background Art

[0002] In energy storage systems, EMS (Energy Management System, referred to as EMS) is responsible for energy scheduling. In industrial and commercial energy storage systems, the main energy scheduling strategies include peak shaving and valley filling, load tracking, reverse flow prevention, overcapacity prevention, etc. In order to accurately complete these control strategies, accurate battery power is required.

[0003] However, the detection of battery state of charge (SOH) and subsequent calculation of battery power in existing energy storage systems are generally implemented in the cloud, resulting in a certain lag and poor reliability when using traditional methods to calculate battery power. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a method, system, medium and electronic device for calculating power based on LMS adaptation, which solves the problems of certain lag and poor reliability when calculating battery power in the prior art.

[0005] At least one embodiment of the present invention provides a method for calculating power based on LMS adaptation, including:

[0006] The third resistance value is obtained in real time by using the EMS, and the initial health state of the battery at the current moment is determined according to the preset first resistance value, the second resistance value and the third resistance value at the current moment; wherein the first resistance value represents the internal resistance corresponding to the battery when the retirement condition is met, the second resistance value represents the nominal internal resistance of the battery, and the third resistance value represents the measured internal resistance of the battery;

[0007] Acquire the initial health state at consecutive preset moments within a preset time period as the original signal, perform adaptive filtering, and obtain the target health state at each preset moment;

[0008] The target power that the system can discharge within the preset time period is determined based on the target health status at each preset time, the total power of the battery when it is fully charged, the discharge efficiency of the system, the depth of battery discharge and the state of charge at each preset time.

[0009] The technical solution disclosed in the present invention has at least the following beneficial effects:

[0010] The present invention utilizes the third resistance value obtained in real time by the local EMS, combined with the first resistance value and the second resistance value, to obtain the target health state of the battery, and determines the target amount of electricity that the system can discharge within a preset time period according to the target health state, so that the EMS breaks the tradition and can accurately calculate the remaining power of the local battery while fully considering the impact of the battery health, making the feedback of the remaining power more timely and reliable.

[0011] In one embodiment of the present invention, the method for calculating power based on LMS adaptation further includes:

[0012] Obtaining the load power of the transformer within the preset time period according to the change rate of the target electric quantity within the preset time period;

[0013] According to the difference between the load power and the rated output power of the transformer, the output power of the transformer is compensated according to a preset strategy.

[0014] The technical solution disclosed in the present invention has at least the following beneficial effects:

[0015] According to the obtained load power, the corresponding control unit can be used to compensate for the output power state of the transformer and optimize the output of the transformer, so that the power output can be more accurately controlled when the transformer is reduced in capacity and supplemented.

[0016] In the LMS adaptive power calculation method provided in one embodiment of the present invention, the calculation formula of the initial health state of the battery is as follows:

[0017] SOH0=[(R E -R N ) / (R E -R B )]×100%

[0018] Wherein, SOH0 represents the initial health state of the battery, R E Represents the first resistance, R B Represents the second resistance, R N Indicates the third resistance value.

[0019] The technical solution disclosed in the present invention has at least the following beneficial effects:

[0020] Using the above formula, we can combine the internal resistance of the battery when it meets the retirement conditions, the nominal internal resistance of the battery, and the measured internal resistance of the battery to obtain the initial health status of the battery, and refer to the calculation of subsequent methods to make the assessment of battery health more accurate and reliable.

[0021] In the LMS adaptive power calculation method provided in one embodiment of the present invention, the adaptive filtering process includes:

[0022] S1, obtaining the output of the filter at the moment before the current moment and the preset expected output of the filter;

[0023] S2, subtracting the output of the filter at the previous moment from the preset expected output of the filter to determine the error of the filter;

[0024] S3, obtaining the filter weight and initial health state at the previous moment, and presetting the learning coefficient, combining the error, the filter weight and initial health state at the previous moment, to determine the filter weight at the current moment;

[0025] S4. According to the filter weight and initial health state at the current moment, determine the output of the filter at the current moment as the target health state at the current moment.

[0026] The technical solution disclosed in the present invention has at least the following beneficial effects:

[0027] After adopting the above steps, the weight of the filter can be adjusted continuously by feedback, and then the output of the filter can be adjusted adaptively, so as to filter out the interference noise in the original signal more accurately and effectively.

[0028] In the LMS adaptive power calculation method provided in one embodiment of the present invention, a variable coefficient method is used to preset the range of the learning coefficient, and a value within the range that meets a preset condition is used as the learning coefficient.

[0029] The technical solution disclosed in the present invention has at least the following beneficial effects:

[0030] By using the variable coefficient method, a relatively large learning coefficient can be used in the early stage of adaptive filtering, that is, when the algorithm is not stable, to speed up the stabilization of the algorithm. In the later stage when the algorithm tends to be stable, a relatively small learning technology is used to fine-tune the weight of the filter.

[0031] In the LMS adaptive power calculation method provided in one embodiment of the present invention, a value within the range that meets a preset condition is used as the learning coefficient, specifically including:

[0032] Gaussian white noise is doped into the original signal, and the convergence curves of the corresponding multiple filter outputs are determined by setting different learning coefficients, and the specific values ​​of the learning coefficients are determined according to the convergence curves.

[0033] The technical solution disclosed in the present invention has at least the following beneficial effects:

[0034] Using the above method, the specific value of the learning coefficient can be accurately determined.

[0035] At least one embodiment of the present invention further provides an LMS-based adaptive power calculation system, comprising:

[0036] The data acquisition module uses the EMS to acquire the third resistance value in real time, and determines the initial health state of the battery at the current moment according to the preset first resistance value, the second resistance value and the third resistance value at the current moment; wherein the first resistance value represents the internal resistance corresponding to the battery when the retirement condition is met, the second resistance value represents the nominal internal resistance of the battery, and the third resistance value represents the measured internal resistance of the battery;

[0037] A filtering module obtains the initial health state at consecutive preset moments within a preset time period as an original signal, performs adaptive filtering, and obtains the target health state at each preset moment;

[0038] The target power determination module determines the target power that the system can discharge within a preset time period based on the target health status at each preset time, the total power when the battery is fully charged, the discharge efficiency of the system, the depth of battery discharge and the state of charge at each preset time.

[0039] In one embodiment of the present invention, the LMS adaptive power calculation system further includes:

[0040] A load power determination module, which obtains the load power of the transformer within a preset time period according to the change rate of the target power within the preset time period;

[0041] The compensation module compensates the output power of the transformer according to a preset strategy based on the difference between the load power and the rated output power of the transformer.

[0042] At least one embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, and when the instructions are executed on a terminal device, the terminal device executes the steps of the above-mentioned LMS-based adaptive power calculation method.

[0043] At least one embodiment of the present invention further provides an electronic device, including a memory, a processor, and a program stored in the memory and running on the processor, wherein when the processor executes the program, the steps of the above-mentioned LMS-based adaptive power calculation method are implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A schematic diagram of the flow chart of the LMS-based adaptive power calculation method provided by the present invention;

[0045] Figure 2 A schematic diagram of the flow chart of the LMS-based adaptive power calculation system provided by the present invention. DETAILED DESCRIPTION

[0046] The principles and features of the present invention are described below. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention.

[0047] Furthermore, those skilled in the art may combine and associate different embodiments or examples and features of different embodiments or examples described in this specification without mutual contradiction.

[0048] The present invention provides a method for calculating power based on LMS self-adaptation. Figure 1 As shown, including:

[0049] In the energy storage system, the EMS (Energy Management System) communicates with the Battery Management System (BMS) through RS485 or CAN (Controller Area Network, CAN) or Ethernet, and can obtain information such as battery cell voltage, battery temperature, and battery internal resistance from the BMS.

[0050] The internal resistance of the battery obtained in real time by the EMS when the battery meets the retirement condition is used as the first resistance value, the nominal internal resistance of the battery is used as the second resistance value, the measured internal resistance of the battery is used as the third resistance value, and the initial health state of the battery at the current moment is determined according to the preset first resistance value, the second resistance value and the third resistance value at the current moment;

[0051] In order to prevent the battery operating conditions and environmental factors from affecting the calculation accuracy of the initial health state of the battery, it is also necessary to perform adaptive filtering on the obtained initial health state;

[0052] That is, the initial health state at consecutive preset moments within a preset time period is obtained as the original signal, and adaptive filtering is performed to obtain the target health state at each preset moment;

[0053] According to the target health status at each preset time, the total power of the battery when it is fully charged, the discharge efficiency of the system, the depth of battery discharge and the state of charge at each preset time, the target power that the system can discharge within the preset time period is determined. The formula is as follows:

[0054] Q=Q0×SOH×DOD×SOC×η

[0055] Among them, Q represents the target power that the system can discharge, Q0 represents the total power when the battery is fully charged, SOH represents the target health state, DOD represents the depth of discharge of the battery at the current moment, SOC represents the current state of charge of the battery, and η represents the discharge efficiency of the system;

[0056] After adopting the above scheme: using the third resistance value obtained in real time by the local EMS, combined with the first resistance value and the second resistance value, the target health state of the battery can be obtained, and based on the target health state, the target amount of electricity that the system can discharge within a preset time period is determined. This allows EMS to break the tradition and accurately calculate the remaining power of the local battery while fully considering the impact of battery health, making the feedback of the remaining power more timely and reliable.

[0057] Furthermore, the method further comprises:

[0058] According to the change rate of the target electric quantity in the preset time period, the load power of the transformer in the preset time period is obtained;

[0059] According to the difference between the load power and the rated output power of the transformer, the output power of the transformer is compensated according to the preset strategy.

[0060] Specifically, the above preset strategies include:

[0061] Set Pt as the rated output power of the transformer and Pl as the load power.

[0062] (1) When Pt>Pl, it indicates that the transformer capacity is sufficient for the load. At this time, the local EMS only needs to perform charging and discharging according to the peak shaving and valley filling strategy.

[0063] (2) When Pt<Pl, it indicates that the transformer capacity is insufficient for load use. At this time, the local EMS will reduce the capacity of the transformer to prevent the transformer from being damaged by overcapacity. The energy storage system supplement power is at least Pl-Pt. When the capacity reduction demand supplement is performed, the energy storage system power reaches the maximum power, and the transformer output reaches 1.2 times the rated power. At this time, the energy storage system exceeds the upper limit of the transformer capacity reduction supplement, and the EMS issues a transformer overcapacity alarm.

[0064] Combined with the above calculation of the battery health status, the available power of the battery can be calculated more accurately, so that the output power can be controlled more accurately to reduce the capacity of the transformer.

[0065] Specifically, the calculation formula for the initial health state of the battery is as follows:

[0066] SOH0=[(R E -R N ) / (R E -R B )]×100%

[0067] Among them, SOH0 represents the initial health state of the battery, R E Represents the first resistance, R B Represents the second resistance, RN Indicates the third resistance value.

[0068] Using the above formula, we can combine the internal resistance of the battery when it meets the retirement conditions, the nominal internal resistance of the battery, and the measured internal resistance of the battery to obtain the initial health status of the battery, and refer to the calculation of subsequent methods to make the assessment of battery health more accurate and reliable.

[0069] Specifically, the process of adaptive filtering includes:

[0070] S1, obtaining the output of the filter at the moment before the current moment and the preset expected output of the filter;

[0071] S2, subtracting the output of the filter at the previous moment from the preset expected output of the filter to determine the error of the filter;

[0072] S3, obtaining the filter weight and initial health state at the previous moment, and presetting the learning coefficient, combining the error, the filter weight at the previous moment and the initial health state, to determine the filter weight at the current moment;

[0073] S4. According to the filter weight and initial health state at the current moment, determine the output of the filter at the current moment as the target health state at the current moment.

[0074] That is, if the original input signal is set to s, the output of the filter can be expressed as:

[0075]

[0076] In the formula, ω(n) represents the filter weight corresponding to time series n, T represents the transposition operation; s(n) represents the initial health state value corresponding to time series n.

[0077] At this time, the error of the filter can be expressed as:

[0078] e(n)=d(n)-y(n)

[0079] Where d(n) represents the desired output of the filter.

[0080] Then, the filter weights are updated to make the actual output of the filter converge to the expected output. The calculation of the updated weights is:

[0081] ω(n+1)=ω(n)+μ*e(n)*s(n)

[0082] ∫ x |f n (s)-f(s)|dμ→0

[0083] In the formula, ω(n+1) represents the filter weight corresponding to the time series n+1, and μ represents the learning coefficient, which is used to adjust the learning speed of the LMS algorithm. The larger the value of μ, the faster the learning speed, but it will lead to an increase in the filter error e(n); conversely, the smaller the value of μ, the slower the learning speed, but it will reduce the filter error e(n).

[0084] After adopting the above steps, the weight of the filter can be adjusted continuously by feedback, and then the output of the filter can be adjusted adaptively, so as to filter out the interference noise in the original signal more accurately and effectively.

[0085] Specifically, the range of the learning coefficient is preset by using the variable coefficient method, and the value within the range that meets the preset conditions is used as the learning coefficient.

[0086] By using the variable coefficient method, a relatively large learning coefficient can be adopted in the early stage of adaptive filtering, that is, when the algorithm is not stable, to speed up the stability of the algorithm. In the later stage, when the algorithm tends to be stable, a relatively small learning technology is used to fine-tune the weight of the filter. Here, mean convergence is used until the value range of the learning coefficient is determined as follows:

[0087]

[0088] In the formula, λ max It is expressed as the maximum eigenvalue of the correlation matrix of the original signal.

[0089] Furthermore, the values ​​within the range that meet the preset conditions are used as learning coefficients, including:

[0090] The specific value of μ is determined by combining Matlab simulation. The operation is as follows: Gaussian white noise is doped into the original signal, and the convergence curves of the corresponding multiple filter outputs are determined by setting different learning coefficients, and the specific value of the learning coefficient is determined according to the convergence curve;

[0091] Through the above operations, it can be determined that the value of μ in the early stage is 0.01, and the value of μ in the later stage is 0.001.

[0092] The present invention also provides a power calculation system based on LMS self-adaptation, please refer to Figure 2 As shown, including:

[0093] The data acquisition module uses the EMS to acquire the third resistance value in real time, and determines the initial health state of the battery at the current moment according to the preset first resistance value, the second resistance value and the third resistance value at the current moment; wherein the first resistance value represents the internal resistance corresponding to the battery when the retirement condition is met, the second resistance value represents the nominal internal resistance of the battery, and the third resistance value represents the measured internal resistance of the battery;

[0094] A filtering module obtains the initial health state at consecutive preset moments within a preset time period as an original signal, performs adaptive filtering, and obtains the target health state at each preset moment;

[0095] The target power determination module determines the target power that the system can discharge within a preset time period based on the target health status at each preset time, the total power when the battery is fully charged, the discharge efficiency of the system, the depth of battery discharge and the state of charge at each preset time.

[0096] The system also includes:

[0097] The load power determination module obtains the load power of the transformer within a preset time period according to the change rate of the target power within the preset time period;

[0098] The compensation module compensates the output power of the transformer according to a preset strategy based on the difference between the load power and the rated output power of the transformer.

[0099] Furthermore, in the data acquisition module, the calculation formula for the initial health status of the battery is as follows:

[0100] SOH0=[(R E -R N ) / (R E -R B )]×100%

[0101] Among them, SOH0 represents the initial health state of the battery, R E Represents the first resistance, R B Represents the second resistance, R N Indicates the third resistance value.

[0102] Furthermore, the process of adaptive filtering in the filtering module includes:

[0103] S1, obtaining the output of the filter at the moment before the current moment and the preset expected output of the filter;

[0104] S2, subtracting the output of the filter at the previous moment from the preset expected output of the filter to determine the error of the filter;

[0105] S3, obtaining the filter weight and initial health state at the previous moment, and presetting the learning coefficient, combining the error, the filter weight at the previous moment and the initial health state, to determine the filter weight at the current moment;

[0106] S4. According to the filter weight and initial health state at the current moment, determine the output of the filter at the current moment as the target health state at the current moment.

[0107] Furthermore, the filtering module uses a variable coefficient method to preset a range of the learning coefficient, and uses a value within the range that meets a preset condition as the learning coefficient.

[0108] Furthermore, the filtering module uses the values ​​within the range that meet the preset conditions as learning coefficients, specifically including:

[0109] Gaussian white noise is doped into the original signal, and the convergence curves of the corresponding multiple filter outputs are determined by setting different learning coefficients, and the specific values ​​of the learning coefficients are determined according to the convergence curves.

[0110] The present invention also provides a computer-readable storage medium, in which instructions are stored. When the instructions are executed on a terminal device, the terminal device executes the steps of the above-mentioned LMS-based adaptive power calculation method.

[0111] The present invention also provides an electronic device, including a memory, a processor, and a program stored in the memory and running on the processor. When the processor executes the program, the steps of the above-mentioned LMS-based adaptive power calculation method are implemented.

[0112] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. The power calculation method based on LMS self-adaptation is characterized by: include: The third resistance value is obtained in real time by using the EMS, and the initial health state of the battery at the current moment is determined according to the preset first resistance value, the second resistance value and the third resistance value at the current moment; wherein the first resistance value represents the internal resistance corresponding to the battery when the retirement condition is met, the second resistance value represents the nominal internal resistance of the battery, and the third resistance value represents the measured internal resistance of the battery; Acquire the initial health state at consecutive preset moments within a preset time period as the original signal, perform adaptive filtering, and obtain the target health state at each preset moment; Determine the target power that the system can discharge within a preset time period based on the target health status at each preset time, the total power of the battery when it is fully charged, the discharge efficiency of the system, the depth of battery discharge and the state of charge at each preset time; The process of the adaptive filtering includes: S1, obtaining the output of the filter at the moment before the current moment and the preset expected output of the filter; S2, subtracting the output of the filter at the previous moment from the preset expected output of the filter to determine the error of the filter; S3, obtaining the filter weight and initial health state at the previous moment, and presetting the learning coefficient, combining the error, the filter weight and initial health state at the previous moment, to determine the filter weight at the current moment; S4. According to the filter weight and initial health state at the current moment, determine the output of the filter at the current moment as the target health state at the current moment.

2. The method for calculating power based on LMS self-adaptation according to claim 1, characterized in that: Also includes: Obtaining the load power of the transformer within the preset time period according to the change rate of the target electric quantity within the preset time period; According to the difference between the load power and the rated output power of the transformer, the output power of the transformer is compensated according to a preset strategy.

3. The method for calculating power based on LMS self-adaptation according to claim 1, characterized in that: The calculation formula of the initial health state of the battery is as follows: SOH0=[(R E -R N ) / (R E -R B )]×100% Wherein, SOH0 represents the initial health state of the battery, R E Represents the first resistance, R B Represents the second resistance, R N Indicates the third resistance value.

4. The method for calculating power based on LMS self-adaptation according to claim 1, characterized in that: The range of the learning coefficient is preset by using a variable coefficient method, and a value within the range that meets a preset condition is used as the learning coefficient.

5. The method for calculating power based on LMS self-adaptation according to claim 4, characterized in that: The value that meets the preset conditions within the range is used as the learning coefficient, specifically including: Gaussian white noise is doped into the original signal, and the convergence curves of the corresponding multiple filter outputs are determined by setting different learning coefficients, and the specific values ​​of the learning coefficients are determined according to the convergence curves.

6. The LMS-based adaptive power calculation system is characterized by: include: The data acquisition module uses the EMS to acquire the third resistance value in real time, and determines the initial health state of the battery at the current moment according to the preset first resistance value, the second resistance value and the third resistance value at the current moment; wherein the first resistance value represents the internal resistance corresponding to the battery when the retirement condition is met, the second resistance value represents the nominal internal resistance of the battery, and the third resistance value represents the measured internal resistance of the battery; A filtering module obtains the initial health state at consecutive preset moments within a preset time period as an original signal, performs adaptive filtering, and obtains the target health state at each preset moment; The target power determination module determines the target power that the system can discharge within a preset time period according to the target health status at each preset time, the total power of the battery when it is fully charged, the discharge efficiency of the system, the depth of battery discharge and the state of charge at each preset time; The process of adaptive filtering in the filtering module includes: S1, obtaining the output of the filter at the moment before the current moment and the preset expected output of the filter; S2, subtracting the output of the filter at the previous moment from the preset expected output of the filter to determine the error of the filter; S3, obtaining the filter weight and initial health state at the previous moment, and presetting the learning coefficient, combining the error, the filter weight at the previous moment and the initial health state, to determine the filter weight at the current moment; S4. According to the filter weight and initial health state at the current moment, determine the output of the filter at the current moment as the target health state at the current moment.

7. The LMS-based adaptive power calculation system according to claim 6, characterized in that: Also includes: A load power determination module, which obtains the load power of the transformer within a preset time period according to the change rate of the target power within the preset time period; The compensation module compensates the output power of the transformer according to a preset strategy based on the difference between the load power and the rated output power of the transformer.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed on the terminal device, the terminal device executes the steps of the LMS-based adaptive power calculation method as described in any one of claims 1 to 5.

9. An electronic device comprising a memory, a processor, and a program stored in the memory and running on the processor, characterized in that: When the processor executes the program, the steps of the LMS-based adaptive power calculation method according to any one of claims 1 to 5 are implemented.

Images