CN111143973B - Valve-regulated lead-acid storage battery degradation trend prediction method based on Gauss process regression - Google Patents

Valve-regulated lead-acid storage battery degradation trend prediction method based on Gauss process regression Download PDF

Info

Publication number
CN111143973B
CN111143973B CN201911236574.8A CN201911236574A CN111143973B CN 111143973 B CN111143973 B CN 111143973B CN 201911236574 A CN201911236574 A CN 201911236574A CN 111143973 B CN111143973 B CN 111143973B
Authority
CN
China
Prior art keywords
component
internal resistance
voltage
gpr
value
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201911236574.8A
Other languages
Chinese (zh)
Other versions
CN111143973A (en
Inventor
李瑞津
刘斌
邓云书
毕小熊
党军朋
李涛
张学敏
刘祺
郭伟
王斌
胡云
施迎春
岳斌
赵华
叶文华
陈运忠
潘再金
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Yuxi Power Supply Bureau of Yunnan Power Grid Co Ltd
Original Assignee
Yuxi Power Supply Bureau of Yunnan Power Grid Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Yuxi Power Supply Bureau of Yunnan Power Grid Co Ltd filed Critical Yuxi Power Supply Bureau of Yunnan Power Grid Co Ltd
Priority to CN201911236574.8A priority Critical patent/CN111143973B/en
Publication of CN111143973A publication Critical patent/CN111143973A/en
Application granted granted Critical
Publication of CN111143973B publication Critical patent/CN111143973B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01M—PROCESSES OR MEANS, e.g. BATTERIES, FOR THE DIRECT CONVERSION OF CHEMICAL ENERGY INTO ELECTRICAL ENERGY
    • H01M8/00—Fuel cells; Manufacture thereof
    • H01M8/04—Auxiliary arrangements, e.g. for control of pressure or for circulation of fluids
    • H01M8/04298—Processes for controlling fuel cells or fuel cell systems
    • H01M8/04992—Processes for controlling fuel cells or fuel cell systems characterised by the implementation of mathematical or computational algorithms, e.g. feedback control loops, fuzzy logic, neural networks or artificial intelligence
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
    • Y02E60/30—Hydrogen technology
    • Y02E60/50—Fuel cells

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Automation & Control Theory (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Manufacturing & Machinery (AREA)
  • Evolutionary Computation (AREA)
  • Fuzzy Systems (AREA)
  • Medical Informatics (AREA)
  • Health & Medical Sciences (AREA)
  • Secondary Cells (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Sustainable Development (AREA)
  • Sustainable Energy (AREA)
  • Chemical & Material Sciences (AREA)
  • Chemical Kinetics & Catalysis (AREA)
  • Electrochemistry (AREA)
  • General Chemical & Material Sciences (AREA)
  • Theoretical Computer Science (AREA)

Abstract

The invention relates to a degradation trend prediction method of a valve-regulated lead-acid storage battery based on Gaussian process regression, and belongs to the technical field of artificial intelligence for service life prediction of lead-acid storage batteries. The method considers the operation characteristics of uniform charging and floating charging of the storage battery of the transformer substation and the influence of the operation characteristics on the degradation of the storage battery, and adopts GPR to predict the degradation trend of the storage battery. The method provided by the invention realizes accurate prediction of the voltage and the internal resistance of the storage battery of the direct current system of the transformer substation, so that the degradation trend of the storage battery is known, the method can be used for guiding the operation and maintenance of the valve-controlled lead-acid storage battery in the transformer substation of the power grid, and the method is undoubtedly of great significance for improving the safety and reliability of the transformer substation.

Description

Valve-regulated lead-acid storage battery degradation trend prediction method based on Gauss process regression
Technical Field
The invention belongs to the technical field of artificial intelligence for predicting the service life of a lead-acid storage battery, and particularly relates to a valve control type lead-acid storage battery degradation trend prediction method based on Gaussian process regression.
Background
Automation, intellectualization and unattended operation of a power grid transformer substation are main trends of power grid development. Due to the increasingly high degree of automation and intellectualization of the transformer substation and the popularization of unattended operation, the role of the direct-current power supply of the transformer substation is increasingly important. In a transformer substation, a valve-regulated lead-acid storage battery pack of a direct-current system is connected in parallel with a charger and commonly undertakes important direct-current load power supply tasks such as relay protection, an automatic device, automation equipment, a breaker tripping and closing mechanism and the like. Therefore, the valve-regulated lead-acid storage battery pack is the core of a direct-current power supply system, and the performance quality of the valve-regulated lead-acid storage battery pack is related to the safe and stable operation of the whole transformer substation.
The valve-regulated lead-acid storage battery serving as a substation backup power supply is formed by connecting a plurality of valve-regulated lead-acid storage batteries in series, and the most serious condition of the valve-regulated lead-acid storage batteries is that the valve-regulated lead-acid storage batteries are open-circuited in the running process, and the capacity of the valve-regulated lead-acid storage batteries is insufficient. Open circuit will cause the breaker to be unable to trip and close and then burn out transformer or high-voltage chamber, and capacity is not enough then causes direct current load in the transformer substation to drop to very low in voltage in very short time, leads to the power loss and unable work such as relay protection device, automatics, breaker trip and close mechanism, emergency lighting. At present, a fixed valve-regulated lead-acid storage battery is generally used in a transformer substation. The valve-controlled lead-acid storage battery has the advantages of good sealing, high cost performance, no need of supplementing electrolyte and distilled water, no pollution, strong heavy-current discharge capacity and the like. However, the disadvantage of the valve-regulated lead-acid battery is that the requirements for operation maintenance and operation environment are high, and if the environment is severe and the operation maintenance is not timely and not in place, the valve-regulated lead-acid battery often fails in advance. If the severe environment frequently appears and the polling maintenance work is not timely followed, the early failure time limit of the valve-regulated lead-acid storage battery is far ahead of the design value.
With the development of intelligent power grids in China, management targets such as 'monitoring visualization, decision intellectualization, control closed-loop, data platformization' and the like can be achieved for the power grids, but on the aspect of management of a valve-controlled lead-acid storage battery pack of a direct-current system of a transformer substation, the difficulty of pre-judging and pre-processing failure of the battery by evaluating the life cycle of the battery is always a working difficulty of operation and maintenance of the power grids. Grid companies are gradually implementing centralized management of the grid. If a valve-regulated lead-acid storage battery failure analysis model is established by carrying out statistical analysis on the data of the valve-regulated lead-acid storage battery for several years, then the life cycle of the valve-regulated lead-acid storage battery of the transformer substation direct-current system of the whole network is estimated through the model, a set of service life prediction method of the valve-regulated lead-acid storage battery is established, the direct-current valve-regulated lead-acid storage battery of the power grid is guided to operate and maintain, and the method has great significance for cost reduction, efficiency improvement and lean operation of the power grid company and improvement of the safety and reliability of the transformer substation.
Disclosure of Invention
The invention aims to solve the defects of the prior art and provides a degradation trend prediction method of a valve-regulated lead-acid storage battery based on Gaussian process regression. The data of the valve-regulated lead-acid storage battery for years are subjected to statistical analysis, so that the voltage (internal resistance) is predicted, the degradation trend of the storage battery is further known, and the aims of improving the operation and maintenance efficiency and guaranteeing the safe and stable operation of a power grid can be finally fulfilled.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
the degradation trend prediction method of the valve-regulated lead-acid storage battery by Gaussian process regression comprises the following steps:
respectively decomposing the acquired voltage and internal resistance data by adopting an empirical mode decomposition method;
calculating the content of each component obtained by decomposition, and screening the components of the voltage and the internal resistance by taking the content as an index;
taking the value of the screened voltage component at a certain moment, the floating charge time between the moment and the next moment and the value of the average charge times as the input of a GPR model, taking the value of the screened voltage component at the next moment as the output of the GPR model, and training the GPR model of each component to obtain a GPR voltage prediction model of each component;
taking the value of the screened internal resistance component at a certain moment, the floating charge time between the moment and the next moment and the value of the average charge times as the input of a GPR model, taking the value of the screened internal resistance component at the next moment as the output of the GPR model, and training the GPR model of each component to obtain a GPR internal resistance prediction model of each component;
taking each component obtained by EEMD decomposition and content screening of a voltage or a voltage predicted value at a certain moment as input, taking the floating charge duration and the average charge times in a use plan of the storage battery between the moment and the next moment as input, bringing the input to a GPR voltage prediction model of the corresponding component to obtain the voltage predicted value of each component, and accumulating the voltage predicted values of each component to obtain the predicted value of the voltage of the storage battery at the next moment;
and taking each component obtained by EEMD decomposition and content screening of the internal resistance or the predicted value of the internal resistance at a certain moment as input, taking the floating charge time and the average charge times in the use plan of the storage battery from the moment to the next moment as input, bringing the floating charge time and the average charge times to the GPR internal resistance prediction model of the corresponding component to obtain the predicted value of the internal resistance of each component, and accumulating the predicted values of the internal resistance of each component to obtain the predicted value of the internal resistance of the storage battery at the next moment.
Further, preferably, the empirical mode decomposition method is EEMD.
Further, it is preferable that when the content of a component is smaller than the content threshold λ, the component is ignored.
The invention also provides a system for predicting the degradation trend of the valve-regulated lead-acid storage battery by Gaussian process regression, which comprises the following steps:
the first data processing module is used for decomposing the acquired voltage and internal resistance data by adopting an empirical mode decomposition method;
the second data processing module is used for calculating the content rate of each component obtained by decomposition and screening the components of the voltage and the internal resistance by taking the content rate as an index;
the GPR voltage prediction model building module is used for taking the value of the screened voltage component at a certain moment, the floating charge time between the moment and the next moment and the value of the average charge times as the input of the GPR model, taking the value of the screened voltage component at the next moment as the output of the GPR model, and training the GPR model of each component to obtain the GPR voltage prediction model of each component;
the GPR internal resistance prediction model building module is used for taking the value of the screened internal resistance component at a certain moment, the floating charge time between the moment and the next moment and the value of the average charge times as the input of the GPR model, taking the value of the screened internal resistance component at the next moment as the output of the GPR model, and training the GPR model of each component to obtain the GPR internal resistance prediction model of each component;
the voltage prediction module is used for taking the voltage at a certain moment or each component obtained by EEMD decomposition and content screening of the voltage predicted value as input, taking the floating charge duration and the average charge times in the use plan of the storage battery between the moment and the next moment as input, bringing the floating charge duration and the average charge times to a GPR voltage prediction model of the corresponding component to obtain the voltage predicted value of each component, and accumulating the voltage predicted values of each component to obtain the predicted value of the storage battery voltage at the next moment;
and the internal resistance prediction module is used for taking each component obtained by carrying out EEMD decomposition and content screening on the internal resistance or the internal resistance prediction value at a certain moment as input, taking the floating charge duration and the average charge times in the use plan of the storage battery from the moment to the next moment as input, bringing the floating charge duration and the average charge times to the GPR internal resistance prediction model of the corresponding component to obtain the internal resistance prediction value of each component, and accumulating the internal resistance prediction values of each component to obtain the prediction value of the internal resistance of the storage battery at the next moment.
The invention also provides an electronic device comprising a memory, a processor and a computer program stored on the memory and operable on the processor, wherein the processor implements the steps of the method for predicting the degradation trend of a valve-regulated lead-acid battery by gaussian process regression as described above when executing the program.
The invention additionally provides a non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method for predicting a degradation trend of a valve-regulated lead-acid battery as a gaussian process regression as described above.
The method provided by the invention takes the operation characteristics of the uniform charging and floating charging of the storage battery of the transformer substation and the influence of the operation characteristics on the degradation of the storage battery into consideration, and the degradation trend of the storage battery is predicted by adopting GPR. (1) Firstly, an EEMD decomposition method is adopted to decompose the terminal voltage and internal resistance data sequence of the storage battery into a plurality of components, and reasonable components are selected through content analysis. (2) And respectively taking each component of the voltage and internal resistance data and the floating charge duration and the average charge times corresponding to the voltage and internal resistance data as training samples, and training the GPR model to form a plurality of GPR models. (3) And finally, taking the use plan of the storage battery in a period of time in the future, namely the floating charge duration, the equalizing charge times and the like as input, bringing the input into a GPR model for prediction respectively, and accumulating and averaging the obtained results of all groups to obtain the predicted value of the voltage (or the internal resistance) of the storage battery at the next time. The method realizes accurate prediction of the voltage (or internal resistance) of the storage battery of the direct-current system of the transformer substation, and further knows the service life degradation trend of the storage battery.
The method can be used for guiding the operation and maintenance of the valve-controlled lead-acid storage battery in the power grid transformer substation, has great significance for cost reduction, efficiency improvement and lean operation of the power grid company and improvement of the safety and reliability of the transformer substation undoubtedly.
The invention provides a service life prediction method of a valve-regulated lead-acid storage battery based on Gaussian process regression, which considers the influence of uniform charge and floating charge on battery degradation and can predict the voltage and internal resistance of the valve-regulated lead-acid storage battery of a backup power supply of a power grid substation. As shown in fig. 5, the specific method is as follows:
preprocessing of data samples
1. Data decomposition based on empirical mode decomposition (EEMD) method
And respectively carrying out EEMD decomposition on the voltage and internal resistance data, and specifically carrying out the following steps.
The method comprises the following steps: using the collected voltage and internal resistance data samples as original signals x (t), adding amplitude coefficients with uniform frequency k to the original signals x (t) for N times1The gaussian white noise sequence n (t) is as follows:
xj(t)＝x(t)+k1·nj(t) (1)
in the formula (1), j represents the j-th Gaussian white noise addition, and the range of j is 1 to N. n isj(t) represents the j-th added Gaussian white noise sequenceColumn, xj(t) represents the j-th mixed signal.
Step two: finding a mixed signal xj(t) all local maxima and minima points in the signal x, and then constructing an upper envelope and a lower envelope by curve fitting in combination with the respective maxima points, such that the signal xj(t) is enveloped by the upper and lower envelopes.
Step three: their mean curves m (t) can be constructed from the upper and lower envelope curves, and the signal x is usedj(t) subtracting the mean curve m (t) to obtain a decomposition modal component H (t).
Step four: it is determined whether H (t) satisfies imf two constraints: 1) in the whole data segment, the number of the extreme points is equal to or has no more than one difference with the number of the zero crossing points; 2) at any time, the average value of the upper envelope formed by the local maximum point and the lower envelope formed by the local minimum point is zero, that is, the upper and lower envelopes are locally symmetrical with respect to the time axis. If the constraint is satisfied, H (t) is a content modal component (imf); if not, repeating the second to fourth steps by taking H (t) as the mixed signal until the constraint condition is met, so that H (t) is the signal xj(t) a corresponding one of the connotative modal components (imf).
Step five: to obtain xj(t) corresponding ith connotation modal component imfiAfter (t), using the mixed signal xj(t) subtracting imfi(t) recycling steps two to four again with the result obtained as a new signal to obtain xj(t) the corresponding i +1 th connotation modal component imfi+1(t) of (d). And so on, stopping until the residual component r (t) is a monotonic function or the SD (screening threshold, generally 0.2-0.3) is less than the threshold, where SD is:
Figure RE-GDA0002403944630000061
in the formula (2), T represents xj(t) time corresponding to the last data, Hk(t) represents xj(t) the corresponding kth decomposition modal component, in this case
Figure RE-GDA0002403944630000062
Wherein m represents xj(t) number of imf min obtained by decomposition imfj，i(t) represents xj(t) the corresponding i-th connotative modal component, rj(t) represents xj(t) the corresponding margin.
Step six: averaging the imf components after N decompositions, i.e.
Figure RE-GDA0002403944630000063
Figure RE-GDA0002403944630000064
In the expressions (4) and (5), j represents the jth mixed signal and also represents the jth additive white gaussian noise, and ranges from 1 to N. i represents the i-th imf component in the same mixed signal, and ranges from 1 to m. imfj，i(t) denotes the i-th imf component, r, corresponding to the j-th mix signalj(t) represents a margin corresponding to the j-th mix signal. imfi(t) is the average value of the ith imf components obtained by decomposing N different mixed signals, riAnd (t) is the average value of the residuals obtained by decomposing N different mixed signals.
Step seven: and superposing the imf component and the average residue to obtain a final result x' (t):
Figure RE-GDA0002403944630000065
1. content screening
The method comprises the following steps: the original data, imf components and the residual quantity are respectively integrated, and the integration result QV、QR、QHuAs follows:
Figure RE-GDA0002403944630000071
Figure RE-GDA0002403944630000072
Figure RE-GDA0002403944630000073
in the formulae (7), (8), (9), T1The time length corresponding to the voltage data and the internal resistance data is in the unit of days; v is a set of voltage data; r is a set of internal resistance data; huThe first EEMD component corresponds to the voltage data or the internal resistance data.
Step 2: the content ratio of each component is calculated. The content ratio of the component is the proportion of the component to the original signal, and the invention is expressed by the ratio of the integral of the component to the integral of the original signal in the same time. Therefore, the content ratio of each component is:
Figure RE-GDA0002403944630000074
Figure RE-GDA0002403944630000075
in the formulae (10) and (11), ljuRepresenting the content of the u-th component corresponding to the voltage data; lRuRepresenting the content of the u-th component corresponding to the internal resistance data;
Figure RE-GDA0002403944630000076
representing the u-th component corresponding to the voltage data;
Figure RE-GDA0002403944630000077
indicating the u-th component corresponding to the internal resistance data.
And step 3: and (4) component screening. Because the degradation of the storage battery is influenced by various factors such as floating charge time, equalizing charge times, temperature and the like, but the influence of each factor is different, the invention neglects the factor with smaller influence on the degradation of the storage battery, reduces the subsequent calculation amount on the premise of not influencing the precision of the prediction result, and improves the calculation speed of the whole process. For this reason, the present invention sets a content threshold λ (λ is a constant), and when the component content is smaller than λ, it will be ignored.
Voltage (internal resistance) prediction based on Gaussian process regression
GPR model training
Step 1: selecting a training set D ═ (X, Y), where X ═ X1,x2,...,
xM}，Y＝{y1,y2,...,yMIn which xiThe ith 3-dimensional input sample is represented and comprises the value of a certain component of which the voltage (or internal resistance) is preprocessed at a certain moment, the floating charge duration and the value of the average charge times in the next time period, and yiIs the corresponding ith output value, which is the component of voltage (or internal resistance) at ti+1A value of a time of day;
assuming that there is a hidden function f, it is constructed as an f (x)1)，f(x2)，...，f(xM) A set consisting of a mean function m (x) and a kernel function k (x)i,xj) If it is determined that the zero mean function is satisfied with a Gaussian distribution, the process is a Gaussian regression process defined as:
y＝f(xn)+ξn (12)
in the formula, xinAs noise, obeys a mean of 0 and a variance of
Figure RE-GDA0002403944630000081
Normal distribution of (2);
step 2: and (4) selecting a kernel function. In this context, a squared exponential covariance function (SE) is chosen, and the SE kernel function is defined as follows:
Figure RE-GDA0002403944630000082
in this context, the Gaussian process regresses with 3 hyper-parameters, respectively the noise ξnStandard deviation of (a)nAmplitude factor σfLength scale factor l; x is the number ofiAnd xj are two different training input samples.
And step 3: and (4) calculating the hyperparameter. And solving a maximum likelihood estimation equation of the hyper-parameters by using a gradient descent method to obtain the optimal hyper-parameters.
And 4, step 4: calculating covariance matrix elements; input xi、xjSubstitution of K (x)i,xj) The corresponding element K can be obtainedij. The covariance matrix K (X, X) can thus be found:
y～N(0,K(X,X)+σn 2I) (14)
where K (X, X) is the derived kernel covariance matrix and I is the M identity matrix.
Figure RE-GDA0002403944630000083
The variance of the noise, y, is a gaussian process formed by the set of the joint distributions of all observed target values in the sample set.
2. Prediction using GPR
Step 1: given a new sample input x, the corresponding output is y, and according to the bayesian principle, the joint distribution of the output value y and the training set sample y is:
Figure RE-GDA0002403944630000091
k (X, X) is an mxm-order covariance matrix between the test sample X and the training set X, and K (X, X) is an mxm-order covariance matrix between the training set X and the test sample X. k (x, x) equals 1.
Step 2: the corresponding posterior distribution y is calculated, and the predicted output y can be expressed as:
y*|X,y,x*～N(μ,∑) (16)
wherein the content of the first and second substances,
Figure RE-GDA0002403944630000092
Figure RE-GDA0002403944630000093
Figure RE-GDA0002403944630000094
μ is the mean of the predicted distribution, which is actually the predicted value of the test output.
Sigma is the variance of the predicted value mu corresponding to the test point x, and y is the sample of the training set
And step 3: and taking the other components which are reserved as a new group of data as a new data set, and repeating the process. And accumulating the predicted voltage (internal resistance) data to obtain a final result.
Third, prediction of degradation trend of storage battery
Step 1: suppose that the current data of the storage battery corresponds to the time TtI.e. the last set in the training data. When the battery is known to be in the future Tt-Tt+kThe time period is calculated as Tt+1,Tt+2,......,Tt+kFloat charging time T of storage battery between time nodesfAnd the number of uniform charging times Nc。
Step 2: will TtVoltage at time, Tt+1The total floating time and the average charging times of the time are used as the input of a GPR function, and T can be obtainedt+1Voltage at time; will TtInternal resistance at time, Tt+1The floating charge time and the average charge times of the moment are used as the input of a GPR function to obtain Tt+1The internal resistance at that moment;
and step 3: converting the voltage and internal resistance data in the regression function input into the predicted data, updating the floating charge time and the equalizing charge times, substituting the floating charge time and the equalizing charge times into the GPR function again for calculation until T is predictedt+kVoltage and internal resistance at time. And accumulating the predicted values corresponding to the components to obtain an average value, thereby obtaining a final result.
After the model is finished, the state of health of a battery (SOH) is identified and obtained by using the existing data driving method according to the obtained voltage and internal resistance, and when the SOH is less than 80%, the battery is eliminated and replaced
Compared with the prior art, the invention has the beneficial effects that: the invention provides a method for predicting the degradation trend of a storage battery of a direct current system of a transformer substation based on Gaussian process regression. The method takes the influence of floating charge and uniform charge under the direct-current system of the transformer substation on the storage battery into consideration, and the degradation trend of the storage battery is predicted by adopting a GPR method. Compared with the traditional storage battery degradation trend prediction method, the method provided by the invention has higher accuracy in the environment of the transformer substation, is beneficial to the inspection and maintenance of maintenance personnel of the transformer substation, ensures the safe and reliable electricity utilization of equipment in emergency, and can avoid the further deterioration of the performance of the whole storage battery due to the deterioration of individual batteries in the storage battery.
The method realizes accurate prediction of the voltage (internal resistance) of the storage battery of the direct-current system of the transformer substation, further knows the degradation trend of the storage battery, can be used for guiding the operation and maintenance of the valve-controlled lead-acid storage battery in the transformer substation of the power grid, and has great significance for improving the safety and reliability of the transformer substation.
Drawings
FIG. 1 is a flow chart of EEMD decomposition of raw data;
FIG. 2 is a flow chart of component selection based on content;
FIG. 3 is a flow chart for training a GPR model;
FIG. 4 is a GPR based degradation trend prediction flow diagram;
FIG. 5 is a flow chart of the method of the present invention;
FIG. 6 is a schematic diagram of the system of the present invention;
FIG. 7 is a schematic diagram of an electronic device according to the present invention;
FIG. 8 is a graph comparing the actual voltage value with the predicted voltage value according to an embodiment of the present invention;
FIG. 9 is a graph comparing the voltage internal resistance value with the predicted value according to the embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to examples.
It will be appreciated by those skilled in the art that the following examples are illustrative of the invention only and should not be taken as limiting the scope of the invention. The examples do not specify particular techniques or conditions, and are performed according to the techniques or conditions described in the literature in the art or according to the product specifications. The materials or equipment used are not indicated by manufacturers, and all are conventional products available by purchase.
The degradation trend prediction method of the valve-regulated lead-acid storage battery by Gaussian process regression comprises the following steps:
respectively decomposing the acquired voltage and internal resistance data by adopting an empirical mode decomposition method;
calculating the content of each component obtained by decomposition, and screening the components of the voltage and the internal resistance by taking the content as an index;
taking the value of the screened voltage component at a certain moment, the floating charge time between the moment and the next moment and the value of the average charge times as the input of a GPR model, taking the value of the screened voltage component at the next moment as the output of the GPR model, and training the GPR model of each component to obtain a GPR voltage prediction model of each component;
taking the value of the screened internal resistance component at a certain moment, the floating charge time between the moment and the next moment and the value of the average charge times as the input of a GPR model, taking the value of the screened internal resistance component at the next moment as the output of the GPR model, and training the GPR model of each component to obtain a GPR internal resistance prediction model of each component;
taking each component obtained by EEM D decomposition and content screening of a voltage or a voltage predicted value at a certain moment as input, taking the floating charge duration and the average charge times in a use plan of the storage battery between the moment and the next moment as input, bringing the input to a GPR voltage prediction model of the corresponding component to obtain the voltage predicted value of each component, and accumulating the voltage predicted values of each component to obtain the predicted value of the voltage of the storage battery at the next moment;
and taking each component obtained by EEMD decomposition and content screening of the internal resistance or the predicted value of the internal resistance at a certain moment as input, taking the floating charge time and the average charge times in the use plan of the storage battery from the moment to the next moment as input, bringing the floating charge time and the average charge times to the GPR internal resistance prediction model of the corresponding component to obtain the predicted value of the internal resistance of each component, and accumulating the predicted values of the internal resistance of each component to obtain the predicted value of the internal resistance of the storage battery at the next moment.
The implementation steps are as follows:
referring to fig. 1, the EEMD is used for performing multi-scale decomposition, which specifically includes:
s11, adding a Gaussian white noise sequence N (t) with uniform frequency and an amplitude coefficient of k to the original signal x (t) for N times by using the collected voltage and internal resistance data samples as the original signal x (t);
s12, finding out all local maximum values and minimum value points in the mixed signal xj (t), and combining the extreme value points to construct an upper envelope line and a lower envelope line by a curve fitting method, so that the signal xj (t) is enveloped by the upper envelope line and the lower envelope line;
s13, constructing a mean curve m (t) of the upper envelope and the lower envelope, and subtracting the mean curve m (t) from the signal xj (t) to obtain a decomposition modal component H (t);
s14, calculating SD from formula 2 to obtain IMF1, and then using mixed signal xj(t) subtracting IMF1(t), taking the result as a new signal, and stopping the steps until the residual component r (t) is a monotonic function or SD (screening threshold value, generally 0.2-0.3) is smaller than the threshold value;
s15, averaging the imf components after N times of decomposition by the formulas (4) and (5), and finally integrating to obtain the EEMD decomposition result.
Referring to fig. 2, content rate screening is performed, and the specific content includes:
s21, integrating the original data and each component by equations (7) to (9);
s22, calculating the content of each component using the expressions (10) and (11));
s23, the results of equations (10) and (11) are compared with the content threshold λ, and the component having a content greater than λ is retained, while the component having a content less than λ is ignored.
Referring to fig. 3, the GPR model is trained, which comprises the following specific steps:
and S31, obtaining the floating charge duration and the average charge times of the storage battery between the predicted time nodes according to the charging plan.
And S32, taking the current voltage (internal resistance), the floating charge time and the equalizing charge times of the next time node as input, and taking the voltage (internal resistance) of the next time node as output.
S33, initializing three hyper-parameters sigma, sigma f and l of a GPR model. Let σ equal to 0.3, σ f equal to 1.61, and l equal to 0.99769.
S34, solving a maximum likelihood estimation equation of the hyperparameter, namely an equation (16), by using a gradient descent method to obtain the optimal hyperparameter.
S35, the covariance elements are calculated by using the kernel function, and the covariance matrix K (X, X) is obtained.
Referring to fig. 4, the specific steps of the degradation trend prediction are as follows:
step 1: suppose that the current data of the storage battery corresponds to the time TtI.e. the last set in the training data. When the battery is known to be in the future Tt-Tt+kThe time period is calculated as Tt+1,Tt+2,......,Tt+kFloat charging time T of storage battery between time nodesfAnd the number of uniform charging times Nc。
Step 2: will TtInternal resistance at time, Tt—Tt+1Taking the floating charge time and the equalizing charge times as input, bringing the floating charge time and the equalizing charge times to the GPR internal resistance prediction model of the corresponding component to obtain the internal resistance prediction value of each component, and accumulating the internal resistance prediction values of each component to obtain Tt+1Predicting the internal resistance of the storage battery at the moment;
will TtVoltage at time, Tt—Tt+1Taking the floating charge time and the equalizing charge times as input, bringing the input to a GPR voltage prediction model of corresponding components to obtain voltage prediction values of the components, accumulating the prediction values of the components, and averaging to obtain Tt+1A predicted value of the battery voltage at the moment;
step 3, converting the voltage and internal resistance data in the two model inputs into predicted data, updating the floating charge time and the equalizing charge times, and then carrying out prediction on the next moment in the step 2;
step 4, repeating step 3 until T is predictedt+kVoltage and internal resistance at time.
In conclusion, the method and the device have the advantages that the influence of the floating charge duration and the average charge frequency of the storage battery under the direct-current system of the transformer substation is taken into consideration, the degradation trend of the storage battery is predicted by adopting a GPR method, the degradation trend in the use plan time is predicted according to the use plan of the storage battery of the transformer substation, and the prediction precision of the degradation trend of the storage battery of the transformer substation can be effectively improved. The method can find the storage battery which does not meet the use requirement in the use plan in time, ensure the safe and reliable electricity utilization of equipment under emergency, and simultaneously avoid the further deterioration of the performance of the whole storage battery due to the deterioration of individual batteries in the storage battery.
As shown in fig. 6, a system for predicting degradation tendency of a valve-regulated lead-acid battery by gaussian process regression includes:
the first data processing module 101 is configured to decompose the acquired voltage and internal resistance data by using an empirical mode decomposition method;
the second data processing module 102 is configured to calculate a content of each decomposed component, and screen the components of the voltage and the internal resistance by using the content as an index;
a GPR voltage prediction model construction module 103 configured to train the GPR model of each component by using the value of the screened voltage component at a certain time, the floating charge time between the certain time and the next time, and the value of the average charge number as inputs of the GPR model, and using the value of the screened voltage component at the next time as an output of the GPR model, so as to obtain a GPR voltage prediction model of each component;
a GPR internal resistance prediction model construction module 104, configured to train the GPR model of each component by using the value of the screened internal resistance component at a certain time, the floating charge duration from the certain time to the next time, and the value of the average charge number as inputs of the GPR model, and using the value of the screened internal resistance component at the next time as an output of the GPR model, so as to obtain a GPR internal resistance prediction model of each component;
the voltage prediction module 105 is used for taking the voltage at a certain moment or each component obtained by EEMD decomposition and content screening of the voltage predicted value as input, taking the floating charge duration and the average charge times in the use plan of the storage battery from the moment to the next moment as input, bringing the floating charge duration and the average charge times to a GPR voltage prediction model of the corresponding component to obtain the voltage predicted value of each component, and accumulating the voltage predicted values of each component to obtain the predicted value of the voltage of the storage battery at the next moment;
and the internal resistance prediction module 106 is used for taking each component obtained by carrying out EEMD decomposition and content screening on the internal resistance or the internal resistance prediction value at a certain moment as input, taking the floating charge duration and the average charge times in the use plan of the storage battery from the moment to the next moment as input, bringing the floating charge duration and the average charge times to the GPR internal resistance prediction model of the corresponding component to obtain the internal resistance prediction value of each component, and accumulating the internal resistance prediction values of each component to obtain the prediction value of the internal resistance of the storage battery at the next moment.
The system provided by the embodiment of the present invention is used for executing the above method embodiments, and for details of the process and the details, reference is made to the above embodiments, which are not described herein again.
Fig. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, and referring to fig. 7, the electronic device may include: a processor (processor)201, a communication Interface (communication Interface)202, a memory (memory)203 and a communication bus 204, wherein the processor 201, the communication Interface 202 and the memory 203 complete communication with each other through the communication bus 204. The processor 201 may call logic instructions in the memory 203 to perform the following method: respectively decomposing the acquired voltage and internal resistance data by adopting an empirical mode decomposition method; calculating the content of each component obtained by decomposition, and screening the components of the voltage and the internal resistance by taking the content as an index; taking the value of the screened voltage component at a certain moment, the floating charge time between the moment and the next moment and the value of the average charge times as the input of a GPR model, taking the value of the screened voltage component at the next moment as the output of the GPR model, and training the GPR model of each component to obtain a GPR voltage prediction model of each component; taking the value of the screened internal resistance component at a certain moment, the floating charge time between the moment and the next moment and the value of the average charge times as the input of a GPR model, taking the value of the screened internal resistance component at the next moment as the output of the GPR model, and training the GPR model of each component to obtain a GPR internal resistance prediction model of each component; taking each component obtained by EEMD decomposition and content screening of a voltage or a voltage predicted value at a certain moment as input, taking the floating charge duration and the average charge times in a use plan of the storage battery between the moment and the next moment as input, bringing the input to a GPR voltage prediction model of the corresponding component to obtain the voltage predicted value of each component, and accumulating the voltage predicted values of each component to obtain the predicted value of the voltage of the storage battery at the next moment; and taking each component obtained by EEMD decomposition and content screening of the internal resistance or the predicted value of the internal resistance at a certain moment as input, taking the floating charge time and the average charge times in the use plan of the storage battery from the moment to the next moment as input, bringing the floating charge time and the average charge times to the GPR internal resistance prediction model of the corresponding component to obtain the predicted value of the internal resistance of each component, and accumulating the predicted values of the internal resistance of each component to obtain the predicted value of the internal resistance of the storage battery at the next moment.
In addition, the logic instructions in the memory 203 may be implemented in the form of software functional units and stored in a computer readable storage medium when the logic instructions are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, an embodiment of the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program is implemented to, when executed by a processor, perform a degradation trend prediction method for a valve-regulated lead-acid battery by gaussian process regression provided in the foregoing embodiments, for example, the method includes: respectively decomposing the acquired voltage and internal resistance data by adopting an empirical mode decomposition method; calculating the content of each component obtained by decomposition, and screening the components of the voltage and the internal resistance by taking the content as an index; taking the value of the screened voltage component at a certain moment, the floating charge time between the moment and the next moment and the value of the average charge times as the input of a GPR model, taking the value of the screened voltage component at the next moment as the output of the GPR model, and training the GPR model of each component to obtain a GPR voltage prediction model of each component; taking the value of the screened internal resistance component at a certain moment, the floating charge time between the moment and the next moment and the value of the average charge times as the input of a GPR model, taking the value of the screened internal resistance component at the next moment as the output of the GPR model, and training the GPR model of each component to obtain a GPR internal resistance prediction model of each component; taking each component obtained by EEMD decomposition and content screening of a voltage or a voltage predicted value at a certain moment as input, taking the floating charge duration and the average charge times in a use plan of the storage battery between the moment and the next moment as input, bringing the input to a GPR voltage prediction model of the corresponding component to obtain the voltage predicted value of each component, and accumulating the voltage predicted values of each component to obtain the predicted value of the voltage of the storage battery at the next moment; and taking each component obtained by EEMD decomposition and content screening of the internal resistance or the predicted value of the internal resistance at a certain moment as input, taking the floating charge time and the average charge times in the use plan of the storage battery from the moment to the next moment as input, bringing the floating charge time and the average charge times to the GPR internal resistance prediction model of the corresponding component to obtain the predicted value of the internal resistance of each component, and accumulating the predicted values of the internal resistance of each component to obtain the predicted value of the internal resistance of the storage battery at the next moment.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Examples of the applications
Table 1 shows the sample data of voltage, internal resistance, float charge duration, and number of homogeneous charges collected in the experiment for 24 months.
TABLE 1 sample data
Figure RE-GDA0002403944630000171
Figure RE-GDA0002403944630000181
The method comprises the following steps:
1. EEM D decomposition
The method comprises the following steps: the acquired 24 terminal voltage data sequences are used as original signals, and are decomposed by adopting an EEMD algorithm to obtain 8 imf components and 1 margin respectively.
Step two: and decomposing the acquired 24 internal resistance data sequences serving as original signals by adopting an EEMD algorithm to obtain 11 imf components and 1 margin respectively.
2. Content screening
The method comprises the following steps: respectively integrating the components obtained by decomposing the original terminal voltage, the internal resistance data and the EEMD by using the formulas (7) to (9);
step two: calculating the content ratio of each component by using the formulas (10) and (11);
step three: when the component content is smaller than the content threshold λ (λ is a constant), it will be ignored, and the present invention sets the content threshold to 5%. The screened voltage data has 5 IMF components and the internal resistance data has 7 IMF components.
3. GPR model training and prediction
The method comprises the following steps: taking the voltage data of 1-20 times as a training sample, and taking the voltage data of 21-23 times as a test sample;
training input: voltage value of 1-20 times, floating charge duration and equalizing charge times of each time interval;
and (3) training output: voltage values from 2 th to 21 th;
test input: the 21 st to 23 th voltage values, the floating charge duration and the equalizing charge times of each time interval;
and (4) test output: voltage values from 22 th to 24 th;
step two: taking the internal resistance data of 1-20 times as a training sample, and taking the internal resistance data of 21-23 times as a test sample;
training input: 1-20 internal resistance values, floating charge duration and uniform charge times of each time interval;
and (3) training output: internal resistance values from 2 nd to 21 th;
test input: the 21 st to 23 th internal resistance values, the floating charge duration and the uniform charge times of each time interval;
and (4) test output: internal resistance values from 22 th to 24 th;
4. analysis of results
(1) Fig. 8 is a graph comparing the actual value with the predicted value of the voltage data for 21-24 times, and the voltage results predicted by the above steps are 10.601, 10.358 and 10.252. The average absolute error is 3.23%; the root mean square error is 3.84%, and the root mean square error do not exceed 5%, so that the actual requirement is met.
(2) As shown in fig. 9, the comparison graph of the actual value and the predicted value of the internal resistance data is 21-24 times, and the internal resistance results obtained by the prediction in the steps are 1108.622, 1118.925 and 1132.846. The average absolute error is 3.75 percent; the root mean square error is 3.81%, and the root mean square error do not exceed 5%, so that the actual requirement is met.
The foregoing shows and describes the general principles, essential features, and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are described in the specification and illustrated only to illustrate the principle of the present invention, but that various changes and modifications may be made therein without departing from the spirit and scope of the present invention, which fall within the scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.

Claims (5)

1. The degradation trend prediction method of the valve-regulated lead-acid storage battery by Gaussian process regression is characterized by comprising the following steps of:
decomposing the acquired voltage and internal resistance data by adopting the EEMD respectively;
calculating the content of each component obtained by decomposition, and screening the components of the voltage and the internal resistance by taking the content as an index;
taking the value of the screened voltage component at a certain moment, the floating charge time between the moment and the next moment and the value of the average charge times as the input of a GPR model, taking the value of the screened voltage component at the next moment as the output of the GPR model, and training the GPR model of each component to obtain a GPR voltage prediction model of each component;
taking the value of the screened internal resistance component at a certain moment, the floating charge time between the moment and the next moment and the value of the average charge times as the input of a GPR model, taking the value of the screened internal resistance component at the next moment as the output of the GPR model, and training the GPR model of each component to obtain a GPR internal resistance prediction model of each component;
taking each component obtained by EEMD decomposition and content screening of a voltage or a voltage predicted value at a certain moment as input, taking the floating charge duration and the average charge times in a use plan of the storage battery between the moment and the next moment as input, bringing the input to a GPR voltage prediction model of the corresponding component to obtain the voltage predicted value of each component, and accumulating the voltage predicted values of each component to obtain the predicted value of the voltage of the storage battery at the next moment;
and taking each component obtained by EEMD decomposition and content screening of the internal resistance or the predicted value of the internal resistance at a certain moment as input, taking the floating charge time and the average charge times in the use plan of the storage battery from the moment to the next moment as input, bringing the floating charge time and the average charge times to the GPR internal resistance prediction model of the corresponding component to obtain the predicted value of the internal resistance of each component, and accumulating the predicted values of the internal resistance of each component to obtain the predicted value of the internal resistance of the storage battery at the next moment.
2. The method of predicting degradation tendency of a valve-regulated lead-acid battery by gaussian process regression as claimed in claim 1, wherein when the content of a component is less than a threshold value λ of the content, the component is ignored.
3. The system for predicting the degradation trend of the valve-regulated lead-acid storage battery by Gaussian process regression is characterized by comprising the following steps:
the first data processing module is used for decomposing the acquired voltage and internal resistance data by adopting an EEMD method;
the second data processing module is used for calculating the content rate of each component obtained by decomposition and screening the components of the voltage and the internal resistance by taking the content rate as an index;
the GPR voltage prediction model building module is used for taking the value of the screened voltage component at a certain moment, the floating charge time between the moment and the next moment and the value of the average charge times as the input of the GPR model, taking the value of the screened voltage component at the next moment as the output of the GPR model, and training the GPR model of each component to obtain the GPR voltage prediction model of each component;
the GPR internal resistance prediction model building module is used for taking the value of the screened internal resistance component at a certain moment, the floating charge time between the moment and the next moment and the value of the average charge times as the input of the GPR model, taking the value of the screened internal resistance component at the next moment as the output of the GPR model, and training the GPR model of each component to obtain the GPR internal resistance prediction model of each component;
the voltage prediction module is used for taking the voltage at a certain moment or each component obtained by EEMD decomposition and content screening of the voltage predicted value as input, taking the floating charge duration and the average charge times in the use plan of the storage battery between the moment and the next moment as input, bringing the floating charge duration and the average charge times to a GPR voltage prediction model of the corresponding component to obtain the voltage predicted value of each component, and accumulating the voltage predicted values of each component to obtain the predicted value of the storage battery voltage at the next moment;
and the internal resistance prediction module is used for taking each component obtained by carrying out EEMD decomposition and content screening on the internal resistance or the internal resistance prediction value at a certain moment as input, taking the floating charge duration and the average charge times in the use plan of the storage battery from the moment to the next moment as input, bringing the floating charge duration and the average charge times to the GPR internal resistance prediction model of the corresponding component to obtain the internal resistance prediction value of each component, and accumulating the internal resistance prediction values of each component to obtain the prediction value of the internal resistance of the storage battery at the next moment.
4. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the program implements the steps of the method for predicting degradation trend of valve-regulated lead-acid battery regressive with gaussian process according to any of claims 1 to 2.
5. A non-transitory computer-readable storage medium having stored thereon a computer program, wherein the computer program, when being executed by a processor, implements the steps of the method for predicting degradation tendency of a valve-regulated lead-acid battery regressive with gaussian process according to any one of claims 1 to 2.
CN201911236574.8A 2019-12-05 2019-12-05 Valve-regulated lead-acid storage battery degradation trend prediction method based on Gauss process regression Active CN111143973B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911236574.8A CN111143973B (en) 2019-12-05 2019-12-05 Valve-regulated lead-acid storage battery degradation trend prediction method based on Gauss process regression

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911236574.8A CN111143973B (en) 2019-12-05 2019-12-05 Valve-regulated lead-acid storage battery degradation trend prediction method based on Gauss process regression

Publications (2)

Publication Number Publication Date
CN111143973A CN111143973A (en) 2020-05-12
CN111143973B true CN111143973B (en) 2021-01-26

Family

ID=70517574

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911236574.8A Active CN111143973B (en) 2019-12-05 2019-12-05 Valve-regulated lead-acid storage battery degradation trend prediction method based on Gauss process regression

Country Status (1)

Country Link
CN (1) CN111143973B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11099219B2 (en) * 2018-03-26 2021-08-24 Oracle International Corporation Estimating the remaining useful life of a power transformer based on real-time sensor data and periodic dissolved gas analyses
CN112287597B (en) * 2020-09-22 2023-10-03 国网天津市电力公司电力科学研究院 Lead-acid storage battery SOH estimation method based on VPGA-GPR algorithm
CN113406525B (en) * 2021-06-15 2023-06-02 金陵科技学院 Lithium battery pack residual life prediction method based on optimization variation modal decomposition
CN113533986A (en) * 2021-07-16 2021-10-22 厦门爱维达科技工程有限公司 VMD-Hilbert-based storage battery internal resistance measuring method and system

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103336913A (en) * 2013-07-25 2013-10-02 哈尔滨工业大学 Method for monitoring state and predicting cut-off voltage of spatial lithium ion battery
CN103389471A (en) * 2013-07-25 2013-11-13 哈尔滨工业大学 Cycle life indirect prediction method for lithium ion battery provided with uncertain intervals on basis of GPR (general purpose register)
CN105487017A (en) * 2016-01-22 2016-04-13 国网黑龙江省电力有限公司检修公司 Valve-regulation sealed lead acid battery state estimation and prediction method used for transformer substation UPS
CN106125004A (en) * 2016-08-29 2016-11-16 哈尔滨理工大学 Lithium battery health status Forecasting Methodology based on neutral net kernel function GPR
CN106339755A (en) * 2016-08-29 2017-01-18 深圳市计量质量检测研究院 Lithium battery SOH (State of Health) prediction method based on neural network and periodic kernel functions GPR
WO2017146877A1 (en) * 2016-02-03 2017-08-31 General Electric Company Systems and methods for monitoring and diagnosing transformer health
CN107656216A (en) * 2017-11-15 2018-02-02 国网辽宁省电力有限公司鞍山供电公司 A kind of lead-acid accumulator on-line monitoring is safeguarded and early warning system and performance estimating method
CN108303652A (en) * 2018-01-18 2018-07-20 武汉理工大学 A kind of lithium battery method for predicting residual useful life
US10445644B2 (en) * 2014-12-31 2019-10-15 Ebay Inc. Anomaly detection for non-stationary data
CN110633516A (en) * 2019-08-30 2019-12-31 电子科技大学 Method for predicting performance degradation trend of electronic device

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5852399B2 (en) * 2011-10-17 2016-02-03 インターナショナル・ビジネス・マシーンズ・コーポレーションＩｎｔｅｒｎａｔｉｏｎａｌ Ｂｕｓｉｎｅｓｓ Ｍａｃｈｉｎｅｓ Ｃｏｒｐｏｒａｔｉｏｎ Battery state prediction system, method and program
CN102393503B (en) * 2011-11-14 2013-12-04 北京航空航天大学 GaAs solar energy cell performance degeneration prediction method in space environment
CN202696189U (en) * 2012-05-22 2013-01-23 深圳供电局有限公司 Voltage equalizing device
US9236748B2 (en) * 2012-08-30 2016-01-12 Texas Instruments Incorporated Method and apparatus of charging the battery with globally minimized integral degradation possible for predefined charging duration
CN104316879B (en) * 2014-10-13 2018-08-10 珠海许继电气有限公司 A kind of prediction technique in lead-acid batteries service life
KR20160090140A (en) * 2015-01-21 2016-07-29 삼성전자주식회사 Method and apparatus for estimating state of battery
CN110531279A (en) * 2019-05-31 2019-12-03 西安工程大学 Lithium ion battery remaining life prediction technique based on IUPF

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103336913A (en) * 2013-07-25 2013-10-02 哈尔滨工业大学 Method for monitoring state and predicting cut-off voltage of spatial lithium ion battery
CN103389471A (en) * 2013-07-25 2013-11-13 哈尔滨工业大学 Cycle life indirect prediction method for lithium ion battery provided with uncertain intervals on basis of GPR (general purpose register)
US10445644B2 (en) * 2014-12-31 2019-10-15 Ebay Inc. Anomaly detection for non-stationary data
CN105487017A (en) * 2016-01-22 2016-04-13 国网黑龙江省电力有限公司检修公司 Valve-regulation sealed lead acid battery state estimation and prediction method used for transformer substation UPS
WO2017146877A1 (en) * 2016-02-03 2017-08-31 General Electric Company Systems and methods for monitoring and diagnosing transformer health
CN106125004A (en) * 2016-08-29 2016-11-16 哈尔滨理工大学 Lithium battery health status Forecasting Methodology based on neutral net kernel function GPR
CN106339755A (en) * 2016-08-29 2017-01-18 深圳市计量质量检测研究院 Lithium battery SOH (State of Health) prediction method based on neural network and periodic kernel functions GPR
CN107656216A (en) * 2017-11-15 2018-02-02 国网辽宁省电力有限公司鞍山供电公司 A kind of lead-acid accumulator on-line monitoring is safeguarded and early warning system and performance estimating method
CN108303652A (en) * 2018-01-18 2018-07-20 武汉理工大学 A kind of lithium battery method for predicting residual useful life
CN110633516A (en) * 2019-08-30 2019-12-31 电子科技大学 Method for predicting performance degradation trend of electronic device

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
Lithium-ion batteries remaining useful life prediction based on amixture of empirical mode decomposition and ARIMA model;Yapeng Zhou 等;《Microelectronics Reliability》;20160730;第265-273页 *
基于 EEMD 和 MKRVM 的锂电池剩余寿命预测方法;张朝龙 等;《电力系统及其自动化学报》;20180731;第38-44页 *
电网用阀控式铅酸蓄电池寿命预测研究;王文强;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20170315;第17、30页 *

Also Published As

Publication number Publication date
CN111143973A (en) 2020-05-12

Similar Documents

Publication Publication Date Title
CN111143973B (en) Valve-regulated lead-acid storage battery degradation trend prediction method based on Gauss process regression
CN112034356B (en) GP-UKF-based online SOH estimation method for electric vehicle power battery
CN109543317B (en) Method and device for predicting remaining service life of PEMFC
CN113064093B (en) Method and system for jointly estimating state of charge and state of health of energy storage battery
CN111736084B (en) Valve-regulated lead-acid storage battery health state prediction method based on improved LSTM neural network
CN107292502B (en) Power distribution network reliability assessment method
CN111680848A (en) Battery life prediction method based on prediction model fusion and storage medium
CN108847686B (en) Photovoltaic inverter fault prediction method
CN107330573B (en) State evaluation method and device for photovoltaic system key equipment
CN109524982B (en) AC/DC power grid transient stability risk assessment method
CN110378504B (en) Photovoltaic power generation climbing event probability prediction method based on high-dimensional Copula technology
CN111191398B (en) SVR-based method for predicting degradation trend of storage battery of direct-current system of transformer substation
CN113792939B (en) Electric energy meter reliability prediction method and device based on mixed Weibull distribution
CN111210049A (en) LSTM-based method for predicting degradation trend of lead-acid valve-controlled storage battery of transformer substation
CN114636948A (en) Energy storage system service life assessment method and device, electronic equipment and storage medium
CN111091141B (en) Photovoltaic backboard fault diagnosis method based on layered Softmax
CN115146739A (en) Power transformer fault diagnosis method based on stacked time series network
CN115564310A (en) Reliability evaluation method for new energy power system based on convolutional neural network
CN109299552A (en) A kind of appraisal procedure and its assessment system of battery power status
CN113791351B (en) Lithium battery life prediction method based on transfer learning and difference probability distribution
CN114784793A (en) New energy power grid cascading failure prediction method based on machine learning and storage medium
CN107563641B (en) Disaster-resistant multi-scene differentiation planning method for power distribution network considering disaster preference
Zhang et al. Remaining useful life prediction of lithium-ion batteries based on TCN-DCN fusion model combined with IRRS filtering
CN117312797A (en) Composite denoising super capacitor backup power performance degradation rule and residual life LSTM prediction method
CN114186756A (en) Method for predicting energy storage capacity of storage battery of power distribution network terminal

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant