Monte Carlo method-based lithium battery abnormal working condition data self-organizing enhancement methodTechnical Field
The invention belongs to the field of lithium battery detection, and relates to a Monte Carlo method-based lithium battery abnormal working condition data self-organizing enhancement method.
Background
As an energy storage device with high energy density, long service life and clean output, the lithium battery is widely applied to energy systems such as electric automobiles, portable electronic equipment, smart grids with renewable energy sources and the like. In the past decade, research on energy density and power output of lithium batteries has been significantly progressed, but the safe use of lithium batteries still faces significant challenges due to the difficulty in completely observing internal electrochemical reactions and the complex and uncertain operating conditions.
Many research efforts have utilized self-acquired data sets for modeling and validation, and some have also employed NASA, CALCE data sets. The distribution of these measurements is not balanced, e.g., the NASA dataset has 95% of normal operating data, and 5% of abnormal extreme environmental data, and no analysis is made as to why. The distribution of the CALCE dataset is similar to NASA. In most self-acquired data sets, only few studies use outlier data to investigate the robustness of the model. Data imbalance will result in insufficient generalization of the trained ANN model, forcing the prediction framework to be not shared by multiple monomers, i.e., the trained ANN can only be used for one monomer or one operating condition. There are many studies on the lack of sharing, and when monitoring a plurality of lithium batteries in the same batch or replacing some of the batteries, the prediction framework for not sharing needs to be retrained, which causes more resource waste. The root cause of the data imbalance is that the test of the laboratory condition cannot enumerate the abnormal working conditions of the lithium battery (such as extreme weather environment, bad habit of people, accidental impact, accident and the like) infinitely.
Disclosure of Invention
In view of the above, the present invention aims to provide a new data enhancement method for solving the problem of scarcity of high-value data in lithium batteries or other industrial fields from the industrial data perspective, obtain multiscale components of abnormal working condition data through wavelet transformation, and freely combine the multiscale components by adopting a monte carlo method to generate artificial data synthesized by real components under specific distribution.
In order to achieve the above purpose, the present invention provides the following technical solutions:
a lithium battery abnormal working condition data self-organizing enhancement method based on a Monte Carlo method comprises the following steps:
s1: performing wavelet decomposition on the abnormal working condition sequence of the small sample lithium battery to obtain a multi-scale wavelet coefficient;
s2: mapping the multi-scale wavelet coefficient into an m-dimensional hyperspace midpoint;
s3: performing Monte Carlo ad hoc on the multi-scale components;
s4: and carrying out wavelet coefficient reconstruction on the screened components to generate a synthesized lithium battery sequence.
Further, in step S1, for a given abnormal working condition sequence f (t) of the small sample lithium battery, the wavelet packet of the L layer of the abnormal working condition sequence f (t) includes a voltage, a current, a capacity, and an internal resistance curve, where the wavelet packet is decomposed into:
wherein phi (t) and ψ (t) are respectively father and mother wavelets, n represents the position of the wavelet packet in the layer where the wavelet packet is located, hk Represents a low-pass filter, t represents time, k represents position coordinates, gk Representing a high pass filter.
Further, in step S2, let x= [ x ]1 ,x2 ,…,x2L-1 x2L ]＝[Φ1 (t),Ψ1 (t),…ΦL (t),ΨL (t)]The multi-scale wavelet coefficients are mapped into m-dimensional hyperspace midpoints (called particles), and the dimension of a wavelet coefficient of a certain layer is selected as a space reference according to specific sample characteristics.
Further, in step S3, a system of 2L particles is examined, the structure of which is described by vector x, i.e. the particle state is described by { x }i Definition of xi Is the firsti wavelet coefficients; the equilibrium distribution of the system is described as pi (x). Alpha. Exp [ -U (x)/kB T]Wherein k isB The thermal energy coefficient, T is the temperature corresponding to the lithium battery sequence, U (x) is a cost energy function used for describing the sum of interaction among particles, and is defined as:
wherein ,Vij Is the energy of interaction between the ith particle and the jth particle; defining an augmented spatial distribution pi (x, b), wherein { x } is an initial state space, { b } is a state coupling space, and b represents a vector of coupling between particles, namely coupling between wavelet coefficients of respective scales corresponding to a lithium battery sequence, and the coupling mode comprises the following steps of L (L-1)/2 elements:
the augmentation spatial distribution II (x, b) and the equilibrium distribution pi (x) of the system satisfy the following conditions:
Π(x,b)＝π(x)p(b,x) (4)
wherein p (b, x) represents that a coupling state b= { b is formed at a given initial state xij The probability of }, written as:
the q-function representation forms a specific coupling probability, expressed as:
q(bij ,Vij )＝bij +(1-2bij )exp[min(Vij ,0)/kB Θ] (6)
according to formula (6), if the interaction V between particles i and ji,j Positive, they are always uncoupled, bi,j The probability of =0 is 1; if Vi,j Negative, when |Vij When the I is not less than k theta, the probability b of coupling is not less than k thetai,j =1 to 1; Θ represents a "pseudo" controlling the coupling processTemperature "; in practical application, determining a value interval of Θ according to the characteristics of abnormal working conditions;
based on the above definition, the self-organization between particles, i.e. starting from the initial pairing x, b, the algorithm that generates the new pairing, is categorized into two modes:
(1) selecting a new coupling state b': for all b
i,j Zero initialization is performed when i < j and V
i,j <At 0, b
i,j The probability of 1 is set to 1-exp [ V
ij (x
i ,x
j )/k
B Θ]The method comprises the steps of carrying out a first treatment on the surface of the This step groups the particles of the system into groups, i.e. by a series of couplings b
i,j =1, a set of particles, which are interconnected, to screen out the wavelet coefficients that can be reconstructed;
(2) metropolis state transition x→x': selecting a particle j, and forcibly coupling the other particles with the j and moving the particles in a random direction under the constraint of a coupling state space; when the movement is completed, a new energy U (x') is calculated; meanwhile, the probability of the mobile acceptance is expressed as:
the probability ratio in formula (7) is further expressed as:
wherein ,
further, in step S4, the particles selected according to the monte carlo simulation, i.e. the possible coupled wavelet coefficients, are reconstructed according to the following formula to generate a synthesized lithium battery sequence:
wherein ,for the wavelet coefficients that are screened.
The invention has the beneficial effects that: the invention gives consideration to the authenticity of data sources and the reconfigurability of data distribution, provides a new technical means for unbalanced lithium battery data processing, and can also be moved to other industrial fields as one of methods for solving the problem of data unbalance.
Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention. The objects and other advantages of the invention may be realized and obtained by means of the instrumentalities and combinations particularly pointed out in the specification.
Drawings
For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in the following preferred detail with reference to the accompanying drawings, in which:
fig. 1 is a flow chart of a multi-scale lithium battery data self-organizing enhancement method based on a monte carlo method according to the invention.
Detailed Description
Other advantages and effects of the present invention will become apparent to those skilled in the art from the following disclosure, which describes the embodiments of the present invention with reference to specific examples. The invention may be practiced or carried out in other embodiments that depart from the specific details, and the details of the present description may be modified or varied from the spirit and scope of the present invention. It should be noted that the illustrations provided in the following embodiments merely illustrate the basic idea of the present invention by way of illustration, and the following embodiments and features in the embodiments may be combined with each other without conflict.
Wherein the drawings are for illustrative purposes only and are shown in schematic, non-physical, and not intended to limit the invention; for the purpose of better illustrating embodiments of the invention, certain elements of the drawings may be omitted, enlarged or reduced and do not represent the size of the actual product; it will be appreciated by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.
The same or similar reference numbers in the drawings of embodiments of the invention correspond to the same or similar components; in the description of the present invention, it should be understood that, if there are terms such as "upper", "lower", "left", "right", "front", "rear", etc., that indicate an azimuth or a positional relationship based on the azimuth or the positional relationship shown in the drawings, it is only for convenience of describing the present invention and simplifying the description, but not for indicating or suggesting that the referred device or element must have a specific azimuth, be constructed and operated in a specific azimuth, so that the terms describing the positional relationship in the drawings are merely for exemplary illustration and should not be construed as limiting the present invention, and that the specific meaning of the above terms may be understood by those of ordinary skill in the art according to the specific circumstances.
Aiming at the imbalance problem of abnormal working condition data of the lithium battery, a Monte Carlo self-organizing enhancement method of the abnormal working condition data of the lithium battery under multi-scale description is researched. Taking actually measured small sample sequences such as current, voltage, temperature, capacity, internal resistance and the like under abnormal working conditions as objects, designing wavelet packet transformation to obtain sequence multi-scale components and corresponding hyperspace mapping; and taking an ultra-high-dimensional system formed by multi-scale sequence components as an object, researching a Monte Carlo simulation method under the constraint of the working temperature of the lithium battery, and realizing the self-group screening of the sequence components with physical significance so as to synthesize a new lithium battery sequence.
The invention provides a multi-scale lithium battery data self-organizing enhancement method based on a Monte Carlo method, which is shown in figure 1 and comprises four steps of wavelet packet decomposition, hyperspace mapping, monte Carlo simulation and sequence reconstruction of a lithium battery sequence.
1) Wavelet packet decomposition of original sequence
The specific small sample lithium battery abnormal working condition sequence f (t) can be curves of voltage, current, capacity, internal resistance and the like, and the wavelet packet of the L layer is decomposed into:
wherein phi (t) and ψ (t) are parent and parent wavelets respectively, hk and gk Respectively low and high pass filters.
2) Hyperspace mapping of multiscale components
Let x= [ x ]1 ,x2 ,…,x2L-1 x2L ]＝[Φ1 (t),Ψ1 (t),…ΦL (t),ΨL (t)]The multiscale wavelet coefficients are mapped to m-dimensional superspace midpoints (called particles). In practical research, the dimension of a wavelet coefficient of a certain layer is selected as a space reference according to specific sample characteristics.
3) Monte Carlo ad hoc of multi-scale components
Consider a system of 2L particles whose structure is described by a vector x, i.e. the particle state is defined by { x }i Definition (x)i Is the ith wavelet coefficient). The equilibrium distribution of the system is described as pi (x). Alpha. Exp [ -U (x)/kB T]Wherein k isB The thermal energy coefficient, T is the temperature corresponding to the lithium battery sequence, U (x) is a cost energy function used for describing the sum of interaction among particles, and is defined as:
wherein ,Vij Is the energy of the interaction between the ith particle and the jth particle. At the same time, an augmented spatial distribution pi (x, b) is defined, where { x } is the initial state space and { b } is the state coupling space. b represents the vector of coupling between particles (coupling between wavelet coefficients of respective scales corresponding to the sequence of lithium batteries) having L (L-1)/2 elements in the following manner:
the augmentation spatial distribution pi (x, b) and the equilibrium distribution pi (x) of the system satisfy the following conditions:
Π(x,b)＝π(x)p(b,x) (4)
wherein p (b, x) represents that a coupling state b= { b is formed at a given initial state xij The probability of } can be written as:
here, the q-function representation forms a particular coupling probability, which can be expressed as:
q(bij ,Vij )＝bij +(1-2bij )exp[min(Vij ,0)/kB Θ] (6)
according to formula (6), if the interaction V between particles i and ji,j Positive, they are always uncoupled (bi,j Probability of 0 is 1); if Vi,j Negative, when |Vij When the I is not less than kΘ, the probability of coupling (bi,j =1) becomes 1.Θ represents a "pseudo temperature" that controls the coupling process. In practical application, the value interval of theta is determined according to the characteristics of abnormal working conditions.
Based on the above definition, the self-organization between particles, i.e. starting from the initial pairing x, b, the algorithm of generating a new pairing can be generalized to the following two modes:
(1) selecting a new coupling state b': for all b
i,j Zero initialization is performed when i < j and V
i,j <At 0, b
i,j The probability of 1 is set to 1-exp [ V
ij (x
i ,x
j )/k
B Θ]. This step groups the particles of the system into groups, i.e. by a series of couplings (b
i,j =1), the particles are interconnected to form a set of particles, and the wavelet coefficients that can be reconstructed are selected.
(2) Metropolis state transition x→x': selecting a particle j, and forcing the other particles to be coupled with j and move in a random direction under the constraint of a coupling state space. When the movement is completed, a new energy U (x') can be calculated. Meanwhile, the probability that the movement is acceptable is expressed as:
the probability ratio in equation (7) can be further expressed as:
wherein ,
4) Artificial sequence generation of screening components
The screened particles (i.e., the possible coupled wavelet coefficients) are reconstructed according to the Monte Carlo simulation, and the wavelet coefficients are reconstructed according to the following formula to generate a synthesized lithium battery sequence.
wherein ,for the wavelet coefficients that are screened.
Finally, it is noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications and equivalents may be made thereto without departing from the spirit and scope of the present invention, which is intended to be covered by the claims of the present invention.