Power distribution cabinet electrical equipment internal operation state monitoring time sequence analysis system

Abstract

The invention discloses a power distribution cabinet electrical equipment internal operation state monitoring time sequence analysis system. The model is a sensor data detection model based on an HMM (Hidden Markov Model), and the HMM is a dynamic Bayesian network generation model capable of being used for marking problems and describes a process of randomly generating an unobservable state sequence by an HMM chain and then generating an observation sequence by the state sequence; the HMM is determined by initial probability distribution, state transition probability distribution and observation probability distribution, the initial probability distribution and the state transition probability distribution determine an HMM chain, and an unobservable state sequence is generated; observing probability distribution to determine how to generate an observation sequence from the state sequence; and after the HMM is established, performing comparative analysis according to data acquired by the current sensor and an observation sequence, and finally outputting a current state judgment value. The method has the beneficial effects that the processing speed is higher, the hardware cost requirement is lower, and the data detection precision is more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field related to electrical equipment in a power distribution cabinet, and in particular to a timing analysis system for monitoring the internal operating status of electrical equipment in a power distribution cabinet. Background Art

[0002] The distribution cabinet is a very important electrical equipment in the power system. The deterioration of the operating status of the distribution cabinet due to various reasons often leads to failures in the power system. Therefore, real-time monitoring of the operating status of the distribution cabinet, timely detection of hidden faults, and prediction of cumulative faults are of great significance for ensuring the normal operation of the distribution cabinet, reducing the number of maintenance times, and improving the operating reliability and automation of the power system.

[0003] The main types of failures in the power distribution cabinet are as follows:

[0004] (1) Failure to operate or malfunction: Failure to operate or malfunction is the most common failure of the distribution cabinet. The causes can be divided into two categories. One is caused by mechanical reasons of the operating mechanism transmission system, which are manifested as mechanism jamming, component deformation or damage, loose or jammed opening and closing iron core, loose shaft pin, tripping failure, etc. The other is caused by electrical control and auxiliary circuits, which are manifested as poor secondary wiring contact, loose terminals, incorrect wiring, burning of the opening and closing coil due to mechanism jamming or poor conversion switch, ineffective switching of the auxiliary switch, operating power failure, closing contactor failure, and micro switch failure.

[0005] (2) Opening and closing faults: Opening and closing faults are caused by the circuit breaker itself. For low-oil circuit breakers, the main manifestations are oil injection short circuit, arc extinguishing chamber burnout, insufficient breaking capacity, explosion during closing, etc. For vacuum circuit breakers, the manifestations are arc extinguishing chamber and bellows leakage, reduced vacuum, reignition of the cut-off capacitor group, and ceramic tube rupture.

[0006] (3) Insulation failure: Insulation failure is manifested as external insulation to ground flashover breakdown, internal insulation to ground flashover breakdown, phase-to-phase insulation flashover breakdown, lightning overvoltage flashover breakdown, as well as porcelain bottle bushings and capacitor bushings flashover, pollution flashover, breakdown, explosion, etc.

[0007] (4) Cut-off failure: The main cause of the cut-off failure at the 7.2kV to 12kV voltage level is the poor contact of the isolation plug of the distribution cabinet, which causes the contacts to burn.

[0008] (5) External forces and other faults: External forces and other faults include impacts from foreign objects, natural disasters, short circuits caused by small animals, etc.

[0009] Traditional monitoring methods for power distribution cabinets mainly rely on independent sensors to monitor individual monitoring items independently, and the core lies in the sensor type and accuracy. Sensors are mainly divided into temperature sensors, inductive position sensors, displacement sensors, pressure sensors, and vibration sensors. Most state monitoring methods are based on machine learning models, and the model learning process is usually: first, the monitored sensor data is normalized, denoised, and processed; then, feature extraction and feature selection are performed from the time domain, frequency domain, and time-frequency domain; finally, a pre-established machine learning model is used for classification.

[0010] The deep learning method model is an "end-to-end model" that does not require complex data preprocessing of the original data. It can achieve relatively ideal results by adaptively extracting high-dimensional features hidden in sensor signals and reasonable network depth design. However, the use of end-to-end deep learning methods requires high real-time signal processing capabilities of the monitoring terminal, and the hardware cost is very high. For data that has not undergone preliminary signal preprocessing, there are large detection errors, and even often there are problems with poor regression and non-convergence of global functions. Summary of the invention

[0011] The present invention aims to overcome the above-mentioned deficiencies in the prior art and provides a timing analysis system for monitoring the internal operating status of electrical equipment in a distribution cabinet, which can improve processing speed and data detection accuracy.

[0012] In order to achieve the above object, the present invention adopts the following technical solutions:

[0013] A time series analysis system for monitoring the internal operating status of electrical equipment in a power distribution cabinet is disclosed. The model is a sensor data detection model based on a hidden Markov model HMM. The hidden Markov model HMM is a dynamic Bayesian network generation model that can be used for labeling problems. It describes the process of randomly generating an unobservable state sequence by an HMM chain, and then generating an observation sequence from the state sequence; the HMM is determined by an initial probability distribution, a state transition probability distribution, and an observation probability distribution, wherein the initial probability distribution and the state transition probability distribution determine the HMM chain to generate an unobservable state sequence; the observation probability distribution determines how to generate an observation sequence from the state sequence; the specific operations are as follows:

[0014] (1) Importing data monitored by various sensors and fault status samples of electrical equipment in the distribution cabinet into the hidden Markov model HMM to establish the hidden Markov model HMM;

[0015] (2) After the hidden Markov model HMM is established, the data collected by the current sensor is compared and analyzed with the observation sequence, and finally the current state judgment value is output.

[0016] The present invention proposes a sensor data detection model based on a hidden Markov model (HMM), which has faster processing speed, lower hardware cost requirements and more accurate data detection accuracy compared to an end-to-end neural network deep learning algorithm.

[0017] Preferably, the hidden Markov model HMM is defined as follows:

[0018] The set of all possible states is Q, the set of possible observations is V, the state sequence is I, and the observation sequence is O:

[0019] Q={q1,q2,q3,…,q N}

[0020] V={v1,v2,v3,…,v M}

[0021] I={i1,i2,i3,…,i T}

[0022] O={o1,o2,o3,…,o T}

[0023] Where: N is the number of states that may correspond to the HMM chain, M is the number of observations that may be generated for each state, and T is the sequence length; the state sequence I is the internal operating state standard of the electrical equipment in the distribution cabinet, and its internal value i is the state standard value; the state set Q is the actual operating state of the electrical equipment in the distribution cabinet after running for a period of time, and its internal value q is the operating state value; the observation sequence O is the observation standard of each type of sensor under each state, and its internal value o is the observation standard value; the observation set V is the sum of the observation values ​​v generated by each type of sensor in each state.

[0024] Preferably, the initial probability distribution is as follows:

[0025] The initial state probability vector is π:

[0026] π=(π i )

[0027] π i =P(i t =q i ),i=1,2,3,…,N

[0028] π i At time t, it is in state q i The initial state probability of

[0029] The state transition probability distribution is specifically as follows:

[0030] The state transition probability matrix is ​​A:

[0031]

[0032] a ij =P(i t+1 =q j i t =q i ),i=1,2,3,…,N a ij is in state q at time t i Under the condition of j probability;

[0033] The observation probability distribution is as follows:

[0034] The observation probability matrix is ​​B:

[0035]

[0036] b j (k) = P(o t =v k |i t =q j ), k=1,2,3,…,M; j=1,2,3,…,N

[0037] b j (k) is in state q at time t j Generate observations v under the condition k The probability of, k refers to the sequence number of the observation value, j refers to the sequence number of the state value;

[0038] In the process of establishing the hidden Markov model HMM, λ=(π, A, B) is used to represent the hidden Markov model HMM. The input data of the model is the state sequence I, the observation sequence O and the observation set V are preset by the model, and the state set Q is the sample data.

[0039] As a preferred method, when using the Hidden Markov Model HMM for comparison analysis, it is necessary to solve the probability calculation problem, which refers to: given the model λ = (π, A, B), the observation sequence O = {o1, o2, o3, ..., o T}, calculate the probability P(O|λ) of the observation sequence O appearing under the given model λ; specifically, the forward algorithm and the backward algorithm are combined with the agreed algorithm to completely calculate the probability P(O|λ) of the observation sequence O appearing under the given model λ.

[0040] As a preferred method, the forward algorithm and the backward algorithm are used together to agree on the following algorithm: the forward algorithm needs to define the forward probability, and the partial observation sequence at time t is o1, o2, o3, ..., o T And the state is q iThe probability of is the forward probability:

[0041] a t (i)=P(o1, o2, o3,...,o T ,i t =q i |λ)

[0042] The forward probability a is obtained by recursive method t+1 (i) and the observed sequence probability P(O|λ);

[0043] Initial value:

[0044] a1(i)=π i b i (o1),i=1,2,3,...,N

[0045] where π i In state q i The initial state probability of b i (o1) is in state q i The probability of generating the observation value o1 under the condition of ;

[0046] Recursively, for t=1,2,3,…,T-1:

[0047]

[0048] termination:

[0049]

[0050] where a T (i) is the final value of the forward probability;

[0051] The backward algorithm needs to define the backward probability. At time t, the state is q i Under the condition of t+1 , o t+2 , o t+3 ,…,o T The probability of is the backward probability:

[0052] β t (i) = P(o t+1 , o t+2 , o t+3 , ..., o T ,i t =q i |λ)

[0053] The backward probability β is obtained by recursive method t+1 (i) and the observed sequence probability P(O|λ)

[0054] Initial value:

[0055] β T (i)=1, i=1, 2, 3,...,N

[0056] Recursively, for t = T-1, T-2, T-3, ..., 1:

[0057]

[0058] termination:

[0059]

[0060] where β T (i) is the initial value of the backward probability, β t (i) is the final value of the backward probability;

[0061] The observation sequence probability P(O|λ) defined by the forward algorithm and the backward algorithm is:

[0062]

[0063] As a preferred method, when using the hidden Markov model HMM for comparison analysis, a prediction problem is required; the prediction problem is a known HMM model λ = (π, A, B) and an observation sequence O = {o1, o2, o3, ..., o T}, calculate the corresponding state sequence I = {i1, i2, i3, ..., i T}.

[0064] As a preferred method, for the prediction problem, an improved Viterbi algorithm is proposed to find the state sequence corresponding to the probabilistic optimal path. The specific calculation method is:

[0065] initialization:

[0066] δ1(i)=π i b i (o1), i=1,2,3,…,N

[0067] ψ1(i)=0,i=1,2,3,…,N

[0068] Among them, δ1(i) is the initialization optimal path probability value, ψ1(i) is the correction parameter;

[0069] Recursively, for t=2,3,4,…,T:

[0070]

[0071] termination:

[0072]

[0073] Optimal path backtracking, for t = T-1, T-2, T-3, ..., 1:

[0074]

[0075] Where P * To initialize the maximum probability of the optimal path; is the final state value of the optimal path, is the optimal path state value;

[0076] Then the state sequence corresponding to the optimal path is:

[0077]

[0078] As a preferred method, when using the Hidden Markov Model HMM for comparison analysis, a learning problem needs to be solved; the learning problem is that given training data only contains S observation sequences of length T O = {o1, o2, o3, ..., o s}, the goal is to learn the parameters of the HMM model λ=(π,A,B) so that the probability P(O|λ) of the observed sequence O under the model λ is maximized, that is, the parameters are estimated using the maximum likelihood estimation method. At this time, the HMM is P(O|λ)=∑ I P(O|I,λ)P(I|λ).

[0079] As a preferred method, for the learning problem, an unsupervised learning algorithm is proposed to calculate and determine the estimation parameters of the maximum likelihood estimation method. The specific calculation method is:

[0080] Determine the log-likelihood function of the complete data, all observations are O = {o1,o2,o3,…,o T}, all hidden data are I={i1,i2,i3,…,i T}, the complete data is O, I = {o1,o2,o3,…,o T ,i1,i2,i3,…,i T}, the log-likelihood function of the complete data is logP(O|I,λ);

[0081] The first step is to find the Q function

[0082]

[0083] Where: λ is the HMM parameter to be maximized; is the current estimate of the HMM parameters;

[0084]

[0085] The second step is to maximize the Q function Find the model parameters π, A, B;

[0086]

[0087] The third step is to combine the forward algorithm with the backward algorithm and use γ t (i) represents the probability of the forward and backward algorithm optimizing the path, ξ t (i,j) represents the forward and backward algorithm optimization parameters, where:

[0088]

[0089]

[0090] The fourth step is to design the optimization parameter estimation of the EM algorithm. The formula is:

[0091] π i =γ1(i)

[0092]

[0093] The beneficial effects of the present invention are: faster processing speed, lower hardware cost requirements and more accurate data detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0094] Figure 1 It is a schematic diagram of the process of establishing the Hidden Markov Model HMM in the present invention. DETAILED DESCRIPTION

[0095] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments.

[0096] like Figure 1 In the described embodiment, a time series analysis system for monitoring the internal operating status of electrical equipment in a power distribution cabinet is provided. The model is a sensor data detection model based on a hidden Markov model HMM. The hidden Markov model HMM is a dynamic Bayesian network generation model that can be used for labeling problems. It describes the process of randomly generating an unobservable state sequence by an HMM chain, and then generating an observation sequence from the state sequence; different positions in the sequence correspond to different moments. The HMM is determined by the initial probability distribution, the state transition probability distribution, and the observation probability distribution, wherein the initial probability distribution and the state transition probability distribution determine the HMM chain to generate an unobservable state sequence; the observation probability distribution determines how to generate an observation sequence from the state sequence; the specific operations are as follows:

[0097] (1) Importing data monitored by various sensors and fault status samples of electrical equipment in the distribution cabinet into the hidden Markov model HMM to establish the hidden Markov model HMM;

[0098] (2) After the hidden Markov model HMM is established, the data collected by the current sensor is compared and analyzed with the observation sequence, and finally the current state judgment value is output.

[0099] The hidden Markov model HMM is formally defined as follows:

[0100] The set of all possible states is Q, the set of possible observations is V, the state sequence is I, and the observation sequence is O:

[0101] Q={q1,q2,q3,…,q N}

[0102] V={v1,v2,v3,…,v M}

[0103] I={i1,i2,i3,…,i T}

[0104] O={o1,o2,o3,…,o T}

[0105] Where: N is the number of possible states in the HMM chain, M is the number of possible observations for each state, and T is the sequence length.

[0106] The state sequence I is the internal operating state standard of the electrical equipment in the power distribution cabinet. The value i inside is the state standard value, including no fault, refusal to operate fault, malfunction fault, opening fault, closing fault, insulation fault, current cutoff fault, and external force fault; the length is T, T = 8. i1 indicates no fault, i2 indicates refusal to operate fault, and so on.

[0107] The state set Q is the actual operating state of the electrical equipment in the distribution cabinet after a period of operation. The value q inside it is the operating state value. The state values ​​are randomly distributed and have a length of N.

[0108] The observation sequence O is the observation standard of each type of sensor in each state, and the value o inside it is the observation standard value; the observation sequence O includes temperature value, temperature change, position, displacement, pressure value, pressure change, presence or absence of vibration, and vibration frequency.

[0109] The observation set V is the sum of the observation values ​​v generated by each type of sensor in each state.

[0110] For example, assuming that the state value i is a refusal to move fault i2, the observed values ​​v of this fault include position o3, displacement o4, pressure value o5, and pressure change o6.

[0111] The initial probability distribution is as follows:

[0112] The initial state probability vector is π:

[0113] π=(π i )

[0114] π i =P(i t =q i ),i=1,2,3,…,N

[0115] π i is in state q at time t i The initial state probability, the state set Q is the actual operating state of the electrical equipment in the distribution cabinet for a period of time, and the time t is the time relative to time 0;

[0116] The state transition probability distribution is as follows:

[0117] The state transition probability matrix is ​​A:

[0118]

[0119] a ij =P(i t+1 =q j i t =q i ),i=1,2,3,…,N

[0120] a ij Under the condition that the state is qi at time t, the state is transferred to q at time t+1. j probability;

[0121] The observed probability distribution is as follows:

[0122] The observation probability matrix is ​​B:

[0123]

[0124] b j (k) = P(o t =v k |i t =q j ), k=1,2,3,…,M; j=1,2,3,…,N

[0125] b j (k) is in state q at time t j Generate observations v under the condition k The probability of k is the number of the observation value, and j is the number of the state value.

[0126] HMM is determined by the initial state probability vector π, the state transition probability matrix A and the observation probability matrix B. The composition of HMM is as follows: Figure 1As shown. π and A determine the HMM chain, generating an unobservable state sequence; B determines how to generate an observation sequence from the state sequence. HMM is represented by the above triples:

[0127] λ=(π,A,B).

[0128] In the entire process of establishing the hidden Markov model HMM, λ=(π,A,B) is used to represent the hidden Markov model HMM. The input data of the model is the state sequence I, the observation sequence O and the observation set V are preset by the model. The forward and backward algorithms, Viterbi algorithms, and EM algorithms in the following text are all optimizations of the observation sequence, and the state set Q is the sample data. After the model is established, it is applied to the electrical equipment of the distribution cabinet. First, there is a preset process to import the data monitored by various sensors and the fault state samples of the electrical equipment of the distribution cabinet into the HMM model. After the HMM model is established, the model compares the data collected by the current sensor with the observation sequence, and finally outputs the current state judgment value; including one of no fault, refusal to operate fault, malfunction fault, opening fault, closing fault, insulation fault, cut-off fault, and external force fault.

[0129] When using the Hidden Markov Model HMM for comparative analysis, it is necessary to solve the probability calculation (evaluation) problem, which means: given the model λ = (π, A, B), the observation sequence O = {o1, o2, o3, …, o T}, calculate the probability P(O|λ) of the observation sequence O occurring under the given model λ.

[0130] Specifically, the forward algorithm and the backward algorithm are combined with the agreed algorithm to completely calculate the probability P(O|λ) of the observation sequence O appearing under the given model λ. The forward algorithm and the backward algorithm are combined with the agreed algorithm as follows:

[0131] The forward algorithm needs to define the forward probability. At time t, the partial observation sequence is o1, o2, o3, …, o T And the state is q i The probability of is the forward probability:

[0132] a t (i)=P(o1, o2, o3,...,o T ,i t =q i |λ)

[0133] The forward probability a is obtained by recursive method t+1 (i) and the observed sequence probability P(O|λ);

[0134] Initial value:

[0135] a1(i)=π i bi (o1),i=1,2,3,...,N

[0136] where π i For the state q i The initial state probability of b i (o1) is in state q i The probability of generating the observation value o1 under the condition of ;

[0137] Recursively, for t=1, 2, 3, ..., T-1:

[0138]

[0139] termination:

[0140]

[0141] where a T (i) is the final value of the forward probability;

[0142] The backward algorithm needs to define the backward probability. At time t, the state is q i Under the condition of t+1 , o t+2 , o t+3 ,…,o T The probability of is the backward probability:

[0143] β t (i) = P(o t+1 , o t+2 , o t+3 , ..., o T ,i t =q i |λ)

[0144] The backward probability β is obtained by recursive method t+1 (i) and the observed sequence probability P(O|λ)

[0145] Initial value:

[0146] β T (i) = 1, i = 1, 2, 3, ..., N

[0147] Recursively, for t = T-1, T-2, T-3, ..., 1:

[0148]

[0149] termination:

[0150]

[0151] where β T(i) is the initial value of the backward probability, β t (i) is the final value of the backward probability;

[0152] The observation sequence probability P(O|λ) defined by the forward algorithm and the backward algorithm is:

[0153]

[0154] When using the Hidden Markov Model HMM for comparison analysis, a prediction (decoding) problem is required; the prediction (decoding) problem is a known HMM model λ = (π, A, B) and an observation sequence O = {o1, o2, o3, …, o T}, calculate the corresponding state sequence I={i1,i2,i3,…,i T}.

[0155] For the prediction (decoding) problem, an improved Viterbi algorithm is proposed to find the state sequence corresponding to the probabilistic optimal path. The specific calculation method is:

[0156] initialization:

[0157] δ1(i)=π i b i (o1), i=1,2,3,…,N

[0158] ψ1(i)=0,i=1,2,3,…,N

[0159] Among them, δ1(i) is the initialization optimal path probability value, ψ1(i) is the correction parameter;

[0160] Recursively, for t=2,3,4,…,T:

[0161]

[0162] termination:

[0163]

[0164] Optimal path backtracking, for t = T-1, T-2, T-3, ..., 1:

[0165]

[0166] Where P * To initialize the maximum probability of the optimal path; is the final state value of the optimal path, is the optimal path state value;

[0167] Then the state sequence corresponding to the optimal path is:

[0168]

[0169] When using the Hidden Markov Model HMM for comparison analysis, it is necessary to solve the learning (training) problem; the learning (training) problem is that given training data only contains S observation sequences of length T O = {o1, o2, o3, ..., o s}, the goal is to learn the parameters of the HMM model λ=(π,A,B) so that the probability P(O|λ) of the observed sequence O under the model λ is maximized, that is, the parameters are estimated using the maximum likelihood estimation method. At this time, the HMM is P(O|λ)=∑ I P(O|I,λ)P(I|λ).

[0170] In view of the learning problem, an unsupervised learning algorithm is proposed to calculate and determine the estimated parameters of the maximum likelihood estimation method. The specific calculation method is:

[0171] Determine the log-likelihood function of the complete data, all observations are O = {o1,o2,o3,…,o T}, all hidden data are I={i1,i2,i3,…,i T}, the complete data is O, I = {o1,o2,o3,…,o T ,i1,i2,i3,…,i T}, the log-likelihood function of the complete data is logP(O|I,λ);

[0172] The first step is to find the Q function

[0173]

[0174] Where: λ is the HMM parameter to be maximized; is the current estimate of the HMM parameters;

[0175]

[0176] The second step is to maximize the Q function Find the model parameters π, A, B;

[0177]

[0178] The third step is to combine the forward algorithm with the backward algorithm and use γ t (i) represents the probability of the forward and backward algorithm optimizing the path, ξ t (i,j) represents the forward and backward algorithm optimization parameters, where:

[0179]

[0180] The fourth step is to design the optimization parameter estimation of the EM algorithm. The formula is:

[0181] p i =γ1(i)

[0182]

Claims

1. A time series analysis system for monitoring the internal operating status of electrical equipment in a power distribution cabinet, characterized in that: This model is a sensor data detection model based on the hidden Markov model HMM. The hidden Markov model HMM is a dynamic Bayesian network generation model that can be used for labeling problems. It describes the process of randomly generating an unobservable state sequence by an HMM chain, and then generating an observation sequence from the state sequence; the HMM is determined by the initial probability distribution, the state transition probability distribution and the observation probability distribution, where the initial probability distribution and the state transition probability distribution determine the HMM chain and generate an unobservable state sequence; the observation probability distribution determines how to generate an observation sequence from the state sequence; the specific operations are as follows: (1) Importing data monitored by various sensors and fault status samples of electrical equipment in the distribution cabinet into the hidden Markov model HMM to establish the hidden Markov model HMM; (2) After the hidden Markov model HMM is established, the data collected by the current sensor is compared and analyzed with the observation sequence, and finally the current state judgment value is output.

2. According to claim 1, a timing analysis system for monitoring the internal operating status of electrical equipment in a power distribution cabinet is characterized in that: The hidden Markov model HMM is defined as follows: The set of all possible states is Q, the set of possible observations is V, the state sequence is I, and the observation sequence is O: Q={q1,q2,q3,…,q N } V={v1,v2,v3,…,v M } I={i1,i2,i3,…,i T } O={o1,o2,o3,…,o T } Where: N is the number of states that may correspond to the HMM chain, M is the number of observations that may be generated for each state, and T is the sequence length; the state sequence I is the internal operating state standard of the electrical equipment in the distribution cabinet, and its internal value i is the state standard value; the state set Q is the actual operating state of the electrical equipment in the distribution cabinet after running for a period of time, and its internal value q is the operating state value; the observation sequence O is the observation standard of each type of sensor under each state, and its internal value o is the observation standard value; the observation set V is the sum of the observation values ​​v generated by each type of sensor in each state.

3. A timing analysis system for monitoring the internal operating status of electrical equipment in a power distribution cabinet according to claim 2, characterized in that: The initial probability distribution is as follows: The initial state probability vector is π: π＝(π i ) π i =P(i t =q i ),i=1,2,3,…,N π i At time t, it is in state q i The initial state probability of The state transition probability distribution is specifically as follows: The state transition probability matrix is ​​A: A=[a ij ] N×N a ij =P(i t+1 =q j |i t =q i ),i=1,2,3,…,N a ij At time t, it is in state q i Under the condition of j probability; The observation probability distribution is as follows: The observation probability matrix is ​​B: B=[b j (k)] N×M b j (k)=P(o t =v k |i t =q j ),k=1,2,3,…,M;j=1,2,3,…,N b j (k) is in state q at time t j Generate observations v under the condition k The probability of, k refers to the sequence number of the observation value, j refers to the sequence number of the state value; In the process of establishing the hidden Markov model HMM, λ=(π,A,B) is used to represent the hidden Markov model HMM. The input data of the model is the state sequence I, the observation sequence O and the observation set V are preset by the model, and the state set Q is the sample data.

4. A time series analysis system for monitoring the internal operating status of electrical equipment in a power distribution cabinet according to claim 2 or 3, characterized in that: When using the Hidden Markov Model HMM for comparison analysis, it is necessary to solve the probability calculation problem. The probability calculation problem refers to: given the model λ=(π,A,B), the observation sequence O={o1,o2,o3,…,o T }, calculate the probability P(O|λ) of the observation sequence O appearing under the given model λ; specifically, the forward algorithm and the backward algorithm are combined with the agreed algorithm to completely calculate the probability P(O|λ) of the observation sequence O appearing under the given model λ.

5. A timing analysis system for monitoring the internal operating status of electrical equipment in a power distribution cabinet according to claim 4, characterized in that: The forward algorithm and the backward algorithm are used together to agree on the following algorithm: the forward algorithm needs to define the forward probability, and the partial observation sequence at time t is o1, o2, o3, …, o T And the state is q i The probability of is the forward probability: a t (i)=P(o1,o2,o3,...,o T ,i t =q i |λ) The forward probability a is obtained by recursive method t+1 (i) and the observed sequence probability P(O|λ); Initial value: a1(i)=π i b i (o1),i=1,2,3,...,N where π i For the state q i The initial state probability of b i (o1) is in state q i The probability of generating the observation value o1 under the condition of ; Recursively, for t=1,2,3,…,T-1: termination: where a T (i) is the final value of the forward probability; The backward algorithm needs to define the backward probability. Under the condition that the state at time t is qi, the partial observation sequence from t+1 to T is o t+1 ,o t+2 ,o t+3 ,…,o T The probability of is the backward probability: β t (i)=P(o t+1 ,o t+2 ,o t+3 ,...,o T ,i t =q i |λ) The backward probability β is obtained by recursive method t+1 (i) and the observed sequence probability P(O|λ) Initial value: β T (i)=1,i=1,2,3,…,N Recursively, for t=T-1,T-2,T-3,…,1: termination: where β T (i) is the initial value of the backward probability, β t (i) is the final value of the backward probability; The observation sequence probability P(O|λ) defined by the forward algorithm and the backward algorithm is:

6. A time series analysis system for monitoring the internal operating status of electrical equipment in a power distribution cabinet according to claim 5, characterized in that When using the hidden Markov model HMM for comparative analysis, a prediction problem is required; the prediction problem is a known HMM model λ=(π,A,B) and an observation sequence O={o1,o2,o3,…,o T }, calculate the corresponding state sequence I={i1,i2,i3,…,i T }.

7. A time series analysis system for monitoring the internal operating status of electrical equipment in a power distribution cabinet according to claim 6, characterized in that: In view of the prediction problem, an improved Viterbi algorithm is proposed to find the state sequence corresponding to the probabilistic optimal path. The specific calculation method is: initialization: δ1(i)=π i b i (o1), i=1,2,3,…,N ψ1(i)=0,i=1,2,3,…,N Among them, δ1(i) is the initialization optimal path probability value, ψ1(i) is the correction parameter; Recursively, for t=2,3,4,…,T: termination: Optimal path backtracking, for t = T-1, T-2, T-3, ..., 1: Where P * To initialize the maximum probability of the optimal path; is the final state value of the optimal path, is the optimal path state value; Then the state sequence corresponding to the optimal path is:

8. The timing analysis system for monitoring the internal operating status of electrical equipment in a power distribution cabinet according to claim 5 is characterized in that When using the Hidden Markov Model HMM for comparative analysis, a learning problem needs to be solved; the learning problem is that given training data only contains S observation sequences of length T O = {o1, o2, o3, ..., o s }, the goal is to learn the parameters of the HMM model λ=(π,A,B) so that the probability P(O|λ) of the observed sequence O under the model λ is maximized, that is, the parameters are estimated using the maximum likelihood estimation method. At this time, the HMM is P(O|λ)=∑ I P(O|I,λ)P(I|λ).

9. A time series analysis system for monitoring the internal operating status of electrical equipment in a power distribution cabinet according to claim 8, characterized in that: In view of the learning problem, an unsupervised learning algorithm is proposed to calculate and determine the estimated parameters of the maximum likelihood estimation method. The specific calculation method is: Determine the log-likelihood function of the complete data, all observations are O = {o1,o2,o3,…,o T }, all hidden data are I={i1,i2,i3,…,i T }, the complete data is O, I = {o1,o2,o3,…,o T ,i1,i2,i3,…,i T }, the log-likelihood function of the complete data is logP(O|I,λ); The first step is to find the Q function Where: λ is the HMM parameter to be maximized; is the current estimate of the HMM parameters; The second step is to maximize the Q function Find the model parameters π, A, B; The third step is to combine the forward algorithm with the backward algorithm and use γ t (i) represents the probability of the forward and backward algorithm optimizing the path, ξ t (i,j) represents the forward and backward algorithm optimization parameters, where: The fourth step is to design the optimization parameter estimation of the EM algorithm. The formula is: p i =γ1(i)

Images