A time sequence prediction method, device, storage medium and apparatus of a disconnector

Abstract

The application discloses a kind of time sequence prediction method, equipment, storage medium and device of isolator, comprising: obtaining the fault data of different types of isolator, and according to fault data Hidden Semi-Markov Model is constructed, according to the time distribution of fault occurrence determined according to fault data, according to the time distribution Hidden Semi-Markov Model determines fault parameter set, according to the fault of isolator is time sequence predicted according to forward-backward iteration algorithm and the fault parameter set of Hidden Semi-Markov Model;Since the application is constructed Hidden Semi-Markov Model, according to Hidden Semi-Markov Model and forward-backward iteration algorithm, the probability of isolator fault occurrence is predicted, compared with the way that the existing strain gauge is installed on the operating lever of isolator to predict the fault of isolator, the prediction accuracy of the probability of isolator fault occurrence is improved.

Description

Technical Field

[0001] This invention belongs to the field of fault prediction technology, and particularly relates to a timing prediction method, device, storage medium and apparatus for disconnecting switches. Background Technology

[0002] Disconnect switches are mainly used to isolate high-voltage maintenance equipment from live equipment during power equipment maintenance to ensure the personal safety of maintenance personnel. However, disconnect switches are exposed to the outdoors for a long time, and various factors can cause damage to the disconnect switches themselves. Once a disconnect switch fails, serious consequences will occur.

[0003] Testing of disconnect switches typically involves installing strain gauges on the operating lever of the disconnect switch to monitor the switch. However, the installation position of the strain gauges is extremely critical; if the position is incorrect, it can lead to significant deviations in the prediction results. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a timing prediction method, device, storage medium and apparatus for disconnecting switches, so as to solve the technical problem that in the prior art, monitoring the fault of disconnecting switches by strain gauges on the operating rod of the disconnecting switch is subject to extremely high requirements for the installation position of the strain gauges. If the installation position is incorrect, it will lead to a large deviation in the prediction result, resulting in low accuracy of the strain gauge in predicting the probability of disconnecting switch faults.

[0005] The technical solution of this invention is:

[0006] A timing prediction method for disconnecting switches, the method comprising:

[0007] Obtain fault data for different types of disconnect switches, and construct a hidden semi-Markov model based on the fault data;

[0008] Determine the time distribution of fault occurrences based on fault data;

[0009] Determine the fault parameter set of the hidden semi-Markov model based on the time distribution;

[0010] The timing prediction of faults in disconnecting switches is performed based on the fault parameter set of the forward and backward iterative algorithm and the hidden semi-Markov model.

[0011] The steps for time-series prediction of faults in disconnecting switches based on the fault parameter set of the forward and backward iterative algorithm and the hidden semi-Markov model include:

[0012] The backward recursive model is determined based on the fault parameter set of the forward and backward iterative algorithm and the hidden semi-Markov model;

[0013] Preprocessing auxiliary variables are determined based on the hidden semi-Markov model;

[0014] The initial conditions of the auxiliary variables are determined based on the backward recursive model and the preprocessed auxiliary variables;

[0015] The failure probability of the disconnecting switch is predicted based on the initial conditions of the auxiliary variables and the timing prediction probability algorithm.

[0016] The timing prediction probability algorithm includes a timing prediction probability formula and a step of predicting the fault probability of the disconnecting switch based on the initial conditions of the auxiliary variables and the timing prediction probability algorithm, including:

[0017] The probability of a fault occurring in the disconnector at the current moment is calculated based on the initial conditions of the auxiliary variables and the time-series prediction probability formula.

[0018] The time series prediction probability formula is as follows:

[0019]

[0020] In the formula, S T and H i This represents the state value of the i-th fault occurrence, O 1:T Let λ represent the observed value at time t, λ represent the auxiliary parameter, and γ represent the value at time t. T (i) represents the initial condition of the auxiliary variable, P(O 1:T |λ) represents the conditional probability;

[0021] Calculate the probability of a fault occurring in the disconnector at the current time step based on the probability of a fault occurring in the disconnector at the current moment.

[0022] The steps for determining the backward recursive model based on the iterative algorithm and the fault parameter set of the hidden semi-Markov model include:

[0023] The backward variables are determined based on the fault parameter set of the hidden semi-Markov model;

[0024] The backward recursive model is determined based on the forward and backward iterative algorithm and the backward variable.

[0025] The steps for determining the initial conditions of auxiliary variables based on the backward recursive model and preprocessed auxiliary variables include:

[0026] The forward variables are determined based on the fault parameter set of the hidden semi-Markov model;

[0027] The preprocessing auxiliary variables are transformed based on the forward variables to obtain the auxiliary variables to be derived;

[0028] Based on the forward and backward recursive models, the auxiliary variables to be derived are obtained, along with their initial conditions.

[0029] The steps for acquiring fault data of different types of disconnect switches and constructing a hidden semi-Markov model based on the fault data include:

[0030] A training set is constructed based on fault data of different types of disconnect switches;

[0031] The model is trained using the training set to obtain a hidden semi-Markov model.

[0032] The steps for determining the fault parameter set of the hidden semi-Markov model based on the time distribution include:

[0033] Determine the remaining time probability of the disconnecting switch under the same fault condition based on the time distribution;

[0034] The remaining time probability of the disconnecting switch under the same fault state is added to the parameter set of the model training, and the fault parameter set of the hidden semi-Markov model is constructed based on the parameter set of the model training.

[0035] A timing prediction device for a disconnecting switch, the timing prediction device for the disconnecting switch comprising: a memory, a processor, and a timing prediction program for the disconnecting switch stored in the memory and executable on the processor, wherein the timing prediction program for the disconnecting switch, when executed by the processor, implements a timing prediction method for the disconnecting switch.

[0036] A storage medium storing a timing prediction program for a disconnect switch, wherein the timing prediction program for the disconnect switch is executed by a processor to implement a timing prediction method for the disconnect switch.

[0037] A timing prediction device for a disconnecting switch, the device comprising: a model determination module, a time distribution module, a parameter determination module, and a timing prediction module;

[0038] The model determination module is used to acquire fault data of different types of disconnect switches and construct a hidden semi-Markov model based on the fault data.

[0039] The time distribution module is used to determine the time distribution of fault occurrence based on the fault data;

[0040] The parameter determination module is used to determine the fault parameter set of the hidden semi-Markov model based on the time distribution.

[0041] The timing prediction module is used to predict the timing of the faults of the disconnecting switch based on the forward and backward iterative algorithm and the fault parameter set of the hidden semi-Markov model.

[0042] The beneficial effects of this invention are:

[0043] This invention acquires fault data of different types of disconnect switches, constructs a hidden semi-Markov model based on the fault data, determines the time distribution of fault occurrence based on the fault data, determines the fault parameter set of the hidden semi-Markov model based on the time distribution, and performs time-series prediction of disconnect switch faults based on a forward and backward iterative algorithm and the fault parameter set of the hidden semi-Markov model. Because this invention constructs a hidden semi-Markov model and predicts the probability of disconnect switch fault occurrence based on the hidden semi-Markov model and the forward and backward iterative algorithm, compared with the existing method of predicting disconnect switch faults by installing strain gauges on the disconnect switch operating rod, this invention improves the prediction accuracy of disconnect switch fault probability.

[0044] This invention solves the technical problem in existing technologies that monitor disconnect switch faults using strain gauges on the operating lever of the disconnect switch. Due to the extremely high requirements for the installation position of the strain gauges, incorrect installation can lead to significant deviations in the prediction results, resulting in low accuracy in predicting the probability of disconnect switch faults. Attached Figure Description

[0045] Figure 1 This is a schematic diagram of the timing prediction device for the disconnecting switch in the hardware operating environment involved in the embodiments of the present invention;

[0046] Figure 2 This is a flowchart illustrating the first embodiment of the timing prediction method for disconnecting switches according to the present invention.

[0047] Figure 3 This is a flowchart illustrating a second embodiment of the timing prediction method for disconnecting switches according to the present invention.

[0048] Figure 4 This is a flowchart illustrating the third embodiment of the timing prediction method for disconnecting switches according to the present invention.

[0049] Figure 5 This is a structural block diagram of the first embodiment of the timing prediction device for the disconnecting switch of the present invention. Detailed Implementation

[0050] Reference Figure 1 , Figure 1 This is a schematic diagram of the timing prediction device for the disconnecting switch in the hardware operating environment involved in the embodiments of the present invention.

[0051] like Figure 1As shown, the timing prediction device for the disconnecting switch may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to establish communication between these components. The user interface 1003 may include a display screen, and optionally, it may also include a standard wired interface or a wireless interface. In this invention, the wired interface of the user interface 1003 may be a USB interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be a high-speed random access memory (RAM) or a non-volatile memory (NVM), such as a disk storage device. The memory 1005 may also optionally be a storage device independent of the aforementioned processor 1001.

[0052] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the timing prediction device for disconnecting switches, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0053] like Figure 1 As shown, the memory 1005, which is identified as a computer storage medium, may include an operating system, a network communication module, a user interface module, and a timing prediction program for disconnecting switches.

[0054] exist Figure 1 In the timing prediction device for the disconnecting switch shown, the network interface 1004 is mainly used to connect to the backend server and communicate data with the backend server; the user interface 1003 is mainly used to connect to the user equipment; the timing prediction device for the disconnecting switch calls the timing prediction program for the disconnecting switch stored in the memory 1005 through the processor 1001 and executes the timing prediction method for the disconnecting switch provided in this embodiment of the invention.

[0055] Based on the above hardware structure, an embodiment of the timing prediction method for the disconnecting switch of the present invention is proposed.

[0056] Reference Figure 2 , Figure 2 This is a flowchart illustrating the first embodiment of the timing prediction method for disconnecting switches according to the present invention, which presents the first embodiment of the timing prediction method for disconnecting switches according to the present invention.

[0057] Step S10: Obtain fault data of different types of disconnect switches, and construct a hidden semi-Markov model based on the fault data.

[0058] It should be noted that the execution subject of the method in this embodiment can be a computing service device with data processing, network communication and program running functions, such as a mobile phone, tablet computer, personal computer, etc., and this embodiment does not limit it.

[0059] It should be understood that because disconnect switches are exposed to the outdoors for extended periods, various factors can cause them to malfunction. A malfunction can lead to serious consequences; therefore, it is necessary to predict the probability of disconnect switch malfunctions so that personnel can intervene promptly. However, current technology predicts the probability of disconnect switch malfunctions by installing strain gauges on the operating rod. This method places high demands on the installation position of the strain gauges; any deviation in the installation position will result in unsatisfactory prediction results.

[0060] To overcome the above-mentioned defects, this embodiment constructs a hidden semi-Markov model and a hidden semi-Markov model parameter set. The hidden semi-Markov model uses a forward and backward iterative algorithm to calculate the hidden semi-Markov model parameter set to obtain the probability of the disconnecting switch malfunctioning.

[0061] It should be noted that the fault types of disconnecting switches can include overheating of the disconnecting switch contacts, damage or flashover of the disconnecting switch insulators, refusal to open or close the disconnecting switch, automatic tripping and closing of the disconnecting switch, and accidental opening or closing of the disconnecting switch, etc. This embodiment does not impose any limitations on these types of faults. By collecting data when a fault occurs in the disconnecting switch, a hidden semi-Markov model is constructed based on this fault data. For example, when a fault occurs due to overheating of the disconnecting switch contacts, the temperature of the disconnecting switch contacts is used as the fault data.

[0062] It should be noted that the Hidden Semi-Markov Model is a simple dynamic Bayesian network. The Hidden Semi-Markov Model has only one discrete hidden state variable and a set of discrete or continuous observation nodes. That is, the Hidden Semi-Markov Model takes the time that the disconnector stays in each fault state as a discrete hidden state variable and takes the parameters within this fault time period as a discrete set.

[0063] Understandably, the concept proposed by the Hidden Semi-Markov Model is that there are two transitions between fault states: one is to remain in the current fault state, and the other is to enter the next, more severe fault state.

[0064] Furthermore, in order to calculate the probability of disconnector switch failure based on the hidden semi-Markov model and improve prediction accuracy, step S10 in this embodiment also includes:

[0065] A training set is constructed based on the fault data of the different types of disconnect switches;

[0066] The hidden semi-Markov model is obtained by training the model based on the training set.

[0067] It is understandable that different types of fault data already exist and have been validated, and the training set is constructed using these fault data.

[0068] It should be understood that the training set needs to contain a sufficient number of fault parameters. Only when there are enough fault parameters can the quality of the hidden semi-Markov model be evaluated through the degree of fit of the training, thereby reducing the error.

[0069] In practice, a training set is constructed based on different types of fault data. The model is trained by cleaning these fault data and scaling the data features to obtain a hidden semi-Markov model.

[0070] Step S20: Determine the time distribution of the fault occurrence based on the fault data.

[0071] It should be noted that the time distribution of the current fault state is determined based on the distribution of fault data at the current time step.

[0072] Step S30: Determine the fault parameter set of the hidden semi-Markov model based on the time distribution.

[0073] It's important to note that the Hidden Semi-Markov Model (HSM) has only two possible choices at each time step: remain in the same fault state or transition to the next fault state. The probability of remaining time in the same state is generated using a time distribution, and the transition probability can only be used once. Therefore, once the duration of remaining in a fault state ends, the fault state changes. Thus, the transition matrix of the HSM at each time step, considering both possible transitions, the duration of remaining in a state, and the time distribution, shows that the only valid option is to transition to the next state. Therefore, it is assumed that each row of the HSM transition matrix contains only one non-zero element, indicating the transition to the next fault state. Based on these assumptions and the time distribution, the parameter set of the HSM for each fault state is determined.

[0074] Understandably, different fault states under the time distribution have different fault data. Based on different fault data, we can know the time when the disconnector will enter the next fault state, and determine the probability of the remaining time of the disconnector's current state based on the time distribution.

[0075] Step S40: Perform time-series prediction of the fault of the disconnecting switch based on the forward and backward iterative algorithm and the fault parameter set of the hidden semi-Markov model.

[0076] It should be noted that the failure probability of disconnecting switches is predicted by using a time-series probability model. The time-series probability model can be a set of time-expandable probability models. The simplest time-series probability model is the Hidden Markov Model. By proposing a new concept, the Hidden Semi-Markov Model is improved to obtain the Hidden Semi-Markov Model. The failure probability of disconnecting switches is predicted by using the Hidden Semi-Markov Model and the forward and backward iterative algorithm.

[0077] In the specific implementation, the fault parameter set is determined by the front-to-back line iterative algorithm and the hidden semi-Markov model to determine the backward recursive model. The preprocessing auxiliary variables are determined by the hidden semi-Markov model. The initial conditions of the auxiliary variables are determined by the backward recursive model and the preprocessing auxiliary variables. The probability of the disconnector sending a fault is determined by the initial conditions of the auxiliary variables and the timing prediction probability algorithm.

[0078] This invention acquires fault data of different types of disconnect switches, constructs a hidden semi-Markov model based on the fault data, determines the time distribution of fault occurrence based on the fault data, determines the fault parameter set of the hidden semi-Markov model based on the time distribution, and performs time-series prediction of disconnect switch faults based on a forward and backward iterative algorithm and the fault parameter set of the hidden semi-Markov model. Because this invention constructs a hidden semi-Markov model and predicts the probability of disconnect switch fault occurrence based on the hidden semi-Markov model and the forward and backward iterative algorithm, compared with the existing method of predicting disconnect switch faults by installing strain gauges on the disconnect switch operating rod, this invention improves the prediction accuracy of disconnect switch fault probability.

[0079] Reference Figure 3 , Figure 3 This is a flowchart illustrating a second embodiment of the timing prediction method for disconnecting switches according to the present invention.

[0080] In the second embodiment, step S40 includes:

[0081] Step S401: Determine the backward recursive model based on the forward and backward iterative algorithm and the fault parameter set of the hidden semi-Markov model.

[0082] It should be understood that while Hidden Markov Models (HMMs) use a geometric distribution for time distribution, a fixed time distribution can lead to unsatisfactory prediction results when the time distribution is not geometric. Since geometric distributions are not always geometric in many practical applications, a more complex time-series probability model, the Hidden Semi-Markov Model (HSM), is proposed to overcome these shortcomings and address the problem of fixed-duration distributions.

[0083] It should be noted that the forward and backward iteration algorithm is a recursive algorithm. By setting forward and backward variables, the forward and backward iteration algorithm is simplified. The failure probability of the disconnecting switch is calculated by using the forward and backward iteration algorithm and the timing prediction probability model.

[0084] It should be noted that the backward recursive model is simply the backward recursive formula, obtained by simplifying and rewriting the backward variables. The backward variables are determined based on the fault parameter set of the hidden semi-Markov model.

[0085] Furthermore, in order to simplify the timing prediction probability algorithm for disconnecting switches and improve computational efficiency, step S401 in this embodiment may include:

[0086] The backward variables are determined based on the fault parameter set of the hidden semi-Markov model;

[0087] The backward recursive model is determined based on the forward and backward iterative algorithm and the backward variables.

[0088] It should be noted that the backward variable is:

[0089] β t (i,k)≡P(O t+1:T |(s t ,t i )=(H i ,k))

[0090] In the formula, O t+1:T The input is the observed values ​​over time from t+1 to T, (s) t ,t i )=(H i (k) indicates that the fault state should remain in this state for the next k time steps, and then transition to the next state. H i It is the i-th fault state, which is the fault state in the fault parameter set of the hidden semi-Markov model.

[0091] In the specific implementation, the backward variable is recursively processed according to the initial recursion condition of the forward and backward iteration algorithm, and the backward variable is simplified to obtain the backward recursion formula.

[0092] Step S402: Determine the preprocessing auxiliary variables based on the hidden semi-Markov model.

[0093] It should be noted that the preprocessing variables are defined to simplify the time-series prediction probability algorithm for disconnecting switches. The preprocessing variables are estimates of the probability of disconnecting switch failure based on the entire observation sequence, that is, estimates of the current fault state of the disconnecting switch based on the fault parameter set of the entire hidden semi-Markov model.

[0094] In the specific implementation, the state value of the fault parameter at each time step is estimated based on the observations of the hidden semi-Markov model. The time series prediction probability algorithm is simplified by further defining preprocessing auxiliary variables.

[0095] Step S403: Determine the initial conditions of the auxiliary variables based on the backward recursive model and the preprocessing auxiliary variables.

[0096] In the specific implementation, the forward variables are determined based on the hidden semi-Markov model, the joint probability formula of the preprocessing auxiliary variables is determined based on the forward variables, and the preprocessing variables and the initial conditions of the preprocessing variables are derived based on the joint probability formula and the backward recursive formula.

[0097] Furthermore, in order to simplify the timing prediction probability algorithm for disconnecting switches and improve computational efficiency, step S403 in this embodiment may include:

[0098] The forward variables are determined based on the fault parameter set of the hidden semi-Markov model;

[0099] The preprocessing auxiliary variable is transformed based on the forward variable to obtain the auxiliary variable to be derived;

[0100] The auxiliary variable and its initial conditions are obtained by deriving the auxiliary variable to be derived based on the forward variable and the backward recursive model.

[0101] It should be noted that the forward variables are the joint probability of the observed input at time t for the i-th state as the input progresses from 1 to t, and the remaining time in that state at subsequent times. The forward variables are determined based on the fault parameter set of the hidden semi-Markov model and the observed values.

[0102] It should be noted that the process of transforming the preprocessing auxiliary variable into the auxiliary variable to be derived is as follows:

[0103] P(O 1:T ,S t =H i ,S t+1 =H i )=P(O 1:T ,S t =H i )-P(O 1:T ,S t =H i ,S t+1 =H i+1 )=P(O 1:T ,S t+1 =H i )-P(O 1:T ,S t =Hi-1 ,S t+1 =H i )

[0104] O 1:T S represents the observation value at time t. t =H i S represents the i-th fault state value. t+1 =H i S represents the i-th fault state at the current time step. t+1 =H i+1 This represents the i-th fault state value before entering the next fault state.

[0105] Step S404: Predict the failure probability of the disconnecting switch based on the initial conditions of the auxiliary variables and the timing prediction probability algorithm.

[0106] It should be noted that timing prediction of disconnect switches involves predicting the fault state of the disconnect switch at time T, given all the observation values ​​within a time step range of 1 to T.

[0107] It is understandable that time series probability prediction algorithms include time series probability prediction formulas.

[0108] Furthermore, in order to predict the probability of a disconnector switch failure, step S404 in this embodiment may include:

[0109] The probability of the disconnecting switch failing at the current moment is calculated based on the initial conditions of the auxiliary variables and the time-series prediction probability formula.

[0110] The probability of the disconnector switch failing at the current time step is calculated based on the probability of the disconnector switch failing at the current moment.

[0111] It should be noted that the probability formula for time series prediction is:

[0112]

[0113] In the formula, S T and H i This represents the state value of the i-th fault occurrence, O 1:T Let λ represent the observed value at time t, λ represent the auxiliary parameter, and γ represent the value at time t. T (i) represents the initial condition of the auxiliary variable, P(O 1:T |λ) represents the conditional probability.

[0114] It should be noted that the probability of each fault state of the disconnecting switch needs to be calculated. Therefore, based on the initial conditions of the auxiliary variables and the time-series prediction probability algorithm, the probability of each disconnecting switch fault is calculated at the current time t. The probability of each disconnecting switch fault can be expressed as:

[0115]

[0116] In the formula, γ represents the probability of each disconnector switch malfunctioning. T (i) represents the initial condition of the auxiliary variable.

[0117] It should be noted that the continuous prediction of the failure probability of the disconnector switch at the current time step T can be expressed as:

[0118]

[0119] In the formula, C T This indicates that the probability of a disconnector switch failure is continuously predicted at the current time step T. H represents the probability of each disconnector switch malfunctioning. i This represents the state value of the i-th fault state.

[0120] In the second embodiment, a backward recursive model is determined based on the fault parameter set of the forward and backward iterative algorithm and the hidden semi-Markov model. Preprocessing auxiliary variables are determined based on the hidden semi-Markov model. Initial conditions for the auxiliary variables are determined based on the backward recursive model and the preprocessing auxiliary variables. The fault probability of the disconnecting switch is predicted based on the initial conditions of the auxiliary variables and the time-series prediction probability algorithm. Because this embodiment defines the backward recursive model and the preprocessing auxiliary variables, and performs calculations and inferences using the backward recursive model and the preprocessing auxiliary variables, combined with the time-series prediction probability algorithm, the time-series prediction probability algorithm is simplified by defining the backward recursive model and the preprocessing auxiliary variables, thus improving the efficiency of predicting the probability of fault occurrence.

[0121] Reference Figure 4 , Figure 4 This is a flowchart illustrating the third embodiment of the timing prediction method for disconnecting switches according to the present invention. The third embodiment of the timing prediction method for disconnecting switches according to the present invention is presented.

[0122] In the third embodiment, step S30 includes:

[0123] Step S301: Determine the remaining time probability of the disconnecting switch under the same fault state based on the time distribution.

[0124] It is understandable that the remaining time the isolating switch remains in the current fault state at the current time step can be determined based on the time distribution.

[0125] It should be noted that each time step has only two options: maintain the current fault state or proceed to the next fault state. The next fault state can be the same type as the current fault but with more severe damage, or it can be a fault state of a different type. Using the time distribution to determine the remaining time probability of the disconnector under the same fault state is to determine whether the disconnector's fault state will proceed to the next fault state.

[0126] It should be noted that since the fault condition changes after the duration of the fault state ends, the probability of the remaining time of the disconnector in the same fault state is determined based on the time distribution in order to predict the probability of the disconnector malfunctioning.

[0127] Step S302: Add the remaining time probability of the disconnecting switch under the same fault state to the parameter set of the model training, and construct the fault parameter set of the hidden semi-Markov model based on the parameter set of the model training.

[0128] Understandably, based on the iterative algorithm and the time-series prediction probability model, the parameters in the fault parameter set of the hidden semi-Markov model are calculated using the training set to obtain the probability of the disconnecting switch failing.

[0129] In the third embodiment, the remaining time probability of the disconnecting switch under the same fault state is determined according to the time distribution, and the remaining time probability of the disconnecting switch under the same fault state is added to the parameter set of the model training. The fault parameter set of the hidden semi-Markov model is constructed based on the parameter set of the model training. Since this embodiment determines the parameter set of the model training according to the time distribution and constructs the fault parameter set of the hidden semi-Markov model based on the parameter set of the model training, the fault probability of the disconnecting switch is predicted in time series by calling the parameters in the fault parameter set, instead of using strain gauges to predict the fault probability of the disconnecting switch, thus improving the accuracy of calculating the fault probability.

[0130] Furthermore, this embodiment of the invention also proposes a storage medium storing a timing prediction program for a disconnect switch. When the timing prediction program for the disconnect switch is executed by a processor, it implements the steps of the timing prediction method for the disconnect switch as described above.

[0131] In addition, refer to Figure 5 , Figure 5This is a structural block diagram of the first embodiment of the timing prediction device for the disconnecting switch of the present invention. The timing prediction device for the disconnecting switch includes: a model determination module 10, a time distribution module 20, a parameter determination module 30, and a timing prediction module 40.

[0132] The model determination module 10 is used to acquire fault data of different types of disconnect switches and construct a hidden semi-Markov model based on the fault data.

[0133] The time distribution module 20 is used to determine the time distribution of fault occurrence based on the fault data;

[0134] The parameter determination module 30 is used to determine the fault parameter set of the hidden semi-Markov model based on the time distribution.

[0135] The timing prediction module 40 is used to predict the timing of the faults of the disconnecting switch based on the forward and backward iterative algorithm and the fault parameter set of the hidden semi-Markov model.

[0136] In this embodiment, a method is disclosed for acquiring fault data of different types of disconnect switches, constructing a hidden semi-Markov model based on the fault data, determining the time distribution of fault occurrence based on the fault data, determining the fault parameter set of the hidden semi-Markov model based on the time distribution, and performing time-series prediction of disconnect switch faults based on a forward and backward iterative algorithm and the fault parameter set of the hidden semi-Markov model. Because this invention constructs a hidden semi-Markov model and predicts the probability of disconnect switch fault occurrence based on the hidden semi-Markov model and the forward and backward iterative algorithm, compared with the existing method of predicting disconnect switch faults by installing strain gauges on the disconnect switch operating lever, this invention improves the prediction accuracy of disconnect switch fault probability.

[0137] Based on the first embodiment of the transformer reliability assessment device of the present invention, a second embodiment of the transformer reliability assessment device of the present invention is proposed.

[0138] In this embodiment, the time series prediction module 40 is further configured to determine the backward recursive model based on the forward and backward iterative algorithm and the fault parameter set of the hidden semi-Markov model.

[0139] Furthermore, the time series prediction module 40 is also used to determine preprocessing auxiliary variables based on the hidden semi-Markov model.

[0140] Furthermore, the time series prediction module 40 is also used to determine the initial conditions of the auxiliary variables based on the backward recursive model and the preprocessing auxiliary variables.

[0141] Furthermore, the timing prediction module 40 is also used to predict the failure probability of the disconnecting switch based on the initial conditions of the auxiliary variable and the timing prediction probability algorithm.

[0142] Furthermore, the timing prediction module 40 is also used to calculate the probability of the disconnecting switch failing at the current moment based on the initial conditions of the auxiliary variable and the timing prediction probability formula.

[0143] Furthermore, the timing prediction module 40 is also used to calculate the probability of the disconnecting switch failing at the current time step based on the probability of the disconnecting switch failing at the current moment.

[0144] Furthermore, the time-series prediction module 40 is also used to determine backward variables based on the fault parameter set of the hidden semi-Markov model.

[0145] Furthermore, the time series prediction module 40 is also used to determine the backward recursive model based on the forward and backward iterative algorithm and the backward variable.

[0146] Furthermore, the time-series prediction module 40 is also used to determine forward variables based on the fault parameter set of the hidden semi-Markov model.

[0147] Furthermore, the time series prediction module 40 is also used to transform the preprocessing auxiliary variable based on the forward variable to obtain the auxiliary variable to be derived.

[0148] Furthermore, the time series prediction module 40 is also used to derive the auxiliary variable to be derived and the initial conditions of the auxiliary variable based on the forward variable and the backward recursive model.

[0149] Furthermore, the parameter determination module 30 is also used to determine the remaining time probability of the disconnecting switch under the same fault state based on the time distribution.

[0150] Furthermore, the parameter determination module 30 is also used to add the remaining time probability of the disconnecting switch under the same fault state to the parameter set of the model training, and construct the fault parameter set of the hidden semi-Markov model based on the parameter set of the model training.

[0151] Other embodiments or specific implementations of the business system architecture diagram generation device described in this invention can be found in the above-described method embodiments, and will not be repeated here.

Claims

1. A timing prediction method for disconnecting switches, characterized in that: The method includes: Obtain fault data for different types of disconnect switches, and construct a hidden semi-Markov model based on the fault data; Determine the time distribution of fault occurrences based on fault data; Determine the fault parameter set of the hidden semi-Markov model based on the time distribution; Based on the fault parameter set of the forward and backward iterative algorithm and the hidden semi-Markov model, the fault timing prediction of the disconnecting switch is performed, specifically including: The backward recursive model is determined based on the fault parameter set of the forward and backward iterative algorithm and the hidden semi-Markov model; Preprocessing auxiliary variables are determined based on the hidden semi-Markov model; The initial conditions of the auxiliary variables are determined based on the backward recursive model and the preprocessed auxiliary variables; The fault probability of the disconnecting switch is predicted based on the initial conditions of the auxiliary variables and the timing prediction probability algorithm, including: The probability of a fault occurring in the disconnector at the current moment is calculated based on the initial conditions of the auxiliary variables and the time-series prediction probability formula. The time series prediction probability formula is as follows: ； In the formula, This represents the state value of the i-th fault occurrence. Represents the observed value at time t. Indicates auxiliary parameters, This indicates the initial conditions of the auxiliary variable. Represents conditional probability; Calculate the probability of a fault occurring in the disconnector at the current time step based on the probability of a fault occurring in the disconnector at the current moment.

2. The timing prediction method for a disconnecting switch according to claim 1, characterized in that: The steps for determining the backward recursive model based on the iterative algorithm and the fault parameter set of the hidden semi-Markov model include: The backward variables are determined based on the fault parameter set of the hidden semi-Markov model; The backward recursive model is determined based on the forward and backward iterative algorithm and the backward variable.

3. The timing prediction method for a disconnecting switch according to claim 1, characterized in that: The steps for determining the initial conditions of auxiliary variables based on the backward recursive model and preprocessed auxiliary variables include: The forward variables are determined based on the fault parameter set of the hidden semi-Markov model; The preprocessing auxiliary variables are transformed based on the forward variables to obtain the auxiliary variables to be derived; Based on the forward and backward recursive models, the auxiliary variables to be derived are obtained, along with their initial conditions.

4. The timing prediction method for disconnecting switches as described in any one of claims 1 to 3, characterized in that: The steps for acquiring fault data of different types of disconnect switches and constructing a hidden semi-Markov model based on the fault data include: A training set is constructed based on fault data of different types of disconnect switches; The model is trained using the training set to obtain a hidden semi-Markov model.

5. The timing prediction method for disconnecting switches as described in any one of claims 1 to 3, characterized in that: The steps for determining the fault parameter set of the hidden semi-Markov model based on the time distribution include: Determine the remaining time probability of the disconnecting switch under the same fault condition based on the time distribution; The remaining time probability of the disconnecting switch under the same fault state is added to the parameter set of the model training, and the fault parameter set of the hidden semi-Markov model is constructed based on the parameter set of the model training.

6. A timing prediction device for a disconnecting switch, characterized in that, The timing prediction device for the disconnecting switch includes: a memory, a processor, and a timing prediction program for the disconnecting switch stored in the memory and executable on the processor. When the timing prediction program for the disconnecting switch is executed by the processor, it implements the timing prediction method for the disconnecting switch as described in any one of claims 1 to 5.

7. A storage medium, characterized in that, The storage medium stores a timing prediction program for the disconnecting switch, which, when executed by a processor, implements the timing prediction method for the disconnecting switch as described in any one of claims 1 to 5.

8. A timing prediction device for a disconnecting switch, characterized in that: The device includes: a model determination module, a time distribution module, a parameter determination module, and a time series prediction module; The model determination module is used to acquire fault data of different types of disconnect switches and construct a hidden semi-Markov model based on the fault data. The time distribution module is used to determine the time distribution of fault occurrence based on the fault data; The parameter determination module is used to determine the fault parameter set of the hidden semi-Markov model based on the time distribution. The timing prediction module is used to perform timing prediction of the faults of the disconnecting switch based on the forward and backward iterative algorithm and the fault parameter set of the hidden semi-Markov model, specifically including: The backward recursive model is determined based on the fault parameter set of the forward and backward iterative algorithm and the hidden semi-Markov model; Preprocessing auxiliary variables are determined based on the hidden semi-Markov model; The initial conditions of the auxiliary variables are determined based on the backward recursive model and the preprocessed auxiliary variables; The fault probability of the disconnecting switch is predicted based on the initial conditions of the auxiliary variables and the timing prediction probability algorithm, including: The probability of a fault occurring in the disconnector at the current moment is calculated based on the initial conditions of the auxiliary variables and the time-series prediction probability formula. The time series prediction probability formula is as follows: ； In the formula, This represents the state value of the i-th fault occurrence. Represents the observed value at time t. Indicates auxiliary parameters, This indicates the initial conditions of the auxiliary variable. Represents conditional probability; Calculate the probability of a fault occurring in the disconnector at the current time step based on the probability of a fault occurring in the disconnector at the current moment.

Images