Sensitive equipment jump stop probability prediction method and device, electronic equipment and storage medium

Abstract

The invention discloses a sensitive equipment jump-stop probability prediction method and device, electronic equipment and a storage medium, and the method comprises the steps: firstly obtaining first historical voltage sag data of a power system in a first historical time period and first historical jump-stop data of target sensitive equipment in the power system in the first historical time period; the first historical voltage sag data comprises n historical voltage sag states, and the first historical jump data comprises n historical working states of the target sensitive device in the n historical voltage sag states, and then constructing a first target hidden Markov model based on the first historical voltage sag data and the first historical jump data, and finally, inputting the target voltage sag state into the first target hidden Markov model to obtain a first jump stop probability of the target sensitive equipment in the target voltage sag state. By adopting the embodiment of the invention, the accuracy of evaluating the jump stop probability of the sensitive equipment under the voltage sag is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of voltage sag, and in particular to a method, device, electronic device and storage medium for predicting the probability of tripping of sensitive equipment. Background Art

[0002] Voltage sag is one of the most serious power quality problems, which will seriously affect the operation of sensitive equipment, such as personal computers, trip units, adjustable speed drives and programmable logic controllers. Sensitive equipment has extremely limited tolerance to voltage sags and is prone to tripping under voltage sags, causing industrial users to stop production and produce poorly, bringing huge economic losses to users. Obtaining the voltage sag tolerance of sensitive equipment and evaluating the probability of sensitive equipment tripping under voltage sags are important bases for users to formulate voltage sag management plans. At present, the accuracy of evaluating the probability of sensitive equipment tripping under voltage sags is low. Summary of the invention

[0003] The embodiments of the present application provide a method, an apparatus, an electronic device, and a storage medium for predicting the probability of a sensitive device tripping, which improves the accuracy of evaluating the probability of a sensitive device tripping under a voltage sag.

[0004] In a first aspect, an embodiment of the present application provides a method for predicting the probability of a sensitive device tripping, including:

[0005] Acquire first historical voltage sag data of the power system in a first historical time period and first historical trip data of a target sensitive device in the power system in the first historical time period; the first historical voltage sag data includes n historical voltage sag states, and the first historical trip data includes n historical working states of the target sensitive device under the n historical voltage sag states, the historical working states include trip states or normal working states, each historical working state corresponds to a historical voltage sag state, and n is an integer greater than 1;

[0006] Building a first target hidden Markov model based on the first historical voltage sag data and the first historical trip data;

[0007] The target voltage sag state is input into the first target hidden Markov model to obtain a first trip probability of the target sensitive device under the target voltage sag state.

[0008] In a second aspect, an embodiment of the present application provides a device for predicting the probability of a tripping of a sensitive device, the device comprising: an acquisition unit and a processing unit;

[0009] The obtaining unit is configured to obtain first historical voltage sag data of the power system in a first historical time period and first historical tripping data of a target sensitive device in the power system in the first historical time period; the first historical voltage sag data includes n historical voltage sag states, and the first historical tripping data includes n historical operating states of the target sensitive device in the n historical voltage sag states, where the historical operating state includes a tripping state or a normal operating state, and each historical operating state corresponds to a historical voltage sag state, and n is an integer greater than 1;

[0010] The processing unit is configured to construct a first target hidden Markov model based on the first historical voltage sag data and the first historical tripping data;

[0011] Input the target voltage sag state into the first target hidden Markov model to obtain a first tripping probability of the target sensitive device in the target voltage sag state.

[0012] In a third aspect, an embodiment of the present invention provides an electronic device, including: a processor, a memory, a communication interface, and one or more programs, where one or more programs are stored in the memory and are configured to be executed by the processor so that the electronic device executes the method according to the first aspect.

[0013] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method according to the first aspect.

[0014] In a fifth aspect, an embodiment of the present invention provides a computer program product, where the computer program product includes a non-transitory computer-readable storage medium storing a computer program, so that a computer executes the method according to the first aspect.

[0015] Implementing the embodiments of the present invention has the following beneficial effects:

[0016] It can be seen that for the method for predicting the tripping probability of sensitive devices described in the embodiments of the present invention, first, the first historical voltage sag data of the power system in the first historical time period and the first historical tripping data of the target sensitive device in the power system in the first historical time period are obtained. The first historical voltage sag data includes n historical voltage sag states, and the first historical tripping data includes n historical operating states of the target sensitive device in the n historical voltage sag states. The historical operating states include tripping states or normal operating states, and each historical operating state corresponds to a historical voltage sag state. n is an integer greater than 1. Then, a first target hidden Markov model is constructed based on the first historical voltage sag data and the first historical tripping data. Finally, the target voltage sag state is input into the first target hidden Markov model to obtain the first tripping probability of the target sensitive device in the target voltage sag state, improving the accuracy of evaluating the tripping probability of sensitive devices under voltage sags. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the background art, the following will describe the drawings required for use in the embodiments of the present application or the background art.

[0018] Figure 1 is a flowchart of a method for evaluating the reliability of a distribution network provided by an embodiment of the present application;

[0019] Figure 2 is an example diagram of a voltage sag provided by an embodiment of the present application;

[0020] Figure 3 is a flowchart of constructing a first target hidden Markov model provided by an embodiment of the present application;

[0021] Figure 4 is a flowchart of determining the tripping probability provided by an embodiment of the present application;

[0022] Figure 5 is a flowchart of determining the prediction accuracy provided by an embodiment of the present application;

[0023] Figure 6 is another flowchart of determining the prediction accuracy provided by an embodiment of the present application;

[0024] Figure 7 is a schematic structural diagram of a device for predicting the tripping probability of sensitive devices provided by an embodiment of the present application;

[0025] Figure 8 is a schematic structural diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] In order to enable those skilled in the art to better understand the solution of this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of this application.

[0027] The terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products, or devices.

[0028] Referring to "embodiments" herein means that specific features, structures, or characteristics described in connection with the embodiments can be included in at least one embodiment of this application. The phrase appears in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0029] First, the relevant terms involved in this application are explained as follows:

[0030] HMM model: The Hidden Markov Model (HMM) is a statistical model that has been widely used in the fields of speech recognition, natural language processing, bioinformatics, etc. It is a probability model about time series, which describes a Markov process with hidden unknown parameters. The "hidden" means that the internal state of the system is hidden and cannot be directly observed; "Markov" means that the model satisfies the Markov property, that is, the future state of the system only depends on the current state and is independent of the past state. The Hidden Markov Model can use the speech signal as the observation sequence, different speech units (such as phonemes) as the hidden states, and identify the text information corresponding to the speech through the Hidden Markov Model. In tasks such as part-of-speech tagging and named entity recognition, the Hidden Markov Model can also infer the part-of-speech or entity type of each word according to the context information of the text. Taking the various states of the device as the hidden states and the monitoring data of the device as the observation sequence, the Hidden Markov Model can also be used to determine whether the device has a fault and the type of the fault, etc.

[0031] Please refer to Figure 1 , Figure 1 which is a flowchart of a method for evaluating the reliability of a distribution network provided by an embodiment of the present application, including but not limited to the following steps:

[0032] S101: Obtain the first historical voltage sag data of the power system during the first historical time period and the first historical tripping data of the target sensitive equipment in the power system during the first historical time period.

[0033] In this embodiment, the first historical voltage sag data includes n historical voltage sag states, and the first historical tripping data includes n historical working states of the target sensitive equipment under the n historical voltage sag states. The historical working states include tripping states or normal working states, and each historical working state corresponds to a historical voltage sag state. n is an integer greater than 1. Please refer to Figure 2 , Figure 2 which is an example diagram of a voltage sag provided by an embodiment of the present application. The horizontal axis is time, the vertical axis is voltage amplitude, V1 is the voltage amplitude before the voltage sag occurs, V2 is the voltage amplitude after the voltage sag occurs, t1 is the time when the voltage sag just occurs, t2 is the time when the voltage sag ends, so the voltage sag duration is t2 - t1.

[0034] In this embodiment, sensitive equipment refers to equipment that is relatively sensitive to voltage sags and is prone to failure or abnormal operation due to their influence. Common sensitive equipment in the power system includes industrial automation equipment: programmable logic controllers, frequency converters, etc.; computer and data processing equipment: servers in data centers, computers for enterprise office use, etc.; medical equipment: scanners, magnetic resonance imaging equipment, cardiac monitors, etc. in hospitals; communication equipment: base stations, switches, etc.; lighting equipment: especially some fluorescent lamps and light-emitting diode lamps using electronic ballasts.

[0035] The first historical voltage sag data is a set composed of multiple different voltage sag states, and each historical voltage sag state contains information such as voltage sag amplitude and sag duration, which is used to describe the characteristics of a specific voltage sag event.

[0036] When a voltage sag event occurs, there are only two possible working states of the target sensitive equipment, either a tripping state (the equipment stops working due to reasons such as voltage sags) or a normal working state (the equipment can still operate normally during the voltage sag). Each specific voltage sag state corresponds to a definite working state of the target sensitive equipment.

[0037] It can be seen that the hidden Markov model can construct a relationship model between the voltage sag state and the working state of the target sensitive device in the power system based on the first historical voltage sag data and the first historical trip data in the first historical time period. It regards the voltage sag state as a hidden Markov chain, and the working state (trip state or normal working state) of the target sensitive device as the observable output. In this way, the model can capture how the change in the voltage sag state affects the change in the working state of the target sensitive device, that is, establish a state transition model from the historical voltage sag state to the historical working state of the target sensitive device. Inputting the target voltage sag state into the constructed first target hidden Markov model, the model can use the learned state transition rules and probability distributions to predict the first trip probability of the target sensitive device under this target voltage sag state. This helps the power system operators to understand in advance the possibility of the target sensitive device experiencing a trip fault under specific voltage sag conditions, so as to take corresponding preventive measures or formulate emergency plans to ensure the stable operation of the power system.

[0038] S102: Construct a first target hidden Markov model based on the first historical voltage sag data and the first historical trip data.

[0039] In this embodiment, please refer to Figure 3 , Figure 3 is a flowchart for constructing a first target hidden Markov model provided by an embodiment of the present application, including but not limited to the following steps:

[0040] S301: Construct a voltage sag state set based on the first historical voltage sag data.

[0041] In this embodiment, the first historical voltage sag data contains information related to voltage sag events that occurred in the power system in the past period. The purpose of constructing the voltage sag state set is to classify and summarize these complex voltage sag situations for subsequent model processing. The first historical voltage sag data contains specific characteristic information of each voltage sag event, such as the voltage sag amplitude (i.e., the magnitude of voltage drop) and the duration of the voltage sag. The voltage sag situations can be classified according to these characteristics. For example, the voltage sag amplitude is divided into different intervals (such as slight drop, moderate drop, severe drop, etc.), and the duration of the voltage sag is also divided into different intervals (such as short time, medium time, long time, etc.). The combination of different amplitude intervals and duration intervals constitutes different voltage sag states, and the set of all these possible voltage sag states is the voltage sag state set.

[0042] S302: Construct an observation sequence based on the first historical trip data.

[0043] In this embodiment, the first historical trip data records the trip conditions of the target sensitive device within the first historical time period. For each voltage sag event, it is recorded whether the target sensitive device trips (which can be represented by "1" for tripping and "0" for non-tripping). According to the chronological order of the occurrence of voltage sag events, the trip conditions of these devices are recorded in sequence to form a sequence. This sequence is the observation sequence, which reflects the change in the working state of the device under different voltage sag events.

[0044] S303: Determine the probability vector, hidden state matrix, and probability matrix corresponding to the reference hidden Markov model to obtain the reference probability vector, reference hidden state matrix, and reference probability matrix.

[0045] In this embodiment, the hidden Markov model includes three important parameter matrices: the initial state probability vector (representing the probability that the model is in each hidden state at the initial moment), the state transition probability matrix (representing the probability of transitioning from one hidden state to another), and the observation probability matrix (representing the probability of generating different observation results in each hidden state). Here, initial values need to be set for these parameter matrices first, and the matrices and vectors composed of these initial values are called the reference probability vector, reference hidden state matrix, and reference probability matrix. The setting of the initial values can be based on experience or some simple assumptions, such as assuming that the probabilities of each hidden state occurring at the initial moment are equal, or estimating the probabilities of state transition and observation based on the preliminary statistics of historical data.

[0046] S304: Perform iterative operations on the reference probability vector, the reference hidden state matrix, and the reference probability matrix based on the voltage sag state set and the observation sequence to obtain the target probability vector, target hidden state matrix, and target probability matrix.

[0047] In this embodiment, using the already constructed voltage sag state set and observation sequence, the reference probability vector, reference hidden state matrix, and reference probability matrix are adjusted and optimized multiple times through a specific algorithm (such as iterative algorithms like the Baum-Welch algorithm). In each iteration, the probability of the observation sequence occurring is calculated based on the current model parameters, and then the model parameters are updated by maximizing this probability. This process is continuously repeated until the model parameters converge (that is, the change in the parameters in several consecutive iterations is very small, less than a preset threshold), and the parameters obtained at this time are the target probability vector, target hidden state matrix, and target probability matrix.

[0048] S305: Determine the first target hidden Markov model based on the target probability vector, target hidden state matrix, and target probability matrix.

[0049] In this embodiment, the target probability vector, target hidden state matrix, and target probability matrix obtained through iterative optimization are used as the parameters of the hidden Markov model, thus determining a specific hidden Markov model, namely the first target hidden Markov model. This model can be used to predict and analyze the tripping probability of the target sensitive device in future voltage sag events.

[0050] It can be seen that by separately constructing the voltage sag state set based on the first historical voltage sag data and the observation sequence based on the first historical tripping data, the ranges and specific contents of the hidden states and observable states in the model can be clearly defined, providing a clear input and output space for the subsequent construction of the model, enabling the model to accurately model the voltage sags and device operating states in the power system, determining the probability vector, hidden state matrix, and probability matrix corresponding to the reference hidden Markov model, obtaining the reference probability vector, reference hidden state matrix, and reference probability matrix, providing initial parameter estimates for the model. These reference values can be set based on existing experience or prior knowledge, providing a starting point for subsequent iterative optimization, helping to accelerate the convergence speed of the model. Through iterative operations on the reference probability vector, reference hidden state matrix, and reference probability matrix based on the voltage sag state set and the observation sequence, the target probability vector, target hidden state matrix, and target probability matrix are obtained. Such iterative operations can continuously adjust the parameters of the model according to the actual historical data, enabling the model to gradually adapt to the specific characteristics and operating rules of the power system, thereby improving the accuracy and adaptability of the model, and better reflecting the true relationship between the voltage sag state and the operating state of the target sensitive device. By determining the first target hidden Markov model through the above series of steps, the historical data of voltage sags in the power system and the corresponding operating states of the target sensitive device can be comprehensively considered, making full use of the information in the data, enabling the model to accurately capture the influence law of voltage sags on the device operating state, thereby providing a reliable basis for predicting the tripping probability of the target sensitive device under specific voltage sag states, and improving the reliability and practicality of the model in power system fault prediction and analysis.

[0051] For example, the first historical voltage sag data of the power system in the past month (the first historical time period) and the first historical tripping data of the industrial motor in this month are obtained. It is assumed that the first historical voltage sag data contains 10 different historical voltage sag states, and the first historical tripping data records the operating states of the motor under these 10 historical voltage sag states, and the operating states are only two types: tripping and normal operation.

[0052] Next, these 10 historical voltage sag states are analyzed and classified. For example, they are classified according to the amplitude and duration of the voltage sag:

[0053] State S1: The voltage sag amplitude is between 10% - 20%, and the duration is less than 100 milliseconds;

[0054] State S2: The voltage sag amplitude is between 20% - 30%, and the duration is 100 - 200 milliseconds;

[0055] ……

[0056] State S10: The voltage sag amplitude is greater than 50%, and the duration is greater than 500 milliseconds.

[0057] In this way, the voltage sag state set {S1, S2, …, S10} is constructed.

[0058] Assume that based on experience or prior knowledge, we think the probability of the motor being in each voltage sag state at the initial moment is equal. Then the reference probability vector π = [0.1, 0.1, …, 0.1] (a total of 10 elements, and the value of each element is 0.1, indicating that the initial probability of being in each voltage sag state is 0.1).

[0059] The hidden state matrix A represents the probability of transitioning from one voltage sag state to another. We first randomly initialize a 10×10 matrix, for example:

[0060]

[0061] The probability matrix B represents the probability of the motor being in different working states (observations) under a certain voltage sag state. Similarly, randomly initialize a 10×2 matrix, such as:

[0062]

[0063] Among them, the first column of the probability matrix B represents the probability of the motor being in "normal operation" under the corresponding voltage sag state, and the second column represents the probability of being in "tripping".

[0064] Then, use methods such as the Baum - Welch algorithm (a commonly used iterative algorithm for hidden Markov model parameter estimation), and based on the constructed voltage sag state set and observation sequence, perform multiple iterative updates on the reference probability vector π, the reference hidden state matrix A, and the reference probability matrix B.

[0065] After multiple iterations, the parameters gradually converge to obtain the target probability vector π′, the target hidden state matrix A′, and the target probability matrix B′ that are more in line with the actual data. For example, after iteration, the target hidden state matrix A′ may become:

[0066]

[0067] At this time, the first target hidden Markov model is determined by the target probability vector π′, the target hidden state matrix A′, and the target probability matrix B′. Subsequently, this model can be used to input the new target voltage sag state to predict the tripping probability of the industrial motor under the target voltage sag state.

[0068] It can be seen that by constructing a voltage sag state set, classifying and summarizing complex and diverse voltage sag data, dividing features such as voltage sag amplitude and duration into different intervals to form different states, the originally chaotic data becomes more organized, facilitating subsequent analysis and processing, and being able to more clearly present various situations of voltage sags. Constructing an observation sequence based on the first historical tripping data can intuitively reflect the working state changes of the target sensitive device under different voltage sag events, determining the probability vector, hidden state matrix, and probability matrix corresponding to the reference hidden Markov model, providing initial values for model training. The setting of these initial values is based on experience or preliminary statistics, which can provide a basis for subsequent iterative optimization, enabling the model training to have a clear starting point and helping to quickly converge to better model parameters. Using the voltage sag state set and the observation sequence to perform iterative operations on the reference parameters can continuously adjust and optimize the model parameters to make the model better fit the actual data. By maximizing the probability of the observation sequence appearing to update the parameters until the parameters converge, the target probability vector, target hidden state matrix, and target probability matrix that are more accurate and can better reflect the actual situation are obtained, improving the prediction and analysis capabilities of the model. Determining the first target hidden Markov model based on the optimized parameters, this model can be used to predict and analyze the tripping probability of the target sensitive device in future voltage sag events, which helps the power system operators to understand in advance the possible fault situations of the equipment and take corresponding preventive measures, such as strengthening equipment maintenance, adjusting the operation mode, etc., to improve the reliability and stability of the power system and reduce the economic losses and production impacts caused by equipment tripping due to voltage sags.

[0069] S103: Input the target voltage sag state into the first target hidden Markov model to obtain the first tripping probability of the target sensitive device under the target voltage sag state.

[0070] In this embodiment, after constructing the first target hidden Markov model, it is necessary to detect whether the prediction effect of the first target hidden Markov model is good. If the prediction effect is good, then input the target voltage sag state into the first target hidden Markov model to obtain the first tripping probability of the target sensitive device under the target voltage sag state. If the prediction effect is not good, then it is necessary to reconstruct the target hidden Markov model based on data with a longer historical time to determine the tripping probability of the target sensitive device under the target voltage sag state.

[0071] Among them, when the first hop-stop probability is greater than the preset value, the target sensitive device can be controlled to perform a hop-stop operation to ensure the safety of the target sensitive device. On the contrary, when the first hop-stop probability is less than or equal to the preset value, the target sensitive device does not perform a hop-stop operation. Therefore, the hop-stop operation of the target sensitive device can be accurately controlled to ensure the intelligence of the target sensitive device control. The preset value can be set in advance or be the system default.

[0072] Please refer to Figure 4 , Figure 4 which is a flowchart for determining the hop-stop probability provided by an embodiment of the present application, including but not limited to the following steps:

[0073] S401: Input the n historical voltage sag states into the first target hidden Markov model respectively to obtain n predicted hop-stop probabilities of the target sensitive device under the n historical voltage sag states.

[0074] In this embodiment, each historical voltage sag state corresponds to a predicted hop-stop probability. Input these n historical voltage sag states into the first target hidden Markov model one by one. The model will calculate the hop-stop probability of the target sensitive device under each input historical voltage sag state, and finally obtain n predicted hop-stop probabilities, and each historical voltage sag state corresponds to a predicted hop-stop probability.

[0075] S402: Determine the prediction accuracy of the first target hidden Markov model based on the n predicted hop-stop probabilities and the n historical working states.

[0076] In this embodiment, please refer to Figure 5 , Figure 5 which is a flowchart for determining the prediction accuracy provided by an embodiment of the present application, including but not limited to the following steps:

[0077] S501: Determine n predicted working states corresponding to the n historical voltage sag states based on the n predicted hop-stop probabilities.

[0078] In this embodiment, each historical voltage sag state corresponds to a predicted working state.

[0079] Exemplarily, when the first predicted tripping probability is greater than the tripping probability threshold, determine that the predicted operating state corresponding to the first historical voltage sag state is the tripping state. The first historical voltage sag state is any one of the n historical voltage sag states, and the first predicted tripping probability is the predicted tripping probability corresponding to the first historical voltage sag state among the n predicted tripping probabilities. Specifically, among the n predicted tripping probabilities, find the predicted tripping probability corresponding to the first historical voltage sag state and define it as the "first predicted tripping probability". This is because a predicted tripping probability has been calculated for each historical voltage sag state through the first target hidden Markov model before, so the corresponding value can be accurately found. Determine a tripping probability threshold in advance, and this threshold is a standard numerical value for judgment. Compare the first predicted tripping probability with the tripping probability threshold. If the first predicted tripping probability is greater than the tripping probability threshold, then it can be determined that the predicted operating state corresponding to the first historical voltage sag state is the tripping state.

[0080] Exemplarily, when the first predicted tripping probability is less than or equal to the tripping probability threshold, determine that the predicted operating state corresponding to the first historical voltage sag state is the normal operation state. Specifically, compare the first predicted tripping probability with the tripping probability threshold. If the first predicted tripping probability is less than or equal to the tripping probability threshold, then it can be determined that the predicted operating state corresponding to the first historical voltage sag state is the normal operation state. This indicates that according to the prediction of the model, the target sensitive device can operate normally in this corresponding historical voltage sag state and will not trip.

[0081] It can be seen that by setting the tripping probability threshold, the operating state of the target sensitive device can be clearly divided into the tripping state and the normal operation state. This helps to clearly define the performance of the device under different historical voltage sag states, facilitates subsequent analysis and processing, and provides a basis for evaluating the prediction accuracy of the first target hidden Markov model. Comparing the predicted operating state with the actual historical operating state can intuitively understand the accuracy of the model's prediction of the device's operating state, thereby determining whether the model can effectively describe the relationship between the voltage sag state and the device's operating state.

[0082] It should be explained that the first historical voltage sag state is any one of the n historical voltage sag states, and the determination method of the predicted operating state corresponding to the first historical voltage sag state is the same as that of the n predicted operating states corresponding to the n historical voltage sag states. Therefore, according to the determination method of the predicted operating state corresponding to the first historical voltage sag state, the n predicted operating states corresponding to the n historical voltage sag states can be determined based on the n predicted tripping probabilities.

[0083] S502: Determine the predicted operating states that match the n historical operating states among the n predicted operating states, and obtain k predicted operating states.

[0084] In this embodiment, k is an integer less than or equal to n. For each of the n historical voltage sag states, there are corresponding historical operating states and predicted operating states. The predicted operating states under each historical voltage sag state are compared one by one with the corresponding historical operating states. If the predicted operating state under a certain historical voltage sag state is the same as the historical operating state, for example, the predicted operating state is the tripping state and the historical operating state under this state is also the tripping state, or the predicted operating state is the normal operation state and the historical operating state is also the normal operation state, then it is considered that this predicted operating state matches the historical operating state. Count the number of all matching predicted operating states, and this number is k. The value of k is less than or equal to n because there may be some predicted operating states that do not match the historical operating states. In this way, k predicted operating states that match the n historical operating states can be determined from the n predicted operating states.

[0085] S503: Determine the prediction accuracy rate of the first target hidden Markov model based on the k predicted operating states and the n predicted operating states.

[0086] In this embodiment, exemplarily, determine the reference prediction accuracy rate of the first target hidden Markov model based on the k predicted operating states and the n predicted operating states. Specifically, the reference prediction accuracy rate can be determined by calculating the ratio of k to n. This ratio reflects the proportion of the correctly predicted operating states of the model among all the predicted states and is a preliminary evaluation index for the prediction ability of the first target hidden Markov model.

[0087] Exemplarily, obtain the rated voltage value of the target sensitive device. Specifically, the rated voltage of the device determines the voltage range for its normal operation. Devices with a higher rated voltage may be relatively more likely to trip when facing the same voltage sag amplitude because their tolerance to voltage fluctuations may be lower. Obtaining the rated voltage value of the target sensitive device is for adjusting the reference prediction accuracy rate according to the characteristics of the device later. Because devices with different rated voltages may have different sensitivities to voltage sags, it is necessary to further optimize the evaluation of the model's prediction accuracy rate based on its rated voltage.

[0088] Exemplarily, determine the adjustment parameter corresponding to the rated voltage value. Specifically, it can be a mapping relationship between the preset rated voltage value and the adjustment parameter. Based on this mapping relationship, the adjustment parameter corresponding to the rated voltage value can be determined.

[0089] Exemplarily, the reference prediction accuracy is adjusted based on the adjustment parameter to obtain the prediction accuracy of the first target Hidden Markov Model. Specifically, the prediction accuracy of the first target Hidden Markov Model is calculated according to the following formula:

[0090] Prediction accuracy of the first target Hidden Markov Model = Reference prediction accuracy × (1 + Adjustment parameter);

[0091] Based on the above formula, the reference prediction accuracy can be adjusted based on the adjustment parameter to obtain the prediction accuracy of the first target Hidden Markov Model.

[0092] It can be seen that by calculating the proportion of k predicted working states that match the historical working states among n predicted working states to obtain the reference prediction accuracy, it can intuitively reflect the prediction accuracy of the model for the working states, which is a quantitative evaluation of the basic performance of the model, enabling users to understand the prediction ability of the model without considering other factors, obtain the rated voltage value of the target sensitive device, and determine the corresponding adjustment parameter. Considering that devices with different rated voltages have different sensitivities to voltage sags, the reference prediction accuracy is adjusted based on the adjustment parameter, comprehensively considering the basic prediction ability of the model and the characteristics of the device, making the finally obtained prediction accuracy more in line with the actual situation, and being able to more accurately evaluate the prediction performance of the first target Hidden Markov Model on specific devices, providing a more reliable basis for subsequent model improvement, device maintenance, and related decisions.

[0093] Please refer to Figure 6 , Figure 6 which is another flowchart for determining the prediction accuracy provided by the embodiments of the present application, including but not limited to the following steps:

[0094] S601: Obtain the mapping relationship between the working state and the tripping probability.

[0095] In this embodiment, this mapping can be established through the analysis of a large amount of historical data, experiments, or based on the working principle of the device and relevant domain knowledge. For example, it is found through observation that when the device voltage fluctuates within a certain range, its tripping probability will reach a certain value, thus establishing a mapping relationship between the working state (the device working state corresponding to the voltage fluctuation situation) and the tripping probability.

[0096] In this embodiment, when the working state of the target sensitive device is the tripping state, the corresponding tripping probability can be 100%, and when the working state of the target sensitive device is the normal working state, the corresponding tripping probability can be 0.

[0097] S602: Determine the skip-stop probability corresponding to each of the n historical working states based on the mapping relationship, obtaining n skip-stop probabilities.

[0098] In this embodiment, for each of the n historical working states, the corresponding skip-stop probability can be found according to this mapping relationship. For example, if a certain historical working state occurs when the voltage sags by a certain amplitude, based on the previously established mapping relationship, the skip-stop probability of the device under this voltage sag amplitude can be determined. In this way, by processing each of the n historical working states one by one, n skip-stop probabilities are obtained.

[0099] S603: Determine n skip-stop probability differences based on the n predicted skip-stop probabilities and the n skip-stop probabilities.

[0100] In this embodiment, the n predicted skip-stop probabilities and the skip-stop probabilities corresponding to the n historical working states can be sorted respectively to make their orders correspond one by one, and then the probabilities at the corresponding positions are subtracted to obtain n skip-stop probability differences.

[0101] S604: Determine the average skip-stop probability difference corresponding to the n skip-stop probability differences.

[0102] In this embodiment, the n skip-stop probability differences are added up and then divided by n to obtain the average skip-stop probability difference. This value can be used to measure the average difference between the skip-stop probability predicted by the model and the actual skip-stop probability. It is an index comprehensively reflecting the prediction accuracy of the model. The smaller the average skip-stop probability difference, the closer the skip-stop probability predicted by the model is to the actual skip-stop probability, and the better the prediction effect of the target hidden Markov model.

[0103] S605: Determine the prediction accuracy of the first target hidden Markov model based on the average skip-stop probability difference.

[0104] In this embodiment, there is a certain relationship between the average skip-stop probability difference and the prediction accuracy. The prediction accuracy can be determined according to the average skip-stop probability difference through a certain functional relationship or rule. For example, if the average skip-stop probability difference is within a relatively small range, it can be considered that the prediction accuracy of the model is relatively high, and the prediction accuracy may be set to a relatively high value. On the contrary, if the average skip-stop probability difference is large, it indicates that the prediction effect of the model is not good, and the prediction accuracy will be correspondingly low. It can be the difference between the preset average skip-stop probability difference and the prediction accuracy. Based on this mapping relationship, the prediction accuracy of the first target hidden Markov model can be determined based on the average skip-stop probability difference.

[0105] It can be seen that by obtaining the mapping relationship between the working state and the tripping probability, the operating state of the device can be converted into specific probability values, enabling both the prediction results and the actual situation of the model to be quantitatively represented by the tripping probability. On this basis, the tripping probability difference and the average tripping probability difference are calculated, and then the prediction accuracy is determined, providing a specific quantitative index for the accuracy of the model, intuitively reflecting the degree of closeness between the predicted value and the actual value of the model, helping to accurately evaluate the performance of the model. Using the tripping probabilities corresponding to n historical working states to evaluate the model fully considers the past operating conditions of the device. The historical data contains the actual tripping probability information of the device under various conditions. Evaluating the model based on this information can more comprehensively and objectively reflect the prediction ability of the model under different working states, avoiding the one-sidedness of relying solely on a single data point or limited data for evaluation, and making the evaluation results more reliable and persuasive. Calculating the difference between the predicted tripping probability and the actual tripping probability, and measuring the overall deviation through the average tripping probability difference, can clearly show the direction and degree of the difference between the model prediction result and the actual situation. This not only helps to judge the accuracy of the model but also helps to analyze the possible deviation problems of the model, such as whether it overestimates or underestimates the tripping probability, thereby providing targeted information for the improvement and optimization of the model, facilitating the adjustment of model parameters or structure to improve the prediction performance of the model.

[0106] It can be seen that adopting two methods to calculate the prediction accuracy of the first target hidden Markov model can not only more accurately reflect the prediction ability of the model on a specific device, make the calculation of the accuracy more in line with the actual application scenario, improve the accuracy and pertinence of the evaluation, but also do not need to rely too much on prior knowledge or subjective judgment. Instead, it mines information from the actual data to evaluate the model accuracy, can better adapt to the changes of different devices and working environments, and has strong generality and adaptability.

[0107] S403: When the prediction accuracy is greater than or equal to the preset accuracy, perform the operation of inputting the target voltage sag state into the first target hidden Markov model to obtain the first tripping probability of the target sensitive device under the target voltage sag state.

[0108] In this embodiment, the preset accuracy rate is a pre-set standard value used to measure whether the prediction performance of the first target Hidden Markov Model meets the requirements. If the prediction accuracy rate obtained through the previous calculation is greater than or equal to this preset accuracy rate, it indicates that the prediction effect of the first target Hidden Markov Model is good and can be trusted. At this time, a specific target voltage sag state (this state may be a certain voltage sag situation to be concerned about in the future, or a state to be analyzed key) is input into the first target Hidden Markov Model, and the model will process this target voltage sag state and calculate the probability of the target sensitive device tripping and stopping in this state. This probability is called the first tripping and stopping probability. For example, the preset accuracy rate is 50%, and the calculated prediction accuracy rate is 60%, meeting the condition. After inputting the target voltage sag state into the model, the first tripping and stopping probability is obtained.

[0109] S404: When the prediction accuracy rate is less than the preset accuracy rate, obtain the second historical voltage sag data within the second historical time period and the second historical tripping and stopping data of the target sensitive device within the second historical time period.

[0110] In this embodiment, the start time of the second historical time period is earlier than the start time of the first historical time period, and the end time of the second historical time period is later than the end time of the second historical time period.

[0111] If the calculated prediction accuracy rate is less than the preset accuracy rate, it indicates that the prediction effect of the first target Hidden Markov Model is not good enough to meet the requirements of actual applications. To improve the model, more data needs to be obtained. Here, the second historical voltage sag data within the second historical time period (including information related to all voltage sag events occurring within this time period, such as amplitude, duration, etc.) and the second historical tripping and stopping data of the target sensitive device within this time period (recording the actual tripping and stopping situations of the device under these voltage sag events) are obtained. The range of the second historical time period is wider than that of the first historical time period, that is, its start time is earlier and its end time is later. In this way, more abundant historical information can be obtained, which helps to build a more accurate model.

[0112] S405: Construct a second target Hidden Markov Model based on the second historical voltage sag data and the second historical tripping and stopping data.

[0113] In this embodiment, a new model, namely, the second target Hidden Markov Model, is constructed by following a series of steps for constructing a Hidden Markov Model by using the just-obtained second historical voltage sag data and second historical tripping data. The specific steps may include constructing a voltage sag state set (dividing different states according to the voltage sag characteristics in the second historical voltage sag data), constructing an observation sequence (determining the tripping or normal operation conditions of the device in different voltage sag states according to the second historical tripping data to form an observation sequence), initializing model parameters (such as an initial state probability vector, a state transition probability matrix, an observation probability matrix, etc.), and then optimizing the parameters through an iterative algorithm until the model converges. The method for constructing the second target Hidden Markov Model is the same as that for constructing the first target Hidden Markov Model, and thus will not be elaborated herein.

[0114] S406: Input the target voltage sag state into the second target Hidden Markov Model to obtain a second tripping probability of the target sensitive device in the target voltage sag state.

[0115] In this embodiment, the second target Hidden Markov Model has been constructed. Input the previously mentioned target voltage sag state into this second target Hidden Markov Model. The second target Hidden Markov Model will process the target voltage sag state and calculate the probability of the target sensitive device tripping in this state. This probability is called the second tripping probability.

[0116] It can be seen that by inputting the historical voltage sag state into the first target hidden Markov model to obtain the predicted tripping probability and comparing it with the actual historical working state to determine the prediction accuracy rate, the fitting and prediction capabilities of the model for historical data can be quantitatively evaluated, the performance of the model on past data can be understood, and it can be judged whether it can accurately reflect the tripping situation of the target sensitive device under different voltage sag states. When the prediction accuracy rate is greater than or equal to the preset accuracy rate, the first target hidden Markov model is used to predict the target voltage sag state, indicating that the model has been verified by historical data and has high reliability, and can be used to predict new unknown states, providing strong support for actual decision-making. When the prediction accuracy rate is less than the preset accuracy rate, data from a more extensive second historical time period is obtained to construct a new second target hidden Markov model. This can utilize more data information, discover laws and characteristics that may not have been found in the data of the first historical time period, help optimize the model, improve the accuracy and generalization ability of the model. The selection range of the data in the second historical time period is wider, its start time is earlier than the start time of the first historical time period, and the end time is later than the end time of the first historical time period, which can better adapt to possible changes in the device operation environment, characteristics, etc. By continuously constructing models based on new data, the model can dynamically adapt to changes in the actual situation, thereby more accurately predicting the tripping probability of the target sensitive device under various voltage sag states, providing a more accurate basis for the maintenance, operation management, etc. of the devices in the power system.

[0117] In summary, implementing the embodiments of the present invention has the following beneficial effects:

[0118] It can be seen that in the sensitive device tripping probability prediction method described in the embodiments of the present invention, first, the first historical voltage sag data of the power system in the first historical time period and the first historical tripping data of the target sensitive device in the power system in the first historical time period are obtained. The first historical voltage sag data includes n historical voltage sag states, and the first historical tripping data includes n historical working states of the target sensitive device under the n historical voltage sag states. The historical working state includes a tripping state or a normal working state, and each historical working state corresponds to a historical voltage sag state. n is an integer greater than 1. Then, a first target hidden Markov model is constructed based on the first historical voltage sag data and the first historical tripping data. Finally, the target voltage sag state is input into the first target hidden Markov model to obtain the first tripping probability of the target sensitive device under the target voltage sag state, improving the accuracy rate of evaluating the tripping probability of the sensitive device under voltage sags.

[0119] Please refer to Figure 7 , Figure 7It is a schematic structural diagram of a sensitive device trip probability prediction device provided by an embodiment of the present application. The sensitive device trip probability prediction device 700 includes: an acquisition unit 701 and a processing unit 702;

[0120] The acquisition unit 701 is configured to acquire first historical voltage sag data of the power system in a first historical time period and first historical trip data of a target sensitive device in the power system in the first historical time period. The first historical voltage sag data includes n historical voltage sag states, and the first historical trip data includes n historical working states of the target sensitive device in the n historical voltage sag states. The historical working state includes a trip state or a normal working state, and each historical working state corresponds to a historical voltage sag state. n is an integer greater than 1;

[0121] The processing unit 702 is configured to construct a first target hidden Markov model based on the first historical voltage sag data and the first historical trip data;

[0122] Input the target voltage sag state into the first target hidden Markov model to obtain the first trip probability of the target sensitive device in the target voltage sag state.

[0123] In some possible implementation manners, in terms of constructing the first target hidden Markov model based on the first historical voltage sag data and the first historical trip data, the processing unit 702 is specifically configured to:

[0124] Construct a voltage sag state set based on the first historical voltage sag data;

[0125] Construct an observation sequence based on the first historical trip data;

[0126] Determine the probability vector, hidden state matrix, and probability matrix corresponding to the reference hidden Markov model to obtain the reference probability vector, reference hidden state matrix, and reference probability matrix;

[0127] Perform iterative operations on the reference probability vector, the reference hidden state matrix, and the reference probability matrix based on the voltage sag state set and the observation sequence to obtain the target probability vector, target hidden state matrix, and target probability matrix;

[0128] Determine the first target hidden Markov model based on the target probability vector, target hidden state matrix, and target probability matrix.

[0129] In some possible implementation manners, the processing unit 702 is further specifically configured to:

[0130] Input the n historical voltage sag states into the first target hidden Markov model respectively to obtain n predicted tripping probabilities of the target sensitive device under the n historical voltage sag states; each historical voltage sag state corresponds to a predicted tripping probability;

[0131] Determine the prediction accuracy rate of the first target hidden Markov model based on the n predicted tripping probabilities and the n historical working states;

[0132] When the prediction accuracy rate is greater than or equal to the preset accuracy rate, perform the operation of inputting the target voltage sag state into the first target hidden Markov model to obtain the first tripping probability of the target sensitive device under the target voltage sag state;

[0133] When the prediction accuracy rate is less than the preset accuracy rate, obtain the second historical voltage sag data in the second historical time period and the second historical tripping data of the target sensitive device in the second historical time period; the start time of the second historical time period is earlier than the start time of the first historical time period, and the end time of the second historical time period is later than the end time of the second historical time period;

[0134] Construct a second target hidden Markov model based on the second historical voltage sag data and the second historical tripping data;

[0135] Input the target voltage sag state into the second target hidden Markov model to obtain the second tripping probability of the target sensitive device under the target voltage sag state.

[0136] In some possible implementation manners, in terms of determining the prediction accuracy rate of the first target hidden Markov model based on the n predicted tripping probabilities and the n historical working states, the processing unit 702 is specifically configured to:

[0137] Determine n predicted working states corresponding to the n historical voltage sag states based on the n predicted tripping probabilities; each historical voltage sag state corresponds to a predicted working state;

[0138] Determine the predicted working states that match the n historical working states among the n predicted working states to obtain k predicted working states; k is an integer less than or equal to n;

[0139] Determine the prediction accuracy rate of the first target hidden Markov model based on the k predicted working states and the n predicted working states.

[0140] In some possible embodiments, in determining the n predicted operating states corresponding to the n historical voltage sag states based on the n predicted skip-stop probabilities, the processing unit 702 is specifically configured to:

[0141] When the first predicted skip-stop probability is greater than the skip-stop probability threshold, determine that the predicted operating state corresponding to the first historical voltage sag state is the skip-stop state; the first historical voltage sag state is any one of the n historical voltage sag states, and the first predicted skip-stop probability is the predicted skip-stop probability corresponding to the first historical voltage sag state among the n predicted skip-stop probabilities;

[0142] When the first predicted skip-stop probability is less than or equal to the skip-stop probability threshold, determine that the predicted operating state corresponding to the first historical voltage sag state is the normal operation state.

[0143] In some possible embodiments, in determining the prediction accuracy of the first target hidden Markov model based on the k predicted operating states and the n predicted operating states, the processing unit 702 is specifically configured to:

[0144] Determine the reference prediction accuracy of the first target hidden Markov model based on the k predicted operating states and the n predicted operating states;

[0145] Obtain the rated voltage value of the target sensitive device;

[0146] Determine the adjustment parameter corresponding to the rated voltage value;

[0147] Adjust the reference prediction accuracy based on the adjustment parameter to obtain the prediction accuracy of the first target hidden Markov model.

[0148] In some possible embodiments, in determining the prediction accuracy of the first target hidden Markov model based on the n predicted skip-stop probabilities and the n historical operating states, the processing unit 702 is specifically configured to:

[0149] Obtain the mapping relationship between the operating state and the skip-stop probability;

[0150] Determine the skip-stop probability corresponding to each historical operating state among the n historical operating states based on the mapping relationship to obtain n skip-stop probabilities;

[0151] Determine n skip-stop probability differences based on the n predicted skip-stop probabilities and the n skip-stop probabilities;

[0152] Determine the average skip-stop probability difference corresponding to the n skip-stop probability differences;

[0153] Determine the prediction accuracy of the first target hidden Markov model based on the average difference in skip-stop probabilities.

[0154] Please refer to Figure 8 , Figure 8 which is a schematic structural diagram of an electronic device provided by an embodiment of the present application. As Figure 8 shown, the electronic device 800 includes a transceiver 801, a processor 802, and a memory 803. They are connected through a bus 804. The memory 803 is used to store computer programs and data, and the transceiver 801 can transmit the data stored in the memory 803 to the processor 802. The above program includes instructions for performing the following steps:

[0155] Obtain the first historical voltage sag data of the power system in the first historical time period and the first historical skip-stop data of the target sensitive device in the power system in the first historical time period; the first historical voltage sag data includes n historical voltage sag states, and the first historical skip-stop data includes n historical working states of the target sensitive device in the n historical voltage sag states, the historical working states including skip-stop states or normal working states, and each historical working state corresponds to a historical voltage sag state, where n is an integer greater than 1;

[0156] Construct a first target hidden Markov model based on the first historical voltage sag data and the first historical skip-stop data;

[0157] Input the target voltage sag state into the first target hidden Markov model to obtain the first skip-stop probability of the target sensitive device in the target voltage sag state.

[0158] In some possible implementation manners, in terms of constructing a first target hidden Markov model based on the first historical voltage sag data and the first historical skip-stop data, the above program includes instructions for performing the following steps:

[0159] Construct a voltage sag state set based on the first historical voltage sag data;

[0160] Construct an observation sequence based on the first historical skip-stop data;

[0161] Determine the probability vector, hidden state matrix, and probability matrix corresponding to the reference hidden Markov model to obtain the reference probability vector, reference hidden state matrix, and reference probability matrix;

[0162] Perform iterative operations on the reference probability vector, the reference hidden state matrix, and the reference probability matrix based on the voltage sag state set and the observation sequence to obtain the target probability vector, target hidden state matrix, and target probability matrix;

[0163] Determine the first target hidden Markov model based on the target probability vector, the target hidden state matrix, and the target probability matrix.

[0164] In some possible implementation manners, the above program includes instructions for performing the following steps:

[0165] Input the n historical voltage sag states into the first target hidden Markov model respectively to obtain n predicted tripping probabilities of the target sensitive device under the n historical voltage sag states; each historical voltage sag state corresponds to a predicted tripping probability;

[0166] Determine the prediction accuracy rate of the first target hidden Markov model based on the n predicted tripping probabilities and the n historical operating states;

[0167] When the prediction accuracy rate is greater than or equal to the preset accuracy rate, perform the operation of inputting the target voltage sag state into the first target hidden Markov model to obtain the first tripping probability of the target sensitive device under the target voltage sag state;

[0168] When the prediction accuracy rate is less than the preset accuracy rate, obtain the second historical voltage sag data in the second historical time period and the second historical tripping data of the target sensitive device in the second historical time period; the start time of the second historical time period is earlier than the start time of the first historical time period, and the end time of the second historical time period is later than the end time of the second historical time period;

[0169] Construct a second target hidden Markov model based on the second historical voltage sag data and the second historical tripping data;

[0170] Input the target voltage sag state into the second target hidden Markov model to obtain the second tripping probability of the target sensitive device under the target voltage sag state.

[0171] In some possible implementation manners, in terms of determining the prediction accuracy rate of the first target hidden Markov model based on the n predicted tripping probabilities and the n historical operating states, the above program includes instructions for performing the following steps:

[0172] Determine n predicted operating states corresponding to the n historical voltage sag states based on the n predicted tripping probabilities; each historical voltage sag state corresponds to a predicted operating state;

[0173] Determine the predicted operating states that match the n historical operating states among the n predicted operating states to obtain k predicted operating states; k is an integer less than or equal to n;

[0174] Determine the prediction accuracy rate of the first target Hidden Markov Model based on the k predicted working states and the n predicted working states.

[0175] In some possible implementation manners, in terms of determining the n predicted working states corresponding to the n historical voltage sag states based on the n predicted skip-stop probabilities, the above program includes instructions for performing the following steps:

[0176] When the first predicted skip-stop probability is greater than the skip-stop probability threshold, determine that the predicted working state corresponding to the first historical voltage sag state is the skip-stop state; the first historical voltage sag state is any one of the n historical voltage sag states, and the first predicted skip-stop probability is the predicted skip-stop probability corresponding to the first historical voltage sag state among the n predicted skip-stop probabilities;

[0177] When the first predicted skip-stop probability is less than or equal to the skip-stop probability threshold, determine that the predicted working state corresponding to the first historical voltage sag state is the normal operation state.

[0178] In some possible implementation manners, in terms of determining the prediction accuracy rate of the first target Hidden Markov Model based on the k predicted working states and the n predicted working states, the above program includes instructions for performing the following steps:

[0179] Determine the reference prediction accuracy rate of the first target Hidden Markov Model based on the k predicted working states and the n predicted working states;

[0180] Obtain the rated voltage value of the target sensitive device;

[0181] Determine the adjustment parameter corresponding to the rated voltage value;

[0182] Adjust the reference prediction accuracy rate based on the adjustment parameter to obtain the prediction accuracy rate of the first target Hidden Markov Model.

[0183] In some possible implementation manners, in terms of determining the prediction accuracy rate of the first target Hidden Markov Model based on the n predicted skip-stop probabilities and the n historical working states, the above program includes instructions for performing the following steps:

[0184] Obtain the mapping relationship between the working state and the skip-stop probability;

[0185] Determine the skip-stop probability corresponding to each historical working state among the n historical working states based on the mapping relationship to obtain n skip-stop probabilities;

[0186] Determine n hop-stop probability differences based on the n predicted hop-stop probabilities and the n hop-stop probabilities;

[0187] Determine the average hop-stop probability difference corresponding to the n hop-stop probability differences;

[0188] Determine the prediction accuracy of the first target hidden Markov model based on the average hop-stop probability difference.

[0189] It should be understood that the electronic devices in this application may include a sensitive device hop-stop probability prediction device, a smart phone (such as an Android phone, an iOS phone, a Windows Phone, etc.), a tablet computer, a palm computer, a notebook computer, a mobile Internet device MID (Mobile Internet Devices, abbreviated as: MID) or a wearable device, or a server, an edge computing node, etc. The above electronic devices are only examples and not an exhaustive list, including but not limited to the above electronic devices.

[0190] The embodiments of the present application also provide a computer-readable storage medium. The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement some or all of the steps of any one of the sensitive device hop-stop probability prediction methods described in the above method embodiments.

[0191] The embodiments of the present application also provide a computer program product. The computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute some or all of the steps of any one of the sensitive device hop-stop probability prediction methods described in the above method embodiments.

[0192] It should be noted that for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that this application is not limited by the described action sequence, because according to this application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0193] In the above embodiments, the descriptions of the respective embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0194] In several embodiments provided by the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces. The indirect coupling or communication connection of devices or units can be in electrical or other forms.

[0195] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0196] In addition, in each embodiment of the present application, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software program modules.

[0197] If the integrated unit is implemented in the form of a software program module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a memory and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in each embodiment of the present application. And the aforementioned memory includes: USB flash drives, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), mobile hard disks, magnetic disks, or optical disks and other media that can store program codes.

[0198] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing relevant hardware through a program. The program can be stored in a computer-readable memory. The memory can include: flash drives, read-only memories (abbreviation: ROM), random access memories (abbreviation: RAM), magnetic disks, or optical disks, etc.

[0199] The above has introduced the embodiments of the present application in detail. Specific examples are used herein to illustrate the principle and embodiments of the present application. The description of the above embodiments is only for helping to understand the method and its core idea of the present application; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific embodiments and application scope. In summary, the content of this specification should not be construed as a limitation to the present application.

Claims

1. A method for predicting the tripping probability of a sensitive device, characterized in that Including: Obtaining first historical voltage sag data of a power system during a first historical time period and first historical tripping data of a target sensitive device in the power system during the first historical time period; the first historical voltage sag data includes n historical voltage sag states, and the first historical tripping data includes n historical working states of the target sensitive device under the n historical voltage sag states, where the historical working state includes a tripping state or a normal working state, and each historical working state corresponds to a historical voltage sag state, and n is an integer greater than 1; Constructing a first target hidden Markov model based on the first historical voltage sag data and the first historical tripping data; Inputting a target voltage sag state into the first target hidden Markov model to obtain a first tripping probability of the target sensitive device under the target voltage sag state.

2. The method according to claim 1, characterized in that, The constructing a first target hidden Markov model based on the first historical voltage sag data and the first historical tripping data includes: Constructing a voltage sag state set based on the first historical voltage sag data; Constructing an observation sequence based on the first historical tripping data; Determining a probability vector, a hidden state matrix, and a probability matrix corresponding to a reference hidden Markov model to obtain a reference probability vector, a reference hidden state matrix, and a reference probability matrix; Performing iterative operations on the reference probability vector, the reference hidden state matrix, and the reference probability matrix based on the voltage sag state set and the observation sequence to obtain a target probability vector, a target hidden state matrix, and a target probability matrix; Determining the first target hidden Markov model based on the target probability vector, the target hidden state matrix, and the target probability matrix.

3. The method according to claim 2, wherein The method further includes: Inputting the n historical voltage sag states into the first target hidden Markov model respectively to obtain n predicted tripping probabilities of the target sensitive device under the n historical voltage sag states; each historical voltage sag state corresponds to a predicted tripping probability; Determining a prediction accuracy rate of the first target hidden Markov model based on the n predicted tripping probabilities and the n historical working states; When the prediction accuracy rate is greater than or equal to a preset accuracy rate, performing the operation of inputting the target voltage sag state into the first target hidden Markov model to obtain a first tripping probability of the target sensitive device under the target voltage sag state; When the prediction accuracy rate is less than the preset accuracy rate, obtaining second historical voltage sag data during a second historical time period and second historical tripping data of the target sensitive device during the second historical time period; the start time of the second historical time period is earlier than the start time of the first historical time period, and the end time of the second historical time period is later than the end time of the second historical time period; Constructing a second target hidden Markov model based on the second historical voltage sag data and the second historical tripping data; Input the target voltage sag state into the second target Hidden Markov Model to obtain the second tripping probability of the target sensitive device under the target voltage sag state.

4. The method according to claim 3, wherein The determining the prediction accuracy rate of the first target Hidden Markov Model based on the n predicted tripping probabilities and the n historical operating states includes: Determine the n predicted operating states corresponding to the n historical voltage sag states based on the n predicted tripping probabilities; each historical voltage sag state corresponds to one predicted operating state; Determine the predicted operating states that match the n historical operating states among the n predicted operating states to obtain k predicted operating states; k is an integer less than or equal to n; Determine the prediction accuracy rate of the first target Hidden Markov Model based on the k predicted operating states and the n predicted operating states.

5. The method according to claim 4, wherein The determining the n predicted operating states corresponding to the n historical voltage sag states based on the n predicted tripping probabilities includes: When the first predicted tripping probability is greater than the tripping probability threshold, determine that the predicted operating state corresponding to the first historical voltage sag state is the tripping state; the first historical voltage sag state is any one of the n historical voltage sag states, and the first predicted tripping probability is the predicted tripping probability corresponding to the first historical voltage sag state among the n predicted tripping probabilities; When the first predicted tripping probability is less than or equal to the tripping probability threshold, determine that the predicted operating state corresponding to the first historical voltage sag state is the normal operating state.

6. The method according to claim 4 or 5, characterized in that, The determining the prediction accuracy rate of the first target Hidden Markov Model based on the k predicted operating states and the n predicted operating states includes: Determine the reference prediction accuracy rate of the first target Hidden Markov Model based on the k predicted operating states and the n predicted operating states; Obtain the rated voltage value of the target sensitive device; Determine the adjustment parameter corresponding to the rated voltage value; Adjust the reference prediction accuracy rate based on the adjustment parameter to obtain the prediction accuracy rate of the first target Hidden Markov Model.

7. The method according to claim 3, characterized in that The determining the prediction accuracy rate of the first target Hidden Markov Model based on the n predicted tripping probabilities and the n historical operating states includes: Obtain the mapping relationship between the operating state and the tripping probability; Determine the tripping probability corresponding to each historical operating state among the n historical operating states based on the mapping relationship to obtain n tripping probabilities; Determine the n tripping probability differences based on the n predicted tripping probabilities and the n tripping probabilities; Determine the average tripping probability difference corresponding to the n tripping probability differences; Determine the prediction accuracy rate of the first target Hidden Markov Model based on the average tripping probability difference.

8. A sensitive device trip probability prediction device, characterized in that, The device includes: an acquisition unit and a processing unit; The obtaining unit is configured to obtain first historical voltage sag data of the power system in a first historical time period and first historical tripping data of a target sensitive device in the power system in the first historical time period; the first historical voltage sag data includes n historical voltage sag states, and the first historical tripping data includes n historical working states of the target sensitive device under the n historical voltage sag states, where the historical working states include a tripping state or a normal working state, each historical working state corresponds to a historical voltage sag state, and n is an integer greater than 1; The processing unit is configured to construct a first target hidden Markov model based on the first historical voltage sag data and the first historical tripping data; Input the target voltage sag state into the first target hidden Markov model to obtain a first tripping probability of the target sensitive device under the target voltage sag state.

9. An electronic device, characterized in that, It includes a processor, a memory, a communication interface, and one or more programs, where the one or more programs are stored in the memory and are configured to be executed by the processor, and the one or more programs include instructions for performing the steps in the method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method according to any one of claims 1-7.

Images