A multi-task risk assessment method for offshore wind power grid-connected system

Abstract

The embodiment of the application discloses a kind of for offshore wind power grid-connected system's multitask risk assessment method, it is related to electric power grid technical field, the method includes: obtaining the operation time series data sequence of offshore wind power grid-connected system, pre-processing operation time series data sequence, adds various symbols and obtains the sample data sequence to be encoded;Carry out digital coding, based on digital sequence construction sample data set, and train multitask processing model, output trained multitask processing model;Based on the multitask processing model of trained sample data sequence to be processed is handled to operation time series, according to output result, the future state probability of offshore wind power grid-connected system is predicted, and risk assessment result is output.The application can unify and simplify the state management of offshore wind power grid-connected system by integrating multitask into a multitask processing model, which is beneficial to improve the efficiency and reliability of the system.

Description

Technical Field

[0001] This application relates to the field of power grid technology, and in particular to a multi-task risk assessment method for offshore wind power grid-connected systems. Background Technology

[0002] Offshore wind power grid-connected systems are playing an increasingly important role in the global renewable energy sector, particularly in wind-rich marine areas where they are widely used. However, the operating environment of offshore wind power grid-connected systems is susceptible to several factors, including extreme weather conditions and the marine environment. These include extreme weather events such as high temperatures, high humidity, severe salt spray corrosion, thunderstorms, and typhoons, as well as marine environmental factors such as ocean currents and waves. These factors not only accelerate equipment aging but can also trigger complex fault chains, leading to system instability. Therefore, offshore wind power grid-connected systems and their equipment need to adapt to the continuously changing marine environment. To ensure the stable and safe operation of offshore wind power grid-connected systems, higher requirements are placed on related systems, necessitating real-time monitoring of the operating status of offshore wind power grid-connected systems and their equipment.

[0003] Currently, in the monitoring of offshore wind power grid-connected systems and their equipment, the lower accessibility of offshore wind power equipment leads to significantly higher maintenance and fault repair costs compared to onshore wind power systems. Therefore, developing a method capable of accurately predicting potential risks and performance changes in grid-connected systems is crucial. This predictive method needs to consider the uncertainties and variability of the marine environment, assess risks in real time, and provide reliable data support to optimize system operation and maintenance strategies. Current research mainly focuses on internal factors of equipment and failure mode analysis in onshore environments, while our understanding of equipment status and failure development mechanisms in the unique marine environment remains limited. Furthermore, fault diagnosis and remote monitoring technologies for offshore wind power systems are not yet fully developed; various tasks have independent models, resulting in overall system redundancy and complexity.

[0004] Therefore, there is currently a lack of a method that can accurately predict the performance and safety risks of offshore wind power systems. Summary of the Invention

[0005] This application provides a multi-task risk assessment method for offshore wind power grid-connected systems to address the shortcomings of the aforementioned related technologies. The technical solution is as follows:

[0006] In a first aspect, embodiments of this application provide a multi-task risk assessment method for offshore wind power grid-connected systems, including:

[0007] Obtain the runtime sequence data of the offshore wind power grid-connected system and preprocess the runtime sequence data;

[0008] Add task header symbols, start symbols, separator symbols, and end symbols to the preprocessed runtime sequence data to obtain the sample data sequence to be encoded.

[0009] The data sequence to be encoded is digitally encoded, and a sample dataset is constructed based on the digitally encoded digital sequence;

[0010] The multi-task processing model is trained based on the sample dataset, and the trained multi-task processing model is output.

[0011] The training multi-task processing model is used to process the unprocessed runtime sequence data of the offshore wind power grid-connected system. Based on the output of the training multi-task processing model, the future state probability of the offshore wind power grid-connected system is predicted, and the risk assessment result is output.

[0012] In one alternative embodiment of the first aspect, the acquisition of the runtime sequence data of the offshore wind power grid-connected system and the preprocessing of the runtime sequence data sequence include:

[0013] Obtain the runtime sequence data X of the offshore wind power grid-connected system t =[x1,x2,…,x t ];

[0014] Anomaly detection is performed on the data in the runtime sequence data. If abnormal data is found, the original abnormal data is replaced with the symbol corresponding to the anomaly type. If missing data is found, the original missing data is replaced with the symbol corresponding to the missing data.

[0015] The runtime sequence data is discretized to determine the discrete level, thus converting the runtime sequence data into a time-series discrete data sequence, using the formula:

[0016]

[0017] Where, x t x is an element in the runtime sequence data. d Let x be an element in the time-series discrete data sequence. max x is the maximum value in the runtime sequence data. min It is the minimum value in the runtime sequence data. For rounding down to the nearest integer, N is the discrete series number of the continuous time series data, and ∈ is the parameter.

[0018] In one alternative of the first aspect, adding a task header symbol, a start symbol, a separator symbol, and an end symbol to the preprocessed runtime sequence data to obtain the sample data sequence to be encoded includes:

[0019] The time-series discrete data sequence is segmented to obtain sub-data sequences of a preset length;

[0020] For each sub-data sequence, if there is abnormal data, obtain the result data sequence after processing the abnormal data for each task type, and pair the original data sequence of the sub-data sequence before the anomaly detection and the result data sequence of each task type respectively.

[0021] Based on the paired original data sequence and result data sequence, a task header symbol corresponding to the task type is added to the beginning of the original data sequence, and a start symbol is added after the task header symbol. A separator symbol is added to the end of the original data sequence, and a result data sequence corresponding to the task type is added after the separator symbol. An end symbol is added to the end of the result data sequence.

[0022] Output the sequence of sample data to be encoded.

[0023] In one alternative embodiment of the first aspect, the step of digitally encoding the data sequence to be encoded and constructing a sample dataset based on the digitally encoded digital sequence includes:

[0024] Each data point in the sample data sequence to be encoded is treated as a sub-word, and each data point is digitally encoded according to a preset encoding method to obtain a digitally encoded sequence.

[0025] Based on the original data sequence and the result data sequence after processing the original data sequence according to the corresponding task type, the digital sequence is misaligned to generate input sample data corresponding to the original data sequence and output sample data corresponding to the result data sequence.

[0026] The sample dataset is constructed based on all the input and output sample data.

[0027] In one alternative embodiment of the first aspect, training the multi-task processing model based on the sample dataset and outputting the trained multi-task processing model includes:

[0028] The multi-task processing model is trained in batches based on the sample dataset, and the network parameters of the multi-task processing model are optimized by gradient descent.

[0029] For each batch of sample data, calculate the value of the cross-entropy loss function;

[0030] Whether the multi-task processing model has converged is determined based on the value of the cross-entropy loss function for each batch;

[0031] If convergence is achieved, the trained multi-task processing model is output.

[0032] In one alternative embodiment of the first aspect, the processing of the unprocessed runtime sequence data of the offshore wind power grid-connected system based on the trained multi-task processing model includes:

[0033] Obtain the runtime sequence data to be processed from the offshore wind power grid-connected system, and preprocess the runtime sequence data to be processed.

[0034] Add task header symbols, start symbols, and delimiter symbols to the preprocessed runtime sequence data to obtain the data sequence to be encoded.

[0035] The data sequence to be encoded is digitally encoded to output a digital sequence to be processed;

[0036] Input the digital sequence to be processed into the trained multi-task processing model;

[0037] The trained multi-task processing model executes the task corresponding to the task header symbol, generates a predicted value, adds the predicted value after the separator symbol, and outputs the updated sequence of numbers to be processed.

[0038] Determine whether the updated digital sequence to be processed meets the preset conditions. If the preset conditions are met, then conduct a risk assessment of the offshore wind power grid-connected system based on the updated digital sequence to be processed.

[0039] Otherwise, the updated sequence of numbers to be processed is input into the trained multi-task processing model, and the process proceeds to the step of executing the task corresponding to the task header symbol through the trained multi-task processing model.

[0040] In one alternative embodiment of the first aspect, determining whether the updated sequence of numbers to be processed satisfies a preset condition includes:

[0041] If the trained multi-task processing model outputs an updated sequence of numbers to be processed, and then generates an end symbol at the end of the updated sequence of numbers to be processed, or if the number of digits in the updated sequence of numbers to be processed exceeds a preset digit threshold, then the updated sequence of numbers to be processed satisfies the preset condition.

[0042] Otherwise, the updated sequence of numbers to be processed does not meet the preset conditions.

[0043] Secondly, embodiments of this application also provide a multi-task risk assessment device for offshore wind power grid-connected systems, comprising:

[0044] A data processing unit is used to acquire the runtime sequence data sequence of the offshore wind power grid-connected system and to preprocess the runtime sequence data sequence.

[0045] The symbol processing unit is used to add task header symbols, start symbols, separator symbols and end symbols to the preprocessed runtime sequence data to obtain the sample data sequence to be encoded.

[0046] A sample construction unit is used to digitally encode the data sequence to be encoded, and to construct a sample dataset based on the digitally encoded digital sequence.

[0047] The model training unit is used to train the multi-task processing model based on the sample dataset and output the trained multi-task processing model.

[0048] The risk assessment unit is used to process the unprocessed runtime sequence data of the offshore wind power grid-connected system based on the trained multi-task processing model, predict the future state probability of the offshore wind power grid-connected system based on the output results of the trained multi-task processing model, and output the risk assessment results.

[0049] Thirdly, embodiments of this application also provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method provided by the first aspect or any implementation thereof of the embodiments of this application.

[0050] Fourthly, this application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method provided by the first aspect of the embodiments of this application or any implementation thereof.

[0051] The beneficial effects of the technical solutions provided in some embodiments of this application include at least the following:

[0052] This application provides a multi-task risk assessment method for offshore wind power grid-connected systems. By collecting runtime sequence data of the offshore wind power grid-connected system, a symbol system is constructed to discretize multiple tasks and data. By constructing the symbol system and generating corresponding multi-task processing sample sets, the multi-task processing model is trained based on the sample sets, thereby obtaining a multi-task processing model applicable to multiple tasks such as probability prediction, anomaly detection, and data imputation.

[0053] By integrating multiple tasks into a single multi-task processing model, the state management of offshore wind power grid-connected systems can be unified and simplified, which is beneficial to improving system efficiency and reliability.

[0054] The data output by the multi-task processing model can provide reliable data support for risk assessment of offshore wind power grid-connected systems and can predict the status of offshore wind power grid-connected systems. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in this application or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 This is a flowchart illustrating a multi-task risk assessment method for offshore wind power grid-connected systems provided in an embodiment of this application;

[0057] Figure 2 This is a flowchart illustrating a multi-task risk assessment method for offshore wind power grid-connected systems provided in an embodiment of this application;

[0058] Figure 3 This is a schematic diagram of the structure of a multi-task risk assessment device for an offshore wind power grid-connected system provided in an embodiment of this application;

[0059] Figure 4 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0060] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0061] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or modules is not limited to the steps or modules listed, but may optionally include steps or modules not listed, or may optionally include other steps or modules inherent to such process, method, product, or apparatus.

[0062] It should be noted that the terms "first" and "second" used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects. It is understood that "first" and "second" can be interchanged in a specific order or sequence where permitted. It should be understood that the objects distinguished by "first" and "second" can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in an order other than those described or illustrated herein.

[0063] The present application will now be described in detail with reference to specific embodiments.

[0064] Next, combine Figure 1 This application introduces a multi-task risk assessment method for offshore wind power grid-connected systems, provided by embodiments of this application. For details, please refer to... Figure 1 , Figure 1 This illustration shows a flowchart of a multi-task risk assessment method for offshore wind power grid-connected systems provided in an embodiment of this application. Figure 1 As shown, the method includes the following steps:

[0065] S101, Obtain the runtime sequence data sequence of the offshore wind power grid-connected system, and preprocess the runtime sequence data sequence;

[0066] S102, add task header symbols, start symbols, separator symbols and end symbols to the preprocessed runtime sequence data to obtain the sample data sequence to be encoded;

[0067] S103, digitally encode the data sequence to be encoded, and construct a sample dataset based on the digitally encoded digital sequence;

[0068] S104, Train the multi-task processing model based on the sample dataset and output the trained multi-task processing model.

[0069] S105, Based on the trained multi-task processing model, process the unprocessed runtime sequence data of the offshore wind power grid-connected system, predict the future state probability of the offshore wind power grid-connected system according to the output result of the trained multi-task processing model, and output the risk assessment result.

[0070] In some embodiments, S101, acquiring the runtime sequence data of the offshore wind power grid-connected system and preprocessing the runtime sequence data includes:

[0071] Obtain the runtime sequence data X of the offshore wind power grid-connected system t =[x1,x2,…,x t ];

[0072] Anomaly detection is performed on the data in the runtime sequence data. If abnormal data is found, the original abnormal data is replaced with the symbol corresponding to the anomaly type. If missing data is found, the original missing data is replaced with the symbol corresponding to the missing data.

[0073] The runtime sequence data is discretized to determine the discrete level, thus converting the runtime sequence data into a time-series discrete data sequence, using the formula:

[0074]

[0075] Where, x t x is an element in the runtime sequence data. d Let x be an element in the time-series discrete data sequence. max x is the maximum value in the runtime sequence data. min It is the minimum value in the runtime sequence data. For rounding down to the nearest integer, N is the discrete series number of the continuous time series data, and ∈ is the parameter.

[0076] Specifically, the monitoring system of an offshore wind farm can be used to collect runtime data of various data from the offshore wind farm, such as power time-series data.

[0077] Specifically, anomaly detection can be performed manually, identifying data that significantly deviates from the normal range as anomalous data, and marking the anomalous data with symbols. <error>Replace the existing data; identify missing data due to various measurement causes, and use symbols to represent the missing data. <none>Replace the original data.

[0078] In some embodiments, in S102, a task header symbol, a start symbol, a separator symbol, and an end symbol are added to the preprocessed runtime sequence data to obtain a sample data sequence to be encoded, including:

[0079] The time-series discrete data sequence is segmented to obtain sub-data sequences of a preset length;

[0080] For each sub-data sequence, if there is abnormal data, obtain the result data sequence after processing the abnormal data for each task type, and pair the original data sequence of the sub-data sequence before the anomaly detection and the result data sequence of each task type respectively.

[0081] Based on the paired original data sequence and result data sequence, a task header symbol corresponding to the task type is added to the beginning of the original data sequence, and a start symbol is added after the task header symbol. A separator symbol is added to the end of the original data sequence, and a result data sequence corresponding to the task type is added after the separator symbol. An end symbol is added to the end of the result data sequence.

[0082] Output the sequence of sample data to be encoded.

[0083] For example, in a prediction task, this can be achieved by cutting discrete data X. d =[x1,x2,…,x t ], thus obtaining a sub-data sequence X of a preset length. dnm =[x n ,x n+1 ,…,x m By adding task headers <predict>and the start symbol <s>Symbols such as the closing symbol <\s> are used to obtain a form like <predict> <s>x n ,x n+1 ,…,x m <\s> sample.

[0084] For example, a sub-data sequence that has been manually labeled to identify anomalous data can be selected, and the corresponding original data sequence that has not undergone anomaly detection can be determined, such as data sequence 1: <0> <1> <8> <none><4> After the anomaly detection task, the data corresponding to 8 was determined to be anomalous. Therefore, the example of the sub-data sequence is data sequence 2. <0> <1> <error> <none><4> Data sequence 2 is the result data sequence of the anomaly detection task; after performing data imputation on the anomaly data, data sequence 3 is obtained: <0> <1> <2> <3> <4> Data sequence 3 is the result data sequence of the data imputation task.

[0085] Data sequence 1 and data sequence 2 can be paired as samples for the anomaly detection task, and data sequence 2 and data sequence 3 can be paired as samples for the data imputation task.

[0086] Specifically, this can be achieved through the task header. <e-detect>and <e-repair>These represent the anomaly detection task and the data imputation task, respectively, using... As a separator, used <s>The '\s' symbol indicates the start of the action, and '<\s>' indicates the end of the action.

[0087] For example, a task header symbol for the anomaly detection task can be added to the beginning of data sequence 1. <e-detect>Then, a start symbol is added after the task header symbol, a separator symbol is added to the end of data sequence 1, the result data sequence of the anomaly detection task (i.e., data sequence 2) is added after the separator symbol, and an end symbol is added to the end of data sequence 2 to obtain the sample data sequence to be encoded for the anomaly detection task: <e-detect> <s><0><1><8> <n> <4> <0><1> <e> <n><4> < / s>.

[0088] For example, similarly, a task header symbol for the data completion task can be added to the beginning of data sequence 2. <e-repair>Then, a start symbol is added after the task header symbol, a separator symbol is added to the end of data sequence 2, the result data sequence of the data imputation task (i.e., data sequence 3) is added after the separator symbol, and an end symbol is added to the end of data sequence 3 to obtain the sample data sequence to be encoded for the data imputation task: <e-repair> <s><0><1><8> <n> <4> <0> <1> <2> <3> <4> < / s>.

[0089] Understandably, during training, samples can be constructed according to different task types, including raw data and corresponding result data. During model application, only the collected data to be processed is used; therefore, it's only necessary to add a task header symbol and a start symbol at the beginning of the data to be processed, and a separator symbol at the end.

[0090] In some embodiments, S103, the step of digitally encoding the data sequence to be encoded and constructing a sample dataset based on the digitally encoded digital sequence includes:

[0091] Each data point in the sample data sequence to be encoded is treated as a sub-word, and each data point is digitally encoded according to a preset encoding method to obtain a digitally encoded sequence.

[0092] Based on the original data sequence and the result data sequence after processing the original data sequence according to the corresponding task type, the digital sequence is misaligned to generate input sample data corresponding to the original data sequence and output sample data corresponding to the result data sequence.

[0093] The sample dataset is constructed based on all the input and output sample data.

[0094] Specifically, for the data input to the model, each discretization level or a special character can be considered a subword, and each subword can correspond to a unique numerical code. For time-series data, it is first discretized using a discretizer, and its discrete levels are used as codes, with a range of 0 to N-1. Then, the required special characters are encoded. For example... <predict> 、 <e-detect>and <e-repair>They can be encoded as N, N+1, and N+2 respectively. <s>、<\s>、 They can be encoded as N+3, N+4, and N+5 respectively. <error>and <none>They can be encoded as N+6 and N+7 respectively.

[0095] For example, the sequence of sample data to be encoded: <predict> <s><0> <1> <2> <3> <4> After being encoded digitally, the corresponding number sequence can be obtained as N N+3 0 1 2 3 4N+4.

[0096] For example, a sample dataset for model training can be constructed using the digit sequences corresponding to the sample data sequence to be encoded. A multi-task training dataset can be constructed using the digit sequences corresponding to anomaly detection and data imputation, respectively. The original data and the processed data are misaligned and used as input and output, respectively. For example, the predicted sequence N,N+3,0,1,2,3,4,N+4 is shifted and aligned, with N N+3 0 1 2 3 4 as input and N+3 0 1 2 3 4N+4 as output.

[0097] In some embodiments, S104, training the multi-task processing model based on the sample dataset and outputting the trained multi-task processing model includes:

[0098] The multi-task processing model is trained in batches based on the sample dataset, and the network parameters of the multi-task processing model are optimized by gradient descent.

[0099] For each batch of sample data, calculate the value of the cross-entropy loss function;

[0100] Whether the multi-task processing model has converged is determined based on the value of the cross-entropy loss function for each batch;

[0101] If convergence is achieved, the trained multi-task processing model is output.

[0102] In some embodiments, the construction process of the multi-task processing model specifically includes:

[0103] First, an embedding layer is constructed to transform the discrete input representation into a continuous vector representation. The input vocabulary size is N+7, and the embedding dimension is d. The embedding layer can be represented by a weight matrix W(N+7xd).

[0104] For a given input x, whose index in the vocabulary is i, then the embedding vector e corresponding to input x is... x It can be extracted from the embedding matrix W through a lookup table operation, calculated as follows:

[0105] e x =W[i];

[0106] Furthermore, a causal self-attention module is constructed, including:

[0107] A single causal attention module mainly consists of two parts, the first part being a multi-head attention mechanism. The three basic computational elements in multi-head attention are the question, the key, and the value, represented by Q, K, and V respectively, which are filled in differently in different scenarios.

[0108] Q, K, and V can be mapped to multiple different semantic representation subspaces, that is:

[0109]

[0110] Among them W i Q W i K W i V These represent the parameter matrices used to map Q, K, and V to the i-th subspace, respectively. The attention mechanism in each subspace is defined as follows:

[0111]

[0112] Here, `mask` is a lower triangular matrix used to implement causal relationships, ensuring that attention calculations only consider the current and previous time steps.

[0113] Then, the outputs of multiple subspaces are concatenated, that is:

[0114]

[0115] in, For vector concatenation operations, W o This is the parameter matrix required for linear transformation after splicing, where h is the number of heads (subspaces) of the model.

[0116] The latter half of the causal attention module undergoes layer normalization and residual connections, namely:

[0117] out=LayerNorm(Add(Q+MultiHead(Q,K,V)));

[0118] Among them, layer normalization ensures the stability of data feature distribution, and residual connections prevent network degradation.

[0119] Furthermore, positional encoding is introduced above each attention layer to preserve the temporal information of the time series data, which is achieved by adding the positional encoding vector to the input embedding vector.

[0120] Specifically, positional coding uses a combination of sine and cosine functions to generate a unique code for each position in the sequence, as shown in the following formula:

[0121]

[0122] Where pos represents the position of the token, i represents the dimension index, and d model This is the dimension of the model. This encoding method distinguishes the encoding of each location from the encoding of other locations, and locations that are close to each other have similar encodings, which helps the model capture the relative relationships between locations.

[0123] Finally, a multi-task integrated model is obtained by repeatedly stacking causal self-attention modules through an embedding layer.

[0124] Next, combine Figure 2 This application introduces a multi-task risk assessment method for offshore wind power grid-connected systems, based on embodiments of the present application. Figure 2 A flowchart illustrating a multi-task risk assessment method for offshore wind power grid-connected systems is provided. Specifically, the method involves processing the runtime sequence data of the offshore wind power grid-connected system based on the trained multi-task processing model, including:

[0125] S201, Obtain the runtime sequence data sequence to be processed of the offshore wind power grid-connected system, and preprocess the runtime sequence data sequence to be processed;

[0126] S202, Add task header symbols, start symbols and delimiter symbols to the preprocessed runtime sequence data to be processed to obtain the data sequence to be encoded;

[0127] S203, digitally encode the data sequence to be encoded and output the digital sequence to be processed;

[0128] S204, Input the digital sequence to be processed into the trained multi-task processing model;

[0129] S205, the task corresponding to the task header symbol is executed through the trained multi-task processing model, a predicted value is generated, the predicted value is added after the separator symbol, and the updated numerical sequence to be processed is output.

[0130] S206, determine whether the updated digital sequence to be processed meets the preset conditions;

[0131] If the preset conditions are met, then proceed to step S207:

[0132] S207, Conduct a risk assessment of the offshore wind power grid-connected system based on the updated digital sequence to be processed;

[0133] Otherwise, proceed with step S208:

[0134] S208, input the updated digital sequence to be processed into the trained multi-task processing model, and proceed to the step described in S205 of executing the task corresponding to the task header symbol through the trained multi-task processing model. Repeat steps S205-S206 until the preset conditions are met.

[0135] It should be noted that the runtime sequence data to be processed in S201 can be a real-time data sequence or a data sequence collected in the past. The data for the future period can be predicted based on the data in the past period through a multi-task processing model.

[0136] Specifically, in S202, according to the type of task to be processed, the corresponding task header symbol can be added to the beginning of the data to be processed, and a start symbol can be inserted after the task header symbol. A separator symbol can be added to the end of the data to be processed. For details, please refer to the description of S102, which will not be repeated here.

[0137] Specifically, in S203, digital encoding can be performed using the same method as in S103 to generate a digital sequence to be processed, which will not be elaborated here.

[0138] Specifically, in S205, the multi-task processing model can identify the task header symbol in the input digital sequence to be processed, determine the type of task to be executed, and then generate the corresponding predicted value according to the task type. The predicted value is then filled after the separator symbol to obtain the updated digital sequence to be processed. If the preset conditions are not met at this time, the updated digital sequence to be processed is re-inputted back into the multi-task processing model until the preset conditions are met.

[0139] For example, the input sequence of numbers to be processed corresponds to the sequence of data to be encoded. <e-predict> <s><0><1><2> After the model performs the prediction task, it predicts that the value after 2 is 4. The resulting data sequence is represented as follows: <e-predict> <s><0><1><2> <4> If the preset conditions are not met at this time, then... <e-predict> <s><0><1><2> <4> If the corresponding numerical sequence is re-inputted into the model, and the model predicts that the value after 4 is 6, then the output data sequence is represented as: <e-predict> <s><0><1><2> <4> <6> .

[0140] In some embodiments, determining whether the updated sequence of numbers to be processed meets a preset condition includes:

[0141] If the trained multi-task processing model outputs an updated sequence of numbers to be processed, an end symbol is generated at the end of the updated sequence of numbers to be processed; or, if the number of digits in the updated sequence of numbers to be processed exceeds a preset digit threshold, then the updated sequence of numbers to be processed satisfies the preset condition, and an end symbol is added to the end of the updated sequence of numbers to be processed.

[0142] Otherwise, the updated sequence of numbers to be processed does not meet the preset conditions.

[0143] For example, if the model determines to terminate the prediction task at this time, it can output the end symbol <\s> after the predicted value of 6 after 4. Alternatively, if the prediction task is set to only predict two values, and the number of digits in the updated sequence of numbers to be processed exceeds the upper limit, it is determined that the preset condition is met.

[0144] Finally, output the data sequence. <e-predict> <s><0><1><2> <4> <6> < / s>.

[0145] It should be noted that both the input and output of the model are numerical sequences in the form of numerical codes. After the output results are obtained, the data can be restored to its original form based on the mapping relationship between the numerical values ​​and the numerical codes. This application does not limit this aspect.

[0146] The following are apparatus embodiments of this application, which can be used to execute the method embodiments of this application. For details not disclosed in the apparatus embodiments of this application, please refer to the method embodiments of this application.

[0147] Please see below. Figure 3 This is a schematic diagram of a multi-task risk assessment device for an offshore wind power grid-connected system, provided as an exemplary embodiment of this application. This device can be implemented as all or part of a terminal through software, hardware, or a combination of both, or it can be integrated as an independent module on a server. The multi-task risk assessment device for an offshore wind power grid-connected system in this embodiment can be applied to a terminal or the cloud. The device 30 includes a data processing unit 301, a symbol processing unit 302, a sample construction unit 303, a model training unit 304, and a risk assessment unit 305, wherein:

[0148] The data processing unit 301 is used to acquire the runtime sequence data of the offshore wind power grid-connected system and preprocess the runtime sequence data.

[0149] Symbol processing unit 302 is used to add task header symbols, start symbols, separator symbols and end symbols to the preprocessed runtime sequence data to obtain the sample data sequence to be encoded;

[0150] The sample construction unit 303 is used to digitally encode the data sequence to be encoded, and to construct a sample dataset based on the digitally encoded digital sequence.

[0151] The model training unit 304 is used to train the multi-task processing model based on the sample dataset and output the trained multi-task processing model.

[0152] The risk assessment unit 305 is used to process the unprocessed runtime sequence data of the offshore wind power grid-connected system based on the trained multi-task processing model, predict the future state probability of the offshore wind power grid-connected system according to the output result of the trained multi-task processing model, and output the risk assessment result.

[0153] It should be noted that the apparatus 30 provided in the above embodiments, when executing a multi-task risk assessment method for offshore wind power grid-connected systems, is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the equipment can be divided into different functional modules to complete all or part of the functions described above. Furthermore, the apparatus provided in the above embodiments and the embodiment of a multi-task risk assessment method for offshore wind power grid-connected systems belong to the same concept, and its implementation process is detailed in the method embodiment, which will not be repeated here.

[0154] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of any of the methods described above.

[0155] Please see Figure 4 This is a structural block diagram of an electronic device provided in an embodiment of this application.

[0156] like Figure 4 As shown, the electronic device 400 includes a processor 401 and a memory 402.

[0157] In this embodiment, the processor 401 is the control center of the computer system, and can be a processor of a physical machine or a processor of a virtual machine. The processor 401 may include one or more processing cores, such as a 4-core processor or an 8-core processor. The processor 401 can be implemented using at least one hardware form selected from DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array).

[0158] Processor 401 may also include a main processor and a coprocessor. The main processor is a processor used to process data in the wake-up state, also known as a CPU (Central Processing Unit). The coprocessor is a low-power processor used to process data in the standby state.

[0159] Memory 402 may include one or more computer-readable storage media, which may be non-transitory. Memory 402 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In some embodiments of this application, the non-transitory computer-readable storage media in memory 402 is used to store at least one instruction, which is executed by processor 401 to implement the method in the embodiments of this application.

[0160] In some embodiments, the electronic device 400 further includes a peripheral device interface 403 and at least one peripheral device 404. The processor 401, memory 402, and peripheral device interface 403 can be connected via a bus or signal line. Each peripheral device 404 can be connected to the peripheral device interface 403 via a bus, signal line, or circuit board. Specifically, the peripheral device 404 includes: a display screen, a camera, and audio circuitry. The peripheral device interface 403 can be used to connect at least one I / O (Input / Output) related peripheral device to the processor 401 and memory 402.

[0161] In some embodiments of this application, the processor 401, memory 402, and peripheral device interface 403 are integrated on the same chip or circuit board; in other embodiments of this application, any one or two of the processor 401, memory 402, and peripheral device interface 403 can be implemented on separate chips or circuit boards. This application does not specifically limit the implementation in this regard.

[0162] The block diagram of the electronic device shown in the embodiments of this application does not constitute a limitation on the electronic device 400. The electronic device 400 may include more or fewer components than shown, or combine certain components, or use different component arrangements.

[0163] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the methods in any of the foregoing embodiments. The computer-readable storage medium may include, but is not limited to, any type of disk, including floppy disks, optical disks, DVDs, CD-ROMs, microdrives, as well as magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic cards or optical cards, nanosystems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.

[0164] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of software products. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0165] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application. < / s> < / e-predict> < / s> < / e-predict> < / s> < / e-predict> < / s> < / e-predict> < / s> < / e-predict> < / s> < / predict> < / none> < / error> < / s> < / predict> < / n> < / s> < / e-repair> < / n> < / e> < / n> < / s> < / e-detect> < / s> < / none> < / error> < / none> < / s> < / predict> < / s> < / predict> < / none> < / error>

Claims

1. A multi-task risk assessment method for offshore wind power grid-connected systems, characterized in that, include: Obtain the runtime sequence data of the offshore wind power grid-connected system and preprocess the runtime sequence data; Based on the preprocessed runtime sequence data, a task header symbol, start symbol, separator symbol, and end symbol are added to obtain the sample data sequence to be encoded, including: The time-series discrete data sequence is segmented to obtain sub-data sequences of a preset length; For each sub-data sequence, if there is abnormal data, obtain the result data sequence after processing the abnormal data for each task type, and pair the original data sequence of the sub-data sequence before anomaly detection with the result data sequence of each task type respectively. Based on the paired original data sequence and result data sequence, a task header symbol corresponding to the task type is added to the beginning of the original data sequence, and a start symbol is added after the task header symbol. A separator symbol is added to the end of the original data sequence, and a result data sequence corresponding to the task type is added after the separator symbol. An end symbol is added to the end of the result data sequence. Output the sequence of sample data to be encoded; wherein, the task header symbol is used to identify different task types, including prediction tasks, anomaly detection tasks, and data imputation tasks; The sample data sequence to be encoded is digitally encoded, and a sample dataset is constructed based on the digitally encoded digital sequence; The multi-task processing model is trained based on the sample dataset, and the trained multi-task processing model is output. The trained multi-task processing model is used to process the unprocessed runtime sequence data of the offshore wind power grid-connected system. Based on the output of the trained multi-task processing model, the future state probability of the offshore wind power grid-connected system is predicted, and a risk assessment result is output, including: Obtain the runtime sequence data to be processed from the offshore wind power grid-connected system, and preprocess the runtime sequence data to be processed. Add task header symbols, start symbols, and delimiter symbols to the preprocessed runtime sequence data to obtain the data sequence to be encoded. The data sequence to be encoded is digitally encoded to output a digital sequence to be processed; Input the digital sequence to be processed into the trained multi-task processing model; The trained multi-task processing model executes the task corresponding to the task header symbol, generates a predicted value, adds the predicted value after the separator symbol, and outputs the updated sequence of numbers to be processed. Determine whether the updated digital sequence to be processed meets the preset conditions. If the preset conditions are met, then conduct a risk assessment of the offshore wind power grid-connected system based on the updated digital sequence to be processed. Otherwise, the updated sequence of numbers to be processed is input into the trained multi-task processing model, and the process proceeds to the step of executing the task corresponding to the task header symbol through the trained multi-task processing model.

2. The multi-task risk assessment method for offshore wind power grid-connected systems according to claim 1, characterized in that, The acquisition of the runtime sequence data of the offshore wind power grid-connected system and the preprocessing of the runtime sequence data include: Obtain the runtime sequence data of the offshore wind power grid-connected system ; Anomaly detection is performed on the data in the runtime sequence data. If abnormal data is found, the original abnormal data is replaced with the symbol corresponding to the anomaly type. If missing data is found, the original missing data is replaced with the symbol corresponding to the missing data. The runtime sequence data is discretized to determine the discrete level, thus converting the runtime sequence data into a time-series discrete data sequence, using the formula: ； in, For elements in the runtime sequence data, For the elements in the time-series discrete data sequence, The maximum value in the runtime sequence data. It is the minimum value in the runtime sequence data. To perform integer rounding down, For discrete series of continuous time data, For parameters.

3. The method according to claim 2, characterized in that, The step of digitally encoding the data sequence to be encoded, and constructing a sample dataset based on the digitally encoded digital sequence, includes: Each data point in the sample data sequence to be encoded is treated as a sub-word, and each data point is digitally encoded according to a preset encoding method to obtain a digitally encoded sequence. Based on the original data sequence and the result data sequence after processing the original data sequence according to the corresponding task type, the digital sequence is misaligned to generate input sample data corresponding to the original data sequence and output sample data corresponding to the result data sequence. The sample dataset is constructed based on all the input and output sample data.

4. The method according to claim 1, characterized in that, The step of training the multi-task processing model based on the sample dataset and outputting the trained multi-task processing model includes: The multi-task processing model is trained in batches based on the sample dataset, and the network parameters of the multi-task processing model are optimized by gradient descent. For each batch of sample data, calculate the value of the cross-entropy loss function; Whether the multi-task processing model has converged is determined based on the value of the cross-entropy loss function for each batch; If convergence is achieved, the trained multi-task processing model is output.

5. The method according to claim 1, characterized in that, The step of determining whether the updated sequence of numbers to be processed meets the preset conditions includes: If the trained multi-task processing model outputs an updated sequence of numbers to be processed, and then generates an end symbol at the end of the updated sequence of numbers to be processed, or if the number of digits in the updated sequence of numbers to be processed exceeds a preset digit threshold, then the updated sequence of numbers to be processed satisfies the preset condition. Otherwise, the updated sequence of numbers to be processed does not meet the preset conditions.

6. An apparatus for a multi-task risk assessment method for offshore wind power grid-connected systems based on any one of claims 1-5, characterized in that, The device includes: A data processing unit is used to acquire the runtime sequence data sequence of the offshore wind power grid-connected system and to preprocess the runtime sequence data sequence. The symbol processing unit is used to add task header symbols, start symbols, separator symbols and end symbols to the preprocessed runtime sequence data to obtain the sample data sequence to be encoded. A sample construction unit is used to digitally encode the data sequence to be encoded, and to construct a sample dataset based on the digitally encoded digital sequence. The model training unit is used to train the multi-task processing model based on the sample dataset and output the trained multi-task processing model. The risk assessment unit is used to process the unprocessed runtime sequence data of the offshore wind power grid-connected system based on the trained multi-task processing model, predict the future state probability of the offshore wind power grid-connected system based on the output results of the trained multi-task processing model, and output the risk assessment results.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method as described in any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 5.

Images