CN113887820A - Method and device for predicting fault of electric power spot business system, computer equipment and storage medium - Google Patents
Method and device for predicting fault of electric power spot business system, computer equipment and storage medium Download PDFInfo
- Publication number
- CN113887820A CN113887820A CN202111222721.3A CN202111222721A CN113887820A CN 113887820 A CN113887820 A CN 113887820A CN 202111222721 A CN202111222721 A CN 202111222721A CN 113887820 A CN113887820 A CN 113887820A
- Authority
- CN
- China
- Prior art keywords
- equipment parameter
- equipment
- neural network
- deep neural
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The application discloses a method and a device for predicting faults of a power spot business system, computer equipment and a storage medium, wherein the method for predicting faults of the power spot business system comprises the steps of receiving real-time equipment parameter time sequence data of hardware equipment in the power spot business system, and inputting a first-class deep neural network to obtain the probability of faults of the hardware equipment; the first deep neural network is obtained by pre-construction and comprises the following steps: receiving equipment parameter time sequence data of hardware equipment in the electric power spot business system before failure, and constructing an equipment parameter dictionary according to the equipment parameter time sequence data; vectorizing the equipment parameters by using the equipment parameter dictionary to obtain vectorized equipment parameter time sequence data and training data; and utilizing a deep neural network to finely adjust, migrate and learn the training data to obtain the first type of deep neural network. The method and the device ensure safe operation and reliable operation of the electric power spot market and the service system.
Description
Technical Field
The present application relates to the field of deep learning, and in particular, to a method, an apparatus, a computer device, and a storage medium for predicting a failure of a power spot business system.
Background
Since the uk power industry revolution in the 90's last century, the power market construction has gone around the world for decades. Currently, the power market is established in several countries of the world, and the spot market is a very critical link that establishes a market organization form closely coupled with the operation of the power system in a short time sequence. The electric power spot market mainly comprises auxiliary service trading markets of day-ahead, day-in and real-time electric energy and reserve and the like.
In order to ensure the normal operation of the electric power spot market business and the safe and stable operation of the power grid, the construction of the electric power spot business system receives more and more national attention and investment. The electric power spot business system mainly comprises an electric power dispatching automation system and an electric power spot transaction system. At present, the established electric power spot-goods service system can realize the unified coordination of three-level dispatching plans of China, network and province, and exert the optimal configuration capacity of extra-large power grid resources. Meanwhile, the system can support the organic connection of scheduling operation and market transaction, and promotes the safe operation of the power system and the reliable operation of the market. However, because the power grid equipment is complex, the equipment is closely associated, the trading frequency of the power spot market is high, the trading varieties are multiple, the trading system is complex, the running relation with the power grid is close, and the like, a large number of uncertain faults are often brought to the power spot business system, and the running risk of the power market and the business system is greatly increased. In summary, the method can accurately and reliably predict the uncertain faults of the electric power spot business system, can effectively ensure the safe operation and reliable operation of the electric power spot market and the business system, and has very important significance.
Disclosure of Invention
In view of the above, it is necessary to provide a method for predicting a fault of an electric power spot business system.
The method for predicting the fault of the electric power spot business system comprises the steps of receiving real-time equipment parameter time sequence data of hardware equipment in the electric power spot business system, and inputting the real-time equipment parameter time sequence data into a first-class deep neural network to obtain the probability of the fault of the hardware equipment;
the first deep neural network is obtained by pre-construction and specifically comprises the following steps:
receiving equipment parameter time sequence data of hardware equipment in the electric power spot business system before failure, and constructing an equipment parameter dictionary according to the equipment parameter time sequence data;
vectorizing the equipment parameters by using the equipment parameter dictionary to obtain vectorized equipment parameter time sequence data, wherein at least one part of the vectorized equipment parameter time sequence data is used as training data;
and utilizing a deep neural network to finely adjust, migrate and learn the training data to obtain the first type of deep neural network.
Optionally, at least a portion of the vectorized equipment parameter timing data is used as test data;
and after the first deep neural network is obtained, evaluating the first deep neural network by using the test data.
Optionally, constructing an equipment parameter dictionary according to the equipment parameter time series data specifically includes:
and reading the parameter value of the equipment parameter, and generating a corresponding index number according to the parameter value, wherein the parameter value and the index number of the equipment parameter jointly form an equipment parameter dictionary.
Optionally, the vectorizing processing is performed on the equipment parameter by using the equipment parameter dictionary to obtain vectorized equipment parameter time sequence data, and the vectorizing processing specifically includes:
acquiring an equipment parameter value of each time step in equipment parameter time sequence data;
searching index numbers corresponding to equipment parameter values of each time step according to the equipment parameter dictionary;
establishing an all-zero vector with the vector length being the same as the equipment parameter dictionary length;
and searching the position of the index number according to the equipment parameter dictionary, and setting the value of the corresponding position in the all-zero vector as one to obtain vectorized equipment parameter time sequence data.
The method for predicting the fault of the electric power spot business system comprises the steps of receiving real-time state parameters of a software system in the electric power spot business system, converting the state parameters into attribute measurement vectors of the electric power software system, and inputting the attribute measurement vectors into a second-class deep neural network to obtain the fault probability of the electric power software system;
the second type of deep neural network is obtained by pre-construction and specifically comprises the following steps:
receiving fault data of a software system in a power spot business system, and constructing a power software system attribute measurement vector data set by performing software measurement on the fault data;
acquiring a comparison software system attribute measurement vector, wherein at least one part of the comparison software system attribute measurement vector is used as training data;
training a second-class deep neural network by using the training data;
and utilizing the trained two classified deep neural networks to finely tune, transfer and learn the attribute metric vector data set of the power software system to obtain the second class of deep neural networks.
Optionally, at least a part of the control software system attribute measurement vector is used as test data;
and after the second type of deep neural network is obtained, evaluating the second type of deep neural network by using the test data.
Optionally, the two-class deep neural network can output a class label indicating whether the deep neural network is a normal class or a fault class according to the input software system attribute measurement vector.
Optionally, the software metrics include product metrics and/or process metrics.
Optionally, the power software system attribute metric data set includes a power normal software system attribute metric vector and a power failure software system attribute metric vector.
Optionally, the two-classification deep neural network is a fully-connected classification network or a convolutional neural network.
The method for predicting the fault of the electric power spot business system comprises a hardware device and a software system, and comprises the following steps:
based on the hardware equipment, adopting the corresponding method to carry out fault prediction;
and based on the software system, adopting the corresponding method to predict the fault.
The present application further provides a device for predicting a failure of a power spot service system, including:
the system comprises a first module, a second module and a third module, wherein the first module is used for receiving equipment parameter time sequence data of hardware equipment in the electric power spot business system before failure and constructing an equipment parameter dictionary according to the equipment parameter time sequence data;
the second module is used for carrying out vectorization processing on the equipment parameters by using the equipment parameter dictionary to obtain vectorized equipment parameter time sequence data, and at least one part of the vectorized equipment parameter time sequence data is used as training data;
the third module is used for utilizing the deep neural network to finely adjust, migrate and learn the training data to obtain a first class of deep neural network;
and the fourth module is used for receiving real-time equipment parameter time sequence data of the hardware equipment in the electric power spot business system and inputting the real-time equipment parameter time sequence data into the first-class deep neural network to obtain the failure probability of the hardware equipment.
The present application further provides a device for predicting a failure of a power spot service system, including:
the system comprises a fifth module, a second module and a third module, wherein the fifth module is used for receiving fault data of a software system in the electric power spot business system and constructing an electric power software system attribute measurement vector data set by carrying out software measurement on the fault data;
a sixth module, configured to obtain a comparison software system attribute measurement vector, where at least a portion of the comparison software system attribute measurement vector is used as training data;
a seventh module for training a second-class deep neural network using the training data;
the eighth module is used for utilizing the trained two-class deep neural network to finely tune, migrate and learn the attribute metric vector data set of the power software system to obtain a second class deep neural network;
and the ninth module is used for receiving real-time state parameters of a software system in the electric power spot business system, converting the state parameters into attribute measurement vectors of the electric power software system, and inputting the attribute measurement vectors into the second-class deep neural network to obtain the fault probability of the electric power software system.
The present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method of the present application when executing the computer program.
The present application also provides a computer-readable storage medium having stored thereon a computer program which, when being executed by a processor, carries out the steps of the method as set forth in the present application.
The method for predicting the fault of the electric power spot service system at least has the following effects:
an equipment parameter dictionary is constructed through equipment parameter time sequence data, a first-class deep neural network is obtained through deep learning training of the equipment parameter dictionary, vectors of the equipment parameter time sequence data are detected through the first-class deep neural network, and the probability of equipment parameter failure is output.
According to the fault prediction method of the electric power spot business system based on deep learning, through deep learning of hardware equipment fault data and software system fault data, accurate and reliable prediction can be conducted on uncertain faults of the electric power spot business system, and therefore safe operation and reliable operation of the electric power spot market and the business system are effectively guaranteed.
Drawings
FIG. 1 is a flow diagram illustrating a method for predicting a failure of hardware equipment in a power spot business system according to an embodiment;
FIG. 2 is a flow diagram illustrating a method for predicting a failure of hardware equipment in a power spot business system according to an embodiment;
FIG. 3 is a block diagram of an embodiment of an apparatus for predicting a failure of a power spot business system;
FIG. 4 is a block diagram of an embodiment of an apparatus for predicting a failure of a power spot business system;
FIG. 5 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
In the prior art, the electric power spot business system has high operation risk and more uncertain faults, and accurate and reliable prediction cannot be carried out on the uncertain faults of the electric power spot business system.
As shown in fig. 1, an embodiment of the present application provides a method for predicting a failure of an electric power spot service system, including a method for predicting a failure of a hardware equipment class, where the method for predicting a failure of a hardware equipment class specifically includes step S100, step S200, step S300, and step S400, where:
s100, receiving equipment parameter time sequence data of hardware equipment in the electric power spot business system before failure, and constructing an equipment parameter dictionary according to the equipment parameter time sequence data;
it can be appreciated that pre-failure equipment parameter timing data serves as a sample source for deep learning training. Because the equipment parameter time sequence data is acquired in a period of time before the fault, the subsequent time is considered to be that the fault is determined, and the final first-class deep neural network can be obtained by training by utilizing the equipment parameter time sequence data before the fault based on the corresponding relation.
In one embodiment, constructing the equipment parameter dictionary by using the equipment parameters specifically includes: and reading the parameter values of the equipment parameters, and generating corresponding index numbers according to the parameter values, wherein the parameter values and the index numbers of the equipment parameters jointly form an equipment parameter dictionary.
In this embodiment, the equipment parameter dictionary is used to locate the matching position of the recording parameter value and the corresponding index number, and the location, comparison, and replacement can be performed according to the equipment parameter dictionary.
In one embodiment, step S100 specifically includes step S110 and step S120:
step S110, collecting equipment parameter time series data of a plurality of times before and after the occurrence of the fault. In this embodiment, the plurality of times may be determined according to actual conditions, for example, may be time series data of equipment parameters of a day before a failure occurs; the time series data sampling interval may be determined according to practical conditions, such as half an hour.
And step S120, constructing an equipment parameter dictionary by using the equipment parameters in the equipment parameter time sequence data.
Specifically, each different equipment parameter value in the equipment parameter dictionary corresponds to a different index number, for example, parameter value 1 corresponds to index number 1.
In this embodiment, the equipment parameter time sequence data is obtained by collecting historical fault record data of the electric power spot service system. Specifically, the historical fault record data is divided into two categories: including hardware equipment-like fault data and software system-like fault data. The method disclosed in steps S100-S400 is based on hardware equipment, and steps S500-S900 are based on software system.
Step S200, vectorizing equipment parameters by using the equipment parameter dictionary to obtain vectorized equipment parameter time sequence data, wherein at least one part of the vectorized equipment parameter time sequence data is used as training data;
in one embodiment, the vectorizing of the equipment parameter by using the equipment parameter dictionary to obtain vectorized equipment parameter time sequence data specifically includes: acquiring an equipment parameter value of each time step in equipment parameter time sequence data; searching index numbers corresponding to equipment parameter values of each time step according to the equipment parameter dictionary; establishing an all-zero vector with the vector length being the same as the equipment parameter dictionary length; and searching the position of the index number according to the equipment parameter dictionary, and setting the value of the corresponding position in the all-zero vector as one to obtain vectorized equipment parameter time sequence data.
In one embodiment, step S200 specifically includes step S210 and step S220:
step S210, using the equipment parameter dictionary to represent the equipment parameter data at each time in the equipment parameter time sequence data by one-hot vector, so as to obtain vectorized representation of the equipment parameter time sequence data.
Specifically, a one-hot vector refers to a vector in which only one position has a value of 1 and all other positions have a value of 0. And searching the corresponding index number in the equipment parameter dictionary for the equipment parameter value of each time step in the equipment parameter time sequence data, then establishing an all-zero vector with the same vector length as the equipment parameter dictionary length, setting the value of the position corresponding to the searched index number in the vector as 1, and obtaining a new vector, namely the one-hot vector code of the equipment parameter data of the current time step in the equipment parameter time sequence data. And converting the equipment parameter data of each time step in the equipment parameter time sequence data into one-hot vector coding, wherein the obtained one-hot vector coding matrix is vectorized representation of the equipment parameter time sequence data.
Step S220, dividing the vectorized equipment parameter timing data into training data and test data.
In one embodiment, at least a portion of the vectorized equipment parameter timing data is used as test data; and after the first deep neural network is obtained, evaluating the first deep neural network by using the test data.
Specifically, the ratio of the training data to the test data amount may take 8 to 2 or 7 to 3. The test data is used for detecting a first type of deep learning model. It can be understood that the deep neural network of the first kind obtained in the embodiment of the present application is used for predicting the possibility of future failure during the normal operation of the electric power off-the-shelf service system. And the source of the first deep neural network training data is vectorized equipment parameter time sequence data (a sample of the training data is taken before the fault), so that the reliability of the first deep neural network can be detected by using the vectorized equipment parameter time sequence data as test data.
Step S300, utilizing a deep neural network to finely tune, migrate and learn the training data to obtain the first type of deep neural network;
specifically, the training data is finely tuned and migrated by using a Transformer deep neural network based on an attention mechanism, and the performance is tested on the test data;
in this embodiment, a transform deep neural network based on an attention mechanism is used, and according to the vectorized equipment parameter time sequence data, an evolution rule of the power equipment fault is learned, and the occurrence probability of the power equipment fault is predicted.
Among them, the transform deep Neural Network Is a model for a sequence-to-sequence (sequence-to-sequence) task proposed by Google researchers in 2017 in the article "Attention Is All You Need", and it has no cyclic structure of RNN (Recurrent Neural Network) or Convolutional structure of CNN (Convolutional Neural Network), and has achieved certain promotion in tasks such as machine translation. The method avoids a circular model structure and completely relies on an attention mechanism to model the global dependency of input and output. Attention mechanism (attention) has become an important component of sequence modeling and transduction models in various types of tasks, allowing the modeling of dependencies of input and output sequences without considering their distance in the sequence, so that long-range dependencies can be better captured. The equipment parameter time sequence data are learned and predicted by using a transducer deep neural network based on an attention mechanism, and the dependency relationship among the equipment parameter data in different time periods can be better captured, so that the evolution rule of the power equipment fault can be better learned, and the occurrence probability of the power equipment fault can be more accurately predicted.
And S400, receiving real-time equipment parameter time sequence data of hardware equipment in the electric power spot business system, and inputting the real-time equipment parameter time sequence data into the first-class deep neural network to obtain the fault probability of the hardware equipment.
Specifically, after the deep neural network is trained, the equipment parameter time sequence data vector of a certain period of time can be input, the evolution process of the equipment parameters at the subsequent time is output, and the probability of possible failure of the equipment parameters is predicted.
According to the method for predicting the faults of the electric power spot business system, the equipment parameter dictionary is constructed by receiving the fault data of hardware equipment, vectorized equipment parameter time sequence data is obtained, and then the first-class deep neural network is obtained through training. Through deep learning of hardware equipment fault data, the uncertain fault of the electric power spot business system can be accurately and reliably predicted, and therefore safe operation and reliable operation of the electric power spot market and the business system are effectively guaranteed.
In one embodiment, as shown in fig. 2, there is provided a method for power spot business system fault prediction, including:
step S500, receiving fault data of a software system in the electric power spot business system, and constructing an attribute measurement vector data set of the electric power software system by performing software measurement on the fault data;
specifically, aiming at received fault data of a software system, software measurement is carried out on each module of the power software system, and a power software system attribute measurement vector data set is constructed; the software measurement mode comprises product measurement and/or process measurement of the software project, and the product measurement comprises a software cladding layer, a program class layer and an object method layer; process metrics include demand metrics, personnel metrics, and code change metrics. The software measurement method is only used as an example, and the specific software measurement method may be other software measurement methods in the prior art.
Further, the constructed power software system attribute metric data set may include a power health software system attribute metric vector and a power failure software system attribute metric vector.
Step S600, acquiring a comparison software system attribute measurement vector, wherein at least one part of the comparison software system attribute measurement vector is used as training data;
specifically, collecting attribute measurement vectors of a large number of other software systems (comparison software systems) with similar software quality attributes to the power software system, and constructing attribute measurement training data sets and test data sets of the other software systems;
the comparison software system can be understood as software systems of other prior art spot business systems such as a cotton spot business system, a milk spot business system, a fruit spot business system and the like, besides other comparison software systems which are used for predicting the fault of the electric power spot business system.
Specifically, the attribute metrics of the other software systems include attribute metric vectors of other software systems of a normal class and attribute metric vectors of other software systems of a fault class.
Step S700, training a two-class deep neural network by using the training data;
specifically, a two-class deep neural network is trained by using the other software system attribute measurement training data set, and classification performance is tested on the other software system attribute measurement test data set; the two-classification deep neural network can output the class label of the normal class or the fault class according to the input software system attribute measurement vector. The deep neural network of the second class may adopt a general fully-connected classification network, and may also adopt a mainstream convolutional neural network, for example, the convolutional neural network may be other convolutional neural networks such as ResNet, VGG, DenseNet, and the like.
In one embodiment, at least a portion of the vector is measured against the software system attributes as test data; and after the second type of deep neural network is obtained, evaluating the second type of deep neural network by using the test data.
It is understood that the source of the second deep neural network training data is a comparison software system attribute metric vector (derived from a comparison software system), so that at least a portion of the comparison software system attribute metric vector can be used as test data to test the reliability of the second deep neural network.
Step S800, fine tuning, migrating and learning the attribute metric vector data set of the power software system by using the trained two-class deep neural network to obtain the second class deep neural network;
and S900, receiving real-time state parameters of a software system in the electric power spot business system, converting the state parameters into attribute measurement vectors of the electric power software system, and inputting the attribute measurement vectors into a second-class deep neural network to obtain the fault probability of the electric power software system.
Specifically, after the second type of deep neural network is trained, the attribute measurement vector of the power software system can be input, and the probability of possible failure of the power software system can be predicted.
It can be understood that the deep neural network of the second type obtained in the present embodiment is used for predicting the possibility of future failure according to the real-time (current) state parameters during the normal operation of the power spot business system. Therefore, in a specific application process, the real-time state parameters (state parameters in a period of time in the current working state) of the software system in the electric power spot business system are received, and when the real-time state parameters of the software system are converted into the attribute measurement vector of the electric power software system and then input into the second type deep neural network, the probability that the electric power software system possibly fails in the future time can be detected.
According to the method for predicting the faults of the electric power spot business system, the second type of deep neural network is obtained by collecting other software systems similar to the electric power spot business system, and the second type of deep neural network can accurately and reliably predict the uncertain faults of the electric power spot business system, so that the safe operation and the reliable operation of the electric power spot market and the business system are effectively guaranteed. It can be understood that when different embodiments are combined, the accuracy of the fault prediction of the electric power spot business system can be further improved by performing deep learning on the hardware equipment fault data and the software system fault data.
It should be understood that although the various steps in the flow charts of fig. 1-2 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 1-2 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performance of the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternating with other steps or at least some of the sub-steps or stages of other steps.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program: s100, receiving equipment parameter time sequence data of hardware equipment in the electric power spot business system before failure, and constructing an equipment parameter dictionary according to the equipment parameter time sequence data; step S200, vectorizing equipment parameters by using the equipment parameter dictionary to obtain vectorized equipment parameter time sequence data, wherein at least one part of the vectorized equipment parameter time sequence data is used as training data; step S300, utilizing a deep neural network to finely tune, migrate and learn the training data to obtain the first type of deep neural network; and S400, receiving real-time equipment parameter time sequence data of hardware equipment in the electric power spot business system, and inputting the real-time equipment parameter time sequence data into the first-class deep neural network to obtain the fault probability of the hardware equipment.
In one embodiment, the processor, when executing the computer program, further performs the steps of: step S500, receiving fault data of a software system in the electric power spot business system, and constructing an attribute measurement vector data set of the electric power software system by performing software measurement on the fault data; step S600, acquiring a comparison software system attribute measurement vector, wherein at least one part of the comparison software system attribute measurement vector is used as training data; step S700, training a two-class deep neural network by using the training data; step S800, fine tuning, migrating and learning the attribute metric vector data set of the power software system by using the trained two-class deep neural network to obtain the second class deep neural network; and S900, receiving real-time state parameters of a software system in the electric power spot business system, converting the state parameters into attribute measurement vectors of the electric power software system, and inputting the attribute measurement vectors into a second-class deep neural network to obtain the fault probability of the electric power software system.
In an embodiment of the present application, as shown in fig. 3, there is further provided an apparatus for predicting a failure of an electric power spot business system, including:
the system comprises a first module, a second module and a third module, wherein the first module is used for receiving equipment parameter time sequence data of hardware equipment in the electric power spot business system before failure and constructing an equipment parameter dictionary according to the equipment parameter time sequence data;
the second module is used for carrying out vectorization processing on the equipment parameters by using the equipment parameter dictionary to obtain vectorized equipment parameter time sequence data, and at least one part of the vectorized equipment parameter time sequence data is used as training data;
the third module is used for utilizing a deep neural network to finely adjust, migrate and learn the training data to obtain the first type of deep neural network;
and the fourth module is used for receiving real-time equipment parameter time sequence data of the hardware equipment in the electric power spot business system and inputting the real-time equipment parameter time sequence data into the first-class deep neural network to obtain the failure probability of the hardware equipment.
In one embodiment, as shown in fig. 4, there is also provided an apparatus for predicting a failure of a power spot business system, including:
the system comprises a fifth module, a second module and a third module, wherein the fifth module is used for receiving fault data of a software system in the electric power spot business system and constructing an electric power software system attribute measurement vector data set by carrying out software measurement on the fault data;
a sixth module, configured to obtain a comparison software system attribute measurement vector, where at least a portion of the comparison software system attribute measurement vector is used as training data;
a seventh module for training a second-class deep neural network using the training data;
an eighth module, configured to trim, migrate and learn the attribute metric vector data set of the power software system by using the trained two-class deep neural network, to obtain the second class deep neural network;
and the ninth module is used for receiving real-time state parameters of a software system in the electric power spot business system, converting the state parameters into attribute measurement vectors of the electric power software system, and inputting the attribute measurement vectors into the second-class deep neural network to obtain the fault probability of the electric power software system.
For specific limitations of the device for predicting the fault of the power spot business system in the embodiments of the present application, reference may be made to the above limitations of the method for predicting the fault of the power spot business system, and details are not described herein again. The modules in the device for predicting the fault of the electric power spot business system can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, there is also provided an apparatus for power off-the-shelf business system failure prediction, the apparatus comprising a hardware equipment failure prediction based apparatus and a software system failure prediction based apparatus;
the hardware equipment fault prediction-based device comprises a computer memory, a computer processor and a computer program which is stored in the computer memory and can be executed on the computer processor, wherein when the computer processor executes the computer program, the method based on hardware equipment fault prediction is realized; the software system fault prediction-based device comprises a computer memory, a computer processor and a computer program which is stored in the computer memory and can be executed on the computer processor, and when the computer processor executes the computer program, the method based on the software system fault prediction is realized in the embodiments.
In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in fig. 5. The computer device includes a processor, a memory, a network interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method of power spot business system failure prediction. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
Those skilled in the art will appreciate that the architecture shown in fig. 5 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of: s100, receiving equipment parameter time sequence data of hardware equipment in the electric power spot business system before failure, and constructing an equipment parameter dictionary according to the equipment parameter time sequence data; step S200, vectorizing equipment parameters by using the equipment parameter dictionary to obtain vectorized equipment parameter time sequence data, wherein at least one part of the vectorized equipment parameter time sequence data is used as training data; step S300, utilizing a deep neural network to finely tune, migrate and learn the training data to obtain the first type of deep neural network; and S400, receiving real-time equipment parameter time sequence data of hardware equipment in the electric power spot business system, and inputting the real-time equipment parameter time sequence data into the first-class deep neural network to obtain the fault probability of the hardware equipment.
In one embodiment, the computer program when executed by the processor further performs the steps of: step S500, receiving fault data of a software system in the electric power spot business system, and constructing an attribute measurement vector data set of the electric power software system by performing software measurement on the fault data; step S600, acquiring a comparison software system attribute measurement vector, wherein at least one part of the comparison software system attribute measurement vector is used as training data; step S700, training a two-class deep neural network by using the training data; step S800, fine tuning, migrating and learning the attribute metric vector data set of the power software system by using the trained two-class deep neural network to obtain the second class deep neural network; and S900, receiving real-time state parameters of a software system in the electric power spot business system, converting the state parameters into attribute measurement vectors of the electric power software system, and inputting the attribute measurement vectors into a second-class deep neural network to obtain the fault probability of the electric power software system.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features. When technical features in different embodiments are represented in the same drawing, it can be seen that the drawing also discloses a combination of the embodiments concerned.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. The method for predicting the fault of the electric power spot business system is characterized by comprising the steps of receiving real-time equipment parameter time sequence data of hardware equipment in the electric power spot business system, and inputting the real-time equipment parameter time sequence data into a first-class deep neural network to obtain the fault probability of the hardware equipment;
the first deep neural network is obtained by pre-construction and specifically comprises the following steps:
receiving equipment parameter time sequence data of hardware equipment in the electric power spot business system before failure, and constructing an equipment parameter dictionary according to the equipment parameter time sequence data;
vectorizing the equipment parameters by using the equipment parameter dictionary to obtain vectorized equipment parameter time sequence data, wherein at least one part of the vectorized equipment parameter time sequence data is used as training data;
and utilizing a deep neural network to finely adjust, migrate and learn the training data to obtain the first type of deep neural network.
2. The method of power spot business system failure prediction according to claim 1,
at least a portion of the vectorized equipment parameter timing data as test data;
and after the first deep neural network is obtained, evaluating the first deep neural network by using the test data.
3. The method for predicting the fault of the electric power spot business system according to claim 1, wherein the step of constructing an equipment parameter dictionary according to the equipment parameter time sequence data specifically comprises the following steps:
and reading the parameter value of the equipment parameter, and generating a corresponding index number according to the parameter value, wherein the parameter value and the index number of the equipment parameter jointly form an equipment parameter dictionary.
4. The method for predicting the fault of the electric power spot business system according to claim 3, wherein the equipment parameter dictionary is used for vectorizing the equipment parameters to obtain vectorized equipment parameter time sequence data, and the method specifically comprises the following steps:
acquiring an equipment parameter value of each time step in equipment parameter time sequence data;
searching index numbers corresponding to equipment parameter values of each time step according to the equipment parameter dictionary;
establishing an all-zero vector with the vector length being the same as the equipment parameter dictionary length;
and searching the position of the index number according to the equipment parameter dictionary, and setting the value of the corresponding position in the all-zero vector as one to obtain vectorized equipment parameter time sequence data.
5. The method for predicting the fault of the electric power spot service system is characterized by comprising the steps of receiving real-time state parameters of a software system in the electric power spot service system, converting the state parameters into attribute measurement vectors of the electric power software system, and inputting the attribute measurement vectors into a second-class deep neural network to obtain the fault probability of the electric power software system;
the second type of deep neural network is obtained by pre-construction and specifically comprises the following steps:
receiving fault data of a software system in a power spot business system, and constructing a power software system attribute measurement vector data set by performing software measurement on the fault data;
acquiring a comparison software system attribute measurement vector, wherein at least one part of the comparison software system attribute measurement vector is used as training data;
training a second-class deep neural network by using the training data;
and utilizing the trained two classified deep neural networks to finely tune, transfer and learn the attribute metric vector data set of the power software system to obtain the second class of deep neural networks.
6. The method of power spot business system failure prediction according to claim 5,
at least a portion of the control software system attribute metric vector as test data;
and after the second type of deep neural network is obtained, evaluating the second type of deep neural network by using the test data.
7. An apparatus for predicting a failure of a power spot service system, the apparatus comprising:
the system comprises a first module, a second module and a third module, wherein the first module is used for receiving equipment parameter time sequence data of hardware equipment in the electric power spot business system before failure and constructing an equipment parameter dictionary according to the equipment parameter time sequence data;
the second module is used for carrying out vectorization processing on the equipment parameters by using the equipment parameter dictionary to obtain vectorized equipment parameter time sequence data, and at least one part of the vectorized equipment parameter time sequence data is used as training data;
the third module is used for utilizing the deep neural network to finely adjust, migrate and learn the training data to obtain a first class of deep neural network;
and the fourth module is used for receiving real-time equipment parameter time sequence data of the hardware equipment in the electric power spot business system and inputting the real-time equipment parameter time sequence data into the first-class deep neural network to obtain the failure probability of the hardware equipment.
8. An apparatus for predicting a failure of a power spot service system, the apparatus comprising:
the system comprises a fifth module, a second module and a third module, wherein the fifth module is used for receiving fault data of a software system in the electric power spot business system and constructing an electric power software system attribute measurement vector data set by carrying out software measurement on the fault data;
a sixth module, configured to obtain a comparison software system attribute measurement vector, where at least a portion of the comparison software system attribute measurement vector is used as training data;
a seventh module for training a second-class deep neural network using the training data;
the eighth module is used for utilizing the trained two-class deep neural network to finely tune, migrate and learn the attribute metric vector data set of the power software system to obtain a second class deep neural network;
and the ninth module is used for receiving real-time state parameters of a software system in the electric power spot business system, converting the state parameters into attribute measurement vectors of the electric power software system, and inputting the attribute measurement vectors into the second-class deep neural network to obtain the fault probability of the electric power software system.
9. Computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor realizes the steps of the method according to any of claims 1 to 7 when executing the computer program.
10. Computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111222721.3A CN113887820A (en) | 2021-10-20 | 2021-10-20 | Method and device for predicting fault of electric power spot business system, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111222721.3A CN113887820A (en) | 2021-10-20 | 2021-10-20 | Method and device for predicting fault of electric power spot business system, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113887820A true CN113887820A (en) | 2022-01-04 |
Family
ID=79003787
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111222721.3A Pending CN113887820A (en) | 2021-10-20 | 2021-10-20 | Method and device for predicting fault of electric power spot business system, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113887820A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116128323A (en) * | 2023-04-07 | 2023-05-16 | 阿里巴巴达摩院(杭州)科技有限公司 | Power transaction decision processing method, storage medium and electronic equipment |
-
2021
- 2021-10-20 CN CN202111222721.3A patent/CN113887820A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116128323A (en) * | 2023-04-07 | 2023-05-16 | 阿里巴巴达摩院(杭州)科技有限公司 | Power transaction decision processing method, storage medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113792825B (en) | Fault classification model training method and device for electricity information acquisition equipment | |
| CN103257921B (en) | Improved random forest algorithm based system and method for software fault prediction | |
| CN110489314B (en) | Model anomaly detection method and device, computer equipment and storage medium | |
| CN109783785B (en) | Method and device for generating experiment detection report and computer equipment | |
| Punitha et al. | Software defect prediction using software metrics-A survey | |
| CN111639783A (en) | Line loss prediction method and system based on LSTM neural network | |
| CN113741362A (en) | Method, system, medium, and computing device for optimizing operation of liquefied natural gas receiving station | |
| CN111709548A (en) | Power consumer load prediction method, device, equipment and storage medium based on support vector machine | |
| CN114330934A (en) | Model parameter self-adaptive GRU new energy short-term power generation power prediction method | |
| CN115034519A (en) | Method and device for predicting power load, electronic equipment and storage medium | |
| CN111124898B (en) | Question-answering system testing method and device, computer equipment and storage medium | |
| US20180307218A1 (en) | System and method for allocating machine behavioral models | |
| CN112215398A (en) | Power consumer load prediction model establishing method, device, equipment and storage medium | |
| CN112506757A (en) | Automatic test method, system, computer device and medium thereof | |
| CN113887820A (en) | Method and device for predicting fault of electric power spot business system, computer equipment and storage medium | |
| CN113110961B (en) | Equipment abnormality detection method and device, computer equipment and readable storage medium | |
| Meritxell et al. | On the evaluation, management and improvement of data quality in streaming time series | |
| CN111290953B (en) | Method and device for analyzing test logs | |
| CN112684396A (en) | Data preprocessing method and system for electric energy meter operation error monitoring model | |
| CN116757650A (en) | Project management and resource scheduling method based on machine learning | |
| CN109829745A (en) | Business revenue data predication method, device, computer equipment and storage medium | |
| CN112700065A (en) | Business process completion time interval prediction method and system based on deep learning | |
| CN110865939A (en) | Application program quality monitoring method and device, computer equipment and storage medium | |
| CN110177006A (en) | Node test method and device based on interface prediction model | |
| CN112801372B (en) | Data processing method, device, electronic equipment and readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |