Equipment fault identification method and device and storage medium

Abstract

The invention discloses an equipment fault identification method and device and a storage medium. The method comprises the steps of obtaining device parameters corresponding to multiple energy devices in a target energy scene, wherein the target energy scene is used for collecting energy through the multiple energy devices; fault recognition is performed on the equipment parameters corresponding to the multiple energy equipment through a composite fault recognition model, and the composite fault recognition model is obtained by training the training data obtained in the simulation energy scene. The training data comprises simulation signal data sets generated under N equipment fault types simulated based on a plurality of simulation energy equipment in the simulation energy scene, and simulation state probability vectors observed based on the simulation signal data sets; and obtaining fault recognition confidence matched with the plurality of energy devices and output by the composite fault recognition model. According to the method and the device, the problem of relatively low accuracy of a fault identification result caused by omission of some non-key faults is solved.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular, to a method and device for identifying equipment failures and a storage medium. Background Art

[0002] As wind turbines tend to be designed in larger sizes, the power generation of a single wind turbine has also increased significantly compared to previous wind turbine models. However, once these large wind turbines have non-critical faults, such as failures of control equipment (i.e. sensors, actuators, control units, etc.), they are often more likely to cause greater power generation losses. Among them, since the characteristics of these non-critical faults are not obvious in the early stage and will not cause downtime after the fault occurs, they are often ignored by operation and maintenance personnel.

[0003] However, if no measures are taken to eliminate non-critical faults for a long time, not only will the performance of the wind turbine decline, but it may also cause the wind turbine to shut down, making it impossible to effectively ensure the safety of wind turbine operation. In other words, the fault identification methods currently provided by the prior art for equipment in industrial application scenarios often miss some non-critical faults whose manifestations are not obvious, and there is a problem of low accuracy of fault identification results.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present application provide a method and apparatus for identifying equipment faults and a storage medium, so as to at least solve the technical problem that the accuracy of fault identification results is low due to the omission of some non-critical faults.

[0006] According to one aspect of an embodiment of the present application, a method for identifying equipment faults is provided, comprising: obtaining equipment parameters corresponding to each of a plurality of energy devices within a target energy scenario, wherein the target energy scenario is used to collect energy through the plurality of energy devices; performing fault identification on the equipment parameters corresponding to each of the plurality of energy devices through a composite fault identification model, wherein the composite fault identification model is trained using training data obtained in a simulated energy scenario, wherein the simulated energy scenario is configured with a plurality of simulated energy devices, and the training data comprises a simulation signal data set generated based on N types of equipment fault types simulated by the plurality of simulated energy devices, and a simulation state probability vector corresponding to the operating state of the simulated equipment observed based on the simulation signal data set, wherein N is a positive integer; obtaining a fault identification confidence output by the composite fault identification model that matches the plurality of energy devices, wherein the fault identification confidence is used to indicate the probability of the N types of equipment faults occurring in the plurality of energy devices.

[0007] As an optional implementation, before obtaining the device parameters corresponding to each of the multiple energy devices in the target energy scenario, it also includes: constructing the above-mentioned simulated energy scenario based on preset scenario parameters, wherein the above-mentioned preset scenario parameters include the environmental layout parameters of the above-mentioned simulated energy scenario, and the device parameters of each of the above-mentioned multiple simulated energy devices contained in the above-mentioned simulated energy scenario; determining the energy reference interval for model operation in the above-mentioned simulated energy scenario, and the fault simulation control parameters, wherein the above-mentioned fault simulation control parameters are control parameters used to control the above-mentioned multiple simulated energy devices to switch to a healthy state or a faulty state in the process of simulating the i-th device fault type in the above-mentioned simulated energy scenario, and i is a natural number greater than or equal to 1 and less than or equal to N; obtaining the above-mentioned simulation signal data set generated in the process of simulating N types of device fault types. , wherein the above-mentioned simulation signal data set includes: in the process of simulating each type of equipment failure, the equipment parameters of the above-mentioned multiple simulated energy devices in a healthy state and the equipment parameters of the above-mentioned multiple simulated energy devices in a faulty state; in the process of simulating the operation of each type of equipment failure, obtaining the health observation values ​​observed when the above-mentioned multiple simulated energy devices are in a healthy state, and the fault observation values ​​observed when the above-mentioned multiple simulated energy devices are in a faulty state; based on the comparison results between the above-mentioned health observation values ​​and the above-mentioned fault observation values, determining the above-mentioned simulation state probability vector corresponding to the operation state of the above-mentioned simulation equipment observed by the above-mentioned multiple simulated energy devices; using the above-mentioned simulation signal data set and the above-mentioned simulation state probability vector as training data, training the neural network model in the initial state until the above-mentioned composite fault identification model that meets the training convergence conditions is obtained.

[0008] As an optional implementation, the above-mentioned acquisition of the simulation signal data set generated in the process of simulating N types of equipment failure types includes: determining the first time ratio of the above-mentioned multiple simulated energy devices in a healthy state and the second time ratio of the above-mentioned multiple simulated energy devices in a faulty state in the process of simulating the i-th equipment failure type according to the above-mentioned fault simulation control parameters; inserting the fault operation time determined based on the above-mentioned second time ratio into the healthy operation time determined by the above-mentioned first time ratio according to preset conditions; controlling the above-mentioned multiple simulated energy devices to operate according to healthy device parameters during the above-mentioned healthy operation time, and adjusting the device parameters of the above-mentioned multiple simulated energy devices to operate in the above-mentioned fault operation time to the i-th group of fault parameters matching the i-th equipment failure type; acquiring the behavioral performance parameters of the above-mentioned multiple simulated energy devices in the process of simulating the i-th equipment failure type as the i-th simulation signal; and generating the above-mentioned simulation signal data set using the N types of simulation signals acquired.

[0009] As an optional implementation, after obtaining the above-mentioned simulation signal data set generated in the process of simulating N types of equipment failures, it also includes: scaling the above-mentioned simulation signal data set to obtain a scaled simulation signal data set; performing data enhancement processing and segmentation processing on the above-mentioned scaled simulation signal data set to obtain signal segments, and saving the above-mentioned signal segments in a matrix form.

[0010] As an optional implementation, the above-mentioned determination of the simulation state probability vector corresponding to the operation state of the above-mentioned simulation device observed by the above-mentioned multiple simulated energy devices based on the comparison results between the above-mentioned health observation values ​​and the above-mentioned fault observation values ​​includes: determining K types of simulation device operation states based on the number of the above-mentioned multiple simulated energy devices and the number N of the above-mentioned N types of device fault types; using the comparison results between the above-mentioned health observation values ​​and the above-mentioned fault observation values, creating a state list with a length of K, wherein the above-mentioned comparison results are used to indicate the ratio between the above-mentioned health observation values ​​and the above-mentioned fault observation values ​​corresponding to the j-th simulation energy device under the i-th device fault type; and constructing the above-mentioned simulation state probability vector based on the above-mentioned state list.

[0011] As an optional implementation, the above-mentioned acquisition of equipment parameters corresponding to each of the multiple energy devices within the target energy scenario includes: acquiring the wind turbine operating parameters corresponding to each of the multiple wind turbines within the target wind farm; the above-mentioned fault identification of the equipment parameters corresponding to each of the above-mentioned multiple energy devices through a composite fault identification model includes: fault identification of the wind turbine operating parameters corresponding to each of the above-mentioned multiple wind turbines through a composite fault identification model, wherein the above-mentioned training data used by the above-mentioned composite fault identification model includes a simulated wind turbine signal data set and a simulated wind turbine state probability vector generated under at least two sensor-type equipment failure types simulated based on the multiple simulated wind turbines within the constructed simulated wind farm.

[0012] According to another aspect of an embodiment of the present application, a device for identifying equipment faults is also provided, including: a first acquisition unit, used to acquire equipment parameters corresponding to each of multiple energy devices within a target energy scenario, wherein the target energy scenario is used to collect energy through the multiple energy devices; an identification unit, used to perform fault identification on the equipment parameters corresponding to each of the multiple energy devices through a composite fault identification model, wherein the composite fault identification model is trained using training data acquired in a simulated energy scenario, wherein the simulated energy scenario is configured with multiple simulated energy devices, and the training data includes a simulation signal data set generated based on N types of equipment fault types simulated by the multiple simulated energy devices, and a state probability vector corresponding to the operating state of the simulated device observed based on the simulation signal data set, where N is a positive integer; a second acquisition unit, used to acquire a fault identification confidence output by the composite fault identification model that matches the multiple energy devices, wherein the fault identification confidence is used to indicate the probability of the N types of equipment faults occurring in the multiple energy devices.

[0013] As an optional implementation, the apparatus further includes: a construction unit, for constructing the simulated energy scenario based on preset scenario parameters before obtaining the device parameters corresponding to each of the multiple energy devices in the target energy scenario, wherein the preset scenario parameters include the environmental layout parameters of the simulated energy scenario and the device parameters of each of the multiple simulated energy devices contained in the simulated energy scenario; a first determination unit, for determining an energy reference interval for model operation in the simulated energy scenario, and a fault simulation control parameter, wherein the fault simulation control parameter is a control parameter for controlling the multiple simulated energy devices to switch to a healthy state or a faulty state in the process of simulating the i-th device fault type in the simulated energy scenario, and i is a natural number greater than or equal to 1 and less than or equal to N; a third acquisition unit, for acquiring the simulation signal number generated in the process of simulating N types of device faults. A data set, wherein the above-mentioned simulation signal data set includes: in the process of simulating each type of equipment failure, the equipment parameters of the above-mentioned multiple simulated energy devices in a healthy state and the equipment parameters of the above-mentioned multiple simulated energy devices in a faulty state; a fourth acquisition unit, used to obtain the health observation values ​​observed when the above-mentioned multiple simulated energy devices are in a healthy state, and the fault observation values ​​observed when the above-mentioned multiple simulated energy devices are in a faulty state in the process of simulating the operation of each type of equipment failure; a second determination unit, used to determine the above-mentioned simulation state probability vector corresponding to the operation state of the above-mentioned simulation device observed by the above-mentioned multiple simulated energy devices based on the comparison result between the above-mentioned health observation value and the above-mentioned fault observation value; a training unit, used to use the above-mentioned simulation signal data set and the above-mentioned simulation state probability vector as training data to train the neural network model in an initial state until the above-mentioned composite fault recognition model that meets the training convergence condition is obtained.

[0014] As an optional implementation, the third acquisition unit includes: a first determination module, used to determine the first time ratio of the multiple simulated energy devices in a healthy state and the second time ratio of the multiple simulated energy devices in a faulty state in the process of simulating the i-th device fault type according to the fault simulation control parameters, where i is a natural number greater than or equal to 1 and less than or equal to N; a first insertion module, used to insert the fault operation time determined based on the second time ratio into the healthy operation time determined by the first time ratio according to preset conditions; an adjustment module, used to control the multiple simulated energy devices to operate according to healthy device parameters during the healthy operation time, and adjust the device parameters of the multiple simulated energy devices to the i-th group of fault parameters matching the i-th device fault type during the fault operation time; a first acquisition module, used to obtain the behavioral performance parameters of the multiple simulated energy devices in the process of simulating the i-th device fault type as the i-th simulation signal; a generation module, used to generate the above-mentioned simulation signal data set using the N types of simulation signals obtained.

[0015] As an optional implementation, it also includes: a scaling module, which is used to scale and adjust the above-mentioned simulation signal data set after obtaining the above-mentioned simulation signal data set generated in the process of simulating N types of equipment failures, so as to obtain a scaled simulation signal data set; a processing module, which is used to perform data enhancement processing and segmentation processing on the above-mentioned scaled simulation signal data set to obtain signal segments, and save the above-mentioned signal segments in a matrix form.

[0016] As an optional implementation, the second determination unit includes: a second determination module, used to determine K types of simulated equipment operating states based on the number of the above-mentioned multiple simulated energy devices and the number N of the above-mentioned N types of equipment failure types; a creation module, used to create a state list with a length of K using the comparison results between the above-mentioned health observation values ​​and the above-mentioned fault observation values, wherein the above-mentioned comparison results are used to indicate the ratio between the above-mentioned health observation values ​​and the above-mentioned fault observation values ​​corresponding to the j-th simulated energy device under the i-th equipment failure type, j is a natural number greater than or equal to 1 and less than or equal to M, and M is the number of multiple simulated energy devices configured in the above-mentioned simulated energy scenario; a construction module, used to construct the above-mentioned simulation state probability vector based on the above-mentioned state list.

[0017] As an optional implementation, the first acquisition unit includes: a second acquisition module, used to acquire the wind turbine operating parameters corresponding to each of the multiple wind turbines in the target wind farm; the identification unit includes: an identification module, used to perform fault identification on the wind turbine operating parameters corresponding to each of the multiple wind turbines through a composite fault identification model, wherein the training data used by the composite fault identification model includes a simulated wind turbine signal data set and a simulated wind turbine state probability vector generated under at least two sensor-type equipment failure types simulated by multiple simulated wind turbines in the constructed simulated wind farm.

[0018] According to another aspect of the embodiments of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the above-mentioned device failure identification method when running.

[0019] According to another aspect of the embodiments of the present application, a computer program product or a computer program is provided, the computer program product or the computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the above device fault identification method.

[0020] In an embodiment of the present application, after obtaining the equipment parameters corresponding to each of the multiple energy devices in the target energy scenario, the equipment parameters corresponding to each of the multiple energy devices are identified through a composite fault identification model, where the composite fault identification model is obtained by training using the training data obtained in the simulated energy scenario, where multiple simulated energy devices are configured in the simulated energy scenario, and the training data includes a simulation signal data set generated based on N types of equipment failure types simulated by multiple simulated energy devices, and a simulation state probability vector corresponding to the operating state of the simulated device observed based on the simulation signal data set. In the case where the composite fault identification model outputs a fault identification confidence that matches the multiple energy devices, it is determined based on the fault identification confidence here whether each of the multiple energy devices has an equipment failure, and which type of equipment failure it is. The composite fault identification model is obtained by training the training data produced by simulating different equipment failure types in the constructed simulated energy scenario, and the correlation between various equipment type failures and equipment parameters of energy equipment in the energy scenario is analyzed in a data-driven manner, so that the data relationship between multiple interacting energy devices in the energy scenario is comprehensively analyzed in the composite fault identification model, thereby realizing simultaneous fault identification of multiple energy devices, achieving the effect of accurately and effectively identifying energy equipment failures, and thus overcoming the problem of inaccurate fault identification results caused by missing some features in related technologies. Furthermore, comprehensive analysis of multiple energy devices contained in the entire energy scenario will also achieve the purpose of more comprehensively evaluating the overall performance of a certain regional scenario, and support the refined operation and maintenance development of each energy device in a certain regional scenario. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0022] Figure 1 is a flow chart of an optional method for identifying a device failure according to an embodiment of the present application;

[0023] Figure 2 is a flowchart of another optional device failure identification method according to an embodiment of the present application;

[0024] Figure 3 is a flowchart of another optional method for identifying a device failure according to an embodiment of the present application;

[0025] Figure 4 is a flowchart of another optional method for identifying a device failure according to an embodiment of the present application;

[0026] Figure 5 is a schematic structural diagram of an optional device for identifying equipment failure according to an embodiment of the present application;

[0027] Figure 6 is a schematic structural diagram of an optional device for identifying equipment failure according to an embodiment of the present application;

[0028] Figure 7 It is a schematic diagram of the structure of an optional computer system according to an embodiment of the present application. DETAILED DESCRIPTION

[0029] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0030] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0031] According to one aspect of the embodiments of the present application, a method for identifying a device failure is provided, such as Figure 1 As shown, the method includes:

[0032] S102, obtaining device parameters corresponding to each of a plurality of energy devices in a target energy scenario, wherein the target energy scenario is used to collect energy through a plurality of energy devices;

[0033] Optionally, in this embodiment, the target energy scenario may be, but is not limited to, a regional scenario for collecting natural energy, such as a wind farm, a photovoltaic electric field, etc., and the target energy scenario will be configured with a plurality of energy devices, each of which is used to collect natural energy and convert it into electrical energy. For example, a wind farm is configured with wind power generation equipment for collecting wind energy, and a photovoltaic electric field is configured with photovoltaic equipment for collecting solar energy.

[0034] In addition, in this embodiment, the above-mentioned equipment parameters may include but are not limited to equipment operation parameters, environmental layout parameters of the environment where the equipment is located, etc. Among them, the environmental layout parameters here may include but are not limited to the scene layout of the target energy scene, environmental temperature and humidity, geographical climate, etc. The equipment operation parameters here may include but are not limited to the stable operation time of the equipment, equipment temperature, equipment performance parameters, such as equipment model, sensor model in the equipment, sensor power parameters, etc. The parameter types listed here are all examples, and this embodiment does not limit this.

[0035] S104, performing fault identification on device parameters corresponding to each of the plurality of energy devices through a composite fault identification model, wherein the composite fault identification model is obtained by training using training data obtained in a simulated energy scenario, wherein the simulated energy scenario is configured with a plurality of simulated energy devices, and the training data includes a simulation signal data set generated under N types of device faults simulated based on the plurality of simulated energy devices, and a simulation state probability vector corresponding to the operation state of the simulated device observed based on the simulation signal data set, where N is a positive integer;

[0036] Optionally, in this embodiment, the above-mentioned composite fault identification model can be, but is not limited to, obtained through multiple iterations of training using training data obtained in a simulated energy scenario, wherein the simulated energy scenario can be, but is not limited to, obtained by digitally modeling the energy scenario. In the simulated energy scenario, multiple simulated energy devices configured are simulated to run to collect the simulated signal data sets generated under the simulated N types of equipment failures, and the simulated state probability vectors corresponding to the simulated device operating states observed based on the simulated signal data sets, so as to obtain the training data for training the initialized composite identification model based on this.

[0037] It should be noted that, in this embodiment, the above-mentioned composite fault identification model can be, but is not limited to, a neural network model that supports simultaneous fault identification processing of multiple energy devices. For example, still taking wind power generation equipment as an example, the composite fault identification model here can be a deep convolutional network model (Deep Neural Network, referred to as DNN) for multi-wind turbine fault identification, which is composed of multiple neural network layers, each layer contains multiple neurons, and extracts complex features of data through multi-level nonlinear transformations, thereby achieving efficient learning and prediction of data.

[0038] That is to say, in this embodiment, the above-mentioned composite fault identification model can, but is not limited to, discover the data relationship between multiple simulated energy devices under a specific equipment failure type by learning the data generated by multiple simulation scenarios under simulated energy scenarios through multiple levels of multiple neurons, so as to obtain the data change rules or characteristics therein, so that the composite fault identification model has the equipment parameters corresponding to each of the multiple energy devices input to analyze and determine the fault identification confidence of each energy device.

[0039] S106, obtaining the fault identification confidences matched with the multiple energy devices output by the composite fault identification model, wherein the fault identification confidences are used to indicate the probabilities of N types of device faults occurring in the multiple energy devices.

[0040] Optionally, in this embodiment, the above-mentioned fault identification confidence may be, but is not limited to, the probability of a certain type of equipment fault occurring in the energy device obtained by calculation by the composite fault identification model. For example, the above-mentioned fault identification confidence may be, but is not limited to, a matrix expression form or a vector expression form. It may represent the probability of the same type of equipment fault occurring in multiple energy devices, or it may represent the probability of different types of equipment fault occurring in multiple energy devices, and this embodiment does not impose any limitation on this.

[0041] Through the embodiments provided by the present application, after obtaining the equipment parameters corresponding to each of the multiple energy devices in the target energy scenario, the equipment parameters corresponding to each of the multiple energy devices are identified through a composite fault identification model, where the composite fault identification model is obtained by training using the training data obtained in the simulated energy scenario, where multiple simulated energy devices are configured in the simulated energy scenario, and the training data includes a simulation signal data set generated under N types of equipment failure types simulated by multiple simulated energy devices, and a simulation state probability vector corresponding to the operating state of the simulated device observed based on the simulation signal data set. In the case where the composite fault identification model outputs a fault identification confidence that matches the multiple energy devices, it is determined based on the fault identification confidence here whether each of the multiple energy devices has an equipment failure, and which type of equipment failure it is. The composite fault identification model is obtained by training the training data produced by simulating different equipment failure types in the constructed simulated energy scenario, and the correlation between various equipment type failures and equipment parameters of energy equipment in the energy scenario is analyzed in a data-driven manner, so that the data relationship between multiple interacting energy devices in the energy scenario is comprehensively analyzed in the composite fault identification model, thereby realizing simultaneous fault identification of multiple energy devices, achieving the effect of accurately and effectively identifying energy equipment failures, and thus overcoming the problem of inaccurate fault identification results caused by missing some features in related technologies. Furthermore, comprehensive analysis of multiple energy devices contained in the entire energy scenario will also achieve the purpose of more comprehensively evaluating the overall performance of a certain regional scenario, and support the refined operation and maintenance development of each energy device in a certain regional scenario.

[0042] As an optional solution, before obtaining the device parameters corresponding to each of the multiple energy devices in the target energy scene in step S102, such as Figure 2 As shown, it also includes:

[0043] S202, constructing a simulated energy scene based on preset scene parameters, wherein the preset scene parameters include environment layout parameters of the simulated energy scene and device parameters of each of a plurality of simulated energy devices included in the simulated energy scene;

[0044] Optionally, in this embodiment, the above-mentioned environmental layout parameters may include but are not limited to: the scene layout of the target energy scene, the environmental temperature and humidity, the geographical climate, etc., wherein the scene layout of the target energy scene may include but are not limited to: the area and regional location of the target energy scene, the geographical location of each energy device configured in the target energy scene, etc. The above-mentioned equipment parameters may include but are not limited to the stable operation time of the equipment, the equipment temperature, and the equipment performance parameters, such as the equipment model, the sensor model in the equipment, the sensor power parameters, etc. For example, still taking the wind farm as an example, the data obtained here for constructing the simulated energy scene may include the following: wind farm layout, wind farm parameters, wind turbine model, target detection fault type and related parameters, wherein the data sources of the above-mentioned data are wind turbine design drawings, wind farm resource assessment reports, design drawings, completion drawings and other materials.

[0045] It should be noted that, in the process of simulation operation based on the above-mentioned simulated energy scenario, the environment of the target to be detected can be frequently stimulated but is not limited to being the main focus. Assuming that the simulated energy scenario is a simulated wind farm, for example, in the simulation of the control system failure study, the wind speed v wind The value range can be set to v cut-in <v wind <v cut-out , v cut-in is the cut-in wind speed of the fan, v cut-out is the cut-out wind speed of the wind turbine. The model is simulated for n m seconds, and 10 data points are collected per second. n is generally greater than 100. The value of n is 500 in the experiment of this method. This value can be freely controlled. The larger the value, the more data.

[0046] S204, determining an energy reference interval for running the model in the simulated energy scenario and a fault simulation control parameter, wherein the fault simulation control parameter is a control parameter for controlling a plurality of simulated energy devices to switch to a healthy state or a faulty state during a process of simulating the i-th device fault type in the simulated energy scenario, and i is a natural number greater than or equal to 1 and less than or equal to N;

[0047] Optionally, in this embodiment, the above-mentioned fault simulation control parameters may include but are not limited to the time control ratio of the stable operation time of the simulated energy scenario and the fault occurrence time during the simulation operation based on the above-mentioned simulated energy scenario.

[0048] For example, assume that the fault simulation control parameter is represented by x. Each simulation includes the stable operation time of x and the total fault occurrence time of (1 - x) (0 < x < 1). The fault occurrence time of 1 - x is randomly divided into y parts, and then randomly inserted into the stable operation time. Further assume that when conducting experiments on this method, the simulation duration is 2 seconds and x = 0.8. Therefore, there is a total of 1.6 seconds of stable operation time and 0.4 seconds of total fault occurrence time, with a total of 8 faults occurring, each with a random duration.

[0049] S206, obtain the simulation signal data set generated during the simulation of N types of device fault types, where the simulation signal data set includes: during the simulation of each type of device fault type, the device parameters of multiple simulation energy devices in a healthy state and the device parameters of multiple simulation energy devices in a faulty state;

[0050] Optionally, in this embodiment, the device parameters of the simulation energy devices in the healthy state and the device parameters of the simulation energy devices in the faulty state in the above simulation signal data set. For example, assume that the simulation energy scenario takes the simulation wind farm as an example, and the signal data set is represented by represented by N represents the number of samples, and a represents the fan behavior representative parameter included in each sample. Among them, the fan behavior representative parameter can be adjusted correspondingly based on different simulation scenarios. For example, during the simulation of the wind energy control system fault, if judging the pitch angle sensor fault, the wind speed, wind direction, pitch angle of the blade, power, generator reference torque, and wind force at the top of the nacelle can be selected.

[0051] S208, during the operation of simulating each type of device fault type, obtain the healthy observation values observed by multiple simulation energy devices in a healthy state respectively, and the fault observation values observed by multiple simulation energy devices in a faulty state;

[0052] It should be noted that assume that the stable operation time and the total fault occurrence time are set with the fault simulation control parameter x (0 < x < 1), and then the total fault occurrence time is randomly divided into y durations. Then, during the simulation operation of a certain fault in the simulation energy scenario, the above y durations can be inserted into the stable operation process respectively by using, but not limited to, an offset constant, so as to realize the offset fault setting of the simulated device fault. Further, the device data observed during the stable operation time is used as the healthy observation value, and the device data observed during the fault occurrence time will be used as the fault observation value. For example, assume that the offset setting formula is expressed as:

[0053] x f [t] = x h [t] + δ(1)

[0054] Where δ represents the offset constant, x h [t] and x f [t] represents the healthy observation value and the fault observation value at time t respectively.

[0055] S210, based on the comparison result between the healthy observation value and the fault observation value, determining a simulation state probability vector corresponding to the operation state of the simulation device observed based on the plurality of simulation energy devices;

[0056] For example, based on the above hypothetical scenario, the number of healthy observations n is weakly counted h and the number of fault observations n f , then according to the ratio of healthy observation values ​​to fault observation values, the simulation state probability vector can be constructed Each element Represents the probability that the system is in this state.

[0057] S212, using the simulation signal data set and the simulation state probability vector as training data, training the neural network model in the initial state until a composite fault recognition model that meets the training convergence condition is obtained.

[0058] Optionally, in this embodiment, the above-mentioned composite fault identification model can be but is not limited to a DNN network, which includes: an input layer, a convolution layer, a batch normalization layer, a maximum pooling layer, a flattening layer and a fully connected layer.

[0059] For example, suppose the training process is Figure 3 Take the following content as an example:

[0060] S302, training starts;

[0061] S304, inputting the training data into the input layer;

[0062] S306, performing n convolution processing in the convolution layer and performing normalization processing in the batch normalization layer;

[0063] S308, performing global maximum pooling processing on the normalized result through a maximum pooling layer;

[0064] S310, flattening the result after the maximum pooling processing into one-dimensional data so as to be passed to the fully connected layer for processing.

[0065] S312, repeat the process m times in the fully connected layer until the training is completed, as shown in step S314.

[0066] Optionally, in this embodiment, in the process of training the neural network model in the initial state using training data, the training convergence conditions may include but are not limited to: 1) the number of training times reaches a preset iteration number threshold; 2) the output results of multiple consecutive trainings are less than or equal to a preset threshold. The examples listed here are not limited to the specific content of the training convergence conditions in this embodiment.

[0067] Through the embodiment provided by the present application, through the constructed simulated energy scenario, the simulated signal data set generated in the process of simulating N types of equipment fault types is obtained, and the health observation values ​​observed by multiple simulated energy devices in a healthy state and the fault observation values ​​observed by multiple simulated energy devices in a faulty state are obtained in the process of simulating the operation of each type of equipment fault, and the simulated state probability vector corresponding to the operation state of the simulated device is obtained based on the comparison results of the health observation value and the fault observation value. Thus, the training data constructed based on the simulated state probability vector of the simulated signal data set is obtained, and the training data is applied to the neural network model in the initial state for training until the composite fault identification model that supports fault identification of multiple energy devices is obtained to meet the training convergence conditions.

[0068] As an optional solution, obtaining a simulation signal data set generated in the process of simulating N types of equipment failures includes:

[0069] S1, determining, according to the fault simulation control parameters, a first time ratio of a plurality of simulated energy devices being in a healthy state and a second time ratio of a plurality of simulated energy devices being in a faulty state during a simulation of the i-th device fault type, where i is a natural number greater than or equal to 1 and less than or equal to N;

[0070] S2, inserting the faulty operation time determined based on the second time ratio into the healthy operation time determined by the first time ratio according to a preset condition;

[0071] S3, controlling the multiple simulated energy devices to operate according to healthy device parameters during the healthy operation time, and adjusting the device parameters of the multiple simulated energy devices to the i-th group of fault parameters matching the i-th device fault type during the fault operation time for operation;

[0072] S4, obtaining behavior performance parameters of multiple simulated energy devices during the operation process of simulating the i-th device failure type as the i-th simulation signal;

[0073] S5, generating a simulation signal data set using the acquired N types of simulation signals.

[0074] For example, assuming that the total duration of a simulation is 2 seconds, the first duration ratio is 0.8, and the second duration ratio corresponds to 1-0.8=0.2, it can be determined that there is a total of 1.6 seconds of stable operation time, 0.4 seconds of total fault occurrence time, a total of 8 faults, and the offset constant δ is 0.2. According to the offset expression in the above formula (1), 8 fault durations of 0.4 seconds can be inserted into the 1.6-second stable operation time, so that different equipment states are presented in different time periods during the simulation process, and the occurrence of a certain fault is simulated as realistically as possible. And further obtain the signal data collected from each device after the above duration allocation to obtain the health observation value x h [t] and the fault observation value x f [t].

[0075] Through the embodiments provided in the present application, by flexibly and variably adjusting the stable operation time and fault time of the simulated energy equipment in the simulated energy scenario, and inserting the fault time into the stable operation time, the two are effectively integrated, thereby achieving the most accurate and realistic simulation of different situations of multiple energy equipment in the energy scenario, so as to obtain training data with wider coverage and richer content, thereby improving the accuracy of the training results of the composite fault identification model.

[0076] As an optional solution, after obtaining the simulation signal data set generated in the process of simulating N types of equipment failures, the method further includes:

[0077] S1, scaling and adjusting the simulation signal data set to obtain a scaled simulation signal data set;

[0078] Optionally, in this embodiment, the scaling adjustment may be, but is not limited to, linear scaling adjustment using a linear scaling function.

[0079] For example, suppose that for the simulated signal data set Perform scaling adjustment, that is, apply the determined scaling function to the above simulation signal data set For each signal in The original linear scaling formula is optimized to make the data concentrated in the interval [0.5, 1.5] to further enhance the learning effect of the neural network in the subsequent steps. For example, as shown in Formula 2, and are the minimum and maximum values ​​of the signal respectively.

[0080]

[0081] Among them, the scaled original data set Indicated as x.

[0082] S2, performing data enhancement processing and segmentation processing on the scaled simulation signal data set to obtain signal segments, and saving the signal segments in a matrix form.

[0083] Optionally, in this embodiment, the above-mentioned data enhancement processing is a technology for generating new training data by transforming, rotating, scaling, cropping and other operations on the original data. Data enhancement can improve the generalization ability of the model, reduce the risk of overfitting, and can help the model better learn the invariance and robustness of the data. The above-mentioned segmentation processing can be an overlapping segmentation processing, a data processing method that divides the input data into multiple parts, and there is overlap between these parts. In the overlapping segmentation processing, each part is not completely independent, but has a certain overlapping area with the adjacent part. This method can effectively utilize the repeated information of the data and improve the efficiency and accuracy of data processing. Common application fields include signal processing, image processing and machine learning. Among them, the segmentation processing can also include the following methods: non-overlapping segmentation methods such as block processing, or other methods such as sliding windows, short-time Fourier transforms, and wavelet transforms.

[0084] For example, the above simulation signal data set is segmented by overlapping The scaled signal is subjected to data enhancement, and the segmented signal fragments are saved in matrix form.

[0085] Through the embodiments provided in the present application, data enhancement processing and overlapping segmentation processing are performed on the simulation signal data set, thereby achieving data optimization of the simulation signal data set itself, thereby achieving the purpose of reducing the risk of overfitting.

[0086] As an optional solution, based on the comparison result between the healthy observation value and the fault observation value, the simulation state probability vector corresponding to the operation state of the simulation device observed by the multiple simulation energy devices is determined to include:

[0087] S1, determining K types of simulated equipment operating states based on the number of multiple simulated energy equipment and the number N of N types of equipment failure types;

[0088] S2, using the comparison result between the healthy observation value and the fault observation value, create a state list of length K, wherein the comparison result is used to indicate the ratio between the healthy observation value and the fault observation value corresponding to the jth simulated energy device under the i-th device fault type, j is a natural number greater than or equal to 1 and less than or equal to M, and M is the number of multiple simulated energy devices configured in the simulated energy scenario;

[0089] S3, constructs a simulation state probability vector based on the state list.

[0090] For example, assuming that the target energy scenario is still a wind farm and the energy equipment is still a wind turbine. Assuming that the number of wind turbines is P and the number of detected fault types is N, the calculation method of the simulated equipment operating state K is as follows:

[0091] K=Px N+1 (3)

[0092] Among them, 1 means that all wind turbines in the wind farm are “all units are healthy”.

[0093] Then, create a list of length K, sort the previously determined fault categories in this order, and put the state "all units are healthy" first (that is, the 0th value in the list). Among them, each state k in the above state list K ij The value of is the ratio between the healthy observation value and the fault observation value corresponding to the jth simulated energy device under the i-th device failure type.

[0094] Through the embodiments provided in the present application, by constructing the above-mentioned state list, it is possible to traverse each energy device under various types of equipment failures, so as to ensure the comprehensiveness and diversity of the training data finally obtained, thereby achieving the purpose of improving the recognition accuracy of the composite fault recognition model.

[0095] As an optional solution,

[0096] S102, obtaining device parameters corresponding to each of the multiple energy devices in the target energy scenario includes: S1022, obtaining wind turbine operating parameters corresponding to each of the multiple wind turbines in the target wind farm;

[0097] S104, performing fault identification on equipment parameters corresponding to each of the multiple energy devices through a composite fault identification model includes: S1042, performing fault identification on wind turbine operating parameters corresponding to each of the multiple wind turbines through a composite fault identification model, wherein the training data used by the composite fault identification model includes a simulated wind turbine signal data set and a simulated wind turbine state probability vector generated under at least two types of sensor-type equipment faults simulated by multiple simulated wind turbines in the constructed simulated wind farm.

[0098] Specific combination Figure 4 The process shown is used to illustrate the process of applying the wind farm in the embodiment of the present application:

[0099] S402, start;

[0100] S404, Data acquisition. Training data is obtained by modeling and simulating the operation of a wind farm. The method uses the visualization simulation tool in the modeling calculation software to model the wind farm to be diagnosed. The modeling content includes environmental parameters, wind farm layout, wind field parameters, wind turbine models, fault types to be detected and related parameters, and the relevant information sources are materials such as wind turbine design drawings, resource assessment reports of the wind farm, design drawings, and completion drawings.

[0101] S406, Control system signal generation. The simulation model built in S404 is used for simulation operation. In the design of the simulation environment, the environment that can frequently stimulate the target to be detected is mainly used. For example, in the simulation of the control system fault research, the value range of the wind speed v wind can be set as v cut-in < v wind < v cut-out , v cut-in is the cut-in wind speed of the wind turbine, and v cut-out is the cut-out wind speed of the wind turbine. The model is simulated for n m-seconds, and 10 data points are collected per second. n is generally greater than 100. In the experiment of this method, the value is 500. This value can be freely controlled. The larger the value, the more data, and at the same time, the heavier the calculation burden; each simulation contains x% of the stable operation time and 1 - x% of the total fault occurrence time (0 < x < 1). The 1 - x% of the fault occurrence time is randomly divided into y parts and randomly inserted into the stable operation time; in the experiment of this method, the total duration is 2 seconds. Therefore, there is a total of 1.6 seconds of stable operation time and 0.4 seconds of total fault occurrence time, and a total of 8 faults occur, and the duration of each fault is random. The generated faults are expressed according to the target to be simulated. For example, for one of the problems to be solved when the method is actually used, the offset fault, it can be expressed as the above formula (1), x h [t] and x f [t] respectively represent the healthy and fault observation values at time t.

[0102] The parameters representing the behavior of the wind turbine here are also adjusted according to the target to be detected. For example, in the research on the control system fault, if the pitch angle sensor fault is judged, the wind speed, wind direction, pitch angle of the blade, power, generator reference torque, and wind force at the top of the nacelle can be selected.

[0103] Generate a signal data set containing healthy and fault states N represents the number of samples, a represents the parameters representing the behavior of the wind turbine included in each sample, and is exported.

[0104] S408, Data preprocessing. It includes:

[0105] S4082, Data scaling. In this step, the signal is scaled, that is, the determined scaling function is applied to the data set For each signal in , this method uses a scaling function The original linear scaling formula is optimized to make the data concentrated in the interval [0.5, 1.5] to further enhance the learning effect of the neural network in the subsequent steps. Specifically, as shown in Formula 2, and are the minimum and maximum values ​​of the signal respectively.

[0106] S4084, data segmentation. The signal scaled in S4082 is data enhanced by overlapping segmentation, and the segmented signal segments are saved in a matrix form.

[0107] S4086, constructing a probability vector, i.e., a data set annotation. This includes the following steps:

[0108] 1) Define the state category. According to the determined fault type to be detected, define k states. The calculation formula of k is as follows:

[0109] k=pxq+1 (5)

[0110] Where p is the number of wind turbines simulated by the model, q is the number of fault types detected, and 1 represents “all units are healthy”.

[0111] 2) Make a list of status categories. Create a list of length k, sort the previously determined fault categories in the list, and put the status "all units are healthy" first (that is, the 0th value in the list).

[0112] 3) Label the samples. For each sample, count the number of healthy data points (data points defined in STEP 2, 10 per second) n h and the number of fault observations n f According to the ratio of healthy and faulty observations, a probability vector is constructed Each element It represents the probability that the system is in this state. The logic is shown in Formula 4. i represents the i-th group of data in the data set x, j is the j-th data point, and k has the same meaning as above.

[0113]

[0114] S410, model building. For the DNN network for multiple wind turbine faults, the network input is the signal data set and signal data state annotation obtained in the above steps. The DNN network here is the composite fault identification network (also called diagnostic model) provided in the embodiment of the present application, such as Figure 3 shown.

[0115] S412, model training. The data set is divided, and the DNN network (i.e., the composite fault identification network (also referred to as the diagnosis model)) designed in S410 is used to perform a supervised learning training process.

[0116] S414, use the diagnosis model to diagnose the real-time data (ie, use the composite fault identification model to identify the data for faults). Use the training data obtained in the above manner to start training until the training convergence condition is reached.

[0117] S416, result output: Output the probability of various types of failures of each wind turbine under the current data of the wind turbine.

[0118] S418, end.

[0119] Through the embodiments of the present application, the following technical effects can be achieved:

[0120] 1) Through data-driven approach, non-critical sensor failures are linked to wind turbine operating parameters, providing support for subsequent research on related failures and supporting the development of refined operation and maintenance.

[0121] 2) The important unit production parameters of the pitch angle sensor and the generator speed sensor supported the subsequent research of these two faults.

[0122] 3) The linear scaling formula is used to improve the feature learning effect of the DNN model.

[0123] 4) By integrating multiple interacting wind turbine control systems, a method for fault detection of multiple wind turbines at the same time is proposed, which can more comprehensively evaluate the overall performance of the wind farm and perform more effective fault diagnosis and fault-tolerant control.

[0124] 5) It has universal applicability in the same field, and the method can be transplanted to wind farms of other types and in other regions by replacing the modeling of wind turbines.

[0125] 6) It has cross-domain versatility. By replacing the wind turbine modeling with other equipment modeling and performing target fault simulation, and retraining the deep learning network, it can be applied to equipment fault detection in other industrial scenarios.

[0126] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0127] According to another aspect of the embodiments of the present application, a device for identifying a device failure for implementing the device failure identification method is also provided. Figure 5 As shown, the device comprises:

[0128] A first acquisition unit 502 is used to acquire device parameters corresponding to each of a plurality of energy devices in a target energy scenario, wherein the target energy scenario is used to collect energy through a plurality of energy devices;

[0129] An identification unit 504 is used to identify faults of device parameters corresponding to each of the plurality of energy devices by using a composite fault identification model, wherein the composite fault identification model is obtained by training using training data obtained in a simulated energy scenario, wherein the simulated energy scenario is configured with a plurality of simulated energy devices, and the training data includes a simulation signal data set generated under N types of device faults simulated by the plurality of simulated energy devices, and a state probability vector corresponding to the operating state of the simulated device observed based on the simulation signal data set, where N is a positive integer;

[0130] The second acquisition unit 506 is used to acquire the fault identification confidences output by the composite fault identification model and matching the multiple energy devices, wherein the fault identification confidences are used to indicate the probabilities of N types of device faults occurring in the multiple energy devices.

[0131] Optionally, the embodiments in this solution may refer to but not be limited to the above method embodiments, which will not be described in detail here.

[0132] As an alternative, Figure 6 As shown, the above device also includes:

[0133] A construction unit 602 is used to construct a simulated energy scene based on preset scene parameters before obtaining device parameters corresponding to each of the multiple energy devices in the target energy scene, wherein the preset scene parameters include environment layout parameters of the simulated energy scene and device parameters of each of the multiple simulated energy devices included in the simulated energy scene;

[0134] A first determination unit 604 is used to determine an energy reference interval for running a model in a simulated energy scenario, and a fault simulation control parameter, wherein the fault simulation control parameter is a control parameter for controlling a plurality of simulated energy devices to switch to a healthy state or a faulty state during a process of simulating an i-th device fault type in the simulated energy scenario, and i is a natural number greater than or equal to 1 and less than or equal to N;

[0135] The third acquisition unit 606 is used to acquire a simulation signal data set generated in the process of simulating N types of equipment failures, wherein the simulation signal data set includes: equipment parameters of multiple simulated energy devices in a healthy state and equipment parameters of multiple simulated energy devices in a faulty state in the process of simulating each type of equipment failure;

[0136] The fourth acquisition unit 606 is used to acquire the health observation values ​​observed by each of the multiple simulated energy devices in a healthy state and the fault observation values ​​observed by the multiple simulated energy devices in a fault state during the operation process of simulating each device fault type;

[0137] A second determination unit 608 is used to determine a simulation state probability vector corresponding to the operation state of the simulation device observed by the plurality of simulation energy devices based on the comparison result between the health observation value and the fault observation value;

[0138] The training unit 610 is used to train the neural network model in the initial state using the simulation signal data set and the simulation state probability vector as training data until a composite fault recognition model that meets the training convergence condition is obtained.

[0139] Optionally, the embodiments in this solution may refer to but not be limited to the above method embodiments, which will not be described in detail here.

[0140] As an optional solution, the third acquisition unit includes:

[0141] A first determination module is used to determine, according to the fault simulation control parameter, a first time ratio of multiple simulated energy devices in a healthy state and a second time ratio of multiple simulated energy devices in a fault state during a simulation of an i-th device fault type, where i is a natural number greater than or equal to 1 and less than or equal to N;

[0142] A first insertion module is used to insert the faulty operation time determined based on the second time ratio into the healthy operation time determined by the first time ratio according to a preset condition;

[0143] An adjustment module, used to control the multiple simulated energy devices to operate according to healthy device parameters during the healthy operation time, and to adjust the device parameters of the multiple simulated energy devices to the i-th group of fault parameters matching the i-th device fault type during the fault operation time for operation;

[0144] A first acquisition module is used to acquire the behavior performance parameters of multiple simulated energy devices during the operation process of simulating the i-th device failure type as the i-th simulation signal;

[0145] The generation module is used to generate a simulation signal data set using the acquired N types of simulation signals.

[0146] Optionally, the embodiments in this solution may refer to but not be limited to the above method embodiments, which will not be described in detail here.

[0147] As an optional solution, it also includes:

[0148] A scaling module is used to scale the simulation signal data set after acquiring the simulation signal data set generated in the process of simulating N types of equipment failures to obtain a scaled simulation signal data set;

[0149] The processing module is used to perform data enhancement processing and segmentation processing on the scaled simulation signal data set to obtain signal segments, and save the signal segments in a matrix form.

[0150] Optionally, the embodiments in this solution may refer to but not be limited to the above method embodiments, which will not be described in detail here.

[0151] As an optional solution, the second determining unit includes:

[0152] A second determination module is used to determine K types of simulated device operating states based on the number of multiple simulated energy devices and the number N of N types of device failure types;

[0153] A creation module is used to create a state list of length K using the comparison result between the healthy observation value and the fault observation value, wherein the comparison result is used to indicate the ratio between the healthy observation value and the fault observation value corresponding to the jth simulated energy device under the i-th device fault type, j is a natural number greater than or equal to 1 and less than or equal to M, and M is the number of multiple simulated energy devices configured in the simulated energy scenario;

[0154] A building block for building a simulation state probability vector based on a state list.

[0155] Optionally, the embodiments in this solution may refer to but not be limited to the above method embodiments, which will not be described in detail here.

[0156] As an optional solution, the first acquisition unit includes: a second acquisition module, used to acquire wind turbine operating parameters corresponding to each of the plurality of wind turbines in the target wind farm;

[0157] The identification unit includes: an identification module, which is used to perform fault identification on the wind turbine operating parameters corresponding to each of the multiple wind turbines through a composite fault identification model, wherein the training data used by the composite fault identification model includes a simulated wind turbine signal data set and a simulated wind turbine state probability vector generated under at least two sensor-type equipment failure types simulated by multiple simulated wind turbines in a constructed simulated wind farm.

[0158] Optionally, the embodiments in this solution may refer to but not be limited to the above method embodiments, which will not be described in detail here.

[0159] According to one aspect of the present application, a computer program product is provided. The computer program product includes a computer program / instruction. The computer program / instruction includes a program code for executing the method shown in the above flowchart.

[0160] Figure 7 The structure block diagram of a computer system for implementing an electronic device according to an embodiment of the present application is schematically shown.

[0161] It should be noted that Figure 7 The computer system 700 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0162] like Figure 7 As shown, the computer system includes a central processing unit 701 (CPU), which can perform various appropriate actions and processes according to the program stored in the read-only memory 702 (ROM) or the program loaded from the storage part 708 to the random access memory 703 (RAM). Various programs and data required for system operation are also stored in the random access memory 703. The central processing unit 701, the read-only memory 702 and the random access memory 703 are connected to each other through a bus 704. An input / output interface 705 (Input / Output interface, i.e., I / O interface) is also connected to the bus 704.

[0163] The following components are connected to the input / output interface 705: an input section 706 including a keyboard, a mouse, etc.; an output section 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN card, a modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the input / output interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed so that a computer program read therefrom is installed into the storage section 708 as needed.

[0164] In particular, according to an embodiment of the present application, the process described in each method flow chart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer readable medium, and the computer program contains a program code for executing the method shown in the flow chart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 709, and / or installed from the removable medium 711. When the computer program is executed by the central processor 701, various functions defined in the system of the present application are executed.

[0165] According to one aspect of the present application, a computer-readable storage medium is provided, and a processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the methods provided in various optional implementations of the above-mentioned device failure identification.

[0166] Optionally, in this embodiment, the computer-readable storage medium may be configured to store a computer program for performing the following steps:

[0167] S1, obtaining device parameters corresponding to each of a plurality of energy devices in a target energy scenario, wherein the target energy scenario is used to collect energy through the plurality of energy devices;

[0168] S2, performing fault identification on the equipment parameters corresponding to each of the multiple energy devices through a composite fault identification model, wherein the composite fault identification model is trained using training data obtained in a simulated energy scenario, wherein the simulated energy scenario is configured with multiple simulated energy devices, and the training data includes a simulation signal data set generated based on N types of equipment fault types simulated by the multiple simulated energy devices, and a simulation state probability vector corresponding to the operating state of the simulated device observed based on the simulation signal data set, where N is a positive integer;

[0169] S3, obtaining the fault identification confidence output by the composite fault identification model and matching the multiple energy devices, wherein the fault identification confidence is used to indicate the probability of the N types of device faults occurring in the multiple energy devices.

[0170] Optionally, in the embodiments of the present application, the term "module" or "unit" refers to a computer program or a part of a computer program with a predetermined function, and works together with other related parts to achieve a predetermined goal, and can be implemented in whole or in part by using software, hardware (such as processing circuits or memories), or a combination thereof. Similarly, a processor (or multiple processors or memories) can be used to implement one or more modules or units. In addition, each module or unit can be part of an overall module or unit that includes the function of the module or unit.

[0171] Optionally, in this embodiment, a person of ordinary skill in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing hardware related to the terminal device through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0172] If the integrated units in the above embodiments are implemented in the form of software functional units and sold or used as independent products, they can be stored in the above computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling one or more computer devices (which may be personal computers, servers, or network devices, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application.

[0173] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0174] In the several embodiments provided in the present application, it should be understood that the disclosed client can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0175] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0176] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0177] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for identifying equipment failure, characterized in that: include: Obtaining device parameters corresponding to each of a plurality of energy devices in a target energy scenario, wherein the target energy scenario is used to collect energy through the plurality of energy devices; Fault identification is performed on device parameters corresponding to each of the multiple energy devices through a composite fault identification model, wherein the composite fault identification model is trained using training data obtained in a simulated energy scenario, wherein the simulated energy scenario is configured with multiple simulated energy devices, and the training data includes a simulation signal data set generated based on N types of device fault types simulated by the multiple simulated energy devices, and a simulation state probability vector corresponding to the operating state of the simulated device observed based on the simulation signal data set, where N is a positive integer; Obtaining the fault identification confidences output by the composite fault identification model and matching the multiple energy devices, wherein the fault identification confidences are used to indicate the probabilities of the N types of device faults occurring in the multiple energy devices.

2. The method according to claim 1, characterized in that Before obtaining the device parameters corresponding to the plurality of energy devices in the target energy scenario, the method further includes: Constructing the simulated energy scenario based on preset scenario parameters, wherein the preset scenario parameters include environment layout parameters of the simulated energy scenario and device parameters of each of the plurality of simulated energy devices included in the simulated energy scenario; Determine an energy reference interval for running the model in the simulated energy scenario, and a fault simulation control parameter, wherein the fault simulation control parameter is a control parameter for controlling the plurality of simulated energy devices to switch to a healthy state or a faulty state during the process of simulating the i-th device fault type in the simulated energy scenario, and i is a natural number greater than or equal to 1 and less than or equal to N; Acquire the simulation signal data set generated in the process of simulating N types of equipment failures, wherein the simulation signal data set includes: in the process of simulating each type of equipment failure, equipment parameters of the plurality of simulated energy devices in a healthy state and equipment parameters of the plurality of simulated energy devices in a faulty state; In the operation process of simulating each device fault type, obtaining health observation values ​​observed when each of the plurality of simulated energy devices is in a healthy state, and fault observation values ​​observed when the plurality of simulated energy devices is in a faulty state; Based on the comparison result between the health observation value and the fault observation value, determining the simulation state probability vector corresponding to the operation state of the simulation device observed by the plurality of simulation energy devices; The simulation signal data set and the simulation state probability vector are used as training data to train the neural network model in an initial state until the composite fault identification model that meets the training convergence condition is obtained.

3. The method according to claim 2, characterized in that The obtaining of the simulation signal data set generated in the process of simulating N types of equipment failures includes: Determine, according to the fault simulation control parameter, a first time ratio of the plurality of simulated energy devices being in a healthy state and a second time ratio of the plurality of simulated energy devices being in a faulty state during the process of simulating the i-th device fault type, where i is a natural number greater than or equal to 1 and less than or equal to N; Inserting the faulty operation time determined based on the second time ratio into the healthy operation time determined by the first time ratio according to a preset condition; Controlling the plurality of simulated energy devices to operate according to healthy device parameters during the healthy operation time, and adjusting the device parameters of the plurality of simulated energy devices to operate according to an i-th group of fault parameters matching the i-th device fault type during the fault operation time; Obtaining behavior performance parameters of the plurality of simulated energy devices during operation of simulating the i-th device failure type as the i-th simulation signal; The simulation signal data set is generated using the acquired N types of simulation signals.

4. The method according to claim 2, characterized in that: After acquiring the simulation signal data set generated in the process of simulating N types of equipment failures, the method further includes: Scaling the simulation signal data set to obtain a scaled simulation signal data set; The scaled simulation signal data set is subjected to data enhancement processing and segmentation processing to obtain signal segments, and the signal segments are stored in a matrix form.

5. The method according to claim 2, characterized in that: The step of determining the simulation state probability vector corresponding to the operation state of the simulation device observed by the plurality of simulation energy devices based on the comparison result between the health observation value and the fault observation value comprises: Determining K types of simulated device operating states based on the number of the plurality of simulated energy devices and the number N of the N types of device failure types; Using the comparison result between the health observation value and the fault observation value, create a state list with a length of K, wherein the comparison result is used to indicate the ratio between the health observation value and the fault observation value corresponding to the jth simulated energy device under the i-th device fault type, j is a natural number greater than or equal to 1 and less than or equal to M, and M is the number of multiple simulated energy devices configured in the simulated energy scenario; The simulation state probability vector is constructed based on the state list.

6. The method according to any one of claims 1 to 5, characterized in that The obtaining of equipment parameters corresponding to each of the plurality of energy devices in the target energy scenario comprises: obtaining wind turbine operating parameters corresponding to each of the plurality of wind turbines in the target wind farm; The method of performing fault identification on the equipment parameters corresponding to each of the multiple energy devices through a composite fault identification model includes: performing fault identification on the wind turbine operating parameters corresponding to each of the multiple wind turbines through a composite fault identification model, wherein the training data used by the composite fault identification model includes a simulated wind turbine signal data set and a simulated wind turbine state probability vector generated under at least two types of sensor equipment failures simulated based on multiple simulated wind turbines in the constructed simulated wind farm.

7. A device for identifying equipment failure, characterized in that: include: A first acquisition unit is used to acquire device parameters corresponding to each of a plurality of energy devices in a target energy scenario, wherein the target energy scenario is used to collect energy through the plurality of energy devices; An identification unit, configured to perform fault identification on device parameters corresponding to each of the plurality of energy devices through a composite fault identification model, wherein the composite fault identification model is obtained by training using training data acquired in a simulated energy scenario, wherein the simulated energy scenario is configured with a plurality of simulated energy devices, and the training data includes a simulation signal data set generated based on N types of device faults simulated by the plurality of simulated energy devices, and a state probability vector corresponding to an operating state of the simulated device observed based on the simulation signal data set, where N is a positive integer; The second acquisition unit is used to acquire the fault identification confidence output by the composite fault identification model and matching the multiple energy devices, wherein the fault identification confidence is used to indicate the probability of the N types of device faults occurring in the multiple energy devices.

8. The device according to claim 7, characterized in that Also includes: A construction unit, configured to construct the simulated energy scenario based on preset scenario parameters before obtaining the device parameters corresponding to each of the multiple energy devices in the target energy scenario, wherein the preset scenario parameters include the environment layout parameters of the simulated energy scenario and the device parameters of each of the multiple simulated energy devices included in the simulated energy scenario; A first determination unit is used to determine an energy reference interval for model operation in the simulated energy scenario, and a fault simulation control parameter, wherein the fault simulation control parameter is a control parameter used to control the plurality of simulated energy devices to switch to a healthy state or a faulty state during the process of simulating the i-th device fault type in the simulated energy scenario, and i is a natural number greater than or equal to 1 and less than or equal to N; A third acquisition unit is used to acquire the simulation signal data set generated in the process of simulating N types of equipment failures, wherein the simulation signal data set includes: in the process of simulating each type of equipment failure, the equipment parameters of the plurality of simulated energy devices in a healthy state and the equipment parameters of the plurality of simulated energy devices in a faulty state; a fourth acquisition unit, configured to acquire, during the operation process of simulating each device fault type, health observation values ​​observed when each of the plurality of simulated energy devices is in a healthy state, and fault observation values ​​observed when the plurality of simulated energy devices is in a faulty state; A second determination unit is used to determine the simulation state probability vector corresponding to the operation state of the simulation device observed by the multiple simulation energy devices based on the comparison result between the health observation value and the fault observation value; The training unit is used to use the simulation signal data set and the simulation state probability vector as training data to train the neural network model in an initial state until the composite fault identification model that meets the training convergence condition is obtained.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the program is executed by a processor to perform the method described in any one of claims 1 to 7.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images