Power generation equipment fault diagnosis method and system

Abstract

The invention provides a power generation equipment fault diagnosis method and system, and the method comprises the steps: judging the type of current power generation equipment based on the MAC address of the power generation equipment, and pre-configuring the segmentation ordinal number of the type of the power generation equipment, and enabling different types of the power generation equipment to be different, and enabling the corresponding segmentation ordinal number ranges to be different; constructing a key value pair list of the power generation equipment types and the electric power parameters, obtaining the corresponding power generation equipment types and the corresponding electric power parameters, performing data processing on the electric power parameters, and performing uniform standard feature conversion to obtain electric power feature parameters; filling a corresponding key value pair list with the electric power characteristic parameters, constructing a unified standard electric power characteristic parameter key value pair list, and generating sample sets of different types of power generation equipment by taking the key value pair list as a random search list; and inputting the sample set into a deep neural network classification model for training, and outputting corresponding fault type judgment and a power generation equipment type to which the fault type belongs after training.

Description

Technical Field

[0001] The present invention relates to the technical field of power generation, and in particular to a method and system for diagnosing faults of power generation equipment. Background Art

[0002] At present, the types of power generation equipment are not limited to traditional power generation equipment such as thermal power generation and hydropower generation. Existing new power generation equipment also includes but is not limited to clean energy power generation such as photovoltaic power generation, wind power generation and tidal power generation. However, the components of the above-mentioned different types of power generation equipment are different, and there are certain differences in the power generation effects. However, in the existing mains power network, the above-mentioned power generation equipment are generally connected to the mains power network, so it is difficult to detect the fault types of different power generation equipment. That is to say, the fault detection of power generation equipment in the prior art is a single type of power generation fault detection, which requires a separate device or system to collect and analyze data and obtain a single type of power generation equipment fault detection. Summary of the invention

[0003] One of the invention purposes of the present invention is to provide a method and system for diagnosing faults of power generation equipment, wherein the method and system construct a method for diagnosing faults of multiple types of power generation equipment, wherein the present invention uses segmented ordinal numbers to construct sample numbers of different types of power generation equipment fault types, wherein the sample number is used as one of the characteristic parameters of the sample, and a sample element set of the power characteristic parameters of the corresponding type of power generation equipment is constructed for each sample number, and the sample element set is input into a deep learning classification model for training to obtain the type of power generation equipment fault, and the type of power generation equipment is judged based on the segmented value of the random number of the sample element set. Thus, rapid identification of fault types of multiple types of power generation equipment can be achieved.

[0004] Another inventive object of the present invention is to provide a method and system for diagnosing faults of power generation equipment. The method and system automatically fill in the power characteristic-related sensor data of the corresponding type of power generation equipment and the corresponding equipment type into the corresponding key name and key value in the form of key-value pairs, construct a mapping relationship table including the corresponding key name and key value, and construct sample elements of the power characteristic parameters of the corresponding type of power generation equipment based on the mapping relationship table.

[0005] Another inventive object of the present invention is to provide a method and system for diagnosing faults of power generation equipment, which method and system will perform packet data preprocessing on a sample element set based on the segmented ordinal number as the sample number, and perform data standardization processing on the power characteristic parameters of different types of power generation equipment in the reported sample elements according to their own type data, wherein the standardization processing includes numerical normalization processing, which is beneficial to the deep learning classification model's analysis capability of the power characteristic parameters of different types of power generation equipment, thereby improving the judgment accuracy of the deep learning classification model on the power characteristic parameter faults of different types of power generation equipment.

[0006] In order to achieve at least one of the above-mentioned invention objects, the present invention further provides a method for diagnosing faults of power generation equipment, the method comprising: Determine the type of the current power generation device based on the MAC address of the power generation device, and pre-configure the segment sequence number of the type of the power generation device, wherein different types of power generation devices have different corresponding segment sequence number ranges; Constructing a key-value pair list of the power generation equipment type and the power parameter, obtaining the corresponding power generation equipment type and the corresponding power parameter, performing data processing on the power parameter and performing a feature conversion of a unified standard to obtain the power feature parameter; Fill the power characteristic parameters into the corresponding key-value pair list in sequence according to the segmented ordinal numbers, construct a unified standard power characteristic parameter key-value pair list, and generate a sample set of different types of power generation equipment based on the key-value pair list as a random search list; The sample set is input into a deep neural network classification model for training, and after training, is used to output the corresponding fault type judgment and the type of power generation equipment to which the fault type belongs.

[0007] According to one of the preferred embodiments of the present invention, the method for constructing the key-value pair list of the power generation equipment type and the power parameters includes: randomly generating a first random number in the corresponding segmented ordinal number as the key name of the key-value pair list according to the preset power generation equipment type and the segmented ordinal number, and obtaining the power parameter of the power generation equipment type that matches the first random number, and filling the power parameter that matches the power generation equipment type according to the power parameter obtained at the same detection time into the key-value pair list as the key value, for constructing the key-value pair list of the power generation equipment type and the power parameters.

[0008] According to another preferred embodiment of the present invention, the characteristic conversion method of the unified standard of the power parameters includes: obtaining the power parameters of the corresponding power generation equipment type, and normalizing the power parameters of all power generation equipment types according to the same parameter type, and using linear interpolation to supplement the non-existent power parameter types of the power generation equipment, wherein the normalization method includes: obtaining the maximum value f of the corresponding power parameter type max，n and the minimum value f min,n , where the subscript n represents the power parameter type, max represents the maximum value, min represents the minimum value, and f represents the power parameter value. The normalized power parameter F n = , the normalized power parameter F n as the corresponding key value in the key-value pair list.

[0009] According to another preferred embodiment of the present invention, the power characteristic parameter further includes: after obtaining the corresponding power parameter according to the corresponding power generation equipment type and obtaining the corresponding first random number m, the first random number m is subjected to characteristic conversion according to the following formula: n = lgm, and round off the lgm to obtain the level value of the corresponding power parameter, and convert the level value of the power parameter F n =lgm is input as the power characteristic parameter into the key value of the corresponding key-value pair list.

[0010] According to another preferred embodiment of the present invention, the power characteristic parameters also include: after the power parameters of the corresponding type of power generation equipment are obtained at the same detection time, it also includes a fault status value of the corresponding type of power generation equipment based on the same detection time, wherein the fault status value includes faults of different levels and types, wherein the fault status value of the corresponding type of power generation equipment is obtained by assigning and processing according to a preset assignment rule, and the fault status value of the corresponding type of power generation equipment is stored in the key value of the corresponding key-value pair list.

[0011] According to another preferred embodiment of the present invention, the training method of the deep neural network classification model includes: randomly generating a total of x second random numbers from different ranges in sequence according to the range of the segmented ordinal number, wherein the number of values ​​of the second random numbers in different ranges is x respectively. j , j is the corresponding power generation equipment type, and x= , where the number of values ​​of the second random number in different ranges is x j The corresponding second random number is defined as P j , then according to the second random number P j Used as a key name for key value lookup.

[0012] According to another preferred embodiment of the present invention, according to the second random number P j Input into the key-value pair list of power generation equipment type and power parameter for key name matching. If the second random number P j If there is an identical key name in the key-value pair list, the corresponding key value under the identical key name is searched, and it is determined whether the key value is a null value. If the key value is a null value, a new second random number P is generated again within the corresponding range. j , until the current second random number P j Find x in the corresponding range j Key-value data.

[0013] According to another preferred embodiment of the present invention, obtain x within each segment sequence range. jA key-value data set is taken as a sample set, and the sample set is randomly divided into a training set and a test set, the initialization network layer weights and bias of the DNN network are pre-configured, and the initialization learning rate is set, the DNN is trained through the training set and the network layer weights, bias and learning rate are adjusted, the training effect of the DNN model is judged according to the convergence of the loss function, and the training effect of the DNN model is verified through the test set.

[0014] In order to achieve at least one of the above-mentioned invention purposes, the present invention further provides a power generation equipment fault diagnosis system, wherein the system executes the above-mentioned power generation equipment fault diagnosis method.

[0015] The present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the above-mentioned method for diagnosing faults of power generation equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 Shown is a schematic flow chart of a method for diagnosing faults in power generation equipment according to the present invention. DETAILED DESCRIPTION

[0017] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art can think of other obvious variations. The basic principles of the present invention defined in the following description can be applied to other embodiments, variations, improvements, equivalents, and other technical solutions that do not deviate from the spirit and scope of the present invention.

[0018] It is to be understood that the term "one" should be understood as "at least one" or "one or more", that is, in one embodiment, the number of an element may be one, while in another embodiment, the number of the element may be multiple, and the term "one" should not be understood as a limitation on the quantity.

[0019] Please combine Figure 1The present invention discloses a method and system for diagnosing faults of power generation equipment, the method comprising the following steps: first, the MAC address of the power generation equipment needs to be obtained, wherein the MAC address is used to provide the type of the power generation equipment. It should be noted that the MAC address is the unique identity tag of the Internet of Things device, which is used to identify the type of the Internet of Things device. After obtaining the MAC address, a segmented ordinal number is further configured, wherein the segmented ordinal number, wherein the segmented ordinal number can be set to a segmented range of different levels, for example, 1-99 can be selected as the first segmented range, and 100-999 can be set as the segmented range of the second level, and 1000-9999 can be set as the segmented range of the third level. In some other preferred embodiments of the present invention, data within a specific range can be set as the segmented range, which will not be described in detail in the present invention. The present invention obtains random numbers within different segmented ranges as identifiers of power parameters, wherein the random numbers can be based on a true random number generator of physical noise and a pseudo-random number generator generated based on a random number algorithm, and the random numbers that meet the segmented interval are intercepted to obtain. After completing the segmented ordinal operation, the ordinal within the corresponding segment range is further bound to the corresponding type of power generation equipment to configure the random number identification of the corresponding type of power generation equipment. Therefore, the random number generated in each segmented ordinal in the present invention is used to identify the power parameter of the corresponding type of the segmented ordinal. Therefore, the present invention can distinguish the power parameters of different types of power generation equipment at the mathematical level, which is conducive to data identification, classification and processing. The present invention constructs a key-value pair list based on the segmented ordinal, wherein the key-value pair list represents the mapping relationship between the power parameters detected by the corresponding type of power generation equipment and the corresponding power generation equipment. It should be noted that, since different power generation equipment has different corresponding power parameters under different detection times, and the corresponding detection results under different detection times are different, the present invention can effectively and automatically generate a sample set of association relationships by borrowing the relevant random numbers generated by the segmented ordinal, which is not limited to the way of sample collection and construction. The present invention trains a deep neural network classification model based on the sample set to discover the relationship between power parameter changes and corresponding power generation type faults. Thereby improving the accuracy of the present invention in predicting faults of different types of power generation equipment.

[0020] Specifically, in the present invention, it is necessary to construct a key-value pair list of power generation equipment type and power parameters, wherein the method for constructing the key-value pair list includes: according to the preset power generation equipment type and segmentation sequence number, randomly generate a first random number in the corresponding segmentation sequence number as the key name of the key-value pair list, and obtain the power parameter of the power generation equipment type matched by the first random number, and fill the power parameter of the power generation equipment type matched according to the same detection time into the key-value pair list as the key value, which is used to construct the key-value pair list of the power generation equipment type and power parameter. The power parameters may include but are not limited to voltage, current, power factor, and current imbalance calculated based on the voltage and current, voltage fluctuation and flicker, capture and record transient waveforms, voltage swells and other data; due to different types of power generation equipment, the specific single values ​​and combined values ​​of the above power parameters are different. The present invention needs to configure the first random number of different segmentation sequences as the detection sample identifier according to the detection time or detection number of the corresponding power generation equipment type.

[0021] Furthermore, the present invention needs to perform standardization on the same power parameters. Although the power parameter values ​​of different power generation equipment are different, the power parameters of the same type still conform to the electrical physical characteristics. Therefore, the present invention performs standardization on the power parameters of all different types of power generation equipment based on the power parameters of the same type. The specific method includes: obtaining the power parameters of the corresponding power generation equipment type, and normalizing the power parameters of all types of power generation equipment according to the same parameter type, and using linear interpolation to supplement the power parameter types of non-existent power generation equipment, wherein the normalization method includes: obtaining the maximum value f of the corresponding power parameter type max，n and the minimum value f min,n , where the subscript n represents the power parameter type, max represents the maximum value, min represents the minimum value, and f represents the power parameter value. The normalized power parameter F n = , the normalized power parameter F n As the corresponding key value in the key-value pair list, the power parameter characteristics of the same power parameter type related to the physical characteristics of different types of power generation equipment can be standardized.

[0022] In one of the preferred embodiments of the present invention, in order to provide the classification parameters of the power characteristic parameters, the power characteristic parameters of the present invention further include: after obtaining the corresponding first random number m according to the corresponding power generation equipment type, the first random number m is converted into a characteristic according to the following formula: n = lgm, and round off the lgm to obtain the level value of the corresponding power parameter, and convert the level value of the power parameter F n=lgm is input as the power characteristic parameter into the key value of the corresponding key-value pair list. The level value F of the above power parameter n =lgm indicates the standardization to which the current sample data belongs.

[0023] The power characteristic parameters also include: after the power parameters of the corresponding type of power generation equipment are obtained at the same detection time, it also includes a fault status value of the corresponding type of power generation equipment based on the same detection time, wherein the fault status value includes faults of different levels and types, wherein the fault status value of the corresponding type of power generation equipment is obtained by assigning and processing according to a preset assignment rule, and the fault status value of the corresponding type of power generation equipment is stored in the key value of the corresponding key-value pair list.

[0024] The training method of the deep neural network classification model comprises: randomly generating a total of x second random numbers from different ranges in sequence according to the range of the segmented ordinal number, wherein the number of values ​​of the second random numbers in different ordinal number ranges is respectively x j , j is the corresponding power generation equipment type, and x= , where the number of values ​​of the second random number in different ranges is x j The corresponding second random number is defined as P j , then according to the second random number P j Used as a key name to look up a value.

[0025] According to the second random number P j Input into the key-value pair list of power generation equipment type and power parameter for key name matching. If the second random number P j If there is an identical key name in the key-value pair list, the corresponding key value under the identical key name is searched, and it is determined whether the key value is a null value. If the key value is a null value, a new second random number P is generated again within the corresponding range. j , until the current second random number P j Find x in the corresponding range j Key-value data.

[0026] Get the x in each segmented ordinal range j A key-value data set is taken as a sample set, and the sample set is randomly divided into a training set and a test set, the initialization network layer weights and bias of the DNN network are pre-configured, and the initialization learning rate is set, the DNN is trained through the training set and the network layer weights, bias and learning rate are adjusted, the training effect of the DNN model is judged according to the convergence of the loss function, and the training effect of the DNN model is verified through the test set.

[0027] The embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. The embodiments disclosed in the present invention include 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 flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part, and / or installed from a removable medium. When the computer program is executed by the central processing unit (CPU), the above-mentioned functions defined in the method of the present application are executed. It should be noted that the above-mentioned computer-readable medium of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, but not limited to, a system, device or device of an electrical, magnetic, optical, electromagnetic, infrared segment, or semiconductor, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection with one or more wire segments, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries a computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code embodied on the computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, electrical wire, optical cable, RF, etc., or any suitable combination of the foregoing.

[0028] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0029] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The objectives of the present invention have been fully and effectively achieved. The functional and structural principles of the present invention have been demonstrated and explained in the embodiments. Without departing from the principles, the implementation methods of the present invention may be deformed or modified in any way.

Claims

1. A method for diagnosing faults of power generation equipment, It is characterized in that The method comprises: Determine the type of the current power generation equipment based on the MAC address of the power generation equipment, and pre-configure the segment sequence number of the type of the power generation equipment, wherein different types of power generation equipment have different corresponding segment sequence number ranges; Constructing a key-value pair list of the power generation equipment type and the power parameter, obtaining the corresponding power generation equipment type and the corresponding power parameter, performing data processing on the power parameter and performing a feature conversion of a unified standard to obtain the power feature parameter; Filling the power characteristic parameters into the corresponding key-value pair list according to the segmented ordinal number, constructing a unified standard power characteristic parameter key-value pair list, and generating a sample set of different types of power generation equipment based on the key-value pair list as a random search list; The sample set is input into a deep neural network classification model for training, and after training, is used to output the corresponding fault type judgment and the type of power generation equipment to which the fault type belongs.

2. A method for diagnosing faults of power generation equipment according to claim 1, It is characterized in that The method for constructing the key-value pair list of the power generation equipment type and the power parameters includes: randomly generating a first random number in the corresponding segmented ordinal number as the key name of the key-value pair list according to the preset power generation equipment type and segmented ordinal number, and obtaining the power parameter of the power generation equipment type that matches the first random number, and filling the power parameter that matches the power generation equipment type according to the power parameter obtained at the same detection time into the key-value pair list as the key value, so as to construct the key-value pair list of the power generation equipment type and the power parameters.

3. A method for diagnosing faults of power generation equipment according to claim 1, It is characterized in that The characteristic conversion method of the unified standard of the power parameters includes: obtaining the power parameters of the corresponding power generation equipment type, and normalizing the power parameters of all power generation equipment types according to the same parameter type, and using linear interpolation to supplement the non-existent power parameter types of the power generation equipment, wherein the normalization method includes: obtaining the maximum value f of the corresponding power parameter type max，n and the minimum value f min,n , where the subscript n represents the power parameter type, max represents the maximum value, min represents the minimum value, and f represents the power parameter value. The normalized power parameter F n = , the normalized power parameter F n as the corresponding key value in the key-value pair list.

4. A method for diagnosing faults of power generation equipment according to claim 3, It is characterized in that The power characteristic parameter also includes: after obtaining the corresponding power parameter according to the corresponding power generation equipment type to obtain the corresponding first random number m, the first random number m is converted into a characteristic according to the following formula: n = lgm, and round off the lgm to obtain the level value of the corresponding power parameter, and convert the level value of the power parameter F n =lgm is input as the power characteristic parameter into the key value of the corresponding key-value pair list.

5. A method for diagnosing faults of power generation equipment according to claim 3, It is characterized in that The power characteristic parameters also include: power parameters of the corresponding type of power generation equipment obtained at the same detection time, and also include fault status values ​​of the corresponding type of power generation equipment based on the same detection time, wherein the fault status values ​​include faults of different levels and types, wherein the fault status values ​​of the corresponding type of power generation equipment are obtained by assigning and processing according to preset assignment rules, and the fault status values ​​of the corresponding type of power generation equipment are stored in the key value of the corresponding key-value pair list.

6. A method for diagnosing faults of power generation equipment according to claim 1, It is characterized in that The training method of the deep neural network classification model comprises: randomly generating a total of x second random numbers from different ranges in sequence according to the range of the segmented ordinal number, wherein the number of values ​​of the second random numbers in different ranges is x respectively. j , j is the corresponding power generation equipment type, and x= , where the number of values ​​of the second random number in different ranges is x j The corresponding second random number is defined as P j , then according to the second random number P j Used as a key name to look up a value.

7. A method for diagnosing faults of power generation equipment according to claim 6, It is characterized in that According to the second random number P j Input into the key-value pair list of power generation equipment type and power parameter for key name matching. If the second random number P j If there is an identical key name in the key-value pair list, the corresponding key value under the identical key name is searched, and it is determined whether the key value is a null value. If the key value is a null value, a new second random number P is generated again within the corresponding range. j , until the current second random number P j Find x in the corresponding range j Key-value data.

8. A method for diagnosing faults of power generation equipment according to claim 6, It is characterized in that Get the x in each segmented ordinal range j A key-value data set is taken as a sample set, and the sample set is randomly divided into a training set and a test set, the initialization network layer weights and bias of the DNN network are pre-configured, and the initialization learning rate is set, the DNN is trained through the training set and the network layer weights, bias and learning rate are adjusted, the training effect of the DNN model is judged according to the convergence of the loss function, and the training effect of the DNN model is verified through the test set.

9. A fault diagnosis system for power generation equipment, It is characterized in that The system executes a method for diagnosing faults of power generation equipment as described in any one of claims 1-8.

10. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement a method for diagnosing faults of power generation equipment as described in any one of claims 1 to 8.

Images