Generator set artificial intelligence system machine learning sample screening method and device

Abstract

The application discloses a generator set artificial intelligence system machine learning sample screening method and device, obtains generator set operation data, adopts a preset cleaning method for one or more times of data cleaning on the generator set operation data, obtains a candidate sample set and an invalid sample set, if multiple times of data cleaning are adopted, different cleaning methods are adopted for each time of data cleaning, sorts the candidate sample set by adopting a preset sorting method, obtains a sorted sample set, extracts the sorted sample set according to a preset extraction method, obtains a learning sample subset, repeats the previous step for multiple times, adopts different sorting methods and the same or different extraction methods each time, extracts different learning sample subsets, and combines all the learning sample subsets into a learning sample set. The method is used for screening the generator artificial intelligence system machine learning sample, has low requirements on the professional quality of an operator, has high automation degree in a processing process, and has good extensive distribution of the screened learning sample.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of real-time operation and maintenance and online fault diagnosis of generator sets, and in particular to a machine learning sample screening method and device for an artificial intelligence system of a generator set. BACKGROUND

[0002] Large and medium-sized synchronous generators are one of the core equipment of the power system, which are complex in structure and expensive in price. Once an operation accident or equipment failure occurs, the loss caused is often disastrous. Therefore, it is very important to perform real-time operation and maintenance and online fault diagnosis on the generator.

[0003] The distributed control system (DCS) of a power plant has undergone rapid development for decades and has gradually been perfected, providing a solid foundation for real-time operation and maintenance and online fault diagnosis of the generator. However, the generator is a complex system involving multiple disciplines such as machinery, electromagnetism, materials, and chemical engineering. During operation, various electrical parameters and even non-electrical parameters interact with each other, and the relationship is complex and cannot be accurately modeled using classical physical modeling methods.

[0004] The rise and development of artificial intelligence technology provide another innovative solution for real-time operation and maintenance and online fault diagnosis of the generator. Many previously unachievable functions have been successfully implemented in the artificial intelligence system of the generator. However, the popularity of artificial intelligence technology is also limited by its own limitations. For example, the accuracy of the decision-making results of the machine learning algorithm commonly used in artificial intelligence systems is greatly affected by the correctness and comprehensiveness of the learning samples. Incorrect learning samples can lead to incorrect decision-making results. Invalid learning samples not only lead to decreased learning and decision-making efficiency, but also may lead to decreased decision-making accuracy. Incomplete sample distribution may result in suboptimal decision-making accuracy in certain operating conditions.

[0005] For machine learning algorithms, sample screening is crucial. In general, sample screening needs to be assisted by industry experts, which greatly increases the threshold for system deployment. Human factors also increase the uncertainty of intelligent systems. Manual sample screening is tedious and inefficient, which seriously affects the application and popularization of artificial intelligence technology. SUMMARY

[0006] To solve the above technical problems, the embodiments of the present application propose a machine learning sample screening method and device for an artificial intelligence system of a generator set, which reduces the number of invalid samples in the learning samples and improves the distribution comprehensiveness of the learning samples.

[0007] In a first aspect, the present application provides a method for screening machine learning samples of a generator set artificial intelligence system, comprising the following steps:

[0008] S1, obtaining generator set operation data, performing one or more data cleaning on the generator set operation data by using a preset cleaning method to obtain a candidate sample set and an invalid sample set, if multiple data cleaning is used, different cleaning methods are used for each data cleaning;

[0009] S2, sorting the candidate sample set by using a preset sorting method to obtain a sorted sample set, and extracting the sorted sample set according to a preset extraction method to obtain a learning sample subset;

[0010] S3, repeating step S2 multiple times, using different sorting methods and the same or different extraction methods each time to extract different learning sample subsets;

[0011] S4, combining all learning sample subsets into a learning sample set.

[0012] As a preferred, the cleaning method comprises:

[0013] In response to determining that one or more of the electrical parameters of the generator set is outside the set range, it is classified into the invalid sample set, and the rest is the candidate sample set;

[0014] In response to determining that the sampling time of the generator set operation data is outside the set range, it is classified into the invalid sample set, and the rest is the candidate sample set;

[0015] In response to determining that one or more of the non-electrical parameters of the generator set is outside the set range, it is classified into the invalid sample set, and the rest is the candidate sample set;

[0016] In response to determining that the calculation result according to one or more of the electrical parameters, the sampling time, and the non-electrical parameters is outside the set range, it is classified into the invalid sample set, and the rest is the candidate sample set.

[0017] As a preferred, the sorting method comprises:

[0018] According to the electrical parameters of the generator set;

[0019] According to the sampling time of the generator set operation data;

[0020] According to the non-electrical parameters of the generator set;

[0021] According to the calculation result of one or more of the electrical parameters, the sampling time, and the non-electrical parameters.

[0022] As a preferred, the extraction method comprises one or more of the following:

[0023] random distribution sampling;

[0024] uniform distribution sampling;

[0025] normal or inverse normal sampling;

[0026] triangular or inverse triangular sampling;

[0027] trapezoidal or inverse trapezoidal sampling;

[0028] sine or inverse sine sampling;

[0029] piecewise sampling.

[0030] As preferred, the electrical parameters include active power, reactive power, stator voltage, stator current, rotor current, power factor.

[0031] As preferred, the non-electrical parameters include vibration, noise, pressure, flow, temperature, temperature rise.

[0032] As preferred, the step S1 can be completed in advance in the sample collection stage; if the extraction mode in step S2 is selected as random distribution sampling, the sorting step can be skipped.

[0033] In a second aspect, the present application provides a generator set artificial intelligence system machine learning sample screening device, comprising:

[0034] a cleaning module configured to obtain generator set operation data, perform one or more data cleanings on the generator set operation data using a preset cleaning method, obtain an alternative sample set and an invalid sample set, and if multiple data cleanings are performed, different cleaning methods are used for each data cleaning;

[0035] an ordering and sampling module configured to sort the alternative sample set using a preset sorting method, obtain an ordered sample set, and extract the ordered sample set according to a preset extraction method, to obtain a learning sample subset;

[0036] a repeating module configured to repeatedly execute the ordering and sampling module multiple times, use different sorting methods and the same or different extraction methods each time, and extract different learning sample subsets;

[0037] a merging module configured to merge all the learning sample subsets into a learning sample set.

[0038] In a third aspect, the present application provides an electronic device, comprising one or more processors; a storage device for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any of the implementations of the first aspect.

[0039] In a fourth aspect, the present application provides a computer readable storage medium, having stored thereon a computer program which, when executed by a processor, implements the method as described in any implementation form of the first aspect.

[0040] Compared with the prior art, the present application has the following beneficial effects:

[0041] (1) The generator set artificial intelligence system machine learning sample screening method proposed by the present application adopts different cleaning methods for data cleaning of the generator set operation data. The data cleaning steps can be performed synchronously in the sampling process, which is efficient, reduces invalid samples, and improves learning efficiency, decision efficiency and decision accuracy.

[0042] (2) The generator set artificial intelligence system machine learning sample screening method proposed by the present application adopts multiple sorting extraction steps for data extraction of the candidate sample set. Different sorting methods are used for each sorting, and multiple extraction methods are used for extraction, which ensures the comprehensiveness of the learning sample distribution and the accuracy of the decision result.

[0043] (3) The generator set artificial intelligence system machine learning sample screening method proposed by the present application can screen the learning samples of the generator set artificial intelligence system machine, has low requirements for the professional quality of the operator, high automation degree in the processing process, and good extensive distribution of the screened learning samples. BRIEF DESCRIPTION OF DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0045] Figure 1 is an exemplary device architecture diagram to which an embodiment of the present application can be applied;

[0046] Figure 2 is a flowchart of the generator set artificial intelligence system machine learning sample screening method of the embodiment of the present application;

[0047] Figure 3 is a logic block diagram of the generator set artificial intelligence system machine learning sample screening method of the embodiment of the present application;

[0048] Figure 4 is a schematic diagram of the generator set artificial intelligence system machine learning sample screening device of the embodiment of the present application;

[0049] Figure 5is a structural schematic diagram of a computer device of an electronic device suitable for implementing the embodiments of the present application. DETAILED DESCRIPTION

[0050] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0051] Figure 1 An exemplary device architecture 100 of the generator set artificial intelligence system machine learning sample screening method or the generator set artificial intelligence system machine learning sample screening device to which the embodiments of the present application can be applied is shown.

[0052] As shown in Figure 1 , the device architecture 100 can include a DCS system 101, a server 104, user terminals 107, 108, 109, and the like. The server accesses the DCS system through lines 102 of types such as optical fiber, Ethernet network cable, and RS485 twisted pair, and is generally accessed through a unidirectional isolation device 103. The server accesses an office network 106, and is generally accessed through a firewall or a special gateway device 105.

[0053] The server can be a single physical server or a server cluster composed of multiple physical servers, and can also be a virtual machine system. The server is connected to the DCS system through a wired network, collects samples and acquires real-time data from the DCS system, uses the collected samples for model training, and uses the acquired real-time data for fault diagnosis. The server is installed with a generator online fault diagnosis system program, which can automatically complete sample collection, automatic sample screening provided by the embodiments of the present application, model training, real-time data acquisition, online diagnosis, alarm output, and the like.

[0054] The user terminal can be in the form of a desktop computer, a laptop computer, a tablet computer, or the like, and can also be a cloud virtual computer. The user terminal accesses a WEB service provided by the server through an internal local area network, and can real-time master the working state of the generator set. The user terminal can query and analyze historical data and mine data, and can deeply understand the health indicators of the unit. Authorized users can also manually adjust the sample set automatically screened by the computer, and add or delete specified samples.

[0055] It should be understood that Figure 1The number of terminal devices, networks and servers in the figure is only illustrative. According to the implementation needs, any number of terminal devices, networks and servers can be provided. When the user does not need to query the analysis calculation and diagnosis results of the embodiments of the present application through the network, the firewall or special gateway device, office network, user terminal can not be included in the device architecture diagram, and the user can directly check the related data on the display or large screen provided with the server.

[0056] Figure 2 A generator set artificial intelligence system machine learning sample screening method provided by the embodiments of the present application is shown, which includes the following steps:

[0057] S1, obtain the generator set operation data, and perform one or more data cleaning on the generator set operation data by using a preset cleaning method to obtain a candidate sample set and an invalid sample set. If multiple data cleaning is used, different cleaning methods are used for each data cleaning.

[0058] Specifically, referring to Figure 3 The large number of provided candidate generator set operation data (samples) are automatically screened. First, the invalid samples that do not meet the conditions are cleaned in multiple steps according to the set cleaning method, and then the candidate sample set is sorted and extracted in multiple times according to the set sorting method. The learning samples are extracted according to the set extraction method after sorting.

[0059] In specific embodiments, the cleaning method includes:

[0060] In response to determining that one or more of the electrical parameters of the generator set is outside the set range, it is classified as an invalid sample set, and the rest is a candidate sample set;

[0061] In response to determining that the sampling time of the generator set operation data is outside the set range, it is classified as an invalid sample set, and the rest is a candidate sample set;

[0062] In response to determining that one or more of the non-electrical parameters of the generator set is outside the set range, it is classified as an invalid sample set, and the rest is a candidate sample set;

[0063] In response to determining that the calculation result according to one or more of the electrical parameters, the sampling time, and the non-electrical parameters is outside the set range, it is classified as an invalid sample set, and the rest is a candidate sample set.

[0064] Specifically, the electrical parameters include active power, reactive power, stator voltage, stator current, rotor current, and power factor. Taking the generator rotor inter-turn short circuit intelligent diagnosis system as an example, the cleaning method based on the electrical parameters includes one or more of the following:

[0065] The active power unit value is excluded if it is out of limit;

[0066] If the per-unit value of reactive power exceeds the limit, it is excluded;

[0067] If the stator voltage per unit value exceeds the limit, it is excluded;

[0068] If the stator current per unit value exceeds the limit, it is excluded;

[0069] If the rotor current per-unit value exceeds the limit, then it is excluded;

[0070] If the power factor exceeds the limit, then exclude it;

[0071] If the three-phase voltage imbalance exceeds the limit, it is excluded;

[0072] If the three-phase current imbalance exceeds the limit, it is excluded;

[0073] Apparent power S01, calculated using active and reactive power, is compared to apparent power S02, calculated using three-phase voltage and three-phase current. If the deviation exceeds the limit, it is excluded.

[0074] Specifically, non-electrical parameters include vibration, noise, pressure, flow rate, temperature, and temperature rise. Therefore, cleaning methods based on non-electrical parameters include one or more of the following:

[0075] If the vibration frequency exceeds the limit, then it is excluded;

[0076] If the noise level exceeds the limit in decibels, it is excluded;

[0077] If the pressure exceeds the limit, then eliminate the possibility;

[0078] If the data usage limit is exceeded, it will be excluded;

[0079] If the temperature exceeds the limit, then it is excluded;

[0080] If the temperature rise exceeds the limit, then it is excluded.

[0081] Specifically, the cleaning method also includes: excluding samples if the sampling time exceeds the limit.

[0082] If the calculation results of one or more of the above parameters exceed the set range, then the condition is excluded.

[0083] Depending on the system design and development needs, data cleaning can be completed in the early stages. After starting the sample screening program, you can directly jump to the sorting and sampling steps.

[0084] In one embodiment, the generator set artificial intelligence system of this application employs support vector machine as the core machine learning algorithm. The data cleaning step S1 is completed during the sampling phase, marking the cleaned and excluded samples as invalid and classifying them into an invalid sample set. During automatic sample selection during model training, valid samples from the most recent three months (optional) are read as a candidate sample set. The candidate sample set is then sorted and extracted.

[0085] S2, sort the candidate sample set using a preset sorting method to obtain a sorted sample set, and extract a subset of learning samples from the sorted sample set according to a preset extraction method.

[0086] In specific embodiments, the sorting methods include:

[0087] Sorted according to the electrical parameters of the generator sets;

[0088] Sort the generator set operating data according to the sampling time;

[0089] Sorted according to the non-electrical parameters of the generator sets;

[0090] The results are sorted based on one or more of the calculation results of electrical parameters, sampling time, and non-electrical parameters.

[0091] In specific embodiments, the extraction methods include one or more of the following:

[0092] Random distribution sampling;

[0093] Uniform distribution sampling;

[0094] Sampling based on normal or inverse distribution; sampling based on triangular or inverse triangular distribution; sampling based on trapezoidal or inverse trapezoidal distribution; sampling based on sine or arcsine distribution; and segmented sampling.

[0095] Specifically, if random distribution sampling is selected as the sampling method, the sorting step can be skipped.

[0096] S3. Repeat step S2 multiple times, each time using a different sorting method and the same or different extraction method to extract different subsets of learning samples.

[0097] In one embodiment, both sorting and sampling are performed four times, with the sorting rules being: sorted by rotor current, sorted by average three-phase current, sorted by active power, and sorted by reactive power. Each time, 3000 samples (optional) are drawn, using an inverse distribution (dense at both ends, sparse in the middle) sampling method.

[0098] S4 combines all subsets of the learning samples into a single learning sample set.

[0099] Specifically, the learning sample subsets extracted from multiple sorting processes are aggregated to obtain the learning sample set, which can be used in the training process of the generator set artificial intelligence system.

[0100] Further reference Figure 4 As an implementation of the methods shown in the above figures, this application provides an embodiment of a machine learning sample screening device for a generator set artificial intelligence system. This device embodiment is similar to... Figure 2Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.

[0101] This application provides a machine learning sample screening device for a generator set artificial intelligence system, comprising:

[0102] Cleaning module 1 is configured to acquire generator set operating data, perform one or more data cleaning operations on the generator set operating data using a preset cleaning method, and obtain a candidate sample set and an invalid sample set. If multiple data cleaning operations are performed, a different cleaning method is used for each data cleaning operation.

[0103] The sorting and sampling module 2 is configured to sort the candidate sample set using a preset sorting method to obtain a sorted sample set, and then extract a subset of learning samples from the sorted sample set according to a preset extraction method.

[0104] The repeat module 3 is configured to repeatedly execute the sorting and sampling module multiple times, each time using a different sorting method and the same or different sampling method to extract different subsets of learning samples.

[0105] Merging module 4 is configured to merge all subsets of learning samples into a learning sample set.

[0106] The following is for reference. Figure 5 It illustrates an electronic device suitable for implementing embodiments of this application (e.g., Figure 1 The diagram shows the structure of a computer device 500 (the server shown). Figure 5 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0107] like Figure 5 As shown, the computer device 500 includes a central processing unit (CPU) 501 and a graphics processing unit (GPU) 502, which can perform various appropriate actions and processes according to programs stored in read-only memory (ROM) 503 or programs loaded from storage section 509 into random access memory (RAM) 504. The RAM 504 also stores various programs and data required for the operation of the device 500. The CPU 501, GPU 502, ROM 503, and RAM 504 are interconnected via a bus 505. An input / output (I / O) interface 506 is also connected to the bus 505.

[0108] The following components are connected to I / O interface 506: an input section 507 including a keyboard, mouse, etc.; an output section 508 including an LCD, speakers, etc.; a storage section 509 including a hard disk, etc.; and a communication section 510 including a network interface card, such as a LAN card or modem. The communication section 510 performs communication processing via a network such as the Internet. A drive 511 may also be connected to I / O interface 506 as needed. A removable medium 512, such as a hard disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 511 as needed so that computer programs read from it can be installed into storage section 509 as needed.

[0109] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 510, and / or installed from removable medium 512. When the computer program is executed by central processing unit (CPU) 501 and graphics processing unit (GPU) 502, the functions defined in the methods of this application are performed.

[0110] It should be noted that the computer-readable medium described in this application can be a computer-readable signal medium, a computer-readable medium, or any combination thereof. A computer-readable medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor device, or any combination thereof. More specific examples of a computer-readable medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution device, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than a computer-readable medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution device, apparatus, or apparatus. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0111] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as C#, Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0112] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using dedicated hardware-based means to perform the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0113] The modules described in the embodiments of this application can be implemented in software or hardware. These modules can also be located within a processor.

[0114] In another aspect, this application also provides a computer-readable medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: acquire generator set operating data; perform one or more data cleaning operations on the generator set operating data using a preset cleaning method to obtain a candidate sample set and an invalid sample set; if multiple data cleaning operations are performed, each data cleaning operation uses a different cleaning method; sort the candidate sample set using a preset sorting method to obtain a sorted sample set; extract samples from the sorted sample set according to a preset extraction method to obtain a learning sample subset; repeat the previous step multiple times, each time using a different sorting method and the same or different extraction methods to extract different learning sample subsets; and merge all learning sample subsets into a learning sample set.

[0115] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A method for selecting machine learning samples in an artificial intelligence system for generator sets, characterized in that, Includes the following steps: S1, acquire generator set operating data, and perform one or more data cleaning operations on the generator set operating data using a preset cleaning method to obtain a candidate sample set and an invalid sample set. If multiple data cleaning operations are performed, a different cleaning method is used for each data cleaning operation. The cleaning methods include: In response to the determination that one or more of the electrical parameters of the generator set exceed the set range, it is classified into the invalid sample set, and the rest are the candidate sample set; In response to the determination that the sampling time of the generator set operating data exceeds the set range, it is classified into the invalid sample set, and the rest are candidate sample sets; In response to the determination that one or more of the non-electrical parameters of the generator set exceed the set range, it is classified into the invalid sample set, and the rest are candidate sample sets; In response to determining that the calculation result based on one or more of the electrical parameters, sampling time, and non-electrical parameters exceeds the set range, it is classified into the invalid sample set, and the rest are candidate sample sets; S2, the candidate sample set is sorted using a preset sorting method to obtain a sorted sample set, and the sorted sample set is extracted according to a preset extraction method to obtain a learning sample subset; S3, Repeat step S2 multiple times, each time using a different sorting method and the same or different extraction methods to extract different subsets of learning samples. The sorting methods include: Sorted according to the electrical parameters of the generator sets; Sort the generator set operating data according to the sampling time; Sorted according to the non-electrical parameters of the generator sets; The results are sorted based on one or more of the following: electrical parameters, sampling time, and non-electrical parameters; the extraction method includes one or more of the following: Random distribution sampling; Uniform distribution sampling; Normal or inverse distribution sampling; Sampling using triangular or inverse triangular distributions; Trapezoidal or inverse trapezoidal sampling; Sampling using a sinusoidal or arcsine distribution; Segmented sampling; S4 combines all subsets of the learning samples into a single learning sample set.

2. The method for selecting machine learning samples for an artificial intelligence system for generator sets according to claim 1, characterized in that, The electrical parameters include active power, reactive power, stator voltage, stator current, rotor current, and power factor.

3. The method for selecting machine learning samples for an artificial intelligence system for generator sets according to claim 1, characterized in that, The non-electrical parameters include vibration, noise, pressure, flow rate, temperature, and temperature rise.

4. The method for selecting machine learning samples for an artificial intelligence system for generator sets according to claim 1, characterized in that, Step S1 can be completed synchronously during the sample collection stage; if random distribution sampling is selected as the sampling method in step S2, the sorting step can be skipped.

5. A machine learning sample selection device for a generator set artificial intelligence system, employing the machine learning sample selection method for a generator set artificial intelligence system according to any one of claims 1-4, characterized in that, include: The cleaning module is configured to acquire generator set operating data, perform one or more data cleaning operations on the generator set operating data using a preset cleaning method, and obtain a candidate sample set and an invalid sample set. If multiple data cleaning operations are performed, a different cleaning method is used for each data cleaning operation. The sorting and sampling module is configured to sort the candidate sample set using a preset sorting method to obtain a sorted sample set, and to extract a learning sample subset from the sorted sample set according to a preset extraction method. The repeat module is configured to repeatedly execute the sorting and sampling module multiple times, each time using a different sorting method and the same or different sampling method to extract different subsets of learning samples. The merge module is configured to merge all subsets of the learning samples into a single learning sample set.

6. An electronic device, comprising: One or more processors; Storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-4.

Images