A hydropower station fault diagnosis system based on artificial intelligence

Abstract

The present invention discloses an artificial intelligence-based fault diagnosis system for a hydropower station, which relates to the technical field of fault diagnosis of hydropower stations and solves the technical problem in the prior art that, when the fault diagnosis efficiency cannot meet the current demand, the inadequacy of the fault diagnosis efficiency cannot be compensated by auxiliary monitoring with physical signals. The system comprises a server, and the server is communicatively connected with a fault diagnosis efficiency analysis unit, a real-time physical signal monitoring unit, a diagnostic fault analysis unit and a fault rectification control unit. The present invention performs physical monitoring on the hydropower station during operation, improves the fault diagnosis accuracy of the hydropower station through auxiliary monitoring with physical signals, prevents the fault diagnosis efficiency of the hydropower station from failing to meet the demand, and causes the fault of the hydropower station to be unable to be controlled in time, and also greatly improves the fault diagnosis efficiency through physical signal monitoring, and can compensate in time when the current fault diagnosis method cannot meet the demand, thereby ensuring the operation efficiency of the hydropower station.

Description

Technical Field

[0001] The present invention relates to the technical field of hydropower station fault diagnosis, and in particular to an artificial intelligence-based hydropower station fault diagnosis system. Background Art

[0002] A hydropower station is a comprehensive engineering facility that can convert water energy into electrical energy. It is generally composed of a reservoir formed by water retaining and discharge structures, a water diversion system for the hydropower station, a power plant, and electromechanical equipment. High water from the reservoir flows through the diversion system into the power plant to drive the hydro-generator sets to generate electricity, which is then fed into the power grid through step-up transformers, switch stations, and transmission lines.

[0003] However, in the prior art, when fault diagnosis is performed in a hydropower station, if the fault diagnosis efficiency cannot meet the current demand, it is not possible to use physical signal auxiliary monitoring to compensate for the lack of fault diagnosis efficiency, so that the hydropower station cannot accurately perform fault diagnosis, and it is not possible to carry out targeted management and control according to the type of diagnosed fault, so that the fault diagnosis efficiency is reduced;

[0004] In view of the above technical defects, a solution is now proposed. Summary of the Invention

[0005] The purpose of the present invention is to solve the above-mentioned problems and to propose a hydropower station fault diagnosis system based on artificial intelligence. It analyzes the fault types diagnosed by the hydropower station and accurately analyzes the causes of the current faults in the hydropower station, so as to improve the efficiency of fault maintenance in a targeted manner, prevent the waste of time in fault screening, affect the operating efficiency of the hydropower station, and increase the wear of the equipment of the hydropower station itself, thereby further increasing the risk of failure; the hydropower station is subjected to fault rectification control, and targeted fault rectification is carried out according to different types of sudden faults and non-sudden faults, thereby improving the efficiency of fault rectification, ensuring that the fault can be avoided through rectification, and promoting the operating efficiency of the hydropower station.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] An artificial intelligence-based hydropower station fault diagnosis system includes a server, and the server communication connection has:

[0008] A fault diagnosis efficiency analysis unit is used to analyze the fault diagnosis efficiency during the real-time operation of the hydropower station, obtain the real-time operation time period of the hydropower station and mark it as the analysis time period, obtain the fault diagnosis efficiency analysis coefficient of the hydropower station during the analysis time period, generate a fault diagnosis efficiency failure signal or a fault diagnosis efficiency passing signal based on the comparison of the fault diagnosis efficiency analysis coefficient, and send the signal to the server;

[0009] The real-time physical signal monitoring unit is used to perform physical monitoring of the hydropower station during operation. During the operation of the hydropower station, the operating vibration of the hydropower station mechanical system is monitored to obtain vibration data of the hydropower station mechanical system, wherein the vibration data is represented in the time domain, amplitude domain and frequency domain. In the time domain, a high-risk fault warning signal or a low-risk fault monitoring signal is generated through analysis and sent to the server;

[0010] The diagnostic fault analysis unit is used to analyze the diagnostic faults of the hydropower station during operation, mark the diagnosed faults of the hydropower station as known faults, obtain the operating parameters that fluctuate at the fault location when the known fault occurs based on the diagnosis and maintenance process of the known fault, mark them as influencing parameters, classify the known faults into sudden faults and non-sudden faults through analysis, and send the parameters to the server;

[0011] The fault rectification control unit is used to perform fault rectification control on the hydropower station.

[0012] As a preferred embodiment of the present invention, the operation process of the fault diagnosis efficiency analysis unit is as follows:

[0013] The probability of receiving an early warning before a hydropower station fault occurs during the analysis period is collected, as is the probability of a fault occurring after a warning is issued. The shortest interval between a fault occurring and the recurrence of the same type of fault after diagnosis is completed during the analysis period is collected. The fault diagnosis efficiency coefficient of the hydropower station during the analysis period is obtained through analysis.

[0014] Compare the fault diagnosis efficiency analysis coefficient of the hydropower station within the analysis period with the fault diagnosis efficiency analysis coefficient threshold:

[0015] If the fault diagnosis efficiency analysis coefficient of the hydropower station during the analysis time period exceeds the fault diagnosis efficiency analysis coefficient threshold, then the fault diagnosis efficiency analysis of the hydropower station during the analysis time period is determined to be qualified, a fault diagnosis efficiency qualified signal is generated, and the fault diagnosis efficiency qualified signal is sent to the server; if the fault diagnosis efficiency analysis coefficient of the hydropower station during the analysis time period does not exceed the fault diagnosis efficiency analysis coefficient threshold, then the fault diagnosis efficiency analysis of the hydropower station during the analysis time period is determined to be unqualified, a fault diagnosis efficiency unqualified signal is generated, and the fault diagnosis efficiency unqualified signal is sent to the server.

[0016] As a preferred embodiment of the present invention, the operation process of the real-time physical signal monitoring unit is as follows:

[0017] Based on the fault analysis during the historical operation of the hydropower station, the vibration data of the corresponding mechanical system when the hydropower station has a fault is obtained. The floating value of the vibration data before and after the fault time point is determined according to the corresponding fault time point. If the floating value of the vibration data exceeds the floating value threshold, the corresponding vibration data is marked as fault data. Based on the floating trend of the vibration data and the corresponding value of the fault data, early warning data before the fault data reaches the point of occurrence is obtained;

[0018] After obtaining the fault data and warning data, the real-time operation of the hydropower station is monitored. The real-time vibration data of the mechanical system during the real-time operation of the hydropower station reaches the reciprocating frequency of the warning data and the minimum numerical difference between the real-time vibration data of the mechanical system and the fault data, and they are compared.

[0019] As a preferred embodiment of the present invention, the comparison process is as follows:

[0020] If the reciprocating frequency of the real-time vibration data of the mechanical system reaches the warning data during the real-time operation of the hydropower station and exceeds the reciprocating frequency threshold, or the minimum numerical difference between the real-time vibration data of the mechanical system and the fault data does not exceed the minimum numerical difference threshold, then it is determined that the hydropower station has a high fault risk, and a high fault risk warning signal is generated and sent to the server;

[0021] If the reciprocating frequency of the real-time vibration data of the mechanical system reaches the warning data during the real-time operation of the hydropower station but does not exceed the reciprocating frequency threshold, and the minimum numerical difference between the real-time vibration data of the mechanical system and the fault data exceeds the minimum numerical difference threshold, a low-risk fault monitoring signal is generated and sent to the server. After receiving the low-risk fault monitoring signal, the server evaluates the fault diagnosis efficiency. When the currently matched fault diagnosis method meets the current needs, the physical signal monitoring can be terminated.

[0022] As a preferred embodiment of the present invention, the operation process of the diagnostic fault analysis unit is as follows:

[0023] The maximum instantaneous span affecting the fluctuation of parameter values before a known fault occurs and the frequency of trend changes in the fluctuation of parameter values before a known fault occurs are collected and compared with the maximum instantaneous span threshold and the frequency of trend changes threshold respectively:

[0024] If the maximum instantaneous span of the fluctuation of the value of the influencing parameter before the occurrence of the known fault exceeds the maximum instantaneous span threshold, or the trend change frequency of the fluctuation of the value of the influencing parameter before the occurrence of the known fault exceeds the trend change frequency threshold, the corresponding known fault will be marked as a sudden fault;

[0025] If the maximum instantaneous span affecting the fluctuation of parameter values before the occurrence of the known fault does not exceed the maximum instantaneous span threshold, and the trend change frequency affecting the fluctuation of parameter values before the occurrence of the known fault does not exceed the trend change frequency threshold, the corresponding known fault will be marked as a non-sudden fault; the sudden fault and the non-sudden fault will be sent to the server together.

[0026] As a preferred embodiment of the present invention, the operation process of the fault rectification control unit is as follows:

[0027] The floating span value of the influencing parameter at the moment of sudden fault occurrence and the previous moment as well as the floating speed of the influencing parameter at the moment of sudden fault occurrence and the previous moment are collected and compared with the floating span value threshold and the floating speed threshold respectively:

[0028] If the floating span value of the influencing parameter at the moment of occurrence of the sudden fault of the hydropower station and the corresponding previous moment exceeds the floating span value threshold, or the floating speed of the influencing parameter at the moment of occurrence of the sudden fault and the corresponding previous moment exceeds the floating speed threshold, the early warning value of the sudden fault will be reset, and the interval value corresponding to the early warning value of the influencing parameter of the sudden fault and the value of the influencing parameter at the time of the fault will be expanded;

[0029] If the floating span value of the influencing parameter between the time when the sudden fault occurs and the corresponding previous moment of the hydropower station does not exceed the floating span value threshold, and the floating speed of the influencing parameter between the time when the sudden fault occurs and the corresponding previous moment does not exceed the floating speed threshold, then the interval between the time when the sudden fault occurs and the time when the rectification starts will be shortened.

[0030] As a preferred embodiment of the present invention, for non-sudden faults in hydropower stations, the influencing parameters of the non-sudden faults are controlled and managed. When the floating trend of the non-sudden fault influencing parameters is a fault trend, or the floating time of the influencing parameters exceeds the floating time threshold, it is ensured that when the non-sudden fault influencing parameters reach the warning value, the interval duration of the corresponding non-sudden fault maintenance must be controlled within the interval duration threshold.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] 1. In the present invention, whether the diagnostic efficiency of faults during the operation of a hydropower station meets the requirements is judged, thereby ensuring the operational eligibility of the hydropower station, facilitating timely early warning and maintenance of faults in the hydropower station, minimizing the impact of faults during the operation of the hydropower station, and maximizing the operational efficiency of the hydropower station; the hydropower station is physically monitored during operation, and the accuracy of fault diagnosis of the hydropower station is improved through auxiliary monitoring of physical signals, preventing the fault diagnosis efficiency of the hydropower station from failing to meet the requirements, thereby preventing the failure of the hydropower station from being able to be timely managed and controlled. The fault diagnosis efficiency is greatly improved through physical signal monitoring, and when the current fault diagnosis method fails to meet the requirements, timely compensation can be made, thereby ensuring the operational efficiency of the hydropower station;

[0033] 2. In the present invention, analysis is performed based on the fault types diagnosed corresponding to the hydropower station, and the causes of the current faults in the hydropower station are accurately analyzed, so as to improve the efficiency of fault maintenance in a targeted manner, prevent the waste of time in fault screening, affect the operating efficiency of the hydropower station, and increase the wear of the equipment of the hydropower station itself, thereby further increasing the risk of failure; the hydropower station is subjected to fault rectification control, and targeted fault rectification is performed according to different types of sudden faults and non-sudden faults, thereby improving the efficiency of fault rectification, ensuring that faults can be avoided through rectification, and promoting the operating efficiency of the hydropower station. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] To facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0035] Figure 1 This is a principle block diagram of a hydropower station fault diagnosis system based on artificial intelligence in the present invention. DETAILED DESCRIPTION

[0036] In order to enable those skilled in the art to better understand the solutions of the present invention, the following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0037] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0038] See also Figure 1As shown, an artificial intelligence-based fault diagnosis system for a hydropower station includes a server, which is communicatively connected to a fault diagnosis efficiency analysis unit, a real-time physical signal monitoring unit, a diagnostic fault analysis unit, and a fault rectification control unit. The server is bidirectionally connected to the fault diagnosis efficiency analysis unit, the real-time physical signal monitoring unit, the diagnostic fault analysis unit, and the fault rectification control unit.

[0039] The server generates a fault diagnosis efficiency analysis signal and sends the fault diagnosis efficiency analysis signal to the fault diagnosis efficiency analysis unit. After receiving the fault diagnosis efficiency analysis signal, the fault diagnosis efficiency analysis unit analyzes the fault diagnosis efficiency during the real-time operation of the hydropower station, and determines whether the fault diagnosis efficiency during the operation of the hydropower station meets the requirements, thereby ensuring the operational eligibility of the hydropower station, facilitating timely early warning and maintenance of faults in the hydropower station, minimizing the impact of faults during the operation of the hydropower station, and maximizing the operational efficiency of the hydropower station.

[0040] The real-time operating time period of the hydropower station is obtained and marked as the analysis time period. The probability of an early warning being provided before a hydropower station fault occurs and the probability of a fault still occurring after a fault warning are collected during the analysis time period. These probabilities are marked as YGL and GGL, respectively. The shortest interval between the occurrence of a hydropower station fault and the recurrence of the same type of fault after diagnosis is completed during the analysis time period is collected. This is marked as JGS.

[0041] By formula Obtain the fault diagnosis efficiency analysis coefficient C of the hydropower station during the analysis period, where f1, f2, and f3 are preset proportional coefficients, and f1>f2>f3>0, and β is the error correction factor, which is 0.976;

[0042] Compare the fault diagnosis efficiency analysis coefficient C of the hydropower station within the analysis period with the fault diagnosis efficiency analysis coefficient threshold:

[0043] If the fault diagnosis efficiency analysis coefficient C of the hydropower station during the analysis period exceeds the fault diagnosis efficiency analysis coefficient threshold, it is determined that the fault diagnosis efficiency analysis of the hydropower station during the analysis period is qualified, a fault diagnosis efficiency qualified signal is generated, and the fault diagnosis efficiency qualified signal is sent to the server;

[0044] If the fault diagnosis efficiency analysis coefficient C of the hydropower station during the analysis period does not exceed the fault diagnosis efficiency analysis coefficient threshold, it is determined that the fault diagnosis efficiency analysis of the hydropower station during the analysis period is unqualified, a fault diagnosis efficiency unqualified signal is generated, and the fault diagnosis efficiency unqualified signal is sent to the server;

[0045] After receiving the signal indicating that the fault diagnosis efficiency is unqualified, the server generates a real-time physical signal monitoring signal and sends the real-time physical signal monitoring signal to the real-time physical signal monitoring unit. After receiving the real-time physical signal monitoring signal, the real-time physical signal monitoring unit performs physical monitoring on the hydropower station during operation. The auxiliary monitoring of physical signals improves the fault diagnosis accuracy of the hydropower station, and prevents the fault diagnosis efficiency of the hydropower station from failing to meet the demand, so that the fault of the hydropower station cannot be controlled in time. The fault diagnosis efficiency is greatly improved through physical signal monitoring, and when the current fault diagnosis method cannot meet the demand, it can be compensated in time to ensure the operation efficiency of the hydropower station.

[0046] During the operation of the hydropower station, the vibration of the mechanical system of the hydropower station is monitored to obtain the vibration data of the mechanical system of the hydropower station. The vibration data is represented in the time domain, amplitude domain and frequency domain. The time domain is represented as a random variable or the average value of a group of data in the prior art. The amplitude domain is represented as the probability distribution of the amplitude in the prior art, that is, the probability of describing the instantaneous amplitude of the random vibration being lower than a certain value. The frequency domain is represented as the floating span value of the frequency in the prior art.

[0047] Based on the fault analysis during the historical operation of the hydropower station, the vibration data of the corresponding mechanical system when the hydropower station has a fault is obtained. The floating value of the vibration data before and after the fault time point is determined according to the corresponding fault time point. If the floating value of the vibration data exceeds the floating value threshold, the corresponding vibration data is marked as fault data. Based on the floating trend of the vibration data and the corresponding value of the fault data, early warning data before the fault data reaches the point of occurrence is obtained;

[0048] After acquiring the fault data and the early warning data, the real-time operation of the hydropower station is monitored. The reciprocating frequency at which the real-time vibration data of the mechanical system reaches the early warning data and the minimum numerical difference between the real-time vibration data of the mechanical system and the fault data during the real-time operation of the hydropower station are collected. The reciprocating frequency at which the real-time vibration data of the mechanical system reaches the early warning data and the minimum numerical difference between the real-time vibration data of the mechanical system and the fault data during the real-time operation of the hydropower station are compared with the reciprocating frequency threshold and the minimum numerical difference threshold, respectively:

[0049] It is understandable that vibration data fluctuates in real time and will continue to fluctuate after reaching the warning data. Therefore, the more frequently the warning data is reached, the greater the risk. At the same time, the closer the data is to the fault data after exceeding the warning data, the greater the risk of failure.

[0050] If the reciprocating frequency of the real-time vibration data of the mechanical system reaches the warning data during the real-time operation of the hydropower station and exceeds the reciprocating frequency threshold, or the minimum numerical difference between the real-time vibration data of the mechanical system and the fault data does not exceed the minimum numerical difference threshold, then the hydropower station is judged to have a high fault risk, and a high-risk fault warning signal is generated and sent to the server; after receiving the high-risk fault warning signal, the server will rectify and control the operation of the corresponding hydropower station;

[0051] If the reciprocating frequency of the real-time vibration data of the mechanical system reaches the warning data during the real-time operation of the hydropower station but does not exceed the reciprocating frequency threshold, and the minimum numerical difference between the real-time vibration data of the mechanical system and the fault data exceeds the minimum numerical difference threshold, it is determined that the hydropower station does not have a high fault risk, and a low-fault-risk monitoring signal is generated and sent to the server. After receiving the low-fault-risk monitoring signal, the server evaluates the fault diagnosis efficiency. When the currently matched fault diagnosis method meets the current needs, the physical signal monitoring can be terminated;

[0052] The server generates a diagnostic fault analysis signal and sends it to the diagnostic fault analysis unit. After receiving the diagnostic fault analysis signal, the diagnostic fault analysis unit analyzes the diagnostic faults of the hydropower station during operation. According to the corresponding fault type diagnosed by the hydropower station, the server accurately analyzes the cause of the current fault of the hydropower station, thereby improving the efficiency of fault maintenance in a targeted manner and preventing the waste of time in fault screening, which affects the operating efficiency of the hydropower station and increases the wear of the hydropower station's own equipment, thereby further increasing the risk of failure.

[0053] The faults diagnosed in the hydropower station are marked as known faults. During the diagnosis and maintenance of the known faults, the operating parameters corresponding to the fault location that fluctuated when the known fault occurred are obtained and marked as influencing parameters. The corresponding fault location is the location of the hydropower station where the known fault occurred, and the operating parameters are the operating parameters related to the fault location, such as noise level, temperature, and other parameters.

[0054] The maximum instantaneous span of the numerical fluctuation of the influencing parameter before the occurrence of a known fault and the frequency of trend change in the numerical fluctuation process of the influencing parameter before the occurrence of a known fault are collected, and the maximum instantaneous span of the numerical fluctuation of the influencing parameter before the occurrence of a known fault and the frequency of trend change in the numerical fluctuation process of the influencing parameter before the occurrence of a known fault are compared with the maximum instantaneous span threshold and the trend change frequency threshold respectively: wherein, the maximum instantaneous span is expressed as the maximum span of the numerical fluctuation of the influencing parameter of the hydropower station at adjacent working unit moments, and the trend change frequency is expressed as the frequency of change of the floating trend. For example, if the increasing trend changes to a decreasing trend, it is a change. A low change frequency indicates that the consistency of the trend change is strong. If a fault occurs, it indicates that the cause of the fault persists.

[0055] If the maximum instantaneous span affecting the fluctuation of the parameter value before the occurrence of the known fault exceeds the maximum instantaneous span threshold, or the trend change frequency of the parameter value fluctuation before the occurrence of the known fault exceeds the trend change frequency threshold, then the corresponding known fault will be marked as a sudden fault; if the maximum instantaneous span affecting the fluctuation of the parameter value before the occurrence of the known fault does not exceed the maximum instantaneous span threshold, and the trend change frequency of the parameter value fluctuation before the occurrence of the known fault does not exceed the trend change frequency threshold, then the corresponding known fault will be marked as a non-sudden fault;

[0056] Send sudden faults and non-sudden faults to the server together;

[0057] After receiving sudden faults and non-sudden faults, the server generates a fault rectification control signal and sends the fault rectification control signal to the fault rectification control unit. After receiving the fault rectification control signal, the fault rectification control unit controls the hydropower station and performs targeted fault rectification according to different types of sudden faults and non-sudden faults, thereby improving the efficiency of fault rectification, ensuring that faults can be avoided through rectification, and promoting the operating efficiency of the hydropower station;

[0058] The floating span value of the influencing parameter between the moment of sudden fault occurrence of the hydropower station and the corresponding previous moment and the floating speed of the influencing parameter between the moment of sudden fault occurrence of the hydropower station and the corresponding previous moment are collected, and the floating span value of the influencing parameter between the moment of sudden fault occurrence of the hydropower station and the corresponding previous moment and the floating speed of the influencing parameter between the moment of sudden fault occurrence of the hydropower station and the corresponding previous moment and the floating speed of the influencing parameter between the moment of sudden fault occurrence of the hydropower station and the corresponding previous moment are compared with the floating span value threshold and the floating speed threshold respectively:

[0059] If the floating span value of the influencing parameter at the moment of occurrence of the sudden fault of the hydropower station and the corresponding previous moment exceeds the floating span value threshold, or the floating speed of the influencing parameter at the moment of occurrence of the sudden fault and the corresponding previous moment exceeds the floating speed threshold, the early warning value of the sudden fault will be reset, and the interval value corresponding to the early warning value of the influencing parameter of the sudden fault and the value of the influencing parameter at the time of the fault will be expanded;

[0060] If the floating span value of the influencing parameter between the time when the sudden fault occurs and the corresponding previous time does not exceed the floating span value threshold, and the floating speed of the influencing parameter between the time when the sudden fault occurs and the corresponding previous time does not exceed the floating speed threshold, then the interval between the time when the sudden fault occurs and the time when the rectification starts will be shortened;

[0061] For non-sudden faults of hydropower stations, the influencing parameters of non-sudden faults are controlled and managed. When the floating trend of the non-sudden fault influencing parameters is a fault trend, or the floating time of the influencing parameters exceeds the floating time threshold, it is ensured that when the non-sudden fault influencing parameters reach the warning value, the interval duration of the corresponding non-sudden fault maintenance must be controlled within the interval duration threshold.

[0062] The above formulas are obtained by collecting a large amount of data and performing software simulation to select a formula close to the actual value. The coefficients in the formula are set by those skilled in the art according to actual conditions;

[0063] When the present invention is in use, the fault diagnosis efficiency analysis unit analyzes the fault diagnosis efficiency of the hydropower station during real-time operation, obtains the real-time operation time period of the hydropower station and marks it as the analysis time period, obtains the fault diagnosis efficiency analysis coefficient of the hydropower station within the analysis time period, and compares the fault diagnosis efficiency analysis coefficient to judge the efficiency of the real fault; the real-time physical signal monitoring unit physically monitors the hydropower station during operation, monitors the operating vibration of the mechanical system of the hydropower station during operation, obtains vibration data of the mechanical system of the hydropower station, generates a high-risk fault warning signal or a low-risk fault monitoring signal through analysis, and sends it to the server; the diagnostic fault analysis unit analyzes the diagnostic faults of the hydropower station during operation, divides the known faults into sudden faults and non-sudden faults through analysis, and sends them to the server; the fault rectification control unit performs fault rectification control on the hydropower station.

[0064] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A hydropower station fault diagnosis system based on artificial intelligence, characterized in that: Including the server, the server communication connections are: A fault diagnosis efficiency analysis unit is used to analyze the fault diagnosis efficiency during the real-time operation of the hydropower station, obtain the real-time operation time period of the hydropower station and mark it as the analysis time period, obtain the fault diagnosis efficiency analysis coefficient of the hydropower station during the analysis time period, generate a fault diagnosis efficiency failure signal or a fault diagnosis efficiency passing signal based on the comparison of the fault diagnosis efficiency analysis coefficient, and send the signal to the server; The real-time physical signal monitoring unit is used to perform physical monitoring of the hydropower station during operation. During the operation of the hydropower station, the operating vibration of the hydropower station mechanical system is monitored to obtain vibration data of the hydropower station mechanical system, wherein the vibration data is represented in the time domain, amplitude domain and frequency domain. In the time domain, a high-risk fault warning signal or a low-risk fault monitoring signal is generated through analysis and sent to the server; The diagnostic fault analysis unit is used to analyze the diagnostic faults of the hydropower station during operation, mark the diagnosed faults of the hydropower station as known faults, obtain the operating parameters that fluctuate at the fault location when the known fault occurs based on the diagnosis and maintenance process of the known fault, mark them as influencing parameters, classify the known faults into sudden faults and non-sudden faults through analysis, and send the parameters to the server; Fault rectification control unit, used to perform fault rectification control on the hydropower station; The operation process of the fault diagnosis efficiency analysis unit is as follows: The probability of receiving an early warning before a hydropower station fault occurs during the analysis period is collected, as is the probability of a fault occurring after a warning is issued. The shortest interval between a fault occurring and the recurrence of the same type of fault after diagnosis is completed during the analysis period is collected. The fault diagnosis efficiency coefficient of the hydropower station during the analysis period is obtained through analysis. Compare the fault diagnosis efficiency analysis coefficient of the hydropower station within the analysis period with the fault diagnosis efficiency analysis coefficient threshold: If the fault diagnosis efficiency analysis coefficient of the hydropower station during the analysis time period exceeds the fault diagnosis efficiency analysis coefficient threshold, then the fault diagnosis efficiency analysis of the hydropower station during the analysis time period is determined to be qualified, a fault diagnosis efficiency qualified signal is generated, and the fault diagnosis efficiency qualified signal is sent to the server; if the fault diagnosis efficiency analysis coefficient of the hydropower station during the analysis time period does not exceed the fault diagnosis efficiency analysis coefficient threshold, then the fault diagnosis efficiency analysis of the hydropower station during the analysis time period is determined to be unqualified, a fault diagnosis efficiency unqualified signal is generated, and the fault diagnosis efficiency unqualified signal is sent to the server; The operation process of the real-time physical signal monitoring unit is as follows: Based on the fault analysis during the historical operation of the hydropower station, the vibration data of the corresponding mechanical system when the hydropower station has a fault is obtained. The floating value of the vibration data before and after the fault time point is determined according to the corresponding fault time point. If the floating value of the vibration data exceeds the floating value threshold, the corresponding vibration data is marked as fault data. Based on the floating trend of the vibration data and the corresponding value of the fault data, early warning data before the fault data reaches the point of occurrence is obtained; After obtaining the fault data and warning data, the real-time operation of the hydropower station is monitored. The real-time vibration data of the mechanical system during the real-time operation of the hydropower station reaches the reciprocating frequency of the warning data and the minimum numerical difference between the real-time vibration data of the mechanical system and the fault data, and they are compared.

2. The artificial intelligence-based fault diagnosis system for a hydropower station according to claim 1, characterized in that: The comparison process is as follows: If the reciprocating frequency of the real-time vibration data of the mechanical system reaches the warning data during the real-time operation of the hydropower station and exceeds the reciprocating frequency threshold, or the minimum numerical difference between the real-time vibration data of the mechanical system and the fault data does not exceed the minimum numerical difference threshold, then it is determined that the hydropower station has a high fault risk, and a high fault risk warning signal is generated and sent to the server; If the reciprocating frequency of the real-time vibration data of the mechanical system reaches the warning data during the real-time operation of the hydropower station but does not exceed the reciprocating frequency threshold, and the minimum numerical difference between the real-time vibration data of the mechanical system and the fault data exceeds the minimum numerical difference threshold, a low-risk fault monitoring signal is generated and sent to the server. After receiving the low-risk fault monitoring signal, the server evaluates the fault diagnosis efficiency. When the currently matched fault diagnosis method meets the current needs, the physical signal monitoring can be terminated.

3. The artificial intelligence-based fault diagnosis system for a hydropower station according to claim 1, characterized in that: The operation process of the diagnostic fault analysis unit is as follows: The maximum instantaneous span affecting the fluctuation of parameter values before a known fault occurs and the frequency of trend changes in the fluctuation of parameter values before a known fault occurs are collected and compared with the maximum instantaneous span threshold and the frequency of trend changes threshold respectively: If the maximum instantaneous span of the fluctuation of the value of the influencing parameter before the occurrence of the known fault exceeds the maximum instantaneous span threshold, or the trend change frequency of the fluctuation of the value of the influencing parameter before the occurrence of the known fault exceeds the trend change frequency threshold, the corresponding known fault will be marked as a sudden fault; If the maximum instantaneous span affecting the fluctuation of parameter values before the occurrence of the known fault does not exceed the maximum instantaneous span threshold, and the trend change frequency affecting the fluctuation of parameter values before the occurrence of the known fault does not exceed the trend change frequency threshold, the corresponding known fault will be marked as a non-sudden fault; the sudden fault and the non-sudden fault will be sent to the server together.

4. The artificial intelligence-based fault diagnosis system for a hydropower station according to claim 1, characterized in that: The operation process of the fault rectification control unit is as follows: The floating span value of the influencing parameter at the moment of sudden fault occurrence and the previous moment as well as the floating speed of the influencing parameter at the moment of sudden fault occurrence and the previous moment are collected and compared with the floating span value threshold and the floating speed threshold respectively: If the floating span value of the influencing parameter at the moment of occurrence of the sudden fault of the hydropower station and the corresponding previous moment exceeds the floating span value threshold, or the floating speed of the influencing parameter at the moment of occurrence of the sudden fault and the corresponding previous moment exceeds the floating speed threshold, the early warning value of the sudden fault will be reset, and the interval value corresponding to the early warning value of the influencing parameter of the sudden fault and the value of the influencing parameter at the time of the fault will be expanded; If the floating span value of the influencing parameter between the time when the sudden fault occurs and the corresponding previous moment of the hydropower station does not exceed the floating span value threshold, and the floating speed of the influencing parameter between the time when the sudden fault occurs and the corresponding previous moment does not exceed the floating speed threshold, then the interval between the time when the sudden fault occurs and the time when the rectification starts will be shortened.

5. The artificial intelligence-based fault diagnosis system for a hydropower station according to claim 4 is characterized in that: For non-sudden faults of hydropower stations, the influencing parameters of non-sudden faults are controlled and managed. When the floating trend of the non-sudden fault influencing parameters is a fault trend, or the floating time of the influencing parameters exceeds the floating time threshold, it is ensured that when the non-sudden fault influencing parameters reach the warning value, the interval duration of the corresponding non-sudden fault maintenance must be controlled within the interval duration threshold.

Images