Intelligent Diagnosis and Real-time Fault Early Warning Analysis Method and System for the Top Cover Drainage System

Abstract

The present invention discloses an intelligent diagnosis and real-time fault early warning analysis method and system for a top cover drainage system, including: normalizing each state characteristic quantity to construct a state vector representing the health state of the top cover drainage system; calculating the mean vector of the state vector set in the normal state to obtain the system health state clustering center vector; calculating the Euclidean distance between each vector in the state vector set and the health state clustering center vector, and calculating the average value of consecutive n Euclidean distances as the state early warning index of the top cover drainage system, and taking the sum of the average value and three times the standard deviation as the system fault early warning threshold; calculating the Euclidean distance between each vector in the real-time monitoring signal state vector set and the health state clustering center vector, and calculating the average value of consecutive n Euclidean distances as the real-time state early warning index; The present invention solves the technical problem that the conventional early warning method lacks the ability of automatic and intelligent fault early warning.

Description

Technical Field

[0001] The present invention relates to the technical field of fluid machinery fault diagnosis, and in particular to an intelligent diagnosis and real-time fault early warning analysis method and system for a top cover drainage system. Background Art

[0002] Conventional hydropower and pumped storage have ushered in new development opportunities, with installed capacity growing rapidly and accounting for an increasing proportion in the power grid; the proportion of intermittent renewable energy such as wind and solar energy in the power system is gradually increasing. In order to ensure the good integration of intermittent renewable energy with the existing power system, hydropower energy will undertake more peak-shaving and frequency-regulating tasks, which puts forward higher requirements for the safe operation and management of hydropower energy. As the key equipment for hydropower energy conversion, hydropower units are developing in the direction of large-scale, complex and high-power. The unit structure is becoming more and more complex and the degree of integration is getting higher and higher. Its safety and stability have become the focus of the power industry. Therefore, in order to ensure the safe and stable operation of hydropower units and power grids, improve equipment utilization, and avoid major economic losses and casualties, it is very necessary to carry out fault trend warning for hydropower units.

[0003] The top cover drainage system is part of the automatic control system of the hydropower unit. For hydropower stations with large hydropower generating units, the unit structure is complex and the automatic control unit system is huge. It is becoming increasingly difficult for operation and maintenance personnel to monitor the production and power generation process in real time and quickly and accurately judge various equipment failures. For failures that the staff cannot determine or judge afterwards, they generally have to shut down the machine for inspection to find the failure, which inevitably causes economic losses to the hydropower station. Especially for those failures that are not obvious and cannot be found by sensory organs and instrument detection by power station staff, but can cause major accidents, the economic losses are even greater. Since the automatic control system of the hydropower generating unit itself and related automation components work in a complex and harsh environment, even with the best equipment or components, failures may occur.

[0004] At present, the research objects of fault warning of hydropower stations are mostly main equipment such as turbine generator sets, while the research on fault warning of automatic control systems is relatively small, and is limited to functions such as parameter monitoring of auxiliary equipment of turbine generator sets and command operation control of main equipment, lacking the ability of automated and intelligent fault warning. Summary of the invention

[0005] The object of the present invention is to overcome the above deficiencies and provide an intelligent diagnosis and real-time fault warning analysis method and system for the top cover drainage system, so as to solve the technical problem that in the prior art, the fault warning method for the conventional hydroelectric turbine top cover drainage system is only limited to functions such as parameter monitoring of auxiliary equipment of the hydro-generating unit and command operation control of main equipment, lacking the ability of automatic and intelligent fault warning.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: an intelligent diagnosis and real-time fault warning analysis method for a top cover drainage system, which includes the following steps:

[0007] Step 1: Process the monitored data of the top cover drainage system collected to obtain various state characteristic quantities associated with the state of the top cover drainage system;

[0008] Step 2: Normalize each state characteristic quantity, and then construct a state vector representing the health state of the top cover drainage system. Multiple state vectors form a state vector set;

[0009] Step 3: Calculate the mean vector of the state vector set under the normal state of the top cover drainage system to obtain the system health state clustering center vector;

[0010] Step 4: According to the state vector set of the top cover drainage system in the healthy state obtained in the above Steps 1 - 2, calculate the relative Euclidean distance between each vector in the vector set and the healthy state clustering center vector as the state warning index of the top cover drainage system, calculate the average value (MEAN) and standard deviation (STD) of the Euclidean distance, and use MEAN + 3 * STD as the fault warning threshold;

[0011] Step 5: According to the state vector set of the real-time monitoring signal of the top cover drainage system obtained in the above Steps 1 - 2, calculate the relative Euclidean distance between each vector in the vector set and the healthy state clustering center vector as the real-time state warning index of the top cover drainage system.

[0012] Preferably, in the above Step 2, the normalization processing method uses the maximum-minimum method for linear normalization, and the formula is:

[0013] .

[0014] Furthermore, the process of constructing the state vector and the state vector set in the above Step 2 is as follows:

[0015] If there are n state characteristic quantities in the state characteristic quantity system, after density unification, each state characteristic quantity has m data respectively, then the first state vector will be an n-dimensional vector:

[0016] ;

[0017] Furthermore, the state vector set in step 2 is as follows:

[0018] ;

[0019] That is

[0020] .

[0021] Preferably, the clustering methods in steps 4 and 5 use an Euclidean space clustering model to calculate the relative Euclidean distances between each vector in the vector set and the healthy state clustering center vector, which are used as the state warning indicators for the top cover drainage system.

[0022] Preferably, the threshold in step 4 is determined using the Gaussian threshold rule, calculating the average value (MEAN) and standard deviation (STD) of the Euclidean distances, and using MEAN + 3 * STD as the fault warning threshold.

[0023] In addition, the present invention also discloses an intelligent diagnosis and real-time fault warning analysis system for a top cover drainage system, which includes: a database module, a data cache module, a data processing module, and a real-time fault warning module;

[0024] The database module is used to store the original data of the measuring points of the hydropower unit, as well as various state characteristic quantity data, system healthy state clustering center vector, system fault warning threshold, system historical state warning indicators and other data results obtained through the processing of the data processing module;

[0025] The data cache module is used to store real-time data such as the real-time state warning indicators of the top cover drainage system of the hydropower unit obtained through the processing of the data processing module;

[0026] The data processing module is the background algorithm program of the system. On the one hand, it is used to collect the original data of the measuring points in the database module and perform arithmetic processing on it to obtain the state characteristic quantities of the top cover drainage system, the state vectors in the system healthy state, the system healthy state clustering center vector, the system fault warning threshold, and the system real-time state warning indicators in steps 1 - 5, completing the initialization of the data; on the other hand, it is connected to the database module, the data cache module, and the real-time fault warning module, serving as the information transmission hub between the modules, and sending the processed data to each module through the interface respectively.

[0027] The real-time fault warning module is used to implement the display of an intelligent diagnosis and real-time fault warning analysis method for a top cover drainage system.

[0028] Furthermore, the real-time fault warning module includes the following parts: historical data monitoring part, single state characteristic quantity warning part, Euclidean distance warning index warning part, system comprehensive warning part, and warning information part;

[0029] The historical data monitoring part receives and displays the historical state characteristic quantity data in the database;

[0030] The single state characteristic quantity warning part evaluates whether the single state characteristic quantity is in a normal state. If the real-time value of the state characteristic quantity is not within the normal value range, a warning is issued for the single state characteristic quantity;

[0031] The Euclidean distance warning index warning part compares the system real-time state warning index based on the Euclidean space clustering model with the system fault warning threshold. If the warning index exceeds the threshold, a warning is issued for the overall system state;

[0032] The system comprehensive warning part correlates the single state characteristic quantity warning part and the Euclidean distance warning index warning part, and determines the system comprehensive operation state level (three levels: normal, attention, abnormal) according to the warning results of the above two parts;

[0033] The warning information part correlates the single state characteristic quantity warning part, the Euclidean distance warning index warning part, and the system comprehensive warning part, and displays the system operation state information at the moment of abnormal warning.

[0034] Furthermore,

[0035] The logic for determining the system comprehensive operation state level is as follows:

[0036] When the warning results of all system state characteristic quantities are normal and the system real-time state warning index is normal, the system comprehensive operation state level is: normal;

[0037] When there is one or more abnormal warning results for system state characteristic quantities and the system real-time state warning index is normal, the system comprehensive operation state level is: attention;

[0038] When the warning results of all system state characteristic quantities are normal and the system real-time state warning index is abnormal, the system comprehensive operation state level is: abnormal;

[0039] When the system real-time state warning index exceeds the threshold and there is one or more abnormal warning results for system state characteristic quantities at the same time, the system comprehensive operation state level is: abnormal.

[0040] The present invention also discloses an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the intelligent diagnosis and real-time fault warning analysis method for the above-mentioned top cover drainage system are implemented.

[0041] The present invention also discloses a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the intelligent diagnosis and real-time fault warning analysis method for the above-mentioned top cover drainage system are implemented.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] (1) Aiming at the deficiencies of the current fault warning method for the top cover drainage system of hydropower units, based on the Euclidean space clustering model and Gaussian threshold, the Euclidean space clustering center method is combined with the Gaussian threshold method to conduct real-time risk assessment and fault trend warning on the operation status of the top cover drainage system;

[0044] (2) Utilizing the historical health status monitoring signals of the unit automatic control system, fully considering the complexity of the monitoring values of the top cover drainage system and the variability of operating conditions, designing the judgment logic of each state characteristic quantity, and extracting the state characteristic quantities that can reflect the overall state of the system;

[0045] (3) Introducing a microservice architecture, combining the above-mentioned fault warning method, developing a warning system on the microservice architecture. The system integration follows the principles of security, stability, portability, openness, modularity, etc., and uses the real monitoring data of hydropower stations for development and testing, promoting the transformation of the theoretical results of the fault warning method to the application level;

[0046] (4) Making full use of the massive historical data accumulated by the unit monitoring system, integrating the idea of big data, and at the same time being able to use the unit historical data for self-learning, updating the warning model of the top cover drainage system, and improving the warning accuracy.

[0047] (5) Through the research and application of the fault warning analysis method for the top cover drainage system of hydropower units, before the failure of the automatic control system occurs, most of the operation information, health status, and development trend of the top cover drainage system can be mastered, guiding the operation and maintenance personnel to conduct regulation, extending the equipment operation cycle, ensuring the effective performance of the equipment of the entire hydropower station, minimizing the probability of accidents as much as possible, preventing failures from occurring or reducing the impact and consequences of failures, providing a basis for maintenance decisions, making the maintenance work purposeful and scientific, shortening the maintenance time, and at the same time providing a basis for formulating a reasonable detection and maintenance system, which is the guarantee for the safe and reliable operation of hydropower stations and obtaining great economic and social benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1Flowchart of an intelligent diagnosis and real-time fault warning analysis method for a top cover drainage system provided by the present invention;

[0049] Figure 2 Flowchart of an embodiment of an intelligent diagnosis and real-time fault warning analysis method for a top cover drainage system provided by the present invention;

[0050] Figure 3 Graph of the real-time change trend of the warning indicators of the top cover drainage system in the specific application embodiment provided by the present invention;

[0051] Figure 4 Structural block diagram of an embodiment of an intelligent diagnosis and real-time fault warning analysis system for a top cover drainage system provided by the present invention. Detailed implementation manners

[0052] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0053] As Figure 1 shown, an intelligent diagnosis and real-time fault warning analysis method for a top cover drainage system includes the following steps:

[0054] Step 1: Process the monitored data of the top cover drainage system collected to obtain various state characteristic quantities associated with the state of the top cover drainage system;

[0055] Step 2: Normalize each state characteristic quantity, and then construct a state vector representing the health state of the top cover drainage system. A set of state vectors is composed of multiple state vectors;

[0056] Step 3: Calculate the mean vector of the set of state vectors in the normal state of the top cover drainage system to obtain the system health state clustering center vector;

[0057] Step 4: Obtain the set of state vectors in the healthy state of the top cover drainage system according to the above steps 1-2. Calculate the relative Euclidean distance between each vector in the vector set and the healthy state clustering center vector as the state warning index of the top cover drainage system. Calculate the average value (MEAN) and standard deviation (STD) of the Euclidean distance, and use MEAN + 3*STD as the fault warning threshold;

[0058] Step 5: Obtain the set of state vectors of the real-time monitoring signal of the top cover drainage system according to the above steps 1-2. Calculate the relative Euclidean distance between each vector in the vector set and the healthy state clustering center vector as the real-time state warning index of the top cover drainage system.

[0059] An intelligent diagnosis and real-time fault warning analysis method for the top cover drainage system provided by the present invention, aiming at the deficiencies of the current fault warning methods for the top cover drainage system of hydropower units, combines the Euclidean space clustering center method with the Gaussian threshold method based on the Euclidean space clustering model and the Gaussian threshold to conduct real-time risk assessment and fault trend warning on the operation status of the top cover drainage system; uses the historical health status monitoring signals of the unit automatic control system, fully considers the complexity of the monitoring values of the top cover drainage system and the variability of the operating conditions, designs the judgment logic of each state characteristic quantity, and extracts the state characteristic quantities that can reflect the overall state of the system; introduces the microservice architecture, combines the fault warning method, develops a warning system on the microservice architecture, and the system integration follows the principles of security, stability, portability, openness, modularity, etc., and uses the real monitoring data of the hydropower station for development and testing to promote the transformation of the theoretical results of the fault warning method to the application level; makes full use of the massive historical data accumulated by the unit monitoring system, incorporates the big data concept, and can also use the unit historical data for self-learning to update the warning model of the top cover drainage system and improve the warning accuracy.

[0060] Preferably, the normalization method in step 2 uses the maximum-minimum method for linear normalization, and the formula is:

[0061] 。

[0062] Further, the process of constructing the state vector and the state vector set in step 2 is as follows:

[0063] If there are n state characteristic quantities in the state characteristic quantity system, after density unification, each state characteristic quantity has m data respectively, then the first state vector will be an n-dimensional vector:

[0064] ;

[0065] Further, the state vector set in step 2 is:

[0066] ;

[0067] That is

[0068] 。

[0069] Preferably, the clustering method in steps 4 and 5 uses the Euclidean space clustering model to calculate the relative Euclidean distance between each vector in the vector set and the healthy state clustering center vector as the state warning index of the top cover drainage system.

[0070] Preferably, the threshold value in step 4 is determined using the Gaussian threshold rule, calculating the average value (MEAN) and standard deviation (STD) of the Euclidean distance, and using MEAN + 3 * STD as the fault warning threshold value.

[0071] In addition, the present invention also discloses an intelligent diagnosis and real-time fault warning analysis system for a top cover drainage system, and the system includes: a database module, a data cache module, a data processing module, and a real-time fault warning module;

[0072] The database module is used to store the original data of the measuring points of the hydro-generator set, as well as various state characteristic quantity data, system health state clustering center vectors, system fault warning threshold values, system historical state warning indicators and other data results obtained through processing by the data processing module;

[0073] The data cache module is used to store real-time data such as real-time state warning indicators of the top cover drainage system of the hydro-generator set obtained through processing by the data processing module;

[0074] The data processing module is a background algorithm program of the system. On the one hand, it is used to collect the original data of the measuring points in the database module and perform arithmetic processing on it to obtain the state characteristic quantities of the top cover drainage system, the state vectors in the system health state, the system health state clustering center vectors, the system fault warning threshold values, and the system real-time state warning indicators in steps 1 - 5, and complete the initialization of the data; on the other hand, it is connected to the database module, the data cache module, and the real-time fault warning module, and serves as an information transmission hub between the modules, and sends the processed data to each module through the interface respectively.

[0075] The real-time fault warning module is used to implement the display of an intelligent diagnosis and real-time fault warning analysis method for a top cover drainage system.

[0076] Furthermore, the real-time fault warning module includes the following parts: a historical data monitoring part, a single state characteristic quantity warning part, an Euclidean distance warning index warning part, a system comprehensive warning part, and a warning information part;

[0077] The historical data monitoring part receives and displays the historical state characteristic quantity data of the database;

[0078] The single state characteristic quantity warning part evaluates whether a single state characteristic quantity is in a normal state. If the real-time value of the state characteristic quantity is not within the normal value range, a warning is issued for the single state characteristic quantity;

[0079] The Euclidean distance warning index warning part, by comparing the system real-time state warning index based on the Euclidean space clustering model and the system fault warning threshold value, if the warning index exceeds the threshold value, a warning is issued for the overall state of the system;

[0080] The system comprehensive early warning part is associated with the single state characteristic quantity early warning part and the Euclidean distance early warning index early warning part, and determines the system comprehensive operation state level (three levels: normal, attention, abnormal) according to the early warning results of the above two parts;

[0081] The early warning information part is associated with the single state characteristic quantity early warning part, the Euclidean distance early warning index early warning part and the system comprehensive early warning part, and displays the system operation state information at the moment of abnormal early warning.

[0082] Furthermore,

[0083] The logic for determining the system comprehensive operation state level is as follows:

[0084] When the early warning results of all system state characteristic quantities are normal and the system real-time state early warning index is normal, the system comprehensive operation state level is: normal;

[0085] When there is one or more abnormal early warning results of system state characteristic quantities and the system real-time state early warning index is normal, the system comprehensive operation state level is: attention;

[0086] When the early warning results of all system state characteristic quantities are normal and the system real-time state early warning index is abnormal, the system comprehensive operation state level is: abnormal;

[0087] When the system real-time state early warning index exceeds the threshold and there is one or more abnormal early warning results of system state characteristic quantities at the same time, the system comprehensive operation state level is: abnormal.

[0088] The present invention also discloses an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the intelligent diagnosis and real-time fault early warning analysis method of the above roof drainage system are implemented.

[0089] The present invention also discloses a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the intelligent diagnosis and real-time fault early warning analysis method of the above roof drainage system are implemented.

[0090] The following is illustrated by specific embodiments:

[0091] Embodiment 1

[0092] Embodiment 1 provided by the present invention is an embodiment of the intelligent diagnosis and real-time fault early warning analysis method of a roof drainage system provided by the present invention. As Figure 2 shown is the flowchart of the embodiment of the intelligent diagnosis and real-time fault early warning analysis method of the roof drainage system provided by the present invention. From Figure 2It can be seen that this embodiment includes:

[0093] (1) Determine the fault warning threshold of the top cover drainage system

[0094] Step 1: Determine the unit model, obtain the historical health monitoring data of the top cover drainage system of the unit, and process the collected historical health monitoring data of the top cover drainage system according to the processing logic of each state characteristic quantity in the constructed system state characteristic quantity system to obtain each state characteristic quantity associated with the state of the top cover drainage system.

[0095] Preferably, the construction of the state characteristic quantity system may include: pump pumping efficiency, pump operating state, number of times of top cover water level analog deviation, availability rate of the top cover drainage system, etc.

[0096] The pump pumping efficiency, that is, the value of the water level change per unit time, requires that the pump is in the automatic mode, and the water level reaches the pump start water level when starting the pump and the water level reaches the pump stop water level when stopping the pump. The pump pumping efficiency formula is:

[0097] ;

[0098] In the formula, is the pump start water level, is the pump stop water level, is the pump start time, is the pump stop time.

[0099] The pump operating state, that is, the ratio of the statistical analysis of the pumping time length of a single pump each time during the evaluation period to the average value of all its pumping time lengths during the evaluation period. It requires that the pump is in the automatic mode, and the water level reaches the pump start water level when starting the pump and the water level reaches the pump stop water level when stopping the pump.

[0100] Assume that the pump runs n times during the evaluation period, then the start time of each pump can be extracted through the pump start-stop switch quantity monitoring value , , , ……, and the stop time , , , ……, .

[0101] Then the qualified rate formula for the a-th pump operation is:

[0102] ;

[0103] In the formula, is the start time of the a-th pump operation, is the stop time of the a-th pump operation.

[0104] The operating status of the water pump is the overall qualification rate, and the formula is:

[0105] .

[0106] Step 2: Normalize each state characteristic quantity, and then construct a state vector representing the healthy state of the top cover drainage system. A set of state vectors is composed of multiple state vectors.

[0107] For the normalization method, linear normalization is performed using the maximum-minimum method, and the formula is:

[0108] ;

[0109] If there are n state characteristic quantities in the state characteristic quantity system of the top cover drainage system constructed in Step 1, the constructed state vector is an n-dimensional vector:

[0110] ;

[0111] A set of state vectors is composed of multiple state vectors:

[0112] ;

[0113] That is

[0114] .

[0115] Step 3: Calculate the mean vector of the set of state vectors in the normal state of the top cover drainage system to obtain the system health state clustering center vector :

[0116] ;

[0117] Step 4: According to Steps 1-2, obtain the set of state vectors in the healthy state of the top cover drainage system, calculate the relative Euclidean distance between each vector in the vector set and the healthy state clustering center vector, which is used as the state warning index of the top cover drainage system, and calculate the average value (MEAN) and standard deviation (STD) of the Euclidean distance. Use MEAN + 3 * STD as the fault warning threshold;

[0118] Step 5: According to Steps 1-2, obtain the set of state vectors of the real-time monitoring signal of the top cover drainage system, and calculate the relative Euclidean distance between each vector in the vector set and the healthy state clustering center vector, which is used as the real-time state warning index of the top cover drainage system.

[0119] Embodiment 2

[0120] Example 2 provided by the present invention is a specific application example of the intelligent diagnosis and real-time fault warning analysis method for the top cover drainage system provided by the present invention. In the specific application example of the intelligent diagnosis and real-time fault warning analysis method for the top cover drainage system provided by the present invention, taking a certain power station as an example, the specific implementation cases and effects of the present invention are introduced. The operation status of the top cover drainage system of Unit #1 of this power station in the first ten days of July 2021 is analyzed.

[0121] The top cover drainage system of Unit #1 is equipped with 3 drainage pumps. Since the operating conditions of each drainage pump are relatively stable, according to the state characteristic quantity system of the top cover drainage system, the historical health monitoring data of each pump is processed separately, and the state vector dimension is determined to be 4 dimensions (pump pumping efficiency, pump operating status, number of abnormal pump operations, number of over-limit water level analog quantities), and 252 groups of state characteristic quantity data are constructed, including 85 groups of data for Pump #1, 83 groups of data for Pump #2, and 84 groups of data for Pump #3, which are used to calculate the health state clustering center vector of the top cover drainage system and the system threshold.

[0122] Taking the data of Pump #1 as an example, each group of state characteristic quantity data is linearly normalized by the maximum-minimum method and a state vector is constructed:

[0123] ;

[0124] A state vector set is constructed:

[0125] 。

[0126] Using the Euclidean distance clustering algorithm in data mining technology, each group of state vectors is clustered to obtain the health state clustering center vector 。 The state vector of the top cover drainage system and the health state clustering center vector are shown in Table 1.

[0127] Table 1 State vector of the top cover drainage system and the health state clustering center vector

[0128]

[0129] The state characteristic quantities of the data of Unit #1 of this power station in the first ten days of July 2021 are extracted, with a total of 225 groups, including 76 groups of data for Pump #1, 74 groups of data for Pump #2, and 75 groups of data for Pump #3. The state vectors of each group of data are obtained by linear normalization using the maximum-minimum method , and the real-time state warning index of the top cover drainage system is calculated according to the steps described in item 5 above. The change trend of the warning index over time (signal sequence) is as shown in Figure 3 。

[0130] Through the above specific application embodiments, the method can relatively accurately judge the deterioration trend of the operating state of the top cover drainage system, thereby providing a fault warning function, which has strong guiding significance for realizing preventive maintenance of the top cover drainage system.

[0131] Embodiment 3

[0132] Embodiment 3 provided by the present invention is an embodiment of an intelligent diagnosis and real-time fault warning analysis system for a top cover drainage system provided by the present invention. As Figure 4 shown is the structural block diagram of an embodiment of an intelligent diagnosis and real-time fault warning analysis system for a top cover drainage system provided by the present invention. It can be seen from Figure 4 that this system includes: a database module 101, a data cache module 102, a data processing module 103, and a real-time fault warning module 104.

[0133] The database module 101 is used to store the original data of the measuring points of the hydro-generating unit, as well as various state characteristic quantity data, system health state clustering center vectors, system fault warning thresholds, system historical state warning indicators and other data results obtained through the processing of the data processing module.

[0134] The data cache module 102 is used to store real-time data such as real-time state warning indicators of the top cover drainage system of the hydro-generating unit obtained through the processing of the data processing module.

[0135] The data processing module 103 is the background algorithm program of the system. On the one hand, it is used to collect the original data of the measuring points in the database module and perform arithmetic processing on it to obtain the state characteristic quantities of the top cover drainage system, the state vectors in the system health state, the system health state clustering center vectors, the system fault warning thresholds, and the system real-time state warning indicators in steps 1 - 5, and complete the initialization of the data; on the other hand, it is connected to the database module, the data cache module, and the real-time fault warning module, and serves as the information transmission hub between the modules, and sends the processed data to each module through the interface respectively.

[0136] The real-time fault warning module 104 is used to realize the display of an intelligent diagnosis and real-time fault warning analysis method for a top cover drainage system. The real-time fault warning module 104 includes the following parts: a historical data monitoring part 201, a single state characteristic quantity warning part 202, a Euclidean distance warning index warning part 203, a system comprehensive warning part 204, and a warning information part 205.

[0137] The historical data monitoring part 201 is used to receive and display the historical state characteristic quantity data of the database.

[0138] The single - state characteristic quantity warning part 202 is used to evaluate whether a single - state characteristic quantity is in a normal state. If the real - time value of the state characteristic quantity is not within the normal value range, a warning for the single - state characteristic quantity is given.

[0139] The Euclidean - distance warning index warning part 203 is used to compare the system real - time state warning index based on the Euclidean - space clustering model and the system fault warning threshold. If the warning index exceeds the threshold, a warning for the overall system state is given.

[0140] The system comprehensive warning part 204 is used to associate the single - state characteristic quantity warning part 202 and the Euclidean - distance warning index warning part 203, and determine the system comprehensive operation state level (three levels: normal, attention, abnormal) according to the warning results of 202 and 203.

[0141] System comprehensive operation state level judgment logic:

[0142] 1) When the warning results of all system state characteristic quantities are normal and the system real - time state warning index is normal, the system comprehensive operation state level is: normal;

[0143] 2) When there is one or more abnormal warning results of system state characteristic quantities and the system real - time state warning index is normal, the system comprehensive operation state level is: attention;

[0144] 3) When the warning results of all system state characteristic quantities are normal and the system real - time state warning index is abnormal, the system comprehensive operation state level is: abnormal;

[0145] 4) When the system real - time state warning index exceeds the threshold and there is one or more abnormal warning results of system state characteristic quantities at the same time, the system comprehensive operation state level is: abnormal.

[0146] The warning information part is used to associate the single - state characteristic quantity warning part 202, the Euclidean - distance warning index warning part 203 and the system comprehensive warning part 204, and display the system operation state information at the moment of abnormal warning.

[0147] An embodiment of an intelligent diagnosis and real-time fault warning analysis method and system for a top cover drainage system provided by the present invention can be executed and implemented through a computer program. For example, it includes: Step 1, processing the monitored data of the top cover drainage system collected to obtain various state characteristic quantities associated with the state of the top cover drainage system; Step 2, performing normalization processing on each state characteristic quantity, and then constructing a state vector representing the healthy state of the top cover drainage system, and multiple state vectors form a state vector set; Step 3, calculating the mean vector of the state vector set in the normal state of the top cover drainage system to obtain the system healthy state clustering center vector; Step 4, obtaining the state vector set in the healthy state of the top cover drainage system according to the above Steps 1-2, calculating the relative Euclidean distance between each vector in the vector set and the healthy state clustering center vector as the state warning index of the top cover drainage system, calculating the average value (MEAN) and standard deviation (STD) of the Euclidean distance, and using MEAN + 3*STD as the fault warning threshold; Step 5, obtaining the state vector set of the real-time monitoring signal of the top cover drainage system according to the above Steps 1-2, calculating the relative Euclidean distance between each vector in the vector set and the healthy state clustering center vector as the real-time state warning index of the top cover drainage system.

[0148] The embodiment of the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is implemented to execute the intelligent diagnosis and real-time fault warning analysis method of the top cover drainage system provided by the above embodiments. For example, it includes: Step 1, processing the monitored data of the top cover drainage system collected to obtain various state characteristic quantities associated with the state of the top cover drainage system; Step 2, performing normalization processing on each state characteristic quantity, and then constructing a state vector representing the healthy state of the top cover drainage system, and multiple state vectors form a state vector set; Step 3, calculating the mean vector of the state vector set in the normal state of the top cover drainage system to obtain the system healthy state clustering center vector; Step 4, obtaining the state vector set in the healthy state of the top cover drainage system according to the above Steps 1-2, calculating the relative Euclidean distance between each vector in the vector set and the healthy state clustering center vector as the state warning index of the top cover drainage system, calculating the average value (MEAN) and standard deviation (STD) of the Euclidean distance, and using MEAN + 3*STD as the fault warning threshold; Step 5, obtaining the state vector set of the real-time monitoring signal of the top cover drainage system according to the above Steps 1-2, calculating the relative Euclidean distance between each vector in the vector set and the healthy state clustering center vector as the real-time state warning index of the top cover drainage system.

[0149] The above embodiments are only the preferred technical solutions of the present invention and should not be regarded as limitations on the present invention. In the present application, the embodiments and the features in the embodiments can be arbitrarily combined with each other without conflict. The protection scope of the present invention shall be the technical solutions recorded in the claims, including the equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, the equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. An intelligent diagnosis and real-time fault early warning analysis method for a top cover drainage system, characterized in that: It includes the following steps: Step 1: Process the monitored data of the top cover drainage system to obtain various state characteristic quantities associated with the state of the top cover drainage system; Step 2: Normalize each state characteristic quantity, and then construct a state vector representing the health state of the top cover drainage system. Multiple state vectors form a state vector set; Step 3: Calculate the mean vector of the state vector set in the normal state of the top cover drainage system to obtain the system health state clustering center vector; Step 4: Obtain the state vector set in the healthy state of the top cover drainage system according to Steps 1 - 2. Calculate the relative Euclidean distance between each vector in the vector set and the healthy state clustering center vector as the state warning index of the top cover drainage system. Calculate the average value MEAN and standard deviation STD of the Euclidean distance, and use MEAN + 3 * STD as the fault warning threshold; Step 5: Obtain the state vector set of the real-time monitoring signal of the top cover drainage system according to Steps 1 - 2. Calculate the relative Euclidean distance between each vector in the vector set and the healthy state clustering center vector as the real-time state warning index of the top cover drainage system; The process of constructing the state vector and the state vector set in Step 2 is as follows: If there are n state characteristic quantities in the state characteristic quantity system, after density unification, each state characteristic quantity has m data respectively. Then the first state vector will be an n-dimensional vector: ； Furthermore, the state vector set in Step 2 is: ； That is 。 2. The intelligent diagnosis and real-time fault warning analysis method of the top cover drainage system according to claim 1, characterized in that: The normalization method in Step 2 uses the maximum-minimum method for linear normalization, and the formula is: 。 3. The intelligent diagnosis and real-time fault warning analysis method for the top cover drainage system according to claim 1, characterized in that: The clustering method in Steps 4 and 5 uses the Euclidean space clustering model. Calculate the relative Euclidean distance between each vector in the vector set and the healthy state clustering center vector as the state warning index of the top cover drainage system.

4. The intelligent diagnosis and real-time fault warning analysis method for the top cover drainage system according to claim 1, characterized in that: The threshold in Step 4 is determined using the Gaussian threshold rule. Calculate the average value MEAN and standard deviation STD of the Euclidean distance, and use MEAN + 3 * STD as the fault warning threshold.

5. An intelligent diagnosis and real-time fault warning analysis system for a top cover drainage system applying the intelligent diagnosis and real-time fault warning analysis method of the top cover drainage system according to any one of claims 1 to 4, characterized in that: The intelligent diagnosis and real-time fault warning analysis system of the top cover drainage system includes: a database module, a data cache module, a data processing module, and a real-time fault warning module; The database module is used to store the original data of the water turbine generator unit measurement points, as well as various state characteristic quantity data, system health state clustering center vector, system fault warning threshold, and system historical state warning index data results obtained through the processing of the data processing module; The data cache module is used to store the real-time data of the real-time state warning index of the water turbine generator unit top cover drainage system obtained through the processing of the data processing module; The data processing module is a background algorithm program of the system. On the one hand, it is used to collect the original data of the measuring points in the database module, and perform arithmetic processing on it to obtain the state characteristic quantities of the top cover drainage system, the state vector in the system healthy state, the system healthy state clustering center vector, the system fault warning threshold, and the system real-time state warning index in steps 1-5, completing the initialization of the data. On the other hand, it is connected to the database module, the data cache module, and the real-time fault warning module, serving as an information transmission hub between modules, and sending the processed data to each module through the interface respectively. The real-time fault warning module is used to display an intelligent diagnosis and real-time fault warning analysis method for a top cover drainage system.

6. The intelligent diagnosis and real-time fault warning analysis system of the top cover drainage system according to claim 5, characterized in that: The real-time fault warning module includes the following parts: historical data monitoring part, single state characteristic quantity warning part, Euclidean distance warning index warning part, system comprehensive warning part, and warning information part. The historical data monitoring part receives and displays the historical state characteristic quantity data of the database. The single state characteristic quantity warning part evaluates whether a single state characteristic quantity is in a normal state. If the real-time value of the state characteristic quantity is not within the normal value range, a warning is issued for the single state characteristic quantity. The Euclidean distance warning index warning part compares the system real-time state warning index based on the Euclidean space clustering model with the system fault warning threshold. If the warning index exceeds the threshold, a warning is issued for the overall state of the system. The system comprehensive warning part associates the single state characteristic quantity warning part and the Euclidean distance warning index warning part, and determines the system comprehensive operation state level according to the warning results of the above two parts. The system comprehensive operation state level includes three levels: normal, attention, and abnormal. The warning information part associates the single state characteristic quantity warning part, the Euclidean distance warning index warning part, and the system comprehensive warning part, and displays the system operation state information at the moment of abnormal warning.

7. The intelligent diagnosis and real-time fault warning analysis system for the top cover drainage system according to claim 6, characterized in that: The determination logic of the system comprehensive operation state level is as follows: When the warning results of all state characteristic quantities of the system are normal and the system real-time state warning index is normal, the system comprehensive operation state level is: normal; When there is one or more abnormal warning results of the state characteristic quantities of the system and the system real-time state warning index is normal, the system comprehensive operation state level is: attention; When the warning results of all state characteristic quantities of the system are normal and the system real-time state warning index is abnormal, the system comprehensive operation state level is: abnormal; When the system real-time state warning index exceeds the threshold and there is one or more abnormal warning results of the state characteristic quantities at the same time, the system comprehensive operation state level is: abnormal.

8. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that: When the processor executes the program, it implements the steps of the intelligent diagnosis and real-time fault warning analysis method for the top cover drainage system according to any one of claims 1 to 4.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the steps of the intelligent diagnosis and real-time fault warning analysis method for the top cover drainage system described in any one of claims 1 to 4.

Images