Hydropower station equipment fault detection system and method based on multi-source information fusion

Abstract

The invention discloses a hydropower station equipment fault detection system and method based on multi-source information fusion. The system comprises a central control host, a fault management server, an inspection robot, an environment monitoring sensor and a vibration sensor. The fault management server comprises a historical record database and a fault detection module; the fault detection module is used for performing fault diagnosis on hydropower station equipment; the equipment classification module is used for classifying the hydropower station equipment according to the fault phenomenon; and the inspection frequency setting module is used for setting the inspection frequency of the inspection robot. Different detection modes are set for different types of equipment, and the pertinence of fault detection is improved. The inspection frequency is adjusted according to the historical fault condition of the equipment, high-frequency faults can be found more easily, repeated inspection of low-frequency faults is avoided, and resources are saved. According to the method, the consistency of the frequencies of the multiple input data of the associated equipment fault model can be effectively ensured, and the normal operation of the fault model is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment fault detection, and in particular to a hydropower station equipment fault detection system and method based on multi-source information fusion. Background Art

[0002] Hydropower stations contain a large number of equipment of many types, and the equipment is often in a long-term high-intensity working state and the working environment is complex, so it is easy to malfunction. In the past, in order to deal with equipment failure problems, manual regular fault detection and troubleshooting methods were adopted. At present, with the gradual improvement of the level of intelligence, the relevant technologies for unmanned hydropower stations are gradually being promoted. However, the existing technology often uses a single method to detect faults in hydropower station equipment, such as through inspection robots, or by setting cameras around the equipment. Hydropower stations contain many types of equipment, and the fault characteristics of different equipment are not the same. This single method cannot cope with the actual situation and is prone to missed reports and false reports of faults.

[0003] In addition, inspection robots in the prior art usually have a pre-designed inspection route. When the inspection robot goes around the inspection route, it inspects all the equipment in turn. However, in reality, the failure rates of hydropower station equipment vary. Some equipment is prone to failure, while some equipment rarely fails. If the same inspection frequency is set for all equipment, high-incidence failures cannot be quickly discovered. Repeated inspections of equipment that is not prone to failure are also a waste of resources.

[0004] In addition, the existing technology often uses machine learning algorithms when performing fault diagnosis, and performs fault diagnosis based on deep learning models through the collected device image data, temperature data and / or vibration data. However, the model is often for a single device, and rarely for multiple devices. In fact, some faults are caused by the incoordination of the configuration and working parameters of multiple related devices, rather than a single device. When faced with a model containing multiple devices, a new problem arises that the sampling frequencies of multiple devices are not the same, which leads to inconsistent frequencies of the input data of the diagnostic model, making it difficult for the diagnostic model to work properly.

[0005] In the prior art, the invention patent CN112435361B proposes a substation joint inspection method and system. According to the main equipment code of the linkage signal of the main and auxiliary equipment monitoring system, the associated physical ID and equipment point list are found; according to the equipment point list, the associated linkage strategy list is obtained from the linkage strategy configuration; it is judged whether the linkage strategy list is empty. If the linkage strategy list is not empty, a substation equipment joint inspection task is constructed according to the physical ID, equipment point list and linkage strategy list, and multiple optical acquisition devices and multiple robots perform concurrent inspections in combination with environmental sensing signals; if the linkage strategy list is empty, the linkage ends. This invention realizes the full-coverage inspection operation of indoor and outdoor equipment of the substation by various detection means, and improves the comprehensiveness and intelligence of substation equipment inspection. However, this invention does not solve the technical problems of multi-device classification detection, inspection frequency setting, and multi-device joint fault diagnosis model mentioned above. Summary of the Invention

[0006] Object of the Invention: Aiming at the above problems, the present invention proposes a hydropower station equipment fault detection system and method based on multi-source information fusion.

[0007] Technical Solution:

[0008] In the first aspect, the present invention proposes a hydropower station equipment fault detection system based on multi-source information fusion, including a central control host, a fault management server, an inspection robot, an environmental monitoring sensor, and a vibration sensor;

[0009] The central control host includes an equipment classification module and an inspection frequency setting module;

[0010] The fault management server includes a historical record database and a fault detection module;

[0011] The inspection robot is used to inspect and collect infrared image data and sound data of hydropower station equipment;

[0012] Preferably, the historical record database is used to store historical fault data, and the historical fault data includes a fault equipment number, a fault occurrence time, a fault phenomenon, a fault handling measure, a fault repair cost, and a fault downtime;

[0013] The fault detection module is used to diagnose faults of hydropower station equipment; the equipment classification module is used to classify hydropower station equipment according to fault phenomena;

[0014] The inspection frequency setting module is used to set the inspection frequency of the inspection robot.

[0015] Preferably, the fault detection module includes a number of single-device fault models and associated-device fault models. The single-device fault model is used for fault diagnosis of faults involving a single device, and the associated-device fault model is used for fault diagnosis of faults involving multiple associated devices.

[0016] Preferably, the inspection frequency setting module includes a fault frequency acquisition module, a fault cost acquisition module, a detection priority acquisition module, and a detection frequency acquisition module;

[0017] The fault frequency acquisition module is used to determine the device fault frequency of each device within a preset time period according to the fault device number and the fault occurrence time;

[0018] The fault cost acquisition module is used to determine the fault cost according to the fault repair cost and the fault downtime;

[0019] The detection priority acquisition module is used to determine the detection priority of the inspection robot for each device according to the device fault frequency and the fault cost;

[0020] The detection frequency acquisition module is used to determine the inspection frequency of the inspection robot for each device, as well as the detection frequencies of the environmental monitoring sensor and the vibration sensor, according to the detection priority.

[0021] In a second aspect, the present invention also provides a method for detecting faults in hydropower station equipment based on multi-source information fusion. The method includes:

[0022] Step 1, obtain the historical fault data of the hydropower station equipment;

[0023] The historical fault data includes the fault device number, the fault occurrence time, the fault phenomenon, the fault handling measures, the fault repair cost, and the fault downtime;

[0024] Step 2, classify the hydropower station equipment according to the fault phenomenon in the historical fault data;

[0025] Step 3, set the detection method according to the equipment classification;

[0026] Step 4, obtain the equipment detection priority based on the historical fault data;

[0027] Step 5, determine the equipment detection frequency according to the detection priority;

[0028] Step 6, perform fault diagnosis based on the single-device fault model;

[0029] Step 7, perform fault diagnosis based on the associated-device fault model;

[0030] Step 8, handle the fault and record it in the historical record database.

[0031] Preferably, step 2, classifying the hydropower station equipment according to the fault phenomena in the historical fault data includes:

[0032] Step 21, if the fault phenomenon of the hydropower station equipment is abnormal temperature, it is classified as type A equipment;

[0033] Step 22, if the fault phenomenon of the hydropower station equipment is abnormal sound, it is classified as type B equipment;

[0034] Step 23, if the fault phenomenon of the hydropower station equipment is abnormal vibration, it is classified as type C equipment.

[0035] Preferably, step 3, setting the detection method according to the equipment classification situation includes:

[0036] Step 31, collecting the temperature data and sound data around the equipment through the environmental monitoring sensors beside the equipment;

[0037] Step 32, if the equipment is type A equipment, collecting the infrared image data of the equipment through the regular inspection by the inspection robot;

[0038] Step 33, if the equipment is type B equipment, collecting the sound data of the equipment through the regular inspection by the inspection robot;

[0039] Step 34, if the equipment is type C equipment, collecting the vibration data of the equipment through the vibration sensors set at the equipment site.

[0040] Preferably, step 4, obtaining the equipment detection priority based on the historical fault data includes:

[0041] Step 41, determining the equipment fault frequency F of each equipment within a preset time period according to the fault equipment number and the fault occurrence time;

[0042] Step 42, determining the fault cost C according to the fault repair cost M and the fault shutdown duration T;

[0043] C = α * M + β * T

[0044] where α and β are preset coefficients;

[0045] Step 43, determining the detection priority P of the inspection robot for each equipment according to the equipment fault frequency F and the fault cost C;

[0046] P = [γ * F * C]

[0047] where γ is a preset coefficient and [] is the rounding symbol.

[0048] Preferably, step 5, determining the equipment detection frequency according to the detection priority includes:

[0049] Step 51. Determine the inspection frequency F of the inspection robot for each device R ;

[0050] F R = F0 + P

[0051] where F0 is the preset basic inspection frequency;

[0052] Step 52. Determine the detection frequency F of the environmental monitoring sensor and the vibration sensor S ;

[0053] Statistically analyze the inspection frequency F of each device R to obtain the least common multiple value F RB , then:

[0054] F S = n * F RB

[0055] where n is a positive integer greater than 1.

[0056] Preferably, step 6. Fault diagnosis based on the single-device fault model includes:

[0057] Input the detection data of each device into the single-device fault model of the corresponding device, and output the fault detection result; the detection data of each device includes the inspection data of the inspection robot and / or the detection data of the vibration sensor;

[0058] For type A devices, the input data of the single-device fault model is the temperature characteristic data obtained from the infrared image data;

[0059] For type B devices, the input data of the single-device fault model is the sound characteristic data obtained from the sound data;

[0060] For type C devices, the input data of the single-device fault model is the vibration characteristic data obtained from the vibration data.

[0061] Preferably, step 7. Fault diagnosis based on the associated-device fault model includes:

[0062] Step 71. Determine whether the inspection frequencies of each device in the associated-device fault model are the same; if they are the same, extract the characteristic data from the inspection data of each device as the input data of the associated-device fault model for fault diagnosis; if they are not the same, go to step 72;

[0063] Step 72. Obtain the least common multiple F of the inspection frequencies of each device included in the associated-device model B ;

[0064] Step 73. Expand the equipment inspection data in the associated equipment model through a fitting algorithm, and expand the frequency of the inspection data of each equipment to F B , including:

[0065] For the i-th equipment in the associated equipment failure model, its inspection data within one cycle is (x i,1 , x i,2 , ……, ) where F Ri is the inspection frequency of the i-th equipment, 1 ≤ i ≤ m, and m is the number of equipment included in the associated equipment failure model;

[0066] The inspection data expanded through the fitting algorithm is M = (x i,1 , x i,11 , …, x i,2 , x i,21 , …, x i,k , …, x i,kj , …, …); where x i,k is the inspection data, x i,kj is the fitted data, 1 ≤ k ≤ F Ri ,

[0067] Step 74. Correct the fitted data based on the detection data of the environmental monitoring sensor to obtain the corrected expanded data, including:

[0068] Perform segmented processing on the detection data sequence of the environmental monitoring sensor in chronological order, equally divide it into F B segments of data, each segment of data corresponds to an inspection data in M, and the number of data in each segment is Then the corrected data of the fitted data is:

[0069]

[0070] where t i,k is the mean value of the data segment corresponding to the inspection data x i,k , t i,kj is the mean value of the data segment corresponding to the inspection data x i,kj , and δ is a preset parameter;

[0071] Then the corrected expanded data is M ’ = (x i,1 , x ’ i,11 , …, x i,2 , x ’ i,21 , …, x i,k , …, x ’ i,kj , …, …);

[0072] Step 75: Extract feature data from the corrected extended data as the input data for the associated equipment fault model, and perform fault diagnosis.

[0073] The present invention has the following beneficial effects compared with the prior art:

[0074] 1. The present invention classifies equipment according to fault phenomena for various different types of equipment in a hydropower station, and sets different detection methods for equipment of different classifications, thereby improving the pertinence of fault detection and being able to improve the accuracy of fault diagnosis.

[0075] 2. The present invention adjusts the inspection frequency of equipment according to the fault frequency of the equipment, the cost and time spent on faults. For equipment that is prone to faults and has a greater negative impact after faults, its inspection frequency is increased; for equipment that is not prone to faults and has a smaller impact after faults, the inspection frequency is adjusted to a lower level. This setting can more easily detect high-incidence faults and avoid repeatedly inspecting low-frequency faults, effectively saving resources.

[0076] 3. The fault diagnosis model of the present invention includes a single equipment fault diagnosis model and an associated equipment fault model diagnosis, which can perform fault diagnosis on a single equipment and can also perform fault diagnosis on multiple associated equipment. Moreover, when the present invention performs fault diagnosis on associated equipment, a data supplement and correction mechanism is set up, which can effectively ensure the consistency of the frequencies of multiple input data of the associated equipment fault model and ensure the normal operation of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 It is a schematic structural diagram of a hydropower station equipment fault detection system based on multi-source information fusion provided by an embodiment of the present invention;

[0078] Figure 2 It is a flowchart of a hydropower station equipment fault detection method based on multi-source information fusion provided by an embodiment of the present invention;

[0079] Figure 3 It is a flowchart of a method for performing fault diagnosis based on an associated equipment fault model provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0080] Obviously, many modifications and changes made by those skilled in the art based on the purpose of the present invention fall within the protection scope of the present invention.

[0081] Those skilled in the art can understand that, unless specifically stated otherwise, the singular forms "a", "an", "the" and "said" used herein may also include the plural forms. It should be further understood that the term "comprising" used in the specification of the present invention means the presence of the described features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their groups. It should be understood that when an element or component is referred to as being "connected" to another element or component, it can be directly connected to the other element or component, or there may also be intermediate elements or components. The phrase "and / or" used herein includes any and all combinations of one or more of the associated listed items.

[0082] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0083] Embodiment 1:

[0084] The embodiment of the present invention provides a hydropower station equipment fault detection system based on multi-source information fusion. Specifically, please refer to Figure 1 , Figure 1 which is a schematic structural diagram of a hydropower station equipment fault detection system provided by the embodiment of the present invention. The system includes:

[0085] a central control host, a fault management server, an inspection robot, an environmental monitoring sensor, and a vibration sensor;

[0086] The central control host includes an equipment classification module and an inspection frequency setting module;

[0087] The fault management server includes a historical record database and a fault detection module;

[0088] The inspection robot is used to inspect and collect infrared image data and sound data of the hydropower station equipment;

[0089] Preferably, the historical record database is used to store historical fault data, and the historical fault data includes a fault equipment number, a fault occurrence time, a fault phenomenon, a fault handling measure, a fault repair cost, and a fault shutdown duration;

[0090] The fault detection module is used to diagnose faults of the hydropower station equipment; the equipment classification module is used to classify the hydropower station equipment according to the fault phenomenon;

[0091] The inspection frequency setting module is used to set the inspection frequency of the inspection robot.

[0092] Preferably, the fault detection module includes a number of single-device fault models and associated-device fault models. The single-device fault model is used to diagnose faults of a single device, and the associated-device fault model is used to diagnose faults of multiple associated devices.

[0093] Preferably, the inspection frequency setting module includes a fault frequency acquisition module, a fault cost acquisition module, a detection priority acquisition module, and a detection frequency acquisition module;

[0094] The fault frequency acquisition module is used to determine the device fault frequency of each device within a preset time period according to the fault device number and the fault occurrence time;

[0095] The fault cost acquisition module is used to determine the fault cost according to the fault repair cost and the fault downtime;

[0096] The detection priority acquisition module is used to determine the detection priority of the inspection robot for each device according to the device fault frequency and the fault cost;

[0097] The detection frequency acquisition module is used to determine the inspection frequency of the inspection robot for each device, as well as the detection frequencies of the environmental monitoring sensor and the vibration sensor, according to the detection priority.

[0098] Embodiment 2:

[0099] The embodiment of the present invention also provides a method for detecting faults of hydropower station equipment based on multi-source information fusion. For details, please refer to Figure 2 , Figure 2 is a flowchart of a method for detecting faults of hydropower station equipment based on multi-source information fusion provided by the embodiment of the present invention. The method includes the steps:

[0100] Step 1, obtain the historical fault data of the hydropower station equipment;

[0101] The historical fault data includes the fault device number, the fault occurrence time, the fault phenomenon, the fault handling measures, the fault repair cost, and the fault downtime;

[0102] Step 2, classify the hydropower station equipment according to the fault phenomenon in the historical fault data;

[0103] Preferably, the step 2 of classifying the hydropower station equipment according to the fault phenomenon in the historical fault data includes:

[0104] Step 21, if the fault phenomenon of the hydropower station equipment is abnormal temperature, then classify it as type A equipment;

[0105] Step 22: If the fault phenomenon of the hydropower station equipment is abnormal noise, it is classified as Class B equipment;

[0106] Step 23: If the fault phenomenon of the hydropower station equipment is abnormal vibration, it is classified as Class C equipment.

[0107] It should be noted that there may be an overlap between the classifications of equipment. For example, some equipment in the hydropower station not only has an excessively high temperature but also makes abnormal noises during a fault. Such equipment is classified into both Class A equipment and Class B equipment.

[0108] Step 3: Set the detection method according to the equipment classification;

[0109] Preferably, Step 3: Setting the detection method according to the equipment classification includes:

[0110] Step 31: Collect the temperature data and sound data around the equipment through the environmental monitoring sensors beside the equipment;

[0111] Step 32: If the equipment is Class A equipment, collect the infrared image data of the equipment through regular inspections by the inspection robot;

[0112] Step 33: If the equipment is Class B equipment, collect the sound data of the equipment through regular inspections by the inspection robot;

[0113] Step 34: If the equipment is Class C equipment, collect the vibration data of the equipment through the vibration sensors set at the equipment site.

[0114] The present invention selects the equipment detection method corresponding to the characteristics of the fault phenomenon, which is more targeted and efficient, and avoids useless detections.

[0115] Step 4: Obtain the equipment detection priority based on historical fault data;

[0116] Preferably, Step 4: Obtaining the equipment detection priority based on historical fault data includes:

[0117] Step 41: Determine the equipment fault frequency F of each equipment within a preset time period according to the fault equipment number and the fault occurrence time;

[0118] For example, by analyzing the historical fault database, the number of times each fault equipment number appears within one year can be counted as the fault frequency of each equipment within the preset time period.

[0119] Step 42: Determine the fault cost C according to the fault repair cost M and the fault shutdown duration T;

[0120] C = α * M + β * T

[0121] where α and β are preset coefficients;

[0122] Step 43: Determine the detection priority P of the inspection robot for each device according to the equipment failure frequency F and the failure cost C;

[0123] P = [γ * F * C]

[0124] where γ is a preset coefficient, and [] is the rounding symbol.

[0125] Among them, the failure repair cost is the cost consumed for repairing a single failure of the equipment. For example, for a failure where components in the equipment need to be replaced, the failure repair cost is the purchase cost of the replaced components; however, for some failures, although the repair cost is not high, their repair requires multiple levels of approval and remote delivery, which takes a long time and affects the normal operation of the hydropower station for a long time. Therefore, in this application, the failure stop duration is taken as a consideration factor.

[0126] Step 5: Determine the equipment detection frequency according to the detection priority;

[0127] Preferably, Step 5: Determining the equipment detection frequency according to the detection priority includes:

[0128] Step 51: Determine the inspection frequency F of the inspection robot for each device R ;

[0129] F R = F0 + P

[0130] where F0 is the pre-set basic inspection frequency;

[0131] Step 52: Determine the detection frequency F of the environmental monitoring sensor and the vibration sensor S ;

[0132] Statistically analyze the inspection frequency F of each device R to obtain the least common multiple value F RB , then:

[0133] F S = n * F RB

[0134] where n is a positive integer greater than 1.

[0135] First, the present invention analyzes historical fault data, considers the detection priority of faulty equipment from the perspectives of fault frequency and fault cost, and gives higher-frequency inspections to equipment that is prone to failure or has greater adverse effects after failure. This can improve the efficiency of inspections and avoid waste of energy. In addition, for environmental monitoring sensors and vibration sensors, the data collection and upload frequency is also specially designed so that its detection frequency is an integer multiple of the highest inspection frequency, which is for facilitating the correction processing of data processing in the following text.

[0136] Step 6: Perform fault diagnosis based on the single-device fault model;

[0137] Preferably, step 6: performing fault diagnosis based on the single-device fault model includes:

[0138] Input the detection data of each device into the single-device fault model of the corresponding device to output the fault detection result; the detection data of each device includes the inspection data of the inspection robot and / or the detection data of the vibration sensor;

[0139] For type A devices, the input data of its single-device fault model is the temperature characteristic data obtained according to the infrared image data;

[0140] For type B devices, the input data of its single-device fault model is the sound characteristic data obtained according to the sound data;

[0141] For type C devices, the input data of its single-device fault model is the vibration characteristic data obtained according to the vibration data.

[0142] For different single-device fault models, the input data used is different; there are already many fault diagnosis models based on machine learning algorithms in the prior art. And, some devices have composite characteristics, and multiple abnormal data may occur during their failures. For example, when the temperature is too high and there is an abnormal sound during a failure, the sound characteristic data obtained according to the sound data and the temperature characteristic data obtained according to the infrared image data are used as the input data of the single-device fault model at the same time.

[0143] Step 7: Perform fault diagnosis based on the associated-device fault model;

[0144] Preferably, for performing fault diagnosis based on the associated-device fault model, please refer specifically to Figure 3 , Figure 3 which is a flowchart of a method for performing fault diagnosis based on the associated-device fault model provided by an embodiment of the present invention. This method includes steps:

[0145] Step 71: Determine whether the inspection frequencies of each device in the associated device failure model are the same. If they are the same, extract feature data from the inspection data of each device as the input data of the associated device failure model for failure diagnosis. If they are not the same, proceed to Step 72;

[0146] Optionally, if the associated device failure model also includes Class C devices, downsample the vibration data of Class C devices, and only select the data points corresponding to the inspection times in the vibration data sequence, so as to keep the frequencies of the input data of the associated device failure model consistent. For example, the associated device failure model includes three devices, a, b, and c. The detection data of devices a and b are both detection data (infrared temperature data or sound data) obtained by an inspection robot, and the detection data of device c are detection data collected by a vibration sensor. Moreover, the inspection frequencies of devices a and b are the same, but much lower than the detection frequency of device c. This results in different frequencies of the input data of the associated device failure model and cannot be directly input into the model. Therefore, the present invention processes the detection data sequence of device c and only selects the vibration data corresponding to the detection data of device a (or device b) in the sequence as the input data of the model, thus ensuring the consistency of the data frequencies of the three types of data as the model input.

[0147] Step 72: Obtain the least common multiple F of the inspection frequencies of each device included in the associated device model B ;

[0148] Step 73: Expand the device inspection data in the associated device model through a fitting algorithm, and expand the frequency of the inspection data of each device to F B , including:

[0149] For the i-th device in the associated device failure model, its inspection data in one cycle is (x i,1 , x i,2 , ……, ) where F Ri is the inspection frequency of the i-th device, 1 ≤ i ≤ m, and m is the number of devices included in the associated device failure model;

[0150] The inspection data expanded through the fitting algorithm is M = (x i,1 , x i,11 , …, x i,2 , x i,21 , …, x i,k , …, x i,kj , …, …); where x i,k is the inspection data, x i,kj is the fitted data, 1 ≤ k ≤ F Ri ,

[0151] Optionally, the least squares method is used to fit the inspection data, so that the frequency of the inspection data of each device included in the associated device model is expanded to the same frequency, that is, the least common multiple F B .

[0152] Step 74. Correct the fitted data based on the detection data of the environmental monitoring sensor to obtain the corrected and expanded data, including:

[0153] The detection data sequence of the environmental monitoring sensor is processed in segments in chronological order and equally divided into F B segments of data, each segment of data corresponds to an inspection data in M, and the number of data in each segment is Then the corrected data of the fitted data is:

[0154]

[0155] where t i,k is the mean value of the data segment corresponding to the inspection data x i,k t i,kj is the mean value of the data segment corresponding to the inspection data x i,kj , and δ is a preset parameter;

[0156] Then the corrected and expanded data is M ’ =(x i,1 , x ’ i,11 , …, x i,2 , x ’ i,21 , …, x i,k , …, x ’ i,kj , …, …);

[0157] Those skilled in the art can understand that when the inspection data is temperature characteristic data, the data used for correction is the data of the temperature sensor in the environmental monitoring sensor; when the inspection data is sound characteristic data, the data used for correction is the data of the sound sensor in the environmental monitoring sensor.

[0158] Step 75. Extract the characteristic data from the corrected and expanded data as the input data of the associated device fault model for fault diagnosis.

[0159] Step 8. Process the fault and record it in the historical record database.

[0160] After the fault record is stored in the database, the data in the historical record database is updated, which can ensure that the subsequent execution of this method can maintain an updated and more reasonable setting.

[0161] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the embodiments, reference can be made to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple. For the relevant parts, reference can be made to the description in the method section.

[0162] Those skilled in the art can further realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in the form of hardware or software depends on the specific application and design constraints of the technical solution. Skilled professionals can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0163] Finally, it should also be noted that in this article, relationships such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or further includes elements inherent to such process, method, article or device.

Claims

1. A hydropower station equipment fault detection system based on multi-source information fusion, comprising a central control host, a fault management server, an inspection robot, an environmental monitoring sensor, and a vibration sensor; the central control host includes an equipment classification module and an inspection frequency setting module; the fault management server includes a historical record database and a fault detection module; the inspection robot is used for inspecting and collecting infrared image data and sound data of hydropower station equipment; characterized in that, The historical record database is used to store historical fault data, which includes fault equipment numbers, fault occurrence times, fault phenomena, fault handling measures, fault repair costs, and fault downtime; the fault detection module is used to diagnose faults in hydropower station equipment; the equipment classification module is used to classify hydropower station equipment according to fault phenomena; the inspection frequency setting module is used to set the inspection frequency of inspection robots.

2. The hydroelectric power station equipment fault detection system based on multi-source information fusion according to claim 1, characterized in that The fault detection module includes a number of single-equipment fault models and associated-equipment fault models. The single-equipment fault model is used to diagnose faults involving a single piece of equipment, and the associated-equipment fault model is used to diagnose faults involving multiple associated pieces of equipment.

3. The hydroelectric power station equipment fault detection system based on multi-source information fusion according to claim 2, characterized in that, The inspection frequency setting module includes a fault frequency acquisition module, a fault cost acquisition module, a detection priority acquisition module, and a detection frequency acquisition module; The fault frequency acquisition module is used to determine the equipment fault frequency of each piece of equipment within a preset time period based on the fault equipment number and the fault occurrence time; The fault cost acquisition module is used to determine the fault cost based on the fault repair cost and the fault downtime; The detection priority acquisition module is used to determine the detection priority of the inspection robot for each piece of equipment based on the equipment fault frequency and the fault cost; The detection frequency acquisition module is used to determine the inspection frequency of the inspection robot for each piece of equipment, as well as the detection frequencies of environmental monitoring sensors and vibration sensors, based on the detection priority.

4. A method for detecting faults in hydropower station equipment based on multi-source information fusion, which is applied to the system for detecting faults in hydropower station equipment based on multi-source information fusion according to any one of claims 1-3, characterized in that, The method includes: Step 1, obtain the historical fault data of hydropower station equipment; The historical fault data includes fault equipment numbers, fault occurrence times, fault phenomena, fault handling measures, fault repair costs, and fault downtime; Step 2, classify the hydropower station equipment according to the fault phenomena in the historical fault data; Step 3, set the detection method according to the equipment classification; Step 4, obtain the equipment detection priority based on the historical fault data; Step 5, determine the equipment detection frequency according to the detection priority; Step 6, perform fault diagnosis based on the single-equipment fault model; Step 7, perform fault diagnosis based on the associated-equipment fault model; Step 8, handle the fault and record it in the historical record database.

5. The method for detecting faults in hydropower station equipment based on multi-source information fusion according to claim 4, characterized in that The said Step 2, classifying the hydropower station equipment according to the fault phenomena in the historical fault data includes: Step 21, if the fault phenomenon of the hydropower station equipment is abnormal temperature, then classify it as Class A equipment; Step 22, if the fault phenomenon of the hydropower station equipment is abnormal noise, then classify it as Class B equipment; Step 23, if the fault phenomenon of the hydropower station equipment is abnormal vibration, then classify it as Class C equipment.

6. The method for detecting faults of hydropower station equipment based on multi-source information fusion according to claim 5, characterized in that, The said Step 3, setting the detection method according to the equipment classification includes: Step 31, collect the temperature data and sound data around the equipment through the environmental monitoring sensors beside the equipment; Step 32, if the equipment is Class A equipment, then collect the infrared image data of the equipment through the regular inspection of the inspection robot; Step 33, if the equipment is Class B equipment, then collect the sound data of the equipment through the regular inspection of the inspection robot; Step 34, if the equipment is Class C equipment, then collect the vibration data of the equipment through the vibration sensors set at the equipment site.

7. The method for detecting faults of hydropower station equipment based on multi-source information fusion according to claim 6, characterized in that, Step 4, obtaining the device detection priority based on historical failure data includes: Step 41, determining the device failure frequency F of each device within a preset time period according to the failure device number and the failure occurrence time; Step 42, determining the failure cost C according to the failure repair cost M and the failure downtime T; C = α * M + β * T where α and β are preset coefficients; Step 43, determining the detection priority P of the inspection robot for each device according to the device failure frequency F and the failure cost C; P = [γ * F * C] where γ is a preset coefficient and [] is the rounding symbol.

8. The method for detecting faults of hydropower station equipment based on multi-source information fusion according to claim 7, characterized in that, Step 5, determining the device detection frequency according to the detection priority includes: Step 51, determine the inspection frequency F of the inspection robot for each device R ; F R = F0 + P where F0 is the pre-set basic inspection frequency; Step 52, determine the detection frequency F of the environmental monitoring sensor and the vibration sensor S ; The inspection frequency F for each device R is statistically analyzed to obtain the least common multiple value F RB , then: F S = n * F RB where n is a positive integer greater than 1.

9. The method for detecting faults in hydropower station equipment based on multi-source information fusion according to claim 8, characterized in that, Step 6, performing fault diagnosis based on the single-device fault model includes: Inputting the detection data of each device into the single-device fault model of the corresponding device to output a fault detection result; the detection data of each device includes the inspection data of the inspection robot and / or the detection data of the vibration sensor; For type A devices, the input data of its single-device fault model is the temperature feature data obtained according to the infrared image data; For type B devices, the input data of its single-device fault model is the sound feature data obtained according to the sound data; For type C devices, the input data of its single-device fault model is the vibration feature data obtained according to the vibration data.

10. The method for detecting faults of hydropower station equipment based on multi-source information fusion according to claim 9, characterized in that, Step 7, performing fault diagnosis based on the associated-device fault model includes: Step 71, judging whether the inspection frequencies of each device in the associated-device fault model are the same; if they are the same, extracting feature data from the inspection data of each device as the input data of the associated-device fault model for fault diagnosis; if they are not the same, go to Step 72; Step 72, obtain the least common multiple F of the inspection frequencies of each device included in the associated device model B ; Step 73. Augment the equipment inspection data in the associated equipment model through a fitting algorithm, and augment the frequency of the inspection data of each device to F B ; Step 74, correcting the fitting data based on the detection data of the environmental monitoring sensor to obtain the corrected extended data; Step 75, extracting feature data from the corrected extended data as the input data of the associated-device fault model for fault diagnosis.

Images