Shore power equipment fault detection method and system based on cloud platform and fault model
By adopting the fault detection method of cloud platform and fault model in shore power equipment, real-time monitoring and classification of faults of shore power equipment is realized, the problem of low fault detection efficiency of shore power equipment is solved, and the accuracy and system of fault detection are improved. security.
Patent Information
- Application Number
- CN202411048395.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-01
- Publication Date
- 2024-11-08
Smart Images

Figure CN118914713A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of electric power automation, and in particular to a shore power equipment fault detection method and system based on a cloud platform and a fault model. Background Art
[0002] The electricity for offshore oil and gas fields will mainly come from shore power projects. Under the shore power mode, the offshore oil and gas field power grid areas are highly interconnected, the mutual influence is more profound, the operating conditions are more complex, and the demand for shore power operation automation is increasing. Power dispatching, relay protection, high-voltage submarine cable protection, and key distribution and transformation equipment status monitoring are highly dependent on the construction of data centers.
[0003] Due to the large number of substations, if a monitoring center is established at each site, it will consume a lot of manpower and material resources. Therefore, there is an urgent need for a computing platform that can collect and process relevant data of each power station and substation equipment, and provide real-time processing results, so that the staff can deal with equipment failures in a timely manner and avoid greater economic losses and safety hazards. In order to avoid possible failures during the power generation process of the equipment, as well as transportation, real-time fault detection and diagnosis are required. Fault detection has been a hot topic of research in the last century. At this stage, fault detection is usually based on data transmitted by equipment sensors, and then detected using artificial intelligence algorithms. Fault detection requires massive data collection and complex data analysis. Cloud computing platforms have been widely used in massive data processing in various fields in recent years due to their powerful data storage and computing capabilities. In addition, cloud computing also has the characteristics of versatility, high reliability, high scalability, low price, and automation. Therefore, it is very effective to use cloud computing architecture to build a power grid cloud platform for data aggregation and analysis. Summary of the invention
[0004] In view of the above problems, the purpose of the present invention is to provide a shore power equipment fault detection method and system based on a cloud platform and a fault model, which can improve the maintenance efficiency of shore power equipment and ensure the safe and reliable operation of the entire shore power system.
[0005] To achieve the above-mentioned purpose, the present invention adopts the following technical scheme: a shore power equipment fault detection method based on a cloud platform and a fault model, which comprises: recording and uploading data through a collection of various monitoring devices; preprocessing the data generated by each monitoring device, and uploading it to the cloud service after data preprocessing; using previously stored data to perform offline parallel iterative training on the fault detection model training, and the trained model classifies the faults that occur; inputting the equipment data collected by real-time monitoring into the fault detection model, analyzing the monitoring data, and quickly obtaining the fault category.
[0006] Furthermore, the data generated by each monitoring device includes: partial discharge detection data, humidity sensor data, transformer oil chromatography data, magnetic flux density monitoring data and temperature sensor data.
[0007] Furthermore, the data generated by each monitoring device is preprocessed, including: data cleaning, data integration, data conversion, data standardization and data enrichment; through data preprocessing, the same type of data, consistent format, information refinement and centralized storage are achieved.
[0008] Furthermore, data cleaning uses the idea of backtracking to analyze each process of the data set and extract the rules of data cleaning; these rules are applied to the data set to discover and clean up "dirty data".
[0009] Further, data transformation, data standardization and data enrichment are: Perform data transformation Apply a mathematical function to each attribute value and set the attribute value to , use the z-score standardization function to standardize the data; supplement the field content in the data to complete data enrichment.
[0010] Further, the fault detection model training includes: Establish a fault detection model through offline training using massive sample data to obtain a series of relevant evaluation data sets; Among them, the construction method of a single evaluation data set is: Select any feature in the feature set , randomly select any value corresponding to this feature from all the data ; according to , respectively set up two evaluation data subsets on the left and right. If the time series data Greater than , then it is classified into the right subset, otherwise it is classified into the left subset; Repeat the above process to construct left and right subsets until all time series data The root mean square average of the distances to each element in the set formed after removing the left and right subsets is less than the control threshold , the set formed after eliminating the left and right subsets is the required evaluation data set.
[0011] Furthermore, the monitoring data is analyzed to quickly obtain the fault category, including: After constructing k evaluation data sets, for time series data , the root mean square average of its distance from the elements in the zth evaluation data set is recorded as ; when Greater than the abnormal threshold When the data It is defined as abnormal data, otherwise it is normal data.
[0012] A shore power equipment fault detection system based on a cloud platform and a fault model comprises: an equipment module, which records and uploads data through a collection of various monitoring devices; a field module, which pre-processes the data generated by each monitoring device and uploads the data to the cloud service after the data pre-processing; a cloud service module, which uses previously stored data to perform offline parallel iterative training on the fault detection model training, and the trained model classifies the faults that occur; a visualization and status monitoring module, which inputs the equipment data collected by real-time monitoring into the fault detection model, analyzes the monitoring data, and quickly obtains the fault category.
[0013] A computer-readable storage medium storing one or more programs, wherein the one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any of the above methods.
[0014] A computing device comprises: one or more processors, a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any of the above methods.
[0015] The present invention adopts the above technical solution, which has the following advantages: The present invention combines cloud platform technology with shore power equipment fault detection, directly collects the operating status of the shore power equipment on the shore power equipment side, processes the shore power equipment data, and uses the cloud platform to achieve more powerful applications, thereby optimizing and upgrading services such as shore power equipment fault diagnosis, fault warning, and operation and maintenance, thereby reducing the failure rate of shore power equipment, improving the maintenance efficiency of shore power equipment, and ensuring the safe and reliable operation of the entire shore power system. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 It is an overall framework diagram of the shore power equipment fault detection method in an embodiment of the present invention. DETAILED DESCRIPTION
[0017] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.
[0018] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.
[0019] In order to improve the maintenance efficiency of shore power equipment and ensure the safe and reliable operation of the entire shore power system, the present invention discloses a shore power equipment fault detection method and system based on a cloud platform and a fault model. Based on the cloud computing architecture, it is suitable for fault detection of the power grid cloud platform. A power grid cloud platform is constructed, which is divided into three layers, including: equipment layer, field layer and cloud service layer. The equipment layer is a collection of various monitoring devices, which records and uploads data; the field layer collects data generated by each monitoring system of the equipment layer, and uploads it to the cloud service layer after data preprocessing; the cloud service layer uses previously stored data to perform offline parallel iterative training on the fault detection model training, and the trained model can classify the faults that occur. The equipment data is monitored in real time through the status monitoring module and the visualization module. When a device fails, the fault detection module quickly obtains the fault category by analyzing the monitoring data, and the operator can issue operation commands in time through the human-computer interaction module to prevent the abnormal situation from further expanding and improve the reliability of the overall operation of the shore power system.
[0020] In one embodiment of the present invention, a shore power equipment fault detection method based on a cloud platform and a fault model is provided. In this embodiment, Figure 1 As shown, the method comprises the following steps: 1) Record and upload data through a collection of various monitoring devices; 2) Pre-process the data generated by each monitoring device and upload it to the cloud service after data pre-processing; 3) Using previously stored data, the fault detection model is trained offline in parallel and iteratively, and the trained model classifies the faults that occur; 4) Input the equipment data collected by real-time monitoring into the fault detection model, analyze the monitoring data, and quickly obtain the fault category.
[0021] Specifically, in this embodiment, a power grid cloud platform is constructed based on a cloud computing architecture. The platform is divided into three layers, including: a device layer, a field layer, and a cloud service layer. The device layer performs the above step 1; the field layer performs step 2), and uploads the pre-processed data to the cloud service layer; the cloud service layer performs step 3), and the model trained in step 3) can classify the faults that occur.
[0022] In the above step 1), the data generated by each monitoring device includes: partial discharge detection data, humidity sensor data, transformer oil chromatography data, magnetic flux density monitoring data and temperature sensor data, etc. After the data collection is completed, it is stored and provided to the field layer for data processing or other operations.
[0023] In the above step 2), the data generated by each monitoring device is preprocessed, including: data cleaning, data integration, data conversion, data standardization and data enrichment; through data preprocessing, the same type of data, consistent format, information refinement and centralized storage are achieved.
[0024] In this embodiment, data cleaning uses a backtracking idea to analyze each process of the data set and extract data cleaning rules; these rules are applied to the data set to discover and clean up "dirty data".
[0025] In this embodiment, data conversion, data standardization and data enrichment are: Perform data transformation Apply a mathematical function to each attribute value and set the attribute value to , use the z-score standardization function to standardize the data; supplement the field content in the data, mainly supplement the field content in the data to complete data enrichment.
[0026] Assume the original data set is , then the function formula is as follows:
[0027] in, is the mean of the original data, and is the standard deviation of the original data.
[0028] Optional, supplementary content includes data ID, upload processing server IP, information processing time and entity unified identification, etc.
[0029] Specifically, the field layer mainly includes data processing, data attribute generalization and data upload. Data processing also includes data cleaning, data integration, data conversion, data standardization and data enrichment. Through data processing, the same type of data, consistent format, information refinement and centralized storage are achieved, which prepares for subsequent data upload and makes data analysis at the cloud service layer more efficient.
[0030] In the above step 3), the fault detection model training is specifically: using massive sample data to establish a fault detection model using an offline training method to obtain a series of relevant evaluation data sets.
[0031] Among them, it is assumed that the data set obtained after preprocessing ,in Represents the i-th time series data, whose data feature dimension is d, that is, , is the corresponding label, and n is the number of data. Then the method for constructing a single evaluation data set includes the following steps: 3.1) Select any feature from the feature set , randomly select any value corresponding to this feature from all the data ; 3.2) According to , respectively set up two evaluation data subsets on the left and right. If the time series data Greater than , then it is classified into the right subset, otherwise it is classified into the left subset; 3.3) Repeat the above steps 3.1) and 3.2) to construct left and right subsets until all time series data The root mean square average of the distances to each element in the set formed after removing the left and right subsets is less than the control threshold , the set formed after eliminating the left and right subsets is the required evaluation data set.
[0032] Among them, the control threshold Set according to the normal operation data of the device.
[0033] In this embodiment, the cloud service layer is mainly divided into two parts: service and management. The service part mainly provides users with various cloud-based services, including three levels: software as a service (SaaS), platform as a service (PaaS) and infrastructure as a service (IaaS). The SaaS layer provides web technology services based on HTML, JavaScript and CSS. The PaaS layer provides a multi-tenant core and customized application server. The IaaS layer consists of a relational database. The cloud management layer provides support for user management, billing management, monitoring system, resource management, security management, etc. The management part is mainly used for cloud management. It is the basis of the three-layer cloud service, which can be divided into the user layer, mechanism layer and detection layer.
[0034] In the above step 4), the monitoring data is analyzed to quickly obtain the fault category, including the following steps: 4.1) Construct k evaluation datasets Then, for time series data , the root mean square average of its distance from the elements in the zth evaluation data set is recorded as ; 4.2) When Greater than the abnormal threshold When the data It is defined as abnormal data, otherwise it is normal data.
[0035] In one embodiment of the present invention, a shore power equipment fault detection system based on a cloud platform and a fault model is provided, which includes: The equipment module records and uploads data through a collection of various monitoring devices; The field module pre-processes the data generated by each monitoring device and uploads it to the cloud service after data pre-processing; The cloud service module uses previously stored data to perform offline parallel iterative training on the fault detection model, and the trained model classifies the faults that occur; The visualization and status monitoring module inputs the equipment data collected by real-time monitoring into the fault detection model, analyzes the monitoring data, and quickly obtains the fault category.
[0036] In the above embodiment, the data generated by each monitoring device includes: partial discharge detection data, humidity sensor data, transformer oil chromatography data, magnetic flux density monitoring data and temperature sensor data.
[0037] In the above embodiment, the data generated by each monitoring device is preprocessed, including: data cleaning, data integration, data conversion, data standardization and data enrichment; through data preprocessing, the same type of data, consistent format, information refinement and centralized storage are achieved.
[0038] Among them, data cleaning uses the idea of backtracking to analyze each process of the data set and extract the rules of data cleaning; these rules are applied to the data set to discover and clean up "dirty data".
[0039] Data conversion, data standardization, and data enrichment are: data transformation, applying mathematical functions to each attribute value, setting the attribute value to , use the z-score standardization function to standardize the data; supplement the field content in the data to complete data enrichment.
[0040] In the above embodiment, the fault detection model training includes: Establish a fault detection model through offline training using massive sample data to obtain a series of relevant evaluation data sets; Among them, the construction method of a single evaluation data set is: Select any feature in the feature set , randomly select any value corresponding to this feature from all the data ; according to , respectively set up two evaluation data subsets on the left and right. If the time series data Greater than , then it is classified into the right subset, otherwise it is classified into the left subset; Repeat the above process to construct left and right subsets until all time series data The root mean square average of the distances to each element in the set formed after removing the left and right subsets is less than the control threshold , the set formed after eliminating the left and right subsets is the required evaluation data set.
[0041] In the above embodiment, the monitoring data is analyzed to quickly obtain the fault category, including: After constructing k evaluation data sets, for time series data , the root mean square average of its distance from the elements in the zth evaluation data set is recorded as ; when Greater than the abnormal threshold When the data It is defined as abnormal data, otherwise it is normal data.
[0042] The system provided in this embodiment is used to execute the above-mentioned method embodiments. Please refer to the above-mentioned embodiments for specific processes and detailed contents, which will not be repeated here.
[0043] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A shore power equipment fault detection method based on a cloud platform and a fault model, characterized in that: include: Through the collection of various monitoring devices, data is recorded and uploaded; Pre-process the data generated by each monitoring device and upload it to the cloud service after data pre-processing; The fault detection model training is performed offline and iteratively in parallel using previously stored data, and the trained model classifies the faults that occur; The equipment data collected by real-time monitoring is input into the fault detection model, the monitoring data is analyzed, and the fault category is quickly obtained.
2. The shore power equipment fault detection method based on cloud platform and fault model as claimed in claim 1, characterized in that: The data generated by each monitoring device includes: partial discharge detection data, humidity sensor data, transformer oil chromatography data, magnetic flux density monitoring data and temperature sensor data.
3. The shore power equipment fault detection method based on cloud platform and fault model as claimed in claim 1, characterized in that: The data generated by each monitoring device is preprocessed, including: data cleaning, data integration, data conversion, data standardization and data enrichment; through data preprocessing, the same type of data, consistent format, information refinement and centralized storage are achieved.
4. The shore power equipment fault detection method based on cloud platform and fault model as claimed in claim 3, characterized in that: Data cleaning uses the idea of backtracking to analyze each process of the data set and extract the rules of data cleaning; these rules are applied to the data set to discover and clean up "dirty data".
5. The shore power equipment fault detection method based on cloud platform and fault model as claimed in claim 3, characterized in that: Data transformation, data standardization, and data enrichment for: Perform data transformation Apply a mathematical function to each attribute value and set the attribute value to , use the z-score standardization function to standardize the data; supplement the field content in the data to complete data enrichment.
6. The shore power equipment fault detection method based on cloud platform and fault model as claimed in claim 1, characterized in that: Fault detection model training, including: Establish a fault detection model through offline training using massive sample data to obtain a series of relevant evaluation data sets; Among them, the construction method of a single evaluation data set is: Select any feature in the feature set , randomly select any value corresponding to this feature from all the data ; according to , respectively set up two evaluation data subsets on the left and right. If the time series data Greater than , then it is classified into the right subset, otherwise it is classified into the left subset; Repeat the above process to construct left and right subsets until all time series data The root mean square average of the distances to each element in the set formed after removing the left and right subsets is less than the control threshold , the set formed after eliminating the left and right subsets is the required evaluation data set.
7. The shore power equipment fault detection method based on cloud platform and fault model as claimed in claim 6, characterized in that: Analyze monitoring data and quickly obtain fault categories, including: After constructing k evaluation data sets, for time series data , the root mean square average of its distance from the elements in the zth evaluation data set is recorded as ; when Greater than the abnormal threshold When the data It is defined as abnormal data, otherwise it is normal data.
8. A shore power equipment fault detection system based on a cloud platform and a fault model, characterized in that: include: The equipment module records and uploads data through a collection of various monitoring devices; The field module pre-processes the data generated by each monitoring device and uploads it to the cloud service after data pre-processing; The cloud service module uses previously stored data to perform offline parallel iterative training on the fault detection model, and the trained model classifies the faults that occur; The visualization and status monitoring module inputs the equipment data collected by real-time monitoring into the fault detection model, analyzes the monitoring data, and quickly obtains the fault category.
9. A computer-readable storage medium storing one or more programs, characterized in that: The one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any one of the methods of claims 1 to 7.
10. A computing device, characterized in that: include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any one of the methods described in claims 1 to 7.
