Wind power fire-fighting fault prediction system

Abstract

The invention discloses a wind power fire-fighting fault prediction system, which relates to the field of fire-fighting faults and comprises a management center, the management center is in communication connection with a wind power fire-fighting information acquisition module, a wind power fire-fighting information analysis module, a wind power fire-fighting prediction module and a wind power fire-fighting control module, and wind power edge nodes are arranged. Acquiring historical wind power basic data and wind power basic data of the wind power plant; according to the historical wind power basic data, wind power floating characteristics are obtained, according to the wind power floating characteristics, a wind power characteristic map is constructed, and according to the wind power basic data, wind power stop data and wind power factor data are obtained; constructing a wind power fire protection risk prediction model according to the wind power floating characteristics, and obtaining a wind power fire protection prediction value according to the wind power fire protection risk prediction model and the wind power characteristic map; according to the wind power fire-fighting predicted value, a fire-fighting control prevention strategy is obtained, and wind power fire-fighting fault prevention control is carried out on the wind power plant; according to the invention, the accuracy of fire-fighting fault prediction is improved.

Description

Technical Field

[0001] The invention relates to the field of fire protection faults, and in particular to a wind power fire protection fault prediction system. Background Art

[0002] As global climate change and environmental problems become increasingly serious, countries are actively promoting the transformation of energy structure, reducing the use of fossil fuels and increasing the proportion of renewable energy. Among them, wind energy, as a clean and renewable energy, has become one of the key development areas in my country and has been widely used around the world. Remote monitoring and intelligent operation and maintenance of wind farms have been realized. The development of artificial intelligence, machine learning, big data and other technologies has provided technical support for the realization of wind power fire fault prediction system. Fire safety is an important part of wind power generation system, especially in the area of ​​electrical equipment such as converters and transformers of wind turbines. Fire safety is getting more and more attention. In the prior art, there are many safety hazards in the use of wind energy, such as overheating, overload, insulation damage and other problems. These problems may cause accidents of wind power equipment, affect the normal operation of wind farms, and even cause fire accidents. Traditional fire monitoring methods cannot achieve real-time and accurate prediction of the system. The wind farm environment varies greatly. Wind farms in different regions are affected by various environmental factors, which leads to errors in wind farm fault analysis. There are also problems such as inaccuracy in wind power fire fault prediction, lag in wind farm analysis data transmission, fire safety problems caused by oil leakage in the fan, high wind farm operating costs, and data communication problems in remote wind farms. Therefore, a wind power fire fault prediction system is provided. Summary of the invention

[0003] In order to solve the above technical problems, the object of the present invention is to provide a wind power fire fault prediction system, including a management center, the management center is communicatively connected with a wind power fire information collection module, a wind power fire information analysis module, a wind power fire prediction module and a wind power fire control module; The wind power fire protection information collection module is used to set up wind power edge nodes and obtain historical wind power basic data and wind power basic data of the wind farm; The wind power fire protection information analysis module is used to obtain wind power floating characteristics based on historical wind power basic data, construct a wind power characteristic map based on the wind power floating characteristics, and analyze the wind power basic data to obtain wind power shutdown data and wind power factor data; The wind power fire prediction module is used to construct a wind power fire risk prediction model according to the wind power floating characteristics, and to predict the wind power fire risk of the wind farm according to the wind power fire risk prediction model and the wind power characteristic map to obtain the wind power fire prediction value; The wind power fire control module is used to perform wind power fire fault prediction analysis on the wind farm according to the wind power fire prediction value, obtain the fire control prevention strategy, and perform wind power fire fault prevention control on the wind farm according to the fire control prevention strategy.

[0004] Furthermore, the process of setting up a wind power edge node and obtaining historical wind power basic data and wind power basic data of a wind farm includes: Set up a wind power fire protection IoT platform and obtain historical wind power basic data and wind power basic data collected by wind power edge nodes; A wind power generation unit is provided, and the wind power edge node is provided in the wind power generation unit, and is used to collect historical wind power basic data and wind power basic data of the wind power generation unit; the historical wind power basic data includes historical wind power performance data and historical wind power scene image data; the historical wind power performance data includes historical wind power friction value, historical unit oil image data and historical wind power operation data; the historical wind power operation data includes historical unit temperature value and historical unit wind force value; the historical wind power scene image data includes historical unit scene image data of the location of each wind power generating unit; the wind power basic data includes wind power performance data and wind power scene image data; the wind power performance data includes wind power friction value, unit oil image data and wind power operation data; the wind power operation data includes unit temperature value and unit wind force value; the wind power scene image data includes unit scene image data of the location of each wind power generating unit.

[0005] Furthermore, based on historical wind power basic data, the process of obtaining wind power floating characteristics includes: According to the historical wind power scene image data of the historical wind power basic data, the historical wind speed conditions, historical humidity conditions, historical temperature conditions and historical altitude conditions are obtained; the wind power environment coefficient is set, and the wind power adaptability characteristics are obtained according to the wind power environment coefficient, historical wind speed conditions, historical humidity conditions, historical temperature conditions and historical altitude conditions; the wind resistance coefficient is set, and the wind power friction characteristics are obtained according to the wind resistance coefficient, historical wind power friction value, historical unit wind power value and historical unit temperature value; according to the historical unit oil image data, the oil coverage value and the oil thickness value are obtained; the oil density coefficient is set, and the unit oil overflow characteristics are obtained according to the oil density coefficient, the oil coverage value and the oil thickness value; and the wind power floating characteristics are obtained according to the wind power adaptability characteristics, the wind power friction characteristics and the unit oil overflow characteristics.

[0006] Furthermore, according to the floating characteristics of wind power, the process of constructing a wind power characteristic map includes: A wind power adaptation map is generated based on the wind power floating characteristics and the historical wind power basic data corresponding to the wind power floating characteristics. A wind power friction map is generated based on the historical wind power friction values, historical unit wind force values ​​and historical unit temperature values ​​corresponding to the wind power friction characteristics of the wind power floating characteristics. A unit oil overflow map is generated based on the oil coverage values ​​and oil thickness values ​​corresponding to the unit oil overflow characteristics of the wind power floating characteristics. A wind power characteristic map is constructed based on the wind power adaptation map, the wind power friction map and the unit oil overflow map.

[0007] Furthermore, the process of analyzing the basic wind power data and obtaining wind power outage data and wind power factor data includes: A focused characteristic analysis is performed on the wind power basic data, and scene key image characteristics are obtained based on the wind power scene image data of the wind power basic data, wherein the scene key image characteristics include weather image characteristics and terrain image characteristics, and scene pixel points and wind power edge shutdown data are obtained based on the weather image characteristics and terrain image characteristics of the scene key image characteristics; a wind power basic image is obtained based on the wind power friction value, unit temperature value, unit wind force value and unit oil image data of the wind power performance data; and wind power characteristic points are obtained based on the wind power friction value, unit temperature value, unit wind force value, unit oil image data and the wind power basic image; and characteristic correlation between wind power characteristic points is obtained based on the wind power basic image and the wind power characteristic points, and wind power factor data is obtained based on the characteristic correlation and the wind power characteristic points.

[0008] Furthermore, the process of constructing a wind power fire risk prediction model based on the floating characteristics of wind power includes: Based on the neural network technology, a corresponding wind power fire risk prediction model is constructed. The wind power adaptation characteristics, wind power friction characteristics and unit oil overflow characteristics of the obtained wind power floating characteristics are used to train the corresponding wind power fire risk prediction model until its loss function meets the requirements. The corresponding prediction model is then output to obtain the corresponding wind power fire risk prediction model.

[0009] Furthermore, according to the wind power fire risk prediction model and the wind power characteristic map, the process of predicting the wind power fire risk of the wind farm includes: According to the wind power fire risk prediction model, the predicted wind power adaptability characteristics, predicted wind power friction characteristics and predicted unit oil overflow characteristics are obtained; according to the wind power shutdown data and wind power factor data, the wind power data to be predicted is obtained; according to the wind power data to be predicted, the wind power characteristic points and the wind power fire risk prediction model, the predicted wind power adaptability characteristics, predicted wind power friction characteristics and predicted unit oil overflow characteristics are output; according to the wind power characteristic map, the predicted wind power adaptability characteristics, the predicted wind power friction characteristics and the predicted unit oil overflow characteristics, the corresponding wind power adaptability prediction map, wind power friction prediction map and unit oil overflow prediction map are obtained, and recorded as the wind power fire prediction map; and according to the wind power fire prediction map, the wind power shutdown data and the wind power factor data, the corresponding results and wind power similarity values ​​are obtained, and the similarity threshold is set, and the wind power prediction trend value is obtained according to the wind power similarity value and the similarity threshold, and the wind power fire prediction value is obtained according to the wind power prediction trend value, the wind power adaptation prediction map, the wind power friction prediction map and the unit oil overflow prediction map of the wind power fire prediction map.

[0010] Furthermore, based on the wind power fire prediction value, the wind power fire fault prediction analysis is performed on the wind farm to obtain the fire control prevention strategy. According to the fire control prevention strategy, the process of wind power fire fault prevention control on the wind farm includes: Obtain wind power fire failure events and wind power fire failure factors corresponding to historical wind power basic data, and set wind power fire safety standard values ​​according to wind power fire failure events and wind power fire failure factors; When the wind power fire prediction value is greater than the fire safety standard value, the data in the wind power prediction trend value will be sent to professional maintenance personnel to carry out wind power fire fault prevention and control on the wind farm and generate wind power fire fault prevention measures; when the wind power fire prediction value is less than or equal to the fire safety standard value, the wind farm situation will continue to be predicted.

[0011] Compared with the prior art, the beneficial effects of the present invention are as follows: setting up wind power edge nodes, obtaining historical wind power basic data and wind power basic data of the wind farm, being able to manage the entire wind farm system in real time, reducing the cost of on-site inspection and maintenance, and improving the operational safety and reliability of the wind farm; obtaining wind power floating characteristics based on historical wind power basic data, constructing a wind power characteristic map based on the wind power floating characteristics, and analyzing the wind power basic data to obtain wind power edge outage data and wind power factor data, which can improve the comprehensiveness and accuracy of wind farm fault factor analysis; based on the wind power floating characteristics Construct a wind power fire risk prediction model, and predict the wind power fire risk of the wind farm based on the wind power fire risk prediction model and the wind power characteristic map to obtain the wind power fire prediction value, which can optimize the fault prediction model and improve the prediction accuracy; based on the wind power fire prediction value, conduct wind power fire fault prediction analysis on the wind farm, obtain the fire control prevention strategy, and conduct wind power fire fault prevention control on the wind farm based on the fire control prevention strategy, which can provide early warning, so that the operation and maintenance personnel can take preventive maintenance measures to avoid the occurrence of fire or failure, thereby reducing downtime and losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 This is a schematic diagram of a wind power fire protection fault prediction system according to an embodiment of the present application. DETAILED DESCRIPTION

[0013] like Figure 1 As shown, a wind power fire fault prediction system includes a management center, and the management center is communicatively connected to a wind power fire information collection module, a wind power fire information analysis module, a wind power fire prediction module, and a wind power fire control module; The wind power fire protection information collection module is used to set up wind power edge nodes and obtain historical wind power basic data and wind power basic data of the wind farm; The wind power fire protection information analysis module is used to obtain wind power floating characteristics based on historical wind power basic data, construct a wind power characteristic map based on the wind power floating characteristics, and analyze the wind power basic data to obtain wind power shutdown data and wind power factor data; The wind power fire prediction module is used to construct a wind power fire risk prediction model according to the wind power floating characteristics, and to predict the wind power fire risk of the wind farm according to the wind power fire risk prediction model and the wind power characteristic map to obtain the wind power fire prediction value; The wind power fire control module is used to perform wind power fire fault prediction analysis on the wind farm according to the wind power fire prediction value, obtain the fire control prevention strategy, and perform wind power fire fault prevention control on the wind farm according to the fire control prevention strategy.

[0014] It should be further explained that, in the specific implementation process, the wind power fire information analysis module is used to obtain the wind power floating characteristics according to the historical wind power basic data, to construct the wind power characteristic map according to the wind power floating characteristics, and to analyze the wind power basic data to obtain the wind power shutdown data and wind power factor data. The process is as follows: Extract wind power scene conditions from historical wind power scene image data in historical wind power basic data; the wind power scene condition extraction refers to extracting wind speed conditions, humidity conditions, temperature conditions and altitude conditions of the wind farm; the wind speed conditions, humidity conditions, temperature conditions and altitude conditions refer to the wind speed values, humidity values, temperature values ​​and altitude values ​​in the historical wind power scene image data; obtain historical wind speed conditions, historical humidity conditions, historical temperature conditions and historical altitude conditions; set a wind power environment coefficient, which is related to the degree of mutual influence between the wind speed value, humidity value, temperature value and altitude value; multiply the wind power environment coefficient, historical wind speed conditions, historical humidity conditions, historical temperature conditions and historical altitude conditions to obtain wind power adaptation characteristics; set a wind resistance coefficient, which is related to the wind speed value, humidity value, temperature value and altitude value. The wind power friction value, the historical unit wind power value and the historical unit temperature value are related; the wind resistance coefficient, the historical wind power friction value, the historical unit wind power value and the historical unit temperature value are multiplied to obtain the wind power friction characteristics; the oil characteristics are extracted from the historical unit oil image data; the oil characteristics extraction refers to edge recognition of the oil in the historical unit oil image data to obtain the total coverage area of ​​the oil, which is recorded as the oil coverage value; and the oil coverage thickness is extracted to obtain the maximum value of the oil coverage thickness, which is recorded as the oil thickness value; the oil density coefficient is set, and the oil density coefficient is related to the density and type of the oil; the oil density coefficient, the oil coverage value and the oil thickness value are multiplied to obtain the unit oil overflow characteristics; and the wind power adaptation characteristics, wind power friction characteristics and unit oil overflow characteristics are recorded as wind power floating characteristics; And according to the wind power floating characteristics and the historical wind power basic data corresponding to the wind power floating characteristics, the wind speed conditions, humidity conditions, temperature conditions and altitude conditions corresponding to the wind power adaptation characteristics of the wind power floating characteristics are associated to generate a wind power adaptation map, and the wind power adaptation characteristics are marked in the corresponding wind power adaptation map; the historical wind power friction values, historical unit wind force values ​​and historical unit temperature values ​​corresponding to the wind power friction characteristics of the wind power floating characteristics are associated to generate a wind power friction map, and the wind power floating characteristics are marked in the corresponding wind power friction map; the oil coverage value and the oil thickness value corresponding to the unit oil overflow characteristics of the wind power floating characteristics are associated to generate a unit oil overflow map, and the wind power floating characteristics are marked in the unit oil overflow map; the wind power adaptation map, the wind power friction map and the unit oil overflow map are integrated and associated to construct a wind power characteristic map; Conduct focused characteristic analysis on wind power basic data to obtain wind power outage data and wind power factor data; It should be further explained that, in the specific implementation process, the specific process of focusing on the characteristics of wind power basic data is as follows: based on the natural language recognition technology, the key characteristics of the wind power scene image data of the wind power basic data are extracted, and the key characteristics extraction values ​​are numerically extracted for the wind speed, humidity, temperature and altitude in the wind power scene image data to obtain the scene key image characteristics, and the scene key image characteristics include weather image characteristics and terrain image characteristics, and the weather image characteristics and terrain image characteristics of the scene key image characteristics are converted into characteristic points, recorded as scene pixel points, and the scene pixel points are linked to obtain the wind Power outage data; according to the wind power friction value, unit temperature value, unit wind force value and unit oil image data of the wind power performance data, the corresponding state of the wind turbine generator set corresponding to the wind power basic data is imaged to obtain the wind power basic image; and the corresponding areas of the wind power friction value, unit temperature value, unit wind force value and unit oil image data in the wind power basic image are converted into characteristic points to obtain wind power characteristic points; and based on natural language technology, the wind power basic image and wind power characteristic points are analyzed to obtain the characteristic correlation between the wind power characteristic points, and the wind power characteristic points are linked according to the characteristic correlation to obtain wind power factor data.

[0015] It should be further explained that, in the specific implementation process, the wind power fire prediction module is used to construct a wind power fire risk prediction model based on the wind power floating characteristics, and to predict the wind power fire risk of the wind farm based on the wind power fire risk prediction model and the wind power characteristic map. The process of obtaining the wind power fire prediction value is as follows: Based on the neural network technology, a corresponding wind power fire risk prediction model is constructed, and the wind power adaptation characteristics, wind power friction characteristics and unit oil overflow characteristics of the obtained wind power floating characteristics are used to train the corresponding wind power fire risk prediction model until its loss function meets the requirements, and then the corresponding prediction model is output to obtain the corresponding wind power fire risk prediction model; The wind power fire risk is predicted according to the wind power fire risk prediction model to obtain the predicted wind power adaptation characteristics, predicted wind power friction characteristics and predicted unit oil overflow characteristics; According to the wind power shutdown data and wind power factor data, the wind power data to be predicted is obtained; the characteristics corresponding to the scene pixel points and wind power characteristic points in the wind power data to be predicted are input into the wind power fire risk prediction model, and the predicted wind power adaptation characteristics, predicted wind power friction characteristics and predicted unit oil overflow characteristics are output; the predicted wind power adaptation characteristics, predicted wind power friction characteristics and predicted unit oil overflow characteristics are analyzed through the wind power characteristic map, and the wind power characteristic map is traversed through the corresponding wind power adaptation map, wind power friction map and unit oil overflow map based on the predicted wind power adaptation characteristics, predicted wind power friction characteristics and predicted unit oil overflow characteristics, to obtain the corresponding wind power adaptation prediction map, wind power friction prediction map and unit oil overflow prediction map, and record them as wind power fire prediction maps; And perform conditional trend analysis on the wind power fire prediction map, wind power shutdown data and wind power factor data; traverse and correspond the characteristics in the scene pixel points of the wind power shutdown data with the data information in the wind power adaptation prediction map, and traverse and correspond the characteristics in the wind power characteristic points of the wind power factor data with the data information in the wind power friction prediction map and the unit oil overflow prediction map, respectively, to obtain corresponding results, perform similarity analysis on the corresponding results to obtain wind power similarity values, set a similarity threshold, extract the corresponding results whose wind power similarity values ​​are less than the similarity threshold, and extract the corresponding results belonging to the wind power fire prediction map to obtain the wind power prediction trend value, and perform multiplication and accumulation calculations on the wind power prediction trend value and the characteristics corresponding to the wind power adaptation prediction map, the wind power friction prediction map and the unit oil overflow prediction map of the wind power fire prediction map to obtain the wind power fire prediction value.

[0016] It should be further explained that, in the specific implementation process, the wind power fire control module is used to perform wind power fire fault prediction analysis on the wind farm according to the wind power fire prediction value, obtain the fire control prevention strategy, and the process of wind power fire fault prevention control on the wind farm according to the fire control prevention strategy is as follows: Obtain wind power fire failure events and wind power fire failure factors corresponding to historical wind power basic data, conduct in-depth learning on wind power fire failure events and wind power fire failure factors, and set wind power fire safety standard values; When the wind power fire prediction value is greater than the fire safety standard value, it is predicted that the wind farm will have a fire fault event, and the data in the wind power prediction trend value will be sent to professional maintenance personnel to conduct wind power fire fault prevention and control on the wind farm and generate wind power fire fault prevention measures; It should be further explained that, in the specific implementation process, the wind power fire prevention measures include but are not limited to reducing the friction of the wind turbine generator set, checking the oil leakage in the wind turbine generator set and cleaning the leaked oil, reducing the temperature of the wind turbine generator set and reducing the wind speed around the wind turbine generator set; When the wind power fire prediction value is less than or equal to the fire safety standard value, the wind farm situation will continue to be monitored in real time, and situations that have not occurred in the wind farm will be predicted, and response measures can be taken in a timely manner.

[0017] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A wind power fire protection fault prediction system, comprising a management center, characterized in that: The management center is communicatively connected to a wind power fire information collection module, a wind power fire information analysis module, a wind power fire prediction module and a wind power fire control module; The wind power fire protection information collection module is used to set up wind power edge nodes and obtain historical wind power basic data and wind power basic data of the wind farm; The wind power fire protection information analysis module is used to obtain wind power floating characteristics based on historical wind power basic data, construct a wind power characteristic map based on the wind power floating characteristics, and analyze the wind power basic data to obtain wind power shutdown data and wind power factor data; The wind power fire prediction module is used to construct a wind power fire risk prediction model according to the wind power floating characteristics, and to predict the wind power fire risk of the wind farm according to the wind power fire risk prediction model and the wind power characteristic map to obtain the wind power fire prediction value; The wind power fire control module is used to perform wind power fire fault prediction analysis on the wind farm according to the wind power fire prediction value, obtain the fire control prevention strategy, and perform wind power fire fault prevention control on the wind farm according to the fire control prevention strategy.

2. A wind power fire protection fault prediction system according to claim 1, characterized in that: The process of setting up wind power edge nodes and obtaining historical wind power basic data and wind power basic data of wind farms includes: Set up a wind power fire protection IoT platform and obtain historical wind power basic data and wind power basic data collected by wind power edge nodes; A wind power generation unit is provided, and the wind power edge node is provided in the wind power generation unit, and is used to collect historical wind power basic data and wind power basic data of the wind power generation unit; the historical wind power basic data includes historical wind power performance data and historical wind power scene image data; the historical wind power performance data includes historical wind power friction value, historical unit oil image data and historical wind power operation data; the historical wind power operation data includes historical unit temperature value and historical unit wind force value; the historical wind power scene image data includes historical unit scene image data of the location of each wind power generating unit; the wind power basic data includes wind power performance data and wind power scene image data; the wind power performance data includes wind power friction value, unit oil image data and wind power operation data; the wind power operation data includes unit temperature value and unit wind force value; the wind power scene image data includes unit scene image data of the location of each wind power generating unit.

3. A wind power fire protection fault prediction system according to claim 2, characterized in that: Based on historical wind power basic data, the process of obtaining wind power floating characteristics includes: According to the historical wind power scene image data of the historical wind power basic data, the historical wind speed conditions, historical humidity conditions, historical temperature conditions and historical altitude conditions are obtained; the wind power environment coefficient is set, and the wind power adaptability characteristics are obtained according to the wind power environment coefficient, historical wind speed conditions, historical humidity conditions, historical temperature conditions and historical altitude conditions; the wind resistance coefficient is set, and the wind power friction characteristics are obtained according to the wind resistance coefficient, historical wind power friction value, historical unit wind power value and historical unit temperature value; according to the historical unit oil image data, the oil coverage value and the oil thickness value are obtained; the oil density coefficient is set, and the unit oil overflow characteristics are obtained according to the oil density coefficient, the oil coverage value and the oil thickness value; and the wind power floating characteristics are obtained according to the wind power adaptability characteristics, the wind power friction characteristics and the unit oil overflow characteristics.

4. A wind power fire protection fault prediction system according to claim 3, characterized in that: According to the floating characteristics of wind power, the process of constructing the wind power characteristic map includes: A wind power adaptation map is generated based on the wind power floating characteristics and the historical wind power basic data corresponding to the wind power floating characteristics. A wind power friction map is generated based on the historical wind power friction values, historical unit wind force values ​​and historical unit temperature values ​​corresponding to the wind power friction characteristics of the wind power floating characteristics. A unit oil overflow map is generated based on the oil coverage values ​​and oil thickness values ​​corresponding to the unit oil overflow characteristics of the wind power floating characteristics. A wind power characteristic map is constructed based on the wind power adaptation map, the wind power friction map and the unit oil overflow map.

5. A wind power fire protection fault prediction system according to claim 4, characterized in that: The process of analyzing the basic data of wind power and obtaining the data of wind power failure and wind power factor data includes: A focused characteristic analysis is performed on the wind power basic data, and scene key image characteristics are obtained based on the wind power scene image data of the wind power basic data, wherein the scene key image characteristics include weather image characteristics and terrain image characteristics, and scene pixel points and wind power edge shutdown data are obtained based on the weather image characteristics and terrain image characteristics of the scene key image characteristics; a wind power basic image is obtained based on the wind power friction value, unit temperature value, unit wind force value and unit oil image data of the wind power performance data; and wind power characteristic points are obtained based on the wind power friction value, unit temperature value, unit wind force value, unit oil image data and the wind power basic image; and characteristic correlation between wind power characteristic points is obtained based on the wind power basic image and the wind power characteristic points, and wind power factor data is obtained based on the characteristic correlation and the wind power characteristic points.

6. A wind power fire protection fault prediction system according to claim 5, characterized in that: The process of building a wind power fire risk prediction model based on wind power floating characteristics includes: Based on the neural network technology, a corresponding wind power fire risk prediction model is constructed. The wind power adaptation characteristics, wind power friction characteristics and unit oil overflow characteristics of the obtained wind power floating characteristics are used to train the corresponding wind power fire risk prediction model until its loss function meets the requirements. The corresponding prediction model is then output to obtain the corresponding wind power fire risk prediction model.

7. A wind power fire protection fault prediction system according to claim 6, characterized in that: According to the wind power fire risk prediction model and wind power characteristic map, the process of wind power fire risk prediction for wind farms includes: According to the wind power fire risk prediction model, the predicted wind power adaptability characteristics, predicted wind power friction characteristics and predicted unit oil overflow characteristics are obtained; according to the wind power shutdown data and wind power factor data, the wind power data to be predicted is obtained; according to the wind power data to be predicted, the wind power characteristic points and the wind power fire risk prediction model, the predicted wind power adaptability characteristics, predicted wind power friction characteristics and predicted unit oil overflow characteristics are output; according to the wind power characteristic map, the predicted wind power adaptability characteristics, the predicted wind power friction characteristics and the predicted unit oil overflow characteristics, the corresponding wind power adaptability prediction map, wind power friction prediction map and unit oil overflow prediction map are obtained, and recorded as the wind power fire prediction map; and according to the wind power fire prediction map, the wind power shutdown data and the wind power factor data, the corresponding results and wind power similarity values ​​are obtained, and the similarity threshold is set, and the wind power prediction trend value is obtained according to the wind power similarity value and the similarity threshold, and the wind power fire prediction value is obtained according to the wind power prediction trend value, the wind power adaptation prediction map, the wind power friction prediction map and the unit oil overflow prediction map of the wind power fire prediction map.

8. A wind power fire protection fault prediction system according to claim 7, characterized in that: According to the wind power fire prediction value, the wind power fire fault prediction analysis is carried out on the wind farm to obtain the fire control prevention strategy. According to the fire control prevention strategy, the process of wind power fire fault prevention control on the wind farm includes: Obtain wind power fire failure events and wind power fire failure factors corresponding to historical wind power basic data, and set wind power fire safety standard values ​​according to wind power fire failure events and wind power fire failure factors; When the wind power fire prediction value is greater than the fire safety standard value, the data in the wind power prediction trend value will be sent to professional maintenance personnel to carry out wind power fire fault prevention and control on the wind farm and generate wind power fire fault prevention measures; when the wind power fire prediction value is less than or equal to the fire safety standard value, the wind farm situation will continue to be predicted.

Images