阅读说明：本技术 一种山地风电场应急调度方法 (Emergency dispatching method for mountain wind power plant ) 是由 赵鹏程 刘艳贵 张晓辉 傅望安 王海明 张育钧 高建忠 任鑫 王�华 杜静宇 于 2021-07-14 设计创作，主要内容包括：本发明公开了一种山地风电场应急调度方法,包括以下步骤,步骤1,采集山地风电场的地理信息,绘制场站区域电子地图；步骤2,依据步骤1中的电子地图对风电场运维人员和车辆进行定位；步骤3,采用卷积神经网络判断风电场运维人员的安全状态；步骤4,通过卷积神经网络处理路面图像,自动识别路面缺陷类型和范围,并进行预警；步骤5,在山地风电场上设置区域电子围栏,进行安全防护；步骤6,依据步骤1中的场站区域电子地图,建立救援车辆从场站区域外到内部事故点的快速引导通道。本发明提出的山地风场应急指挥中心建设方法,实现人员车辆的监测管理,有利于提高山地风电场作业的安全性。(The invention discloses a method for dispatching a mountain wind power plant in an emergency mode, which comprises the following steps of 1, collecting geographic information of the mountain wind power plant, and drawing a station area electronic map; step 2, positioning the operation and maintenance personnel and the vehicle of the wind power plant according to the electronic map in the step 1; step 3, judging the safety state of the operation and maintenance personnel of the wind power plant by adopting a convolutional neural network; step 4, processing the road surface image through a convolutional neural network, automatically identifying the type and range of the road surface defect, and early warning; step 5, arranging a regional electronic fence on the mountain wind power plant for safety protection; and 6, establishing a rapid guide channel of the rescue vehicle from the outside of the station area to the internal accident point according to the station area electronic map in the step 1. The method for building the emergency command center of the mountain wind farm provided by the invention realizes monitoring and management of personnel and vehicles, and is beneficial to improving the safety of operation of the mountain wind farm.)

1. A method for dispatching a mountain wind farm in an emergency is characterized by comprising the following steps,

step 1, collecting geographic information of a mountain wind power plant, and drawing a station area electronic map;

step 2, positioning the operation and maintenance personnel and the vehicle of the wind power plant according to the electronic map in the step 1;

step 3, judging the safety state of the operation and maintenance personnel of the wind power plant by adopting a convolutional neural network;

step 4, processing the road surface image through a convolutional neural network, automatically identifying the type and range of the road surface defect, and early warning;

step 5, arranging a regional electronic fence on the mountain wind power plant for safety protection;

and 6, establishing a rapid guide channel of the rescue vehicle from the outside of the station area to the internal accident point according to the station area electronic map in the step 1.

2. The mountain wind farm emergency dispatching method according to claim 1, wherein in step 1, the geographical location comprises wind turbine position information, substation information, road information, valley information and hill information.

3. The method for dispatching the mountain wind farm emergency according to claim 1, wherein in the step 1, the specific process comprises acquiring geographic information of the mountain wind farm by aerial photography of an unmanned aerial vehicle, drawing an off-site map according to the geographic information, and fusing the off-site map and the on-site map to form a geographic information system.

4. The method for dispatching the wind power plant in the mountainous region according to claim 3, wherein in the step 1, a map outside the station is drawn according to geographic information, and the longitude and latitude of the machine body and the surrounding building point position are compiled through conversion calculation of a coordinate system by combining with a CAD drawing of the station; and cutting the plan view of the off-site map into countless tile maps of 1X1 by using a map slicing tool, performing index positioning on the tiles by using a map API (application program interface), and displaying the tiles in the map for fusion to form a complete geographic information system.

5. The mountain wind farm emergency dispatching method according to claim 1, wherein in step 2, positioning is performed by combining a geographic information system through a positioning tag worn by a worker.

6. The mountain wind farm emergency dispatching method according to claim 1, characterized in that in step 3, a Mask R-CNN algorithm training model is established by using a convolutional neural network, unsafe behaviors are identified and an alarm is given.

7. The mountain wind farm emergency dispatching method according to claim 1, characterized in that in step 4, a road surface image is formed by continuously detecting the road surface quality and the structure condition by using radar.

8. The method for dispatching the mountainous wind power plant emergency according to claim 1, wherein in step 6, a fusion navigation model of site internal navigation and site external navigation is constructed according to the site area electronic map in step 1, an optimal driving route of an intersection point and an accident point is automatically generated in real time through an intelligent path planning algorithm, and a rapid guiding channel of a rescue vehicle from the outside of the site area to the inside accident point is established.

Technical Field

The invention belongs to the field of operation and maintenance development of a mountain wind power plant, and particularly belongs to an emergency scheduling method of the mountain wind power plant.

Background

The mountain wind power station road network has many intersections, is similar and complex, and is easy to go wrong. The traditional roadside signboard is easy to damage and difficult to identify; the mountainous wind power plant has a large construction range, remote places and poor peripheral infrastructure, and operators are objectively and subjectively careless for inspection and inspection in complicated work, so that equipment is easy to break down, the operation cost is increased, and economic loss is brought to wind power enterprises; for a wind power plant, the working track of an operator cannot be monitored in real time in the process of inspection operation, various personnel are difficult to classify and manage in different regions, the worker cannot be judged to be on duty, on duty and off duty, the trapped position of the personnel cannot be accurately judged in time, the evacuation route of the personnel in danger is difficult to determine and inform, and if an enterprise cannot accurately master the actual position of the personnel, the production and control efficiency is difficult to improve; the roads in the yard are limited by cost, simple pavements are paved for sand gravel, the road conditions are poor, the running process of the vehicle is lack of safety monitoring and reminding, and the running safety of the vehicle is influenced; the new energy station lacks a digital electronic map due to remote and confidential requirements and cannot navigate. The areas are open and similar, no reference object is used for driving at night, the machine position is easy to miss, the consumed time is long, the on-site rescue vehicle needs to be dispatched for guidance due to unfamiliarity with roads, and the efficiency is low, and the time is long.

Disclosure of Invention

In order to solve the problems in the prior art, the invention provides the emergency scheduling method for the mountain wind farm, which can realize monitoring management of personnel and vehicles, is beneficial to improving the safety of operation of the mountain wind farm, improves the emergency rescue efficiency and reduces personnel and economic losses.

In order to achieve the purpose, the invention provides the following technical scheme:

a method for dispatching a mountain wind farm in an emergency comprises the following steps,

step 1, collecting geographic information of a mountain wind power plant, and drawing a station area electronic map;

step 2, positioning the operation and maintenance personnel and the vehicle of the wind power plant according to the electronic map in the step 1;

step 3, judging the safety state of the operation and maintenance personnel of the wind power plant by adopting a convolutional neural network;

step 4, processing the road surface image through a convolutional neural network, automatically identifying the type and range of the road surface defect, and early warning;

step 5, arranging a regional electronic fence on the mountain wind power plant for safety protection;

and 6, establishing a rapid guide channel of the rescue vehicle from the outside of the station area to the internal accident point according to the station area electronic map in the step 1.

Preferably, in step 1, the geographical location includes wind turbine location information, substation information, road information, pit information, and hill information.

Preferably, in the step 1, the specific process includes acquiring geographic information of the mountainous wind power plant by using aerial photography of the unmanned aerial vehicle, drawing an off-site map according to the geographic information, and fusing the off-site map and the on-site map to form a geographic information system.

Further, drawing a map outside the station according to the geographic information in the step 1, combining the CAD drawing of the station, and compiling and annotating the longitude and latitude of the machine body and the peripheral building point positions through the conversion calculation of a coordinate system; and cutting the plan view of the off-site map into countless tile maps of 1X1 by using a map slicing tool, performing index positioning on the tiles by using a map API (application program interface), and displaying the tiles in the map for fusion to form a complete geographic information system.

Preferably, in step 2, the worker wears the positioning tag to perform positioning in combination with the geographic information system.

Preferably, the specific process in step 3 is as follows, a Mask R-CNN algorithm training model is established by adopting a convolutional neural network, and unsafe behaviors are identified and an alarm is given.

Preferably, in step 4, the road surface image is formed by continuously detecting the road surface quality and the structural condition by using a radar.

Preferably, step 6, a fusion navigation model of station internal navigation and station external navigation is constructed according to the station area electronic map in step 1, an optimal driving route of an intersection point and an accident point is automatically generated in real time through an intelligent path planning algorithm, and a rapid guide channel of the rescue vehicle from the outside of the station area to the internal accident point is established.

Compared with the prior art, the invention has the following beneficial technical effects:

the invention provides an emergency dispatching method for a mountain wind power plant, which is based on a high-precision in-station map and simultaneously utilizes personnel and vehicle positioning technology to position and monitor the positions of personnel and vehicles in real time; non-safety factors in the operation and maintenance process of personnel are identified by utilizing a deep learning algorithm and an alarm is given; automatic detection and intelligent identification of road surface defects are realized by utilizing image identification, alarm reminding of various dangerous factors is realized, and a route is planned again for a driver; the remote command processing of the field emergency is realized, and the safety of field operation and maintenance personnel is comprehensively guaranteed for rescuing vehicles. The method for building the emergency command center of the mountain wind farm provided by the invention realizes monitoring and management of personnel and vehicles, and is beneficial to improving the safety of operation of the mountain wind farm.

Drawings

FIG. 1 is a diagram of a mountain wind farm emergency dispatching system of the invention;

FIG. 2 is a flowchart of a method for identifying a person approaching a hazard source in step 3 of the present invention;

fig. 3 is a flowchart of the method for identifying that the worker does not wear the ppe in step 3 of the present invention.

Detailed Description

The present invention will now be described in further detail with reference to specific examples, which are intended to be illustrative, but not limiting, of the invention.

The invention relates to an emergency dispatching method for a mountain wind power plant, which comprises the following steps,

step 1: collecting geographical information such as a fan position, a transformer substation, road information, a concave land, a hilly land and the like, and drawing a station area electronic map;

step 2: positioning operation and maintenance personnel and vehicles in the wind power plant;

and step 3: the problems that a worker enters a construction site and does not wear a safety helmet, climbing operation under a high wind speed condition and the like are rapidly identified through an intelligent technology by adopting a convolutional neural network;

and 4, step 4: identifying the road defects, providing an alarm for a driver by identifying the defect types and the defect ranges, and giving corresponding driving suggestions;

and 5: setting a regional electronic fence, automatically discriminating external non-admittance persons and vehicles through face and image recognition, and calling and alarming and driving away persons who do not enter, leave and stay in the electronic fence region according to a set rule

Step 6: a rapid guide channel for the rescue vehicle from the outside of the station area to the inside accident point is established, so that the emergency rescue efficiency is improved, and the personnel and economic losses are reduced.

Examples

As shown in fig. 1, step 1: the method comprises the steps of drawing a station area electronic map, collecting geographical information such as power stations, road information, concave lands, hills and the like by adopting an unmanned aerial vehicle aerial photography technology, manufacturing a high-precision scalable plane map, combining CAD (computer-aided design) mapping of stations, and compiling and annotating the longitude and latitude of points such as a machine body and a peripheral building through conversion calculation of a coordinate system. And cutting the plan into a tile map of countless blocks 1X1 by using a map slicing tool, performing index positioning on the tiles by using a map API, and displaying the tiles in the map. And the station map is fused with the off-station map through coordinate system conversion to form a complete geographic information system.

Step 2: and (3) establishing a high-precision station regional map based on the step (1), and realizing real-time positioning of personnel by wearing positioning labels by the personnel and realizing vehicle positioning by using navigation systems such as Beidou adopted by operation and maintenance vehicles.

And step 3: the convolutional neural network is adopted to quickly identify the problems that personnel enter a construction site without wearing a safety helmet, the climbing operation is carried out under the condition of high wind speed and the like through an intelligent technology.

Firstly, the dangerous behaviors of the field operation are divided into: access to sources of danger, improper use of personal protective equipment, etc.:

as shown in fig. 2, the alarm judgment process for approaching a hazard source is as follows: firstly, identifying personnel and a danger source by using a Mask R-CNN algorithm, judging whether the personnel and the danger source have a coexistence relationship, if coexistence continues to judge the positions of the personnel and the danger source, and if the positions are smaller than a set threshold value, further judging the orientation relationship of the personnel and the danger source, and if the personnel and the danger source are in a preset dangerous orientation, giving an alarm; for example, when a danger source (such as a crane) and a person are in the same picture, whether the distance between the person and the danger source is too close or not, whether the person is right below the danger source or not is judged through a convolutional neural network technology, and the like, so that reminding is performed, and the safety of workers is protected.

As fig. 3 does not wear the ppe: firstly, a scene needing to wear the personal protection equipment is calibrated in advance, whether an operator and scene features coexist is judged by using a Mask R-CNN algorithm, if yes, whether the operator is in an operation state is judged by human body joint posture recognition, if yes, whether the personal protection equipment and the scene features coexist is further judged, and if not, alarm prompt is carried out. For example, when a worker welds, the worker and an electric welding tool coexist, whether the worker is in an electric welding posture or not is judged, if yes, whether protective equipment of the worker is worn or not is judged, and if the worker does not wear the protective equipment, an alarm prompt is given.

And 4, step 4: continuously detecting the quality and the structural condition of the road surface by adopting a radar, automatically identifying the type and the range of the road surface defect by utilizing the image processed based on the convolutional neural network, and sending out early warning information to remind a driver of changing lanes or replanning a route;

and 5: setting an electric field area level and a wind turbine generator level electronic fence, counting vehicle license plate information and face information of workers in the wind power field, acquiring the license plate information and the face information of the workers close to the vehicle through cameras arranged outside the electric field area and on the wind turbine generator, automatically screening external non-admittance persons and vehicles, and calling, alarming and driving away the persons who do not enter, leave or stay in the electronic fence area according to the set rules;

step 6: and (2) constructing a fusion navigation model of station internal navigation and external navigation based on the accurate map in the step (1), butting the fusion navigation model with a navigation system commonly used by rescue vehicles, and automatically generating the optimal driving route of an intersection point and an accident point in real time through an intelligent path planning algorithm. The path information is transmitted to the external rescue vehicle through the modes of mobile phone positioning, station internal path static guidance diagrams, marker reminding and the like, and the external rescue vehicle is indicated to rapidly arrive at the accident site.