Safety Risk Assessment of Electric Power Operation Site Based On Variable Precision Rough Set

Abstract

In order to improve the intelligent informatization level of electric power production safety and reduce the accidents, the paper constructs a dynamic perception scheme of electric power production site that utilizes multi-dimensional information such as personnel location, equipment status, and image information. This method uses a multi-sensor network to realize the real-time perception of the image and position information of dynamic power work objects, then uses object identification and intelligent analysis to acquire the dynamic factors. Static factors are selected through questionnaire and historical data, and variable precision fuzzy theory is used to calculate the weight of static factors at dynamic power operation sites. A comprehensive evaluation is established to perceive risk and estimate security probability based on static factors and dynamic scene information. The application system can present the situation of the operation scene, and then realize safety assessment and early warning of the dangerous situation in the case of dynamic power monitoring. This method can prevent safety accidents and enhance the overall safety of power operations.

Safety risk assessment method of key personnel in infrastructure projects based on image data coupling identification

“…For the alarm times of non-dangerous actions, the alarm times of the proposed method are 0, which shows that the proposed method has no false alarm when judging non-dangerous actions. For the alarm times of the first, second and third dangerous actions, the alarm times of the proposed method are less than those of the methods in reference [3] and reference [4], which shows that the proposed method has higher accuracy and sensitivity in judging different dangerous action levels. To sum up, the proposed method can greatly improve the accuracy of dangerous action identification, and has better performance and effect.…”

“…In order to verify the practical application performance of the proposed method, reference [3] method and reference [4] method are introduced for comparative test, and the evaluation result is based on whether the danger alarm signal is issued or not, and the experimental results are shown in Table 2. It can be seen from Table 2 that the alarm times of the proposed method under different dangerous action levels are close to the real results, and it has better performance than the methods in reference [3] and reference [4]. The proposed method can accurately determine whether the action is dangerous or not, while the other two literature methods have the probability to determine the error and send out the wrong alarm signal.…”

“…Reference [3] puts forward an acceptable injury risk assessment method for employees of the Russian Federal Fire Department, and analyzes the cases of employees of the Russian Federal Fire Department who are injured while performing their professional duties. Reference [4] puts forward a safety risk assessment method of power operation site based on variable precision rough set. This method can present the situation of the job site, and then realize the safety assessment and danger warning under the condition of dynamic power monitoring.…”

An operation site security detection method based on point cloud data and improved YOLO algorithm under the architecture of the power internet of things

“…Due to the complexity of power operation sites, traditional detection methods find it difficult to cope with various factors such as colorful safety helmets, ever-changing work scenarios, changes in lighting in the operation environment, and changes in time and small people. Therefore, it is necessary to combine deep learning and object detection algorithms to perform reasonable operational processing on the surveillance video data, in order to monitor the behaviour of workers during construction, such as wearing safety helmets, and greatly reduce the cost of human resources consumed in power system operations, and greatly improve monitoring efficiency [28,29]. However, due to the unique nature of power field scenarios, some deep learning based object detection algorithms still cannot achieve their best results.…”

Research on Electricity Operation Behaviour Recognition Strategy Combined with Intelligent Image Recognition and Its Key Technology

“…Yu, H. et al provided a regional semantic learning and map point recognition method for power plant operation scenarios, which was analyzed by simulation practice and found to have a positive impact on the management and production of the power plant to ensure the stable, efficient, and safe operation of the plant [15]. Chang, Z. et al utilized multi-dimensional information such as personnel location, equipment status, image information, etc., to Design a dynamic sensing scheme for a power production site, which can prevent the occurrence of production accidents and keep the power operation at a steady point [16]. Li, Y. et al proposed a hybrid evaluation model based on Decision Making Experimental and Evaluation Laboratory Combined Method (DEMATEL-ANP) analyzing network process approach and fuzzy sets.…”