Research On Key Technology Of Machinery Safety Risk Early Warning

Mechanical safety originates from the needs of people.It is the practice to achieve the state and condition that the human body is free from the harm of external factors during the use of machinery.At present,research on machinery safety-related technologies and standards mainly focuses on the design stage,that is,how to take measures in this stage to improve the safety level.However,accidents often occur during the usage of machines.The mechanical safety measures taken in the design stage cannot give an early warning of variations in the state of people,machines,and the environment caused by changes in time,application,etc.,when the machine is used,so they cannot avoid the occurrence of accidents or control the expansion of the consequences of accidents.Safety risk early warning technology is a technology that can alert or warn the crisis or dangerous state in advance by analyzing the risk level of the workplace.The intelligent factory has been able to realize the real-time monitoring of the status information of people,machines,and the environment,and sound the alarm according to the real-time monitoring information to reduce the risk and improve the working efficiency of the machine.This paper carries out research on the system building,key technology,and system realization of mechanical safety risk early warning to improve the quality and safety level of machinery equipment in China,lower the occurrence rate of potential mechanical safety accidents,and reduce casualties and property losses caused by mechanical safety accidents.It mainly includes the following aspects:(1)Aiming at the problem that the mechanical safety measures taken in the machine design stage cannot warn people,machines,and the environment in advance of possible unsafe conditions,a mechanical safety risk warning system is built.Based on the accident cause theory,the concept of mechanical safety risk early warning is proposed.A four-dimensional mechanical safety risk early warning model consisting of the logic dimension,time dimension,system dimension,and knowledge dimension is constructed.The key technologies for realizing mechanical safety risk early warning are proposed.It provides theoretical support for the development of mechanical safety risk early warning.(2)Aiming at the problem that the existing machinery safety risk assessment methods cannot be used to determine the monitoring elements of mechanical safety risk early warning,a new method is put forward in this paper.The process for determining the monitoring elements of mechanical safety risks early warning is proposed.The safety investment cost is used to determine the risk reduction measures as safety protection devices or machinery safety risk early warning system,and the annual fee method and fuzzy reasoning method are respectively employed to calculate the safety investment cost.An example is used to verify the determination of monitoring elements of robot mechanical safety risks early warning.With the ability to identify the risk of the system,estimate the level of the risk,and consider the costs,the proposed method can reasonably select the monitoring elements of mechanical safety risk early warning.(3)Human unsafe behavior is the main cause of mechanical injury.The following problems of human unsafe behavior detection and recognition in mechanical safety risk early warning are studied:(1)Aiming at the lack of safety distance calculation method between autonomous mobile machine and the human body,the concept of dynamic safety distance used for the positioning of the safety protection device is proposed.On this basis,the calculation method of dynamic safety distance between an autonomous mobile machine and the human body is presented.The experimental results show that this method can determine the safe distance between autonomous mobile machines such as robots,AGVs,and human bodies.(2)In view of the possibility of being bypassed when using human detection devices such as safety light curtains and safety laser scanners,and the problem of being unable to distinguish people or equipment,a depth camera is used as a mechanical safety human position detection device to achieve the goal of detecting people only and separating people from mobile equipment for early warning.(3)Aiming at the problem that the current personal protective equipment wearing detection system has a single function and cannot detect multiple personal protective equipment at the same time,the deep learning technology is applied to simultaneously detect the wearing of safety helmets,safety gloves,and safety glasses through the personal protective equipment wearing detection model.(4)In view of the lack of a risk value calculation model for machinery safety risk early warning,a new model is constructed in this study considering people,machinery,environment,and their coupling effects.When determining the weight of each monitoring element in the risk value calculation model,two coupling weight determination methods are proposed according to the qualitative or quantitative mechanical safety risk estimation method adopted in monitoring element determination,and an example is verified by calculating the risk value of robot mechanical safety risk early warning.The risk value calculation model and its weight determination method of mechanical safety risk early warning can comprehensively take people,machines,environment,and their coupling effects into consideration,calculate the system risk value based on real-time monitoring data,and achieve hierarchical early warning.(5)The realization of mechanical safety risk early warning depends on the risk early warning system.The mechanical safety risk early warning system is constructed and verified.The functional architecture of the early warning monitoring system,early warning classification system,and early warning measure system is given.A robot mechanical safety risk early warning system consisting of an early warning monitoring system,early warning grading system,early warning measure system,and 3D virtual monitoring system based on digital twins is constructed,and the robot mechanical safety risk early warning experiment is carried out.The warning system can realize automatic monitoring of mechanical safety risk factors,automatic classification of early warning levels,and automatic implementation of early warning measures.