Study On Energy Management Methods In Wireless Sensor Networks For Stress Monitoring In A Hoist Drum Surface

Mine hoisting equipment is an important tool used for the connection between overground and underground,and therefore its running status can affect mine production and personnel security.Mine hoisting equipment will cause a significant change in the surface stress of hoisting drums when overwinding and jamming happen,compared with normal working conditions.Therefore,the hoisting status can be estimated by monitoring the surface stress of drums.Wireless sensor networks have characteristics of flexibility and self-organization,which adapts to drum surface monitoring.However,there is a contradiction between energy constraints and sustainable and high-quality monitoring when wireless sensor networks are applied to monitor the surface stress of drums.In shallow mines,due to the low spinning speed of drums,wireless sensor networks cannot harvest energy effectively.In this situation,the existing energy management strategies can neither predict the remaining capacity of node batteries nor save energy and guarantee data priorities.In deep and ultra-deep mines,wireless sensor networks can harvest energy to supply energy.However,the existing energy management strategies cannot effectively balance energy and quality of service of data,failing to achieve sustainable monitoring.Hence,energy management strategies of wireless sensor networks have been studied to adapt mine hoisting environment.The main contents are as following:1)As the basis of this paper,wireless sensor networks are designed to monitor the surface stress of drums.The categories for abnormal running status are first analyzed,and then the relationship between the change of drum surface stress and the running status of hoisting equipment is analyzed by using an ANSYS Workbench software tool.Then,node hardware and deployment strategies are designed for high-speed and low-speed drums.2)An energy model of wireless sensor networks is built to access energy information.First,a duty-cycling-based energy consumption model is built;second,wind-induced piezoelectric energy harvesting experiments are conducted to build the energy harvesting modeling;next,the energy storage model is built:node battery models under constant current and constant power loading are built respectively.Then,a power-estimation-based state of charge prediction method for lithium batteries is determined by using the‘bridge'function of the equivalent current;in addition,a low-complexity remaining energy estimation algorithm is proposed based on the energy consumption model and the energy harvesting model.The results of energy harvesting experiments show that there is a power function relationship between the harvested piezoelectric energy and the spinning speed of the drum;the results of the state of charge prediction experiments show that the prediction error is less than 4%in the test working conditions,which can provide capacity warning information for monitoring nodes.3)In order to save energy and guarantee data priorities at the same time,a joint optimization scheme combined data transmissions with power allocation is designed for a wireless sensor network.This network is used to monitor the surface stress of low-speed drums in mines.First,a wireless sensor network model for stress monitoring is built.Then a data classification method based on data packet sizes is proposed since sampling rates of nodes vary according to sampling areas.Besides,a joint optimization model is built to minimize a utility function of energy and a data class index.Finally,a Lyapunov-optimization-based distributed network scheduling scheme is proposed by adding a weight factor W.Numerical simulation results show that the weight factor W can balance power consumption and received data priorities in the perfect and imperfect channel state information.4)In order to solve energy unbalance,an energy-aware and quality-of-service-aware protocol is designed for a wireless sensor network.This network is used to monitor the surface stress of a high-speed drum.A wireless sensor network topology,a queue length model and a throughput optimization problem model are built respectively.Then this optimization problem is transformed into two sub-problems.For the energy management sub-problem,a dynamic adaptive duty cycling energy strategy is proposed.For the quality-of-service management sub-problem,Lyapunov optimization method is applied with the addition of a virtual arrival Avir_n~c and a queue weight factor w~c,and a network scheduling algorithm is proposed.Numerical simulation results show that the proposed strategy can help nodes balance energy and monitor the running status of mine hoisting equipment for the long term;in addition,the proposed strategy can guarantee the maximum delay constraint and optimization throughput.5)A wireless sensor network used for surface stress monitoring of drums is built and simulated in Cooja to verify the proposed energy management strategies.First,software of energy management strategies is designed with the Contiki operating system and the strategies are based on the non-energy harvesting model in Chapter 4and the energy harvesting model in Chapter 5 respectively.In addition,both energy management strategies are simulated in Cooja.The simulation results show that the the penalty factor V and the high priority queue weight factor w~H can adjust network performance,which is in accord with previous simulations in Chapter 4 and 5.