Research On Techniques Of Energy Management In Cyber Physical System

The Cyber-Physical System(CPS)is a multi-dimensional complex system for integrated computing,network and physical environments.Through the integration and deep cooperation of 3C(computing,communication,control)technology,it realizes real-time perception,dynamic control and information service for large industrial systems.CPS also realizes computing,the integrated design of communication and physical system which make the system more reliable,efficient and real-time collaborative with important and broad application prospects.CPS including embedded system,IoT and real-time systems technology is fused together to form a new system architecture.As more and more CPS is applied to industrial production for controlling and monitoring,the energy consumption is increasing fast and it's becoming a huge burden for smart grid.For this new significant issue,this dissertation mainly explores the energy management techniques of each module in CPS.Specifically we discuss low-level embedded system,middle-layer wireless sensor network,upper-layer data management entity(Colocation data center)and co-located renewable energy in Colocation.Each entity constitutes a different functional module in a large industrial Cyber-Physical System.Each module interacts with each other through data flow to achieve cooperation.The dissertation expands from the low level entity to high-level entity and introduce the energy management features and energy saving technologies of different entities.(1)In order to minimize energy consumption of real-time parallel tasks for embedded systems,this dissertation proposes a solution for energy minimization problem for parallel task systems with discrete operation modes and under timing constraints based on level-packing.For tasks with fixed(variable)parallel degrees.this dissertation first formulates the problem as a 0-1 Integer Linear Program(0-1 ILP),and then proposes a polynomial-time complexity twostep(rigid task)or three-step(moldable task)heuristic algorithm to determine task schedule,frequency assignment and also parallel degree.Our simulation results show that the heuristics consume nearly the same energy as 0-1 ILPs.(2)In order to extend the life cycle of wireless sensor network's nodes for railway monitoring system of heavy cargo transportation,this dissertation presents a real-life application design of wireless sensor networks in trains to monitor the goods conditions during long-distance transportation,and maximize the life cycle of the nodes.This dissertation studies the wireless sensor network deployment problem in developing a monitoring system with the goal of maximizing the network lifetime under constraints derived from the real application scenario.The key technical problem here is to determine the sensor placement and the transmission level for each sensor node,as well as the appropriate number of sensor nodes.This dissertation first formulates the problem with a realistic discrete power model as a mixed integer linear programming problem.Then,a two-step efficient deployment heuristic algorithm is proposed to satisfy these constraints step by step.The evaluation results indicate that the proposed algorithm performs almost the same as the optimal mixed integer linear solution.This dissertation also discusses a tested experiment in a laboratory environment,as well as the real implementation of the whole monitoring system.(3)Because industrial network transmission delays can seriously affect control performance and in order to maximize control performance in limited wireless resources,this dissertation presents the collaborative design of task control and network scheduling.This dissertation first linearizes task control requirements to provide quantitative optimization guidance for network design.Then it proposes a fixed priority scheduling method based on controlling cost growth trend to solve the problem of link layer access.At last it proposes the control task placement and shortest path method based on the Hungarian algorithm to solve the application and network layers.The simulation results show that comparing with the traditional method,the proposed method improves the control performance with less the wireless resources.(4)In order to maximize the economic profit for Colocation in response to the demand response signal from smart grid,this dissertation first proposes a reward mechanism for tenants who save energy during demand response period in order to incentive more tenants join in saving energy based on fair competition.According to the energy management structure of Colocation data center,the operators and tenants are divided into two levels in the market system.The purpose of this is to divide the relationship of benefit clearly and making better analysis model.For the two level structure,this dissertation proposes an iterative algorithm between two levels and this algorithm converges to an unique equilibrium solution provably.So the operator and tenants reach a win-win situation.At this moment,neither the operator or tenants can improve their individual economic profit by changing their own strategies.The results of the simulation experiments valid that the mechanism effectively incentives tenants to join the DR and receive economic benefits,and the algorithm converges to a unique equilibrium solution.(5)In order to help Colocation operator manages the co-located renewable energy,this dissertation proposes to integrate the output of renewable energy into the operator's decisionmaking,it also presents an algorithm based on stochastic optimization,which uses predictions of collocated renewable energy to update Colocation operator's best strategy.For the uncertainty of renewable energy's output,this dissertation finds the linear relationship between the output of the renewable energy and the prediction error through the competitive analysis method to help the operator maximize the economic benefits.The simulation results show that the more accurate of the prediction of renewable energy,the higher performance this algorithm would reach.In summary,this dissertation proposes energy optimization algorithms for multiple scale cyber physical systems,which effectively reduce the demand of power and improve system performance.The following chapters will detail the specific technical details.