Research On The Key Technologies Of Private Wireless Access Networks For Smart Grid

With the development of power dispatching automation and electric power production and the boost of the information and communication technologies, based on the integrated high speed two-way communication networks and a variety of other advanced technologies, the concept of smart grid is established. The public wireless network and wired communication technology are difficult to meet the communication requirements of the smart grid. In this thesis, the problem of data transmission and base station deployment for the LTE based multi-service bearer private wireless access network for smart grid is investigated.Firstly, the system model of LTE based private wireless access network for smart grid is presented and the research focus of this thesis is to solve the problem that where to deploy base stations and how to alloca-tion wireless resource for smart grid devices to satisfy the uplink communication requirement between base stations and electrical devices. This problem is decomposed into the resource allocation sub-problem and the site selection sub-problem. The resource allocation sub-problem is used to resolve the problem of how to allocation wireless resource for smart grid devices and the site selection sub-problem is used to resolve the problem of where to deploy base stations. Then the resource allocation sub-problem can be solved based on the results of site selection sub-problem, and the payoff provided by resource allocation sub-problem can be used by the site selection sub-problem for guided search of better deployment locations. Finally, the problem can be solved by running these two sub-problems in iterative fashion.Subsequently, we firstly study the resource allocation sub-problem. In this thesis, we proposes two allo-cation algorithms:the interference graph based allocation algorithm and the power control based allocation algorithm. The interference graph based algorithm consists of four steps:interference graph construction, benefit function table construction, priority based channel allocation, and payoff calculation. The power con-trol based allocation algorithm also consists of four steps:device grouping, uplink transmission power control, priority based channel allocation, and payoff calculation. The simulation results show that the power control based algorithm has better performance than the interference graph based algorithm.Then, we study the site selection sub-problem. In this thesis, two site selection algorithms are proposed which are based on simulated annealing and particle swarm optimization, respectively. We compare the performance of the proposed algorithms with the K-means based algorithm. For the simulated annealing based site selection algorithm, we firstly generate the initial BS locations, then calculate the payoff of the BS locations according to the resource allocation algorithm and define the payoff as evaluation function, then we update the BS locations based on the simulated annealing principle. The update of the BS locations proceeds in iterations until reach the end condition. For the site selection algorithm based on particle swarm optimization, we firstly generate a number of initial BS locations as a particle swarm, then calculate the payoff of each BS locations according to the resource allocation algorithm and define the payoff as the evaluation function of each particle, then we calculate the optimal position of each particle and the optimal position of all particles, then update the position and velocity of each particle, then update the optimal position of each particle and optimal position of all particles. The update of the position and velocity proceeds in iterations until reach the end condition.Finally, combined with different resource allocation algorithms, this thesis proposed the overall joint op-timization approach to iterate over resource allocation algorithm and site selection algorithm. The simulation results show that the PSO based site selection algorithm has better performance than the SA base algorithm, and both this two algorithm have better performance than the K-means method. Finally, we deploy base s-tation in practice under the guidance of the simulation results of the overall approach. The results show that the joint optimization algorithm proposed in this thesis can solve the base station deployment and resource allocation problem in private wireless access networks for smart grid.