Utility Theory-based Rush Repair Strategy Of Post-disaster Distribution Networks With The Real-time Adjustment

Our country is vast in territory and with various natural disasters. As the expansion ofthe distribution network and the increasing complex of its frame structure, the probabilityof widespread outage caused by faults in the post-disaster distribution network also greatlyincreases, which has brought huge challenge to the stable running and reliable powersupply of the distribution network. Based on accessing a large number of literatures, thispaper aims at effectively utilizing the limited repair personnel and resource andconstructing a real-time adjustment of rush repair strategy multi-agent system foremergencies during the repair, in order to generate rush repair strategy rapidly andminimize the comprehensive economic losses caused by faults.According to the concept of decision analysis, this paper first fully analyses the rushrepair problem in distribution network, ascribing the fault task allocation problem to thedecision problem of certain situation and ascribing the real-time adjustment of rush repairstrategy problem to the decision problem of uncertain situation. Then the utility theoryused to solve the decision problem with risks is introduced into the fault task allocationproblem, achieving an effective allocation of the fault tasks to repair squads.For the actual situation of the insufficient repair personnel and resource whenallocating fault tasks using the utility theory, the attribute vector is first proposed toquantify the repair information, defining the repair squad ability vector, the fault taskdemand vector and the concept of repair squad preference and failure degree. Then, inorder to quantify faults’ satisfaction degree for accepting the repair of each squad, theutility function model is established based on attribute vectors and revised for the squadcollaboration problem, thus developing the utility function matrix. Finally, the fault taskallocation scheme set based on the optimal utility is obtained via the roulette wheelselection method. Simulation on the33-bus distribution network verifies the effectivenessof the proposed method.For the uncertain problem of emergencies during the repair, this paper puts forwardthe real-time adjustment of rush repair strategies using multi-agent theory after the evaluation of emergencies. First, four types of emergency are defined and thecorresponding evaluation principles are formulated; then, the RA-MAS is built based onthe multi-agent theory. In the real-time adjustment of rush repair strategy multi-agentsystem, the real-time adjustment agent divides the fault levels, develops the rush repairprinciples, establishes the double-level mathematical optimization model of distributionnetwork multi-fault rush repair, and puts forward a new discrete mechanism of thebacterial colony chemotaxis algorithm; the distribution generator agent, the emergentpower-vehicle agent and the switch control agent formulate their own operation rules andcooperate with repair squad agents to implement the rush repair, finally realizing thereal-time adjustment of the rush repair strategy according to the emergencies. Simulationon a distribution network verifies the proposed real-time adjustment of rush repair strategymulti-agent system can deal with emergencies effectively and give the real-time adjustedrush repair strategy rapidly.