| The distribution network,as the hub connecting the energy and user sides,is one of the important links in the power systems.With the continuous emergence and integration of the distributed powers and pluralistic electrical loads such as electrical vehicles and energy storage devices,some significant changes in the aspects as functions and characteristics of the distribution network have been raised.Specifically,the power flows have trend from the traditional one-way mode to the two-way mode,which makes the distribution network present an increasingly complex "multi-source" feature.Such the development trend has posed serious challenges to the safe operation and satisfactory power supply quality of the distribution network.Meanwhile,the distribution network also has some outstanding characteristics such as massive branches and wide supply ranges,flexible and multiple communication modes,large data amounts,complex and diverse data types,lack of a common model semantic foundation,and poor scalability of model architecture.With the deployment of extensive intelligent terminals,integration of distributed resources,and requirement transformantion of equipment operation and maintenance,the contradiction between increasing functional requirements and limited hardware resources has been rasised in the massive terminals in the distribution power Io T.Specifically,the complex security protection,computing capability,communication model,device operation/maintance and advanced business applications have putforwarded high demands for the hardware recources,which has caused too high investment costs on the terminals to popularize the distribution power Io T.Therefore,there is an urgent need to overcome the lightening problem of the terminal resources in the distribution Internet of things.Constructing the power distribution Internet of things with tight integration of traditional distribution automation and Internet of things technologies to achieve comprehensive interconnection and interoperability among the massive terminals in the distribution network has become the development direction and research focus of distribution network.In addition,with application of other advanced technologies such as edge computing and deep learning,comprehensive perception,data fusion and intelligent applications in the distribution network can be realized to improve the intelligent level and operating reliability of the distribution network.In order to ensure the safe and stable operation in the distribution Internet of things under the cloud-edge collaborative framework,robust network intrusion prevention strategy,feasible computing resource allocation mechanism,efficient seamless information interaction scheme,reliable terminal plug and play mechanism,and flexibly deploying various advanced applications are required.However,in current distribution network construction process,there exist some problems including poor security protection capability and lack of effective computing resource management,difficult access of the massive terminals due to different information models,non-standard data management caused by chaotic topology associations among intelligent terminals,etc.Correspondingly,there are many practical engineering pain points in the distribution network,such as difficulty in effectively exerting the value of operation data,low operation and maintenance efficiency,and difficulty in deploying advanced applications.Therefore,aimming at the core functional requirements in the distribution power Io T,as the instrusion prevention,equipment interconnection and plug-and-play mechanism,the novel security defense scheme,equipment interconnection protocol and network topology identification and update method are studied in this thesis.The above studies optimize the storage occupation,data interaction efficiency and other aspects,which can provide the core technologies to reduce the resource occupations and promote the rapid promotion of distribution power Io T.The main contents in this thesis are as follows.(1)Computing resource allocation considering intrusion prevention and the corresponding minimal resource occupation scheduling method: Considering the risk that hackers can easily invade the power distribution Io T via using the potential security vulnerabilities in the case that massive bottom distribution intelligent terminals are connect to the edge-side intelligent fusion terminals,a security intrusion prevention strategy of edge-side intelligent fusion terminals based on differential game theory is proposed.Taking the number of attack nodes,attack frequency and intrusion defense strength into consideration,the process of hacker intrusion into the edge-side intelligent fusion terminals and the defense of the edge-side intelligent fusion terminals is described as a non-cooperative game process.Then,based on the Isaacs-Bellman equation,the security intrusion prevention strength update strategy which adaptively changes according to the attack frequency is obtained.Furthermore,to satisfy the security intrusion prevention strategy,a minimal cloud-edge cooperative resource occupancy scheduling algorithm considering the constraints of request queuing delay,computing resource capacity and the number of servers is porposped.The results show that the proposed method can significantly reduce the frequency of intentional attacks and the number of successful intrusions.Correspondingly,the safety operation and abundant computing capability can be guaranteed for the power distribution Io T.(2)Low resource-occupancy information modeling method for multi distribution power Io T terminal interconnections: For heterogeneous power distribution Io T smart terminals,the differences in communication models and technical details may lead to the information islands.An architecture for information interacting in the cloud-edge collaborative power distribution Io T is designed.Then a unified information modeling method for power distribution Io T terminals based on the lightweight JSON data exchange format is proposed.The typical intelligent sensing terminals of the power distribution Io T are used as the examples to demonstrate informatization modeling,so as to provide a reference for the information modeling of other intelligent sensing terminals.Compared with the traditional information modeling method based on XML data exchange format,the modeling method proposed in this paper has advantages as lower storage space and network bandwidth requirements.The proposed method can provide a lightweight support for the terminal equipments and create the conditions for interconnection and large-scale applications of the multi terminals.(3)Plug-and-play and efficient adaptive topology updating strategy for terminal equipments in distribution power Io T: Considering the difficulty of manual maintenance,verification and upgrade for the access of massive power distribution smart terminals,a plug-and-play mechanism based on timed polling and active registration is proposed to realize remote automatic maintenance based on the information models of the terminals.Secondly,to solve the problems as topology disorder and untimely maintenance/update between the edge-side intelligent fusion terminals and the smart electric energy meters in the distribution station areas in current stage,a topology identification method for the distribution station area based on correlation analysis is proposed.The actual engineering case shows that the proposed plug-and-play mechanism and the station topology identification method have significantly improved the work efficiency of operation and maintenance.(4)Advanced distribution application cases in the distribution power Io T under cloud-edge collaboration framework: Taking orderly charging scheduling of electric vehicles as an example,a feasible implementation scheme is proposed for the new-type business development of the power distribution Io T under the cloud-edge collaborative architecture.Firstly,the power prediction algorithm based on deep belief network is deployed in the edge-side intelligent fusion terminals to predict the traditional residential loads,electrical vehicle charging demends and the active power outputs of renewables.Secondly,for the problems of excessive peak valley difference and frequent voltage fluctuation caused by disorderly charging of electric vehicles,an ordered charging strategy for centralized charging stations based on cloud-edge collaborative architecture is proposed.Specifically,the cloud-side master station sets the next-day power supply plan of each charging station on the basis of collecting the power prediction results from the edge-side intelligent fusion terminals.Moreover,the intelligent fusion terminal in each charging station sets the charging period and charging power for each electric vehicle on the basis of considering the user’s travel plan,charging times and other factors.Simulation results show that the proposed algorithm can reduce the peak-valley difference of load curve effectively and ensure the reliability of distribution network.In summary,under the support of the National Major Creative Program ―Industrial Internet Enabled Energy System and Public Service Platform Construction‖,taking the terminal in the power distribution Internet of things as the carrier,this thesis mainly focuses on the low cloul-edge resource occupancy model construction problem in the distribution power Io T,which includes there dimensions as cyber intrusion prevention,lightweigtht information modeling,intelligent fusion terminal plug-and-play and topology update.The studies in this thesis provide a series of solutions to reduce the resource consumptions in the distribution Io T terminals,and has been verified in the practical engineering cases. |
PDF Full Text Request