Research On Early Warning Model Of Cascading Failure For Power Grid Based On Brittleness Theory Of Complex System

With the rapid development of new energy power generation and deepening of power grid interconnection, the grid structure is increasingly complex and the operation modes become more complicated. All of thoese situations get the safe and stable operation of power grid into unprecedented dangers. Under the circumstances, local disturbance spreads to a large range with faster speed. Some faults of local grid might be propagated and transmitted quickly, which may result in eascading failures and then the blackouts, or even the breakdown of power system, which might bring a huge loss to the socio-economic and national life. Consequently, the effective early warning and control for cascading failures are the key to avoid blackouts. Brittleness theory of complex system is a new approach to analyse the cause, occurrence and propagation of cascading failures. From the perspective of system, it could analyze cascading failures of power system in terms of network, and ensure that the brittleness is not frequently inspired to cause blackouts as much as possible.The cascading failure warning model based on the brittleness theory of complex system is analyzed in this dissertation. The major works are summarized as follows.The brittleness theory of complex system is introduced and given a detail mathematical and physical explanation. Characteristic analysis is proposed with the concepts of brittleness process and brittle element. The dissertation puts detailed description on typical brittle element forms and the brittleness of electric power system, which lays the foundation for the researches.By taking full account of the uncertainty effects in the process of cascading failures, a fuzzy comprehensive evaluation model for brittle element of cascading failures is established. This model calculates the occurrence probability of brittle element based on the quantitative assessment value of the factors that affect cascading failures, and uses severity indexes to assess the severity of brittle element from system transient stability and load losses. At last, it assesses and grades the risk of brittle elements using the fuzzy comprehensive evaluation method, which establishes the basis for cascading failure early warning combined with the interpretation of risk levels.From the view that brittleness is the nature of power system, a brittle source identification model for cascading failures is proposed. This model analyses the mechanism of cascading failure by brittleness relevance and entropy increase, and then makes sure initial faults of cascading failures by brittle degree, lastly, gets brittle risk entropy into assess the impact of component outage from power grid operation and the load removed. Meanwhile, In this model, there is a process to find out the high-risk lines. Making a quick identifying of brittle sources could reduce computation complexity greatly, and bring a more clear purpose for preventing cascading failure.Taking the line failure rate under severe weather into consideration, an early warning model based on brittleness theory of complex system is created combined with fuzzy comprehensive evaluation of brittle element and identifying of brittle sources. The model could give a early warning of brittle element in different risk level effectively, and provide intuitive basis for dispatchers to adjust operating modes and balance the power flow distribution.Seeing that the failure-state information in different stages of cascading failures cannot be mastered, a model to defense cascading failures in complex power grid based on multistage games with incomplete information is proposed, it is proposed to characterize the affects of component faults on power grid under incomplete information by the accident state index. Combining with game theory, the proposed model recognizes the offensive side’s current strategy during different courses of cascading failures and gets the payoff function. According to payment function values, the defensive side takes appropriate defense strategy by calculating the DC power flow sensitivity matrix and load node voltage sensitivity matrix. The proposed model provides direct and accurate evidences for adjustment measures at the initial stage of cascading failures and emergency control measures at the rapid spreading stage.