- Title
- Advanced methods for power system security operations
- Creator
- Zhang, Rui
- Relation
- University of Newcastle Research Higher Degree Thesis
- Resource Type
- thesis
- Date
- 2014
- Description
- Research Doctorate - Doctor of Philosophy (PhD)
- Description
- The fundamental role of an electrical power system is to deliver the requested electricity to customers in a secure and economical manner. In general, to realize the proper functionality of a modern power system, efforts should be devoted to operation and planning stages. This research focuses on the operation stage of power systems. Specifically, advanced methods are developed for three essential problems in power system security operation studies: 1) short-term load forecasting (STLF), 2) security-constrained optimal power flow (SCOPF), and 3) dynamic security assessment (DSA). Where the STLF is the basis for system operation, SCOPF aims to economically dispatch the system while satisfying the security constraints, and DSA focuses on the dynamic security issue of the power system. This research is carried out in a logic way to address the three problems. In the area of STLF, a model based on extreme learning machine (ELM) technique is first developed for the Australian National Electricity Market (NEM). The model is an ensemble of a series of strategically arranged individual ELMs. Compared with some existing methods, it is much faster in model training and can provide more accurate load forecasting results. After this, a composite k-nearest neighbor (kNN) model is developed to deal with the situation that limited temperature input is available. The model uses only the minimum and maximum temperature forecasts as the model input and can provide reasonably accurate load forecast results. At last, a hybrid model for time-varying reactive power load forecasting is developed. This model is among the first in filling the gap between the need for accurate reactive load forecast and current industrial practices which are mainly using a fixed power factor to derive the reactive load. The developed hybrid load forecasting model has been practically applied in a real-world distribution network in U.S. In the area of SCOPF, research efforts are firstly devoted to the corrective security-constrained optimal power flow (CSCOPF) problem. In contrast to traditional OPF models, the CSCOPF considers the use of corrective control (CC) actions to regain the system security in the post-contingency state, and is able to provide the same security level with more economical generator dispatching solutions over the conventional OPF models. This research develops a hybrid computational strategy to solve the much more complex CSCOPF problems. Besides of this, this research proposes a preventive-corrective SCOPF (PCSCOPF) model and corresponding solution method. The PCSCOPF aims to optimally coordinate the preventive control (PC) and CC against the probable contingencies during system operations. In the area of DSA, this research applies the intelligent system (IS) strategy which has been identified in recent years as a promising approach for much faster DSA. First, a comprehensive review and systematic classification of existing methods is conducted. Subsequently, for pre-disturbance DSA, a systematic study on feature selection stage is conducted, and two alternative algorithms are introduced to this area. Then, an IS model for on-line voltage stability margin (VSM) prediction is developed. It employs system loading direction as the input and is much faster in model training and more accurate. After this, a novel IS for post-disturbance transient stability assessment (TSA) is proposed. The proposed IS can be fully integrated in the wide-area protection and control (WAPaC) systems. Compared with existing models, it is radically superior as it is self-adaptive in making the right TSA decision at an appropriate earlier time, hence allowing more time to take the remedial control actions against the instability. The developed methods have been numerically validated on a number of benchmark test systems, and shown satisfactory performance. They can be practically applied for enhanced security operation of modern power systems.
- Subject
- electric power system; power system security; power system stability; load forecasting; optimal power flow
- Identifier
- http://hdl.handle.net/1959.13/1048178
- Identifier
- uon:14884
- Rights
- Copyright 2014 Rui Zhang
- Language
- eng
- Full Text
- Hits: 1732
- Visitors: 2410
- Downloads: 401
| Thumbnail | File | Description | Size | Format | |||
|---|---|---|---|---|---|---|---|
| View Details Download | ATTACHMENT01 | Abstract | 502 KB | Adobe Acrobat PDF | View Details Download | ||
| View Details Download | ATTACHMENT02 | Thesis | 4 MB | Adobe Acrobat PDF | View Details Download |