Use this URL to cite or link to this record in EThOS:
Title: Model predictive based load frequency control studies in a deregulated environment
Author: Ejegi, Eyefujirin Evans
ISNI:       0000 0004 6350 4628
Awarding Body: University of Sheffield
Current Institution: University of Sheffield
Date of Award: 2017
Availability of Full Text:
Access from EThOS:
Access from Institution:
A fundamental objective in power system operations is to ensure reliablity and quality supply, and one key action that aids the accomplishment of this objective is the load frequency control (LFC). Primarily, LFC is an automatic action that aims to restore system frequency and net tie line power between a control area (CA) and its neighbours to their scheduled values; these quantities deviate when there is an imbalance between active power demand and supply in a synchrononus interconnection. This thesis aims to investigate a model predictive control (MPC) technique for LFC problems in a deregulated power system environment which has become a challenging task. In deregulated power interconnections, generation companies (GenCos) and distribution companies (DisCos) exist in each CA, and a transmission system operator (TSO) in each area is responsible for grid reliability. Each TSO handles LFC in its CA and ensures that market participants (GenCos and DisCos) in other CAs have an unbiased and open access to its network. As a result, there has been a rise in cross-border transac- tions between GenCos and DisCos for bulk energy and load matching (LM) and consequently large frequency fluctuations recently. DisCos can participate in LFC by making bilateral LM contracts with GenCos. An extensive review of the LFC literature, in terms of strengths and weaknesses of different control techniques, is presented to identify the key gaps. The review reveals that MPC can bring some benefits in the deregulated environment but its strengths are underexploited. Beginning with a small-scale system to provide insights into deregulated system modelling and predictive control design, a centralised MPC (CMPC)-based LFC scheme is proposed for a 2-area deregulated power system with measured (contracted) and unmeasured (uncontracted) load changes, where the areas are assumed to equally rated. The 2-area deregulated system is developed by incorporating bilateral LM contracts in the well known traditional LFC model as a new set of information. It is assumed that DisCos handle contracted load changes via bilateral LM contracts with GenCos and a TSO handles any variations outside the LM con- tracts (uncontracted) via a supplementary control scheme which represents the CMPC. The CMPC algorithm is developed as a tracking one, with an observer to provide estimates of the system states and uncontracted load changes. Also, input and incremental state constraints, which depict limits on LFC control efforts and generation rate constraints (GRC) respectively, are considered. A simulation comparison of the proposed CMPC solution and optimal linear quadratic regulator (LQR) demonstrates the efficacy of CMPC. Developing deregulated LFC models for larger systems with complex topologies and a large number of CAs/market participants could be laborious. Therefore, a novel generalised modelling framework for deregulated LFC is further proposed. The key benefits of the generalised framework is that it provides a relatively easy and systematic procedure to develop deregulated LFC benchmark systems irrespective of the interconnection size, topology and number of market participants. It also offers the flexibility of accommodating LFC studies where CAs have either equal (often assumed) or unequal (more pragmatic) rated capacities. A 7-area deregulated benchmark model is developed from the generalised framework to illustrate its usage and significance, and the importance of incorporating area rated capacities is demonstrated via simulations. In addition, a 4-area benchmark model is developed to provide a reader with more insight into how the generalised formulation can be applied to develop LFC models for an arbitrary network. Furthermore, to demonstrate the scalability of an MPC design procedure, the CMPC proposed previously is extended to examine the LFC problem of the 7-area system. Key novelties here are CAs are assumed to have unequal rated capacities, some GenCos do not participate in supplementary control, and the control input to each GenCo is computed separately rather than a single lumped input for each CA which is the norm in previous deregulated LFC studies. The separate control inputs is to ensure that the input constraints of each GenCo is accounted for in the CMPC in addition to their GRCs and this is achieved by incorporating the area participation factors of the GenCos explicitly in the CMPC cost function. A test conducted on the 7-area benchmark confirms the benefits of this new approach. CMPC shows great potential for deregulated LFC in terms of multiple inputs coordination, effective disturbance rejection and constraints handling; however it is unrealistic for practical interconnections were CAs are operated by different organisations and have large geographical separations. This limitation is addressed by investigating a distributed MPC (DMPC) technique for rejecting incremental load changes, convenient for a finite number of control areas (subsystems), and therefore represents a more practical control architecture for LFC in multi-area systems. The proposed DMPC is non-cooperative and developed to operate using output feedback, where distributed observers using local measurements are developed to provide uncontracted load changes and subsystem states’ estimates to local MPCs. Moreover, the DMPC, unlike other non-cooperative schemes, is simple and devoid of extensive offline parameter tuning. Using the 4-area and the 7-area benchmarks models developed as test systems for the proposed DMPC, some comparisons of simulations results, regulation cost and discussions are provided between the proposed DMPC and alternative MPC schemes.
Supervisor: Rossiter, J. A. ; Trodden, P. Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available