Name / Th. Bopp et al. / Session / Alpha
Company / UMIST / Block / Discussion Block 4.3
Address / Ferranti Building/A4,Sackville Street / Question n° / 7
PO Box 88,Manchester,M60 1QD,UK / Language used on the floor / English
Phone / +44 (0)161 200-8721 / Accompanying visuals on file ? / Yes
e-mail /
Commercial and Technical Integration of
Distributed Generation into Distribution Networks
The aim of the project described in the presentation is the commercial and technical integration of distributed generation (DG) in distribution networks. It focuses on three main issues: the commercial aspects of DG, the development of a Distribution Management System Controller and system protection issues.
DISTRIBUTION MANAGEMENT SYSTEM CONTROLLER
Active management of complex distribution networks with significant DG can be accomplished using a Distribution Management System Controller (DMSC). Slide 3 shows a possible architecture of a DMSC. The core of the controller consists of a state estimation algorithm and a control-scheduling algorithm. The inputs to the controller consist of real time network measurements, pseudo bus injections and network data including breaker status information. Each measurement has an associated variance that weighs it in the estimator. The control-scheduling block requires network constraints and contracts for taking control actions.
Distribution System State Estimation (DSSE)
In general, distribution networks are large with high R/X ratios, and complete SCADA systems are seldom available at 33kV and below. Hence the availability of real time measurements is limited and the deficiency of them must be compensated using estimates of load consumptions at the distribution transformers as pseudo measurements. In a measurement limited system, location of potential points for network estimation is important. These limited but strategically located measurements must be chosen to supplement the pseudo measurements. A technique to locate critical points for voltage measurement placement has been developed based on a series of load flows. The method locates a given number of measurements to minimise voltage variances on the unmeasured buses.
Example: Uncertainty of Voltage Estimates
A conventional state estimation (SE) algorithm has been applied to the DSSE and tested on a 11 kV generic distribution system. The inputs to the algorithm are the P and Q load estimates at all the nodes and two real time voltage measurements, which are assumed to follow a Gaussian distribution. The load model uses daily load profiles for different types of customers. A more complex load model that provides the variance of the pseudo-measurements depending on the number of customers connected to each transformer was also studied. The SE was run considering a voltage measurement accuracy of 1% and a fixed pseudo-measurement accuracy of 25% at all the load buses. The voltage measurement at the secondary of the 33/11kV substation will normally be available and for this reason it was used as one of the inputs to the estimator. Slide 5 shows the accuracy of voltage estimates obtained at all nodes when the second measurement is placed at one critical (node 51) and at a non-critical (node 29) point. The simulation results show how the voltage estimates are generally improved when a measurement from a critical point is used in the estimator compared to a measurement from a non-critical point.
Control Scheduling and DMSC Outputs
The main objectives of the controller are to take decisions in order to maximise the penetration of DG and to maintain the constraints on voltage limits, lines, generators, transformers and commercial contracts at all times.
Active management
The commercial stream of work is concerned with establishing the value of the integration of DG and active management (AM). AM techniques enable the distribution network operator to maximise the use of the existing circuits by taking full advantage of generator dispatch, control of transformer taps and any reactive compensation devices in an integrated manner. The applied methodology in this paper combines an optimal power flow simulation and cost benefit analysis to determine the added value, which equals the sum of additional benefits minus the additional costs.
Example: Cost benefit Analysis of active management schemes
Different combinations of plant capacity and active management schemes were investigated. An evaluation of a wind farm project at a predetermined location in a rural 11kV network was carried out. Three AM schemes were studied. These included generation curtailment, reactive compensation and on load tap changer voltage control schemes. The study takes into account capital costs, OM&R (operation, maintenance & repair) costs, costs of AM schemes, savings from economies of scale and revenue from both energy sales and environmental incentives. Maximum net present values NPVs, as depicted in slide 7, indicate the optimum DG plant capacity and AM scheme combinations. Positive NPVs represent the added values of AM for DG, which result from increased income streams. The NPV curves start at the maximum DG export capacity that does not require network reinforcement or AM. In this study all three AM schemes enable increased energy exports but have different technical and economic limits.
PROTECTION ISSUES
The increased connection of generation to the distribution network can cause problems with existing protection schemes. This work discusses some of these problems and introduces one new approach to overcome them. Adopting such a scheme can facilitate the connection of larger quantities of generation to the distribution network. More details can be found in the supporting text document “CHILVERS_UK_author_ALPHA4_BLOCK3_Question7”.
Guidelines for Authors' participation in ALPHA sessions1/2