ESKOM DISTRIBUTION NETWORK PLANNING, A …

1 ESKOM DISTRIBUTION NETWORK planning , a quantum LEAP TO. MEET INCREASED LOAD growth AND improved NETWORK . reliability . Author & Presenter: CG Carter-Brown PhD ESKOM , Chief Engineer IARC. Co-author: R Stephen MSc MBA ESKOM , General Manager DISTRIBUTION Capital Program 1. Introduction speed in the shortest time possible. The knowledge/plans of existing planners need to be available for easy reference by new The challenges faced by ESKOM DISTRIBUTION planners. A career path needs to be created to NETWORK planning include the following: retain experienced planners. Distributed Generation (DG): The requirement Load growth : Load growth has doubled as to integrate co-generation and renewable compared to the early 1990s. The number of generation requires a new set of skills as NETWORK planners in ESKOM has not increased DISTRIBUTION planners traditionally do not have proportional to the increase in load growth .

2 Experience with the integration of generation. Planners need to produce more, and need to be Demand Side Management (DSM) and Local more effective. Integrated Resource planning (LIRP): DSM. Universal Access planning (UAP): The target and LIRP require planners to assess both to complete universal access to electricity by supply side and demand side alternatives to 2012 places an additional burden on planners NETWORK constraints. Optimal solutions could to ensure that NETWORK infrastructure can involve DSM and/or utility owned distributed support this additional load. generation. Additional data, skills, models and tools are required for these assessments. reliability : Incentive Based Regulation (IBR). and NETWORK performance targets necessitate This paper describes present initiatives to address that ESKOM DISTRIBUTION improves present these challenges. NETWORK performance levels. A major step change in performance can only be achieved in conjunction with capital expenditure related to 2.

3 Strategy NETWORK redundancy and risk (reducing the number of customers and size of load at risk for As per figure 1, effective NETWORK planning builds NETWORK faults). Planners need to include on base data and systems: reliability implications in NETWORK investment decisions. Base Data: Essential data for power system Note: Capital expenditure (primary plant analysis, load forecasting and alternative planning and design) is a key aspect of NETWORK evaluation. reliability improvement, but it is recognised that Base Systems: Software systems for power there are other business aspects such as system analysis (load-flow and fault-level), load systems and data collection methods, forecasting and need and project registers. performance measurement and reporting standards, operation and maintenance practises, system automation, telecontrol, Research Support, stds, training SCADA visibility and staff skills and training.

4 LIRP. Data: Planners, it is estimated, presently spend 80% of their time searching for data, and 20% Advanced of their time performing actual planning (load Methods forecasting, need identification, alternative evaluation etc). Essential data needs to be Base Systems validated and available for easy export to Line planning tools. Staff turnover and skills: The average Base Data experience of a NETWORK planner is typically less than 2 years. Staff turn over is high. Training systems are required to get new planners up to Figure 1: NETWORK planning dependencies Advanced Methods (probabilistic risk assessment required by planners is captured in SmallWorld and reliability analysis) and LIRP can only be for future NETWORK additions and modifications. implemented once base data and systems are in Standard values: For the purposes of NETWORK place. planning studies, standard (typical) values for equipment attributes can be utilised.

5 Examples Cutting across all of the above mentioned include typical impedances and losses for components is the need for support, standards, transformers, and per-unit length impedances guidelines and training. (R, X and B values) for power lines and cables. A project is nearing completion for the The strategy adopted by ESKOM DISTRIBUTION is as implementation of standard value libraries in follows: SmallWorld such that typical attribute values are available for attributes that have not been Base data and systems are critical, and will populated via data capture. typically be addressed via line projects and Statistical metering: There is renewed focus initiatives (short term within 18-24 months). on remotely downloaded statistical metering Advanced systems and LIRP can only be such that load profile data of active and reactive effective once base data and system issues power (30 minute interval) is available for all have been resolved, and will typically be major power transformers (typically at HV/HV.)

6 Addressed via research (medium term between and HV/MV substations), HV lines/cables and 1 and 3 years). MV feeders. Supporting standards, guidelines and training material will be developed and rolled out in The net objective of these projects is to provide conjunction with the delivery of the planners with easy access to critical NETWORK and components. load data. Overall coordination is performed via the Technology Steering Committee of DISTRIBUTION 4. Base Systems (TESCOD) NETWORK planning Study Committee. Power System Analysis 3. Base Data ESKOM DISTRIBUTION utilise ReticMaster and Present initiatives to address base data DigSilent PowerFactory for power system analysis. requirements include: ReticMaster is used for basic studies on radial MV. and LV networks. PowerFactory is used for sub- Data porting: ESKOM DISTRIBUTION utilise the GE transmission NETWORK analysis and advanced SmallWorld AM/FM GIS system as the primary simulation such as protection coordination, data store for NETWORK information as required dynamic and transient analysis and harmonic for NETWORK schematics, SCADA and power studies.

7 System analysis. HV and MV NETWORK location and connectivity is comprehensively captured in ESKOM 's present SmallWorld system only supports SmallWorld. However, certain attribute data a write-out of MV networks to ReticMaster. A. required for power system analysis, such as project is underway to provide integration between conductor size, is incomplete and the accuracy SmallWorld and PowerFactory. PowerFactory will of certain attributes is uncertain (in Smallworld) then be utilised for power system analysis on both and captured accurately elsewhere. Much of MV and HV networks utilising the NETWORK data this data is still contained in legacy systems and stored in SmallWorld. This facilitates advanced other databases such as Microstation as-built reliability studies (section 5). ReticMaster will drawings, commissioning sheets and test continue to be used for all LV studies, and basic sheets. A project has been initiated to port this radial MV studies.

8 Data from these other systems and data sources into SmallWorld so that this data is A Master Type Library (MTL) has been completed readily available to the planner and can be and provides a single source of type library exported to power system simulation software (standard) values for systems including (see section ). SmallWorld, ReticMaster and PowerFactory. This Data processes: The data porting project will ensures that the same standard values also address the data processes, standards, (impedances etc) are utilised in all systems. commissioning sheets, test sheets etc to ensure that the HV and MV NETWORK data Project Need Register Libraries: Libraries are utilised to provide standard values (load factor, power factor, load The existing Project Need Register (PNR) in profile) for typical customer classes. SmallWorld is being enhanced to provide planners Scenarios: Multiple load forecast scenarios are with the following core functionality within the GIS: supported.

9 PSA link: Load forecast results are linked to Need register: Needs are spatially located ReticMaster and PowerFactory so that manual (GPS coordinates) so that interdependencies population of forecasts within the simulation can be visualised. packages is not required. Project register: A project can have a number of alternatives, one of which is preferred. Each A project is underway to source a commercial off alternative can have jobs, NETWORK , costs, need the shelf solution providing an acceptable fit to the dates, and other attributes associated with it. user requirements. In the interim a locally Alternatives can be written out to ReticMaster developed and supported Microsoft Excel based or PowerFactory for power system analysis. load forecasting tool, PowerGLF, will be utilised. Projects and alternatives can be viewed Rollout is expected to be completed in the fourth spatially and are integrated with K2 project quarter of 2007.

10 A load forecasting guideline has workflow and PowerOffice costing systems. been developed to support the PowerGLF rollout. Projects and alternatives can be linked to needs. Project Evaluation Model The PNR provides the planner with a GIS linked Historically ESKOM DISTRIBUTION NETWORK planning repository for needs, projects and alternatives such selected preferred alternatives based on capital that this data is available throughout the enterprise cost considerations and compliance with minimum and can be integrated with costing engines, standards. Consideration of increased NETWORK workflow and power system simulation tools. reliability and higher cost of technical losses necessitate a new approach. Geo-based load forecasting A Project Evaluation Model (PEM) is being Load forecasting is a critical component of NETWORK developed to supplement the Financial Evaluation planning . Historically there has been no fully Model currently in use, with the intention to standardised system for load forecasting within consider lifetime economic costs associated with ESKOM DISTRIBUTION NETWORK planning .