AMAZON multi-meters discounts AMAZON oscilloscope discounts
1. INTRODUCTION
This section describes the general distribution planning steps that may be taken to estimate the magnitude of the medium and low voltage distribution system loads to be supplied. It presents various load forecasting methods for estimating load development within the time period under review and within the specified geographic area under consideration. Such estimates and forecasts then allow the size of the necessary supply equipment and service overhead lines or cables to be calculated taking into account normal factors such as:
_ continuous current rating;
_ line voltage regulation;
_ fault rating;
_ supply interference (motor starting, harmonic distortion, unbalance, etc.);
_ supply security;
_ construction hazards and standards.
The overall efficiency of the distribution system is as important in load forecasting as energy consumption. Therefore load factor (LDF), maximum demand, diversity, losses and growth characteristics are particularly discussed. Modern distribution planning makes considerable use of computer modeling and equipment reliability statistics to assist with design optimization, and reference is made to such techniques.
Some of the effects of distributed generation on design of distribution networks are also summarized.
FIG. 1 Average hourly loads for the example day. Average hourly
loads (kW); Example peak day
2. DEFINITIONS
This section defines some load definitions and describes the terminology used in distribution planning.
2.1 Demand or Average Demand
The demand of an installation or system is the load at the receiving terminals averaged over a specified interval of time.' The load may be expressed as active power (kW) or reactive power (kVAr).
The period over which the demand is averaged is known as the demand interval and may be governed by the thermal constant of the equipment or the duration of the load. Figure 1 illustrates average hourly loads (kW) over a 24-hour period. The demand interval must always be stated when describing average demand or the figure is meaningless:
Average demand5 Total energy kWh Total period hours From Fig. 1:
Average demand5 371 kWh 24 h 515:45 kW based upon average hourly demands over a 24-h period
FIG. 2 Variation in demand with demand interval (note that a lower
maximum demand results from a larger demand interval because of such
a smoothing effect).
2.2 Maximum Demand (MD)
The maximum demand of an installation or system is the greatest of all demands that have occurred during the specified period of time.' The maximum demand (MD) may be expressed in kW, kVAr, etc. Both the demand interval (average hourly loads, etc.) and the time period (daily, weekly, etc.) must be defined for the expression to be meaningful.
FIG. 2 illustrates the variation in demand with demand interval. Loads normally alter through a 24-hour period with clear peaks occurring. For example, the load increases in the morning as people get up to have breakfast and to go to work. Similarly with the advertisement intervals on the television in the evening, load peaks occur as viewers get up from watching a popular show to use electric kettles to boil water and make a cup of tea. A larger demand interval will have the effect of smoothing out such effects and will therefore normally result in a lower maximum demand.
2.3 Demand Factor
'The demand factor is the ratio of the maximum demand of a system to the total connected load of the system.' The total connected load of the system is defined as 'the sum of the continuous ratings of the load consuming equipment connected to the system.' Both the maximum demand and the total connected load should be expressed in the same units thus making the demand factor dimensionless. Again the demand interval and the period over which the maximum demand applies should be stated. The demand factor is most often used in association with a consumer's services rather than to a complete distribution system:
Demand factor; Maximum demand of the system; Total connected load normally # 1
2.4 Utilization Factor (UF)
'The utilization factor is the ratio of the maximum demand of a system to the rated capacity of the system.' Both the maximum demand and the system rated capacity should both be expressed in the same units to make the utilization factor expression dimensionless. Again the demand interval and the period over which the maximum demand applies should be stated. The utilization factor indicates the degree to which the system is being loaded during peak load periods with respect to its capacity.
'The load factor is the ratio of the average load over a designated period of time to the peak load occurring in that period.' To accurately define the load factor then the demand interval, the period to which the maximum demand and average load apply, the manner in which the maximum demand is measured and the load commodity involved should all be stated. The average and the peak demand loads should be expressed in the same units to make the expression dimensionless. Load factor is usually expressed as a percentage figure or a fraction. Fundamentally, the load factor indicates the degree to which the peak load is sustained during the period. In the United States, the national average load factor is currently approximately 63%, whereas in a developing country it may be as low as 50%.
===
FIG. 3 Coincidental and non-coincidental demands.
Diversified or coincidental demand = demand on the system during demand interval "t " Non-coincidental demand = sum of the demands on the system with no restriction to the demand interval i.e. = 10+9+10+14 = 43 kW (Usually non-coincidental demands are comprised of individual maximum demands.
Therefore, the term is also referred to as the maximum non-coincident demand.)
===
2.6 Diversity Factor (DF)
'The diversity factor is the ratio of the sum of the individual maximum demands of the various subdivisions of a system to the maximum demand of the whole system.' Loads do not normally all peak at the same time. The sum of the individual peak loads will therefore inevitably be greater than the peak load of the composite system. The diversity factor normally has a value greater than unity and is only equal to unity if all the individual demands occur simultaneously. The coincident nature of load demands is of great importance to the distribution planning engineer as it is a key factor in the economic sizing of plant.
FIG. 3 shows the effects of coincidental and non-coincidental demands.
2.7 Coincident Factor (CF)
Some engineers prefer to have a factor which describes the characteristics of loads that have a value equal to or less than unity. The reciprocal of the diversity factor is known as the coincident factor:
Coincident factor; CF 1 DF normally#1
The coincident factor is dependent upon the type of loads connected to the system. Typically, Loads CF Distribution transformers 0.74 0.83 Primary feeders 0.83 0.92 Substations 0.80 0.96 In general, and in the absence of other data:
CF 0: 11 2n13
__ where n is the number of loads connected to the system.
For residential areas in countries with developed economies the coincident factor tends to settle at approximately 0.5. However, caution must be applied and data should be collected to obtain meaningful information, as shown in Fig. 3a, where CF settles at approximately 0.3.
2.8 Load Diversity
'Load diversity is the difference between the sum of the peaks of two or more individual loads and the peak of the combined load.' Since load diversity is the difference between two quantities of similar units (rather than a ratio), it is expressed in the units of the two demands being compared.
FIG. 3a
2.9 Loss Factor (LSF)
'The loss factor is the ratio of the average power loss to the peak load loss, during a specified period of time.' Since power losses are proportional to the square of the load current:
===
FIG. 4a Hypothetical load case where loss factor load factor (LSF LDF
if the load remains at its peak value all the time that it is on, and
zero for the remainder of the time period).
====
FIG. 4b Hypothetical load case where loss factor load factor 2 (LSF
(LDF) 2 if the load has a sharp peak and then a fairly steady value for
the remainder of the period under consideration).
===
The loss factor cannot be determined directly from the loss factor because the loss factor is determined from the losses as a function of time, which in turn are proportional to the time function of the square of the load. However, a relationship has been calculated which gives a reasonable value of the 30-minute, monthly, kW LSF in terms of the corresponding load as shown graphically in Fig. 4c.
FIG. 4c Curves of loss factor LSF as a function of load factor LDF.
2.10 Load Duration
'Load duration is the relationship of demands and the duration of the demands over a specified time period.'
Referring to Fig. 1 the hourly demands have been sorted in descending order and tabulated in Table 1 as shown below to give:
_ Frequency Number of hours of occurrence for each demand
_ Equal/Exceed Summation of frequencies
_ Percent of peak Demand kW Peak kW 3100%
TABLE 1 Load Duration and Loss Table (for the Peak Day Described in Fig. 1)
_ Percent of duration Equal=Exceed Specified time 3100%
_ Square of demands5(Demand) 2 3frequency The load duration parameters for example day have been plotted in Fig. 5 (percent peak load vs. percent duration). Technical losses are a function of the squares of the load current (amps) which is directly related to the squares of the demands. Figure 23.6 is a graph of the squares of the hourly demands for the examples day illustrated in Fig. 1.
FIG. 5 Load duration graph for example, peak day.
2.11 Loss Equivalent Hours
'Loss equivalent hours are the number of hours of peak loads which will pro duce the same total losses as is produced by the actual loads over a specified period of time.'
Both the actual and peak demand values must be chosen from the associated load duration:
Loss equivalent hours5 Square of all actual demands
Square of peak demand With reference to the load duration and loss table (Table 1):
Loss equivalent hours5 6;849 kW2 900 kW2 7:61 hours
[...]
Typical PRF (system) values for different transformers in the system are:
Transformer type PRF (system) PRF2 (system) (~ system loads) (~ system losses) General step-up transformer 1.0 1.0 Transmission substation transformer 0.9 0.81 Distribution substation transformer 0.8 0.64
Distribution feeder pillar transformer (0.46 0.95), say 0.75 Say 0.56 It should be noted that no-load losses are continuous and occur both during system peak demand and at other times. Generation therefore had to be designed to support these no-load losses.
Load losses vary with the load such that peak losses on a particular component of the overall distribution system occur at peak load on that component which may not be at the same as the overall system peak demand. Only a fraction of the individual component losses therefore contribute to the sys tem peak demand.
3. LOAD FORECASTING
3.1 Users of Load Forecasts
Electricity supply authorities plan the capacity of their systems to meet the expected peak demand requirements. They maintain power (kW demand) and energy (kWh) forecasts as a basis for their physical network and financial planning.
In addition to demand forecasts, the projected load curve based upon hour-by-hour demand throughout the planning period, has an influence on the choice of generating capacity and the most economic order in which to bring different generating units onto the grid or distribution system. For example, fast run-up generating units (often with diesel or gas turbine prime movers) may be used to most economically satisfy short peak demands.
The combination of demand and energy forecasts forms the basis for planning generating fuel requirements. They are the starting point for plant capacity and fuel strategies which are, in turn, translated into financial requirements.
Fuel costs may vary substantially between different power stations or even generating plant within a particular station. It is therefore essential to normally employ the most cost effective and fuel efficient plant and only use more expensive plant for short periods (e.g. to deal with short peaks or short-term loss of wind or tidal power). Generating costs may be translated into fixed charges (generally associated with the capital plant and overheads required regardless of the actual power being generated, transmitted and distributed through the network) and variable costs directly associated with the energy demand (additional shift workers, additional fuel, etc.). Energy fore casts are the basis of revenue planning. Forecasts also assist in the compilation of statistical data for the information of the public, government bodies, academic institutions and manufacturers. For example, manufacturers of electric supply equipment are able to gauge their future manufacturing output and marketing strategies from such data. In the short term the distribution planning process allows:
1. Relief of overloads in the distribution system.
2. Voltage control.
3. Reactive compensation (power factor correction).
4. Improvements in service quality for consumers.
5. Short-term system reinforcement and better provision of consumer connection requirements.
And in the longer term:
1. Pre-warning of changes in load and load usage.
2. Selection of the most appropriate primary distribution voltages.
3. Selection of substation capacity.
4. Determination of substation locations (at or near load centers).
5. Sub-transmission system requirements.
6. Long-term budgeting estimates.
3.2 The Preparation of Load Forecasts
It is very important to estimate how the load will grow, the possible load growth rate, the load characteristics and magnitude together with the load location. Macro and micro load-forecasting methods are therefore used and may be checked against one another.
In summary, distribution system planning:
1. Is a continuous process providing rapid evaluation and response to changes.
2. Has a planning period which reflects the lead times associated with project sanction (financial and economic appraisal and approval by the company and funding agencies), equipment procurement, installation and commissioning.
3. Should integrate into other areas of power system expansion as shown in Fig. 7.
4. Provides a framework within which system efficiency (loss reduction campaigns, procurement policies, etc.) may be kept under review.
3.3 The Micro Load Forecast
The micro load forecast is made up from small component parts and a separate forecast for each part is estimated. In microscopic estimations electricity demands are estimated in terms of service classifications or consumer groups for predetermined geographical areas and then integrated to produce peak power and energy demands for each such consumer group. The number of component parts used in the forecast is partly dependent on the value such complexity brings. A typical planning performance index (PPI) might be:
- PPI Quality of the analysis
- Time to perform the analysis
Each part may be subdivided into a number of items and dimensions in which the forecast is made; e.g. maximum demand, energy consumption, numbers of consumers, population, load centers, tariff categories, diversity and losses, etc. Each item may, in turn, be further subdivided (e.g. the number of tariff categories) are inter-related with other items. Normally such micro load forecasts are based upon a combination of the following data:
_ Extrapolations from historical data (sometimes using regression analysis).
_ Data from power market surveys.
_ Forecast of changes in population, housing, commercial, industrial, agricultural and other developments.
_ Stated national, regional and local government policies, the electrical sup ply utility marketing plans and those of other relevant authorities.
_ Expansion plans into currently non-electrified areas (often using results from past comparable experiences).
_ The experience and judgment of the electrical supply utilities' forecasting department.
Since the micro load forecast requires the progressive amalgamation of data it would be ideal to commence with the smallest possible physical area and build such areas up to represent a district or region. In order to prepare the economic input to the production of a micro forecast, it is necessary to assess how a district or branch load grows over a number of years. Table 2 is a useful data collection proforma.
===
FIG. 7 Distribution planning interfacing with other areas of the
power system planning process.
Network and radial feeder analysis Distribution system expansion plans Transmission system expansion plans Generation system expansion plans Substation analysis Transmission network analysis Generation systems Small area load forecast; Total system load forecast
=======
TABLE 2 Micro Forecast District/Branch Data Collection Proforma District/branch
Prosperity Very Prosperous, Prosperous, Average, Below Average, Poor (delete inapplicable categories) Year electrified .........................
Urban/rural ............
Year after connection 1, 2, 3, 4, etc.
Domestic Number of households connected Total number of households
% electrification kWh/consumer Total kWh Commercial Number of consumers Total number of possible consumers kWh/consumer Total kWh List large unconnected consumers Industrial Number of consumers Total number of possible consumers kWh/consumer Total kWh List large unconnected consumers
======
The basic methods employed are:
(a) Scratch pad methods
'Rules of thumb' are employed by experienced distribution system planners.
_ kWh/month or year and load factor information gives a view on the kW demand.
_ kW demand per substation or feeder and power factor at peak demand gives a view on the kVA demand per substation or feeder.
_ Demand per substation or feeder and associated maps allow estimates of:
demand per km of overhead line or underground cables demand per square km demand per connected load (kW or kVA)
_ Demand per square km coupled with information on the population and the customers per unit gives a view on the demand per customer.
Such methods are useful for well-understood areas and for relatively small expansion schemes. Such methods should not be used to support large investment proposals.
(b) Trending
Regression curve fitting analysis is used on historical load growth information to estimate future load growth trends. This is an easy and simple load forecasting method. However, it is not very accurate because it does not take into account new emerging or future dominant factors. The method may be enhanced by carrying out a market survey in order to allow the forecaster to identify and take into account likely changes from past rends and their causes. Linear programming methods may also be employed using multivariate analysis. However, it should be noted that a piecewise linear solution more closely approximates non-linear load growth.
(c) End-use methods
Simulation land-based models are used to take into account such factors as:
_ where people live;
_ where people work;
_ when people want power;
_ how people wish to use the power supplied.
This is a more advanced and accurate method of load forecasting and can be used to forecast the changing character of the load demand over long time periods. Such end-use methods may correlate land use with industrial/ residential/commercial load demand and growth. The method necessitates time consuming data collection and computer analysis.
3.4 The Macro Load Forecast
The macro load forecast is widely used by economic analysts. Such model ling focuses on the relationship between the growth of the national or regional economy and the total energy consumption required to achieve such growth. Energy consumption (as represented by electricity demand) may change linearly with the growth of the economy. In such cases it is valid to relate the growth of electricity consumption to the Gross National Product (GNP), population growth, power consumption in manufacturing, individual consumptive expense, etc.
Detailed information may be collected from a series of countries which are similar in their climate and level of economic development. Alternatively a standard regression model, which has already been tested, may be employed.
The resultant growth trends predicted by these two macro load forecasting methods should be broadly similar. However, in developing countries things are seldom so simple. The load growth is often not governed by demand but by the ability of the electrical supply utility to build and finance the expansion of the network. Pre-investment studies normally take into account a range of confidence levels associated with the load demand forecast.
A typical macro load forecasting study for a developing country would follow the following process:
1. Collect all available data of historical population growth.
2. Analyze available population forecasts.
3. Collect historical Gross Domestic Product (GDP)/Gross Domestic Regional Product (GDRP) data.
4. Discuss and agree future GDP/GDRP forecasts with central planning authorities in country concerned, the World Bank, university economics specialists, etc.
5. Check future forecasts (e.g. elasticity analysis examines the relationship between Gross Nation Product per capita (GNP)/capita and the electricity consumed per capita (kWh/capita).
6. Produce international model based upon countries in the same region.
7. Apply the GDP or GDRP/capita forecast to the international model.
8. Apply an existing regression analysis model and check against the international model.
9. Carry out sensitivity analysis based upon changes to the GDP or GDRP assumption.
10. Produce upper, medium and lower load forecasts.
A major disadvantage of the economics method of load forecasting is that it does not lend itself to forecasting the detailed geographical distribution of demand as required for the practical planning of transmission and distribution facilities. Such work is, however, carried out by economists as part of an overall power distribution project submission to funding agencies.
3.5 Nature of the Load Forecast
It is important to take into account the known constraints, such as technical or financial limitations, on supply expansion. The attainable demand is that portion of the demand for electricity that may be satisfied after taking into account such known constraints on supply. The potential load may be much greater than this attainable demand and there may be underlying demands which it may not be possible to supply because of physical or other constraints. Loads may be suppressed due to:
_ bad voltage conditions at the consumers' terminals;
_ load shedding;
_ voluntary load shedding by selected co-operative consumers.
Potential customers may be placed on waiting lists because:
_ Customers have been refused permission to connect loads to the network for technical or other supply constraint reasons.
_ Lack of supply availability in the short term.
Distributed generation (DG) (or 'embedded generation') may distort the load forecast because:
_ Private electricity generating plant may already be available but not connected or synchronized onto the distribution network.
_ Diesel engines may be currently used to drive pumps and other machinery which could be changed to electrically drive sources if available.
_ Growth of domestic generation (solar power, fuel cells, etc.) may reduce effective demand growth.
The effects of DG are considered further in Section 6 below.
Unsupplied areas
_ May have potential loads not included in the waiting lists.
The transition from underlying demand to satisfied demand may therefore have to take place over several years. Figure 8 shows how some of these factors are incorporated into a forecast prepared in 1994 for load growth projections to 2010.