Annexe 3 to report of working group on specification and negotiation of network services

TRANSMISSION PLANNING CRITERIA

Planning Basis

The design principles and approach used by VPX for planning the transmission network are as follows:

  • The system must be able to sustain any single outage, including disconnection of any single generator, busbar or transmission element, and remain within performance limits as defined in the System Code. For purposes of calculating inter-regional transfer capabilities, busbar faults are not included.
  • Following the forced outage of a single network element, it must be possible to ‘re-adjust’ the system within 30 minutes so that it is capable of tolerating a further forced outage without risk to the security of the overall network. This may involve network switching and/or modification to generation dispatch to account for the new system constraints. Blocks of load may be at risk of disconnection either directly by a second failure (such as for loads that are supplied by two lines where the first line is unavailable and the second line fails) or by load shedding action to reduce flows to plant ratings if the second event occurs.
  • Adequate control measures are in place to secure the power system (ie. retain the system in a stable and controllable condition) and minimise the extent of disruption to customers following low probability events not catered for by the basic network design. Where practicable, load shedding is restricted to interruptable smelter loads.
  • Sufficient periods are available to allow maintenance of critical shared network elements without exposing the network to excessive risk in the event of a further unscheduled outage of a network element. These parts of the network must have sufficient capacity to reliably meet transfer requirements with two elements out of service during light load when maintenance is scheduled.

Level Of Redundancy

The transmission network provides bulk supply of energy between major generation and load centres. While the network is very reliable the consequences of any outages may be very severe both in terms of the amount of load that is lost and the duration of the interruption.

Consequently the transmission network must be designed with some redundancy. Generally this requires that there is sufficient capability built in to the network to allow for the unexpected outage of any plant item under extreme conditions without resulting in immediate overloads on other elements. The consequences of such overloads could be very severe since they may ultimately be controlled by cascade tripping of transmission elements. Hence the redundancy ensures that continued operation of the network is possible and leads to the concept of “firm capability”. That is, the network should have sufficient capacity to allow the unexpected loss of the most critical network element at any time, without any primary transmission plant being overloaded or any normal customer load being shed. This criteria has been adopted for the analysis of future requirements included in this review.

Example

The simplest example is the case of supply to a radial load. Consider that a load of 100 MW maximum demand is supplied by radial transmission lines of 100 km in length. In order to provide firm capability it is necessary to install two lines in parallel since otherwise outage of one line would result in loss of supply to this load. If the outage was a result of equipment failure this outage could be prolonged.

In the example full redundancy is provided to supply a maximum demand of up to 100 MW and this is the firm capability of the network.

Should load increase occur in the case shown, then for the period when load exceeds the 100 MW there is potential for overloading of the network to occur. For example if the load is 105 MW and one transmission line is unexpectedly lost then the flow on the remaining transmission line will increase to 105 MW. If 100 MW is the maximum flow that is allowed on this element without permanent damage occurring then the network must be operated to ensure that flow never increases beyond 100 MW, ie. 5 MW would need to be removed from the load point to maintain the security of the network.

Alternatively the risk could be considered acceptable in which case operation at 105 MW could be allowed on the basis that:

  • the problem only arises if the transmission outage occurs (which has a low probability); and
  • the above outage must occur at a time when the load is at a peak.

Most loads exhibit wide fluctuations within each day and throughout the year. It is likely that peak demands will only be achieved for a particular load point for 10 - 15 hours per annum.

An International Survey showed that many countries use an (n-2) planning basis, where the coincident loss of two network elements is considered in ensuring that there is adequate network capacity available. Consequently, Victoria’s planning basis is less conservative than generally adopted overseas, although in some cases detailed equipment design has been carried out to minimise the possibility and consequences of failures arising from more than a single outage. For example:

  • the Richmond to Brunswick 220 kV underground cable was justified on the basis that a considerable portion of CBD load was supplied on a double circuit tower line;
  • emergency control schemes have been implemented to protect the system from complete shutdown in the event of failure and loss of both Moorabool to Heywood/Portland 500 kV circuits under critical conditions; and
  • major base load power stations have generally been designed with an additional circuit to allow firm support for the entire station with one line out of service.

The (n-1) basis is considered appropriate in Victoria given this attention to those situations where exposures to double failures are higher than normally encountered.

Short Time Ratings

Historically the network was designed and operated in accordance with an (n-1) planning criteria, ie. sufficient redundancy was built into the network to allow an unexpected outage of a single transmission element without the rating of other transmission plant being exceeded. This ensured that no damage occurs to the transmission elements or that no curtailment of load or generation is necessary.

More recently short term equipment ratings have been used to provide increased transmission utilisation. This recognises that the temperature of a transmission element exposed to a sudden loading increase after an outage will take some time to reach its thermal rating. This is due to the mass of metal involved. The use of a ten minute rating has been adopted which means that following a contingency system operators have ten minutes to reduce loading on critical transmission elements before their temperature rating is exceeded.

This approach has become more feasible over the last decade due to the higher level and accuracy of remote monitoring of critical aspects of the transmission system which allow much tighter control. Given the rare occasions in which the firm capability is exceeded it is relatively easy to develop rigorous procedures to apply under critical loading conditions to ensure that response is achieved within the required time frames.

It is recognised that higher utilisation could be achieved if even shorter term ratings were adopted (ie. system operators were required to act in less than 10 minutes). In the event that network loading did exceed the ten minute rating (for example due to higher than forecast load) this approach would be likely to be adopted in preference to taking pre-contingent action (such as load shedding). In planning the network this additional capability is reserved for those rare occasions when the network operation may be more extreme than that anticipated in network planning studies, or to cover the period immediately prior to the augmentation when the energy at risk is insufficient to justify augmentation[1].

The example in the previous section was based on an analysis assuming that immediate and catastrophic failure would occur in the event that the line flow exceeded its rating (of 100 MW). In practice however the determining factor for line ratings is the maximum allowable temperature which is generally dictated by the clearances that need to be maintained to structures and between conductors.

Consequently the 100 MW rating refers to the maximum flow that can be sustained for a continuous period without exceeding the temperature rating of the transmission line. This implies that ratings will depend on ambient temperature since greater heating of the conductor due to the passage of current is allowed at lower ambient temperatures.

The other important feature is that temperature will not rise instantaneously following a sudden increase in flow. Due to the mass of metal that is heated the temperature rises more gradually. Assuming that redundancy is built into the network the flow prior to a contingency will be somewhat less than its continuous rating and the conductor temperature will be at a relatively low value.

Example

Following outage of one line the current flow will increase on the remaining line and the temperature of the line will begin to rise. However there will be time available in which to take action to ensure that this rise is contained within temperature limits. This is shown in the following diagram.

This illustration shows that it is possible to operate beyond the firm capability without compromising on system security providing action is taken to remove overloads immediately following the contingency. In this case the lower curve shows the situation that would exist if no short term overload capability was adopted. It corresponds to an initial line loading of 50 MW which following the outage increases to 100 MW which is just sufficient to increase the conductor temperature to its rating.

The higher curve indicates the temperature that results if initial operation is allowed at higher levels. In this case the conductor temperature rises after the outage and if left unchecked would increase beyond the thermal rating of the line after 10 minutes. However, with facilities in place to remove the overload within the ten minutes, temperature ratings would not be exceeded.

A ten minute response time is considered acceptable. If higher loading was used then temperature ratings and perhaps plant damage would occur prior to the ten minute period allowed for operators to act. Consequently this sets the maximum pre-contingent loading which is allowed, and any load beyond this level should not be transferred since it exposes the load to a security risk.

Deterministic Planning

The deterministic planning approach is to calculate the maximum anticipated transmission loading for expected worst case loading and generation conditions, and the minimum transmission capability for the expected worst case summer temperature. Augmentation is indicated when the required loading exceeds the network capability. This provides an indication of the time when augmentation is required but does not identify the level of risk or the impact on participants.

Probabilistic Planning

A probabilistic planning approach is used to consider the risks associated with transmission constraints. This takes into account the fact that the operation of the market cannot be exactly predicted but is subject to variations in loading (due to forecasting inaccuracies and weather impacts), performance of generating plant and variations in bidding behaviour by generators. A pool simulation model is used (VPX has a very well developed model suitable for this purpose) to determine the hourly generation dispatch for a large number of scenarios to capture the range of variation. Critical transmission line loadings are then determined on an hour by hour basis and compared with the ten minute network capability. This allows the risks associated with the transmission system to be identified.

Critical statistics which can be used to build a more complete picture of the risk exposure of participants include:

  • the total energy beyond the ten minute capability each year;
  • the number of separate occasions in each year in which the network would be required to operate beyond the ten minute capability;
  • the maximum duration of a single event in which the ten minute rating is exceeded; and
  • the maximum amount by which the demand exceeds the ten minute network capability (which equates to the load which would need to be shed).

The energy which cannot be supplied is a critical parameter in justifying any network investment. The network is planned to ensure that an economic balance is struck between the cost of providing additional network capacity to remove any constraints and the cost (and risk) of having some exposure to loading levels beyond the ten minute capability. In other words recognising that very extreme loading conditions may occur for only a few hours in each year there is little economic benefit in providing additional capacity to meet all anticipated loading requirements. Rather the augmentation should take place when loading has increased to the extent that the level of energy at risk justifies expenditure on the transmission system to avoid it.

One implication of this is that for periods immediately prior to new investment, the network will be exposed to some risk. It may be argued that such risk is minimal since system problems will only arise if the critical contingency occurs at a time when the network is operating beyond its limit. While this event has a very low probability it must be recognised that the implications are very severe. There are two possible scenarios which could result if such an event was to occur:

  • Severe damage could be inflicted on critical transmission elements which could require an extended outage for repair, and as a consequence impose long term reductions in network capability in a critical part of the network. This may result in severe constraints on network loading. Such a long term constraint would impose heavy costs on participants if energy could not be supplied or generation had to be run out of merit order as a result.
  • Overloaded elements may be tripped by protection schemes leading to further overloading of remaining elements and cascade tripping of large portions of the network and ultimately to complete shutdown of the entire network. Such an event would involve total loss of supply to all participants in Victoria (and perhaps across the interconnected system). A complete restart of the system will take many hours because of the unavailability of an electricity supply to start auxiliary plant for base load generation. The community costs of such a “system black” event are difficult to estimate but are extremely high, perhaps up to $250 million per event.

The severity of these events means that despite the low probability of such an occurrence it is critical for system security that:

  • The network is planned to ensure that the risks associated with operation of the system beyond 10 minute capabilities are relatively small and manageable; and
  • Control measures are implemented which act to protect the system for the rare occasions on which it operates above the ten minute capability.

If this can be achieved it will result in a practical and economic network design which maintains the required level of system security.

The probabilistic planning approach is more suited to a competitive market environment since it allows an economic assessment to be made on the benefits achieved through network augmentation compared to any financial losses which Participants may sustain in the energy market. However since the cost impacts are generally not confined to a single Participant it is necessary for this assessment to be made across Participants as a whole, with the network planned to a level acceptable across the whole industry.

Example

Even in the simple example considered in the previous two sections, the probabilistic nature of planning can be highlighted since the following factors impact the level of overloads that the network is exposed to:

  • the variation of line ratings with temperature; and
  • the hour by hour and seasonal load variations.

One way of taking this into consideration is to carry out hour by hour modelling for the forecast period. This requires hourly forecasts of load to be used and this is readily achieved through the use of standard load curves established for the station from past records scaled by future demand forecasts for the station. This also needs to take weather effects into account. Hourly temperature forecasts must also be used.

This introduces the probabilistic nature of studies since it is clearly impossible to forecast temperature. Rather it is necessary to consider a large number of scenarios each with temperature and loading profiles within normally encountered values.

With this analysis the load and each hour can be compared to the line rating for each hour (calculated using the temperature information). When repeated over a large number of operating scenarios the results can be averaged to assess the level of risk.