Two Views of Power System Reliability

A science fiction writer once tried to sell the idea of a trip inside the sun as an “opportunity to view sunspots from behind.”  It may not be comfortable, but the observations would be unique and would undoubtedly contribute to a better understanding of the phenomena.  In a more practical sense, being able to examine complex structures from different vantage points — inside and outside, or close-up and from a distance — makes new insight possible, and hopefully, better understanding.

Let’s try this method on the very mundane subject of power system reliability.

The internal observer

When the bulk power system (BPS) is described as “reliable,” it implies that the same system has passed a whole battery of tests, which for the US might mean such standards as NERC Categories A, B, C and D  and other associated NERC standards, as well as standards for regions, regulatory bodies and individual utilities.  In this sense, a “reliable” bulk power system offers a “yes-or-no” qualification that belies the complexity of the power system and its various probable and improbable failure modes, and the tests and criteria applied to make the assessment.

Though the description may imply that the whole system is “reliable”, what it actually means is that the system, taken as a whole, is reliable.  Reliability is not homogeneous, but chunky, or to use a more common term, “locational.”  A thought experiment may help clarify this.  If we allowed load to grow in the BPS, simply assuming more generation from the existing resources, and then applied the reliability tests, we would find … at some point that the BPS would fail the reliability test.  The failure may initially be one slight thermal overload or voltage violation or a generator losing synchronism.  But the number of failures would increase as demand is increased further, perhaps developing into cascading outage, widespread instability or voltage collapse failures. [1]  We might then contend that where the initial failures are observed is where the BPS is least reliable, and where more and more failures are observed we might develop a profile of varying levels of reliability throughout the various locations of the BPS.

Hence, the specific character of reliability within the BPS varies by location, and tends to change over time (as demand grows), regardless of whether the BPS itself is kept “reliable.”  The locational characteristic is primarily a function of the dispatch of generation (which in turn is a function of market rules), outage rates of various BPS equipment such as transmission lines, transformers, power conditioning equipment, maintenance schedules of the same equipment and random external factors that can lead to extended outages.

A hypothetical internal observer, perhaps a system planner, would thus observe power system reliability as a quantitative measure applied to the BPS which has a general value, a yes-or-no reliability, and a locational value, such as frequency and duration of service outages for customers connected to a specific location.

The new external observer

When the bulk power system is described as “reliable”, it is able to accept incoming power from a power plant and transfer the same to a point of delivery.  This implies that the BPS with the power plant has passed the impact study testing for initial operation in accordance with the standards for interconnection of the FERC [2], if in the US, and any of the ruling regional organization such as an RTO or ISO [3], or public service commission or local utility.  This is a generic qualification that allows for approval for construction, interconnection and/or energization without indicating the complexity of the energy market in which the power plant operates.

“Reliable” in this context means meeting specific criteria and standards under certain assumed operating conditions.  In practice, the “reliability” is subject to energy market price fluctuations, demand changes, firm and non-firm transactions, and scheduled outages. [4] A simple example would be to consider power plants delivering power over a single path to a load center.  As more power, and power plants, collocate with the existing power plants to deliver power over the path, the reliability standards would require that the total power eventually be constricted or limited, or to use a more common term, limited by “available transfer capability (ATC).”  ATC would change by time of day, by any scheduled outages, and by any prior reservations made with the owners of the transmission path.  On the other hand, if a power plant locates on the other side of the path, nearer to the load center than the other power resources, it may not have the same constraints on ATC.  In this sense, reliability is “locational” since the point of receipt of power affects the ability to transfer the power to points of delivery.

Hence, to a hypothetical external observer, perhaps a power marketer representing several power generators, the “reliable BPS” is a generic designation which may not impact the actual dispatch of his portfolio.  The key factors, to this observer, would be the price of energy at which he is selling relative to anyone else in the same market, and the specific location of each power plant in his portfolio relative to ATCs.  He would consider this for various timeframes, from long-term sales, to intermediate term, to monthly, to daily and hourly (in some systems, even on a quarter-hour basis).  Each transaction remains subject to reliability specific to the timeframe.  The reliability assessment would establish whether each specific transaction is allowed or not, or if constrained to a lower amount.


We deliberately chose observers to whom a “reliable” BPS may not mean very much, to illustrate a point.  In fact BPS reliability would impact the chosen observers in a less direct manner.  For instance, the reliability of the BPS may be used to determine investments in the transmission grid which would eventually impact the specific reliability concerns of the observers.  Or, BPS reliability may be used to revise standards and criteria, which would also eventually filter down to the observations.

Both observers recognize the locational nature of reliability, although each would measure this in different ways; i.e., the internal viewer measures service interruptions, while the external viewer measures constraints on transfer.  The analytical basis for the measurements also differ in that the dispatch and outage assumptions would be different, even if the actual criteria are the same.  For both observers the impact of local reliability can be measured in cost/benefit terms, so are, in this sense, comparable.  (Although real attempts at making this comparison have been difficult.)


There is a possibility that when different observers refer to “power system reliability”, they may actually mean different things, leading to confusion and miscommunication.  It would be important to recognize that the observations are only common in terms of reliability of the BPS taken as a whole, and to the specific criteria applied to establish it.  Other terms, such as ATC and local reliability, may differ in their basis, analytical character and measurement.


  1. There is a lot of simplification in the preceding illustration, such as the fact that we are looking at different dispatches, or if applying specific rules such as the California ISO’s G-1/N-1 criteria, different generators on outage.  We are also assuming deterministic tests, rather than probabilistic.  But this is only a short article!
  2. FERC – in the United States, the Federal Energy Regulatory Commission.
  3. RTO – regional transmission organization; ISO – independent system operator.  Both industry organizations are common in the United States, tasked with overseeing power system operations and markets over interconnected neighboring utilities.
  4. A curious term arises from having to take all these factors into consideration, “dispatchability.”  A possible definition of dispatchability is the ability to provide energy to the grid both as a sale into the energy market and as a reliable power transfer to point/s of receipt.


  1. Reppen, N.D., “Increasing utilization of the transmission grid requires new reliability criteria and comprehensive reliability assessment,” Probabilistic Methods Applied to Power Systems, 2004 International Conference on, 12-16 Sept. 2004
  2. Clark, H.K.; de Mello, F.P.; Reppen, N.D.; Ringlee, R.J., “The grid in transition – facts or fiction when dealing with reliability?” Power and Energy Magazine, IEEE, Volume 1,  Issue 5,  Sep-Oct 2003
  3. Khan, E.P.; Marnay, C.; Berman, D., “Evaluating dispatchability features in competitive bidding [in power systems],” Power Systems, IEEE Transactions on, Volume 7,  Issue 3,  Aug. 1992

For questions, comments and further discussion, contact us at