Confusion with the IAEA reactor performance data in the PRIS

By: Donald Jones, P.Eng., retired nuclear industry engineer, 2017 July 28.
This article, edited for space, appeared in the 2017 June edition of the Canadian Nuclear Society`s BULLETIN as a Letter to The Editor.

BULLETIN Publisher’s NoteThe CNS Nuclear Canada Yearbook commenced using PRIS data this year. Data was no longer available from the CANDU Owners Group, and data from public sources such as Nuclear Engineering International or Nucleonics Week had become either incomplete or late in publication.

When the author was preparing an article on the performance of Ontario’s CANDU nuclear units (reference 1) he wanted to include some idea of the amount of energy being curtailed by a unit at the Bruce Nuclear Generating Station due to load cycling. To do this required a close look at the  published performance indicators in the Power Reactor Information System (PRIS) database of the International Atomic Energy Agency (IAEA)(reference 2) and this revealed some discrepancies.

The performance of some of Ontario’s nuclear generating stations is affected by the surplus baseload generation (SBG) in the province (reference 1). Some nuclear units saw electricity output reductions during periods of surplus baseload generation (SBG). This means the CFs (Capacity Factors) are not a true performance indicator for those units (reference 3). A better metric of performance in these cases would be the Unit Capability Factor (UCF – used by Ontario Power Generation and by Bruce Power). The Energy Availability Factor (EAF) is another performance indicator and is shown in the Power Reactor Information System (PRIS) database of the International Atomic Energy Agency (IAEA), (reference 2). The EAF adjusts the available energy generation for energy losses attributed to plant management, planned and unplanned, and for external energy losses beyond the control of plant management while the UCF only includes energy losses attributed to plant management and excludes the external losses beyond control of plant management like load cycling/load following, grid failures, earthquakes, cooling water temperature higher than reference temperature, floods, lightning strikes, labour disputes outside the plant etc. The UCF seems a much better indicator of how well the unit is being managed than either CF or EAF. The UCF rather than the CF would also be the more appropriate number to use when calculating the Equivalent Full Power Hours (EFPHs) on the reactor pressure tubes of Bruce units that use steam bypass at constant reactor power since steam bypass operation does not affect the EFPHs on the pressure tubes. Only  reactor power changes would do that.

For Ontario there should be little significant difference between CF,  UCF and EAF for units that do not load cycle (an external energy loss) since other external energy losses will be close to zero. For units that load cycle the UCF will be higher than the EAF and higher than the CF but the EAF should  not be significantly different from the CF. The UCF and the EAF are based on reference ambient conditions so, unlike the CF, they cannot exceed 100 percent. In some cases the CF can be more than the UCF and EAF if the cooling water temperature is lower than the reference temperature and that increases the electrical output of the unit.

Now let’s see where the confusion arises. Take an example from the PRIS data. Bruce B unit 5, a load cycling unit, has a 2016 annual CF of 94 percent and an EAF of 97.4 percent. However, the EAF of 97.4 percent must really be a UCF of 97.4 percent since, as previously stated, the UCF for a load cycling unit must be greater than the CF but the EAF should not be significantly different from the CF since load cycling is accounted for in the EAF, but not in the UCF. For Ontario the only external energy loss would be from load cycling since other external energy losses would be normally near zero. For the Darlington and Pickering units that do not load cycle the CFs and EAFs (really UCFs) in the PRIS are not significantly different.

Let us see if this anomaly in the PRIS database applies only to Ontario and not to other countries. For example the EAFs for Wolsong CANDU units and other Korean units for 2016 are “real” EAF numbers since their shutdown due to the major earthquake (reference 4) must have been included as an external energy loss in the EAF since the EAF is practically the same as the CF. However for France, where load following is necessary because of the preponderance of nuclear on its grid, PRIS shows some French units with EAFs considerably higher than CFs which means that the PRIS is using UCFs and not EAFs, just like Ontario. Hence the confusion. The CFs (or Load Factors as PRIS calls them) have been correctly calculated by PRIS from the PRIS Electricity Supplied (net generation) and the Reference Unit Power (net) so the anomaly is only with the EAFs.

To summarize, the PRIS Capacity Factors seem correct and for Ontario at least the EAFs are really UCFs. Maybe in future the PRIS database should use UCFs instead of EAFs as performance indicators for all countries.


1. Performance of Ontario’s CANDU nuclear generating stations in 2016, Don Jones, 2017 April, https://thedonjonesarticles. performance-of-ontarios-candu- nuclear-generating-stations- in-2016/

2. Power Reactor Information System (PRIS) database of the International Atomic Energy Agency (IAEA),

3. Wind and nuclear and the increasing irrelevance of capacity factor in Ontario, Don Jones, 2014 February,   https://thedonjonesarticles. and-nuclear-and-the- increasing-irrelevance-of- capacity-factor-in-ontario- 2014-february/

4. CANDU 6 Performance in 2016, Don Jones, 2017 March,  https://thedonjonesarticles. candu-6-performance-in-2016/

Comments are closed.

%d bloggers like this: