Wednesday, December 07, 2005

Nuclear Engineering International: An ocean apart?

Nuclear Engineering International: An ocean apart?

Stricker, Laurent
A comparative review covering production performance, costs and human resources of the US and French nuclear power fleets was carried out.

Over 18 months ago, the Nuclear Power Operations and Nuclear Engineering divisions of Electricité de France's (EdF's) Generation branch embarked upon a series of analyses and studies with the aim of improving the economic performance of its nuclear fleet.

Against this backdrop, it became apparent that a comparative review of nuclear generation productivity of the American nuclear power plant fleet (103 units) and that of the French fleet (58 units) would be appropriate. Due to the now fierce competition in Europe, it has become difficult to exchange experiences with other electricity utilities which, in addition, have smaller fleets. Neither was benchmarking with Japan considered relevant at this stage. Japan consolidates data in a way that makes comparison difficult, whereas the USA - through the Nuclear Energy Institute (NEI) with 120 staff at the service of the nuclear industry - regularly produces analyses and data that can be easily compared with French data.

This study was therefore designed to compare productivity in three areas: generation performance, costs and resources employed. Firstly, the changes over the past 20 years were measured, which clearly highlight the differences, shifts in trends and abrupt changes. At the same time, the main features of the organisations and methods implemented were described in detail in order to identify effective practices and begin to quantify improvement goals.

This project, lasting about a year, was carried out by a team made up of representatives from the various divisions involved, coordinated by JAL Consulting. A high-level group of EdF's nuclear executives met with the main nuclear organisations in the USA: the NEI; two nuclear utilities - Exelon in Chicago and Duke Energy in Charlotte; two nuclear vendors - Westinghouse and Framatome ANP; the Electric Power Research Institute (EPRI); and Navigant Consulting, a firm specialising in productivity enhancement studies.


Table 1 summarises the main data for these two nuclear fleets. It should be noted that the production ratio (USA:Francc) in 2001 is 1.92 for a installed capacity ratio (USA:France ) of 1.56. In fact, US plants deliver all they can produce to the grid, whereas the utilisation rate of EdF plants is only 90%, the high percentage of nuclear power in electricity generation (77%) making load following necessary. We shall come back to this point later in connection with Kd (availability factor or capability factor) and Kp (capacity factor or load factor).

It should also be pointed out that 35% of the reactors in the US nuclear fleet are BWRs, and 65% PWRs. Analysis of the relative performance of these two types of reactor shows that, for a number of years, BWR and PWR performance has been equivalent, which justifies an overall approach to the fleet and tends to show that improved performance now depends more on management than design.

As regards the French fleet with 58 PWR reactors, convergence in the performance of the plants of the various standardised series - 90OMWe, 130OMWe and 145OMWe (N4) - can also be noted.

The most significant differences between the fleets are shown in the Panel (bottom left). The standardisation aspect is of particular note, since it seems that the systematic introduction of good practices is one of the main factors in the improvement of US performance. This approach is found in other industries such as aeronautics (aircraft maintenance is done in the same way for a given company, whichever country it is in).

The number of operating companies in the USA, which stood at 54 in 1989, for 113 units in operation, fell to 24 in 2001 for 103 operating units. It should drop to around ten by 2010. The major companies are shown in Figure 1.

The companies operate fleets consisting mostly of PWR and BWR plants but, as already noted, the performance of these two types of plant is comparable. Certain utilities operate in a deregulated context (such as Exclon and Entergy North-East), others in a non-deregulated context (Duke, Entergy South-East, TVA, Southern Nuclear), but all have successful ongoing performance enhancement programmes.

Capability factor (Kd) and capacity factor (Kp)

Figure 2 shows availability factor (or capability factor, Kd) performance for both the US and French fleets.

Over the last 20 years, there has been constant progress in US performance, with the exception of 1996 and 1997 when a sharp fall was due to the Nuclear Regulatory Commission's (NRC's) shutting down often units and limiting generation to 30% for four other units. The 2002 performance of 91.9% was not repeated in 2003 (due to the extended shutdown of Davis-Besse and a number of lengthy shutdowns to carry out programmes related to life extension authorisations for certain units). Overall, US performance should average around 90%, though the stated US goal is a load factor of 93%.

EdF performance, over the last ten years, has been stable at around 82% and the good performance of 82.7% in 2003 is worth highlighting. When compared to the US fleet, as already mentioned, a significant difference lies in the fact that nuclear power accounts for 77% of total electricity generation in France against 20% in the USA. Consequently, under present conditions, where exports - although important - are nevertheless limited, the capacity factor (Kp) is noticeably lower than the availability factor (Kd).

Figure 3 is important for two reasons. Firstly, the capacity factor (Kp) curve corresponds to production sold and therefore to the revenue for the EdF group; it is, in fact, the criterion that all other operators (including US operators) publish when they compare performance. The Kd curve shows the 'available productivity reservoir' if it were possible to sell available production over an area extended to Europe. Although this is unrealistic in the near future, given the inadequate very high voltage grid (in particular interconnection lines), this scenario should sooner or later come about with the growth in electricity needs within the next 20 years, along with a shortfall in generation facilities in most other European countries.


The Three Mile Island incident, which resulted in a considerable increase in generation costs (fuel plus O&M) in the order of 40% between 1980 and 1987, caused a considerable setback for the nuclear industry. In 1997, unit 1 of Three Mile Island was up for sale at around $60 million - the cost of an outage - whereas it must now be worth around 15 times more.

This 'comeback' coincided more or less with the beginning of deregulation (the 1992 Energy Policy Act). In 1994, this resulted in the industry as a whole bringing together its professional decision-making bodies within the NEI in Washington. In addition to carrying out all the performance enhancement programmes, the NEI considers long-term strategic prospects and interfaces with government bodies, particularly the NRC, as well as Congress and the public.

It should be noted here that the relationship with the regulator - the NRC - underwent a major change in 1997. Operators and the NRC embarked upon in-depth dialogue in 1990, which led to the introduction of new approaches: probabilistic risk assessment (PRA) from 1990. The NRC made a policy statement on use of PRA in 1995 and in 1996, all company leaders approved NEI 96-04, a document summarising proposed principles for regulations based on risk and performance assessment.

It would seem that the USA has regained confidence in nuclear energy. In addition to the improvement in performance, which represents an increase of 30% in generation at virtually the same installed capacity, two other factors are contributing to this revival:

* The number of authorisations (30 units at 17 sites) already issued for life extension from 40 to 60 years; on this point, NRC officials have publicly stated that they are expecting 90% of existing units to eventually apply for an extension.

* Power uprate programmes amounting to some 300OMWc in recent years. Power margins can go from 5% to 15% for costs in the region of $500/kWe with a return on investment in around three years. By 2010, an additional 5000MWe could thus be gained.

This recovery in the nuclear industry in the USA seems sustainable, since it has occurred against a backdrop of growing electricity need - a rate of increase of 1.8% per year forecast by the US Department of Energy (DoE) and wider acceptance of nuclear energy. The NEI is trying to promote a 2020 vision for a programme of 50,00OMWe of new units which, with the further improvements expected in productivity and power uprates, would increase the nuclear penetration rate from 20 to 23%.

Whatever the outcome, it is worth pointing out the decisions already made for fuel storage at Yucca Mountain and the funding of preliminary designs for new plant construction. It should also be noted that administrative approval for the APlOOO design has recently been given and that the DoE has a programme to commence building a new reactor by 2010. Finally, the USA plans to renew its enrichment capacity by replacing its Paducah gaseous diffusion plant by 2012 with two projects: LES II (3 million SWU per annum) in a Urenco-led consortium and the USEC American Centrifuge plant (3.5 million SWU/y) in agreement with the DoE.

These medium-term programmes (including Generation IV) set up a favourable background for French projects, in particular EdF's EPR demonstration project and Areva's Georges Besse II enrichment plant.


There are three factors contributing to the difference in US and French capability: cycle length, outage duration and unplanned capability loss factor.

Cycle length

As an overall average, cycles are around 20 months in the USA, whereas under the present fuel management system, they are 13 months long for the EdF fleet. For EdF, a series of programmes has been decided on, which will increase cycle length to an average of 16 months by 2012.

These measures will result in a capability gain of 0.9%, which seems difficult to exceed, unless all the core management policies are revised (new safety studies, and technical and economic calculations).

Average outage duration

Outages last in the region of 37 days in the USA; they are some 18 days longer for the EdF fleet.

It should be noted that average outage duration in the USA was reduced from 105 days to 37 days between in 1990 and 2001. There are two main reasons for this reduction, which should continue down to 25-30 days: stricter and enhanced outage organisation and a significant transfer (estimated at 40%) of the maintenance being carried out while units are at power.

A detailed comparative study was carried out 18 months ago; it shows that an average of 15 days can be made up on the US/French difference of 18 days. A scries of measures has been decided on, which should increase Kd by 3.5% by 2012.

Unplanned capability loss

First analyses show that unplanned capability loss is around 1% in the USA against 2.5% for the EdF fleet, a rate which could be reduced by 1 %.

In all, with the implementation of the measures already embarked upon (82+3.5+0.9), 86.4% should be achieved. Further studies have been undertaken to estimate additional gains which could be made; for example, a five-day gain on outages represents 1.4% of the capability factor.

Even if they do not have a direct impact on capability, other indicators make it possible to assess technical performance of an operational fleet. The World Association of Nuclear Operators (WANO) has defined nine indicators. In addition to Kd and the unplanned capability loss factor that we have already seen, the industrial safety indicator - where the American fleet appears to be making noticeably more headway - and that for radiation exposure were analysed.

Figure 4 shows the considerable progress made both in France and in the USA in the area of radiation protection. At this stage, it seems that any further reductions will go hand in hand with very high costs.


As already mentioned, standardisation is one of the crucial aspects in performance enhancement. It is implemented in each utility, and also at NEI level. It could be seen in the operators visited (Exelon and Duke Power) as well as at the NEI, which uses the SNPM (Standard Nuclear Process Model).

Within Exelon, for example, the standardisation of seven macro-processes was mentioned (outages; training; skills management; operating methods, both for procedures and work routines; preparation of medium-term business plans and annual budgets; reporting using common indicators). The method used is not centralised but purely top-down, and seems to be based on peer groups: these groups draw up conclusions that are submitted to corporate level for approval and applied on all plants. It should be noted that these various improvements, which are the management's responsibility, do not imply identical organisation in each company.

The NEI benchmarks all the processes identified in the SNPM. In all, 25 studies have been carried out since 1995, classified under four headings: culture; skills and expertise; economic practices and ideas; resources. By bringing together experts from the various fields, this systematic analysis mechanism should lead to practical, consensus-based proposals for actions within 3-5 months, a basis for ongoing performance enhancement.

Along with 'standardisation', improvements are driven by a performance-centred management culture in all areas. This change is achieved by alliances with the main contractors, but also initially calls on internal company resources, with the aim of resource optimisation and streamlining.

However, each phase of reorganisation seems to have been prepared very carefully and with proper negotiations, including with staff representatives. Once the decisions have been taken, their implementation and performance are carried out and rigorously checked, and the performance achieved on each site is presented to the plant management.

Overall, the general philosophy consists of bringing all the units in the fleet up to the 'best in class' level.


Various types of operating cost were analysed. Here, operating expenses (O&M costs), fuel costs, and their sum, production costs, are considered. The study covered the 103 US units compared with the 58 French units in 2001 constant dollars, with a conversion rate of euro0.94 to the dollar.

Production costs

The data shown in Figure 5, widely published by the NEI, shows that production cost increased sharply in the USA from 2.4 to 3.4 ¢/kWh from 1981 to 1987, then declined steadily to 1.6¢/kWh in 2002. EdPs production costs, which have always remained below 2.4¢/kWh, have been falling since 1992 due to the drop in fuel cost, and stood at 1.4¢/kWh in 2002. This reduction is due wholly to the decline in the cost of fuel, whereas for the USA it results from a combination of lower fuel cost and direct operating expenditure.

Fuel costs

In terms of fuel costs, front end fuel costs have to be distinguished from those back end. Front end fuel costs are comparable and have dropped on a steady basis, reaching 0.4¢/kWh in 2002. EdF's back end fuel costs are higher, standing at 0.2¢/kWh in 2002 and are decreasing significantly, while US back end costs are valued at a flat rate of 0.1¢/kWh in current value and correspond to a tax paid by utilities to a government waste management fund. The government's delivery, however, has not met expectations and utilities have suffered additional costs (see NEI July 2004, p8). In all, fuel costs with identical definitions are now of the same order, which makes the comparison more favourable for the EdF fleet.

Operating expenses

Operation and maintenance (O&M) costs, or 'direct operating expenses', are lower in France than in the USA (75% of the US production cost in 2001).

The US decrease is certainly impressive (43% since 1987); while EdF has had constantly growing costs, though still much lower than US costs. However, two points should be made. Firstly, comparison in trend over the last five years is more relevant, since fleets have reached stable generating capacity. Over this period, it can be be noted that US expenditure has fallen by around 18% whereas O&M expenses in France remain stable. secondly, for the American fleet, approximately half of this decrease results from improved capability. As a ratio of installed capacity, O&M costs have fallen by 23% since 1987.


The 'resources' study was designed to measure the personnel in quantitative terms, estimate the in-house/outsourcing ratio, and interface between utilities and contractors/suppliers.

The resources available to the EdF fleet (not including the fuel cycle) were analysed from two points of view - internal (EdF) and external (contractors) for which five categories were deemed useful for the purpose of the analysis: manufacture, engineering, maintenance, generation and support. Internal resources were entered as staff directly or indirectly assigned. The analysis of external resources was based on the amount of purchases outside the fuel cycle (around euro1.6 billion in 2002), broken down into ten segments and sales per person, which enabled staff numbers to be determined.

The comparison with the USA has not been completed, since our counterparts became aware of our approach during the mission. However, the NEI did carry out the same type of analysis with the aim of identifying, on the present basis, the needs for the nuclear industry as a whole in ten years, in order to attract more people and skills in the nuclear sector (which is now attractive again) and anticipate the training courses to be set up, with partners in the teaching professions. For benchmarking discussions, engineering, maintenance and support have been grouped together so that the area compared is consistent with that of the operators visited.

Taking internal and external resources for the French fleet gives a total of 48,000 FTE (full time equivalent) staff with 57% EdF and 43% contractors - a ratio of overall resources of 760 persons/100OMWe (p/GWe).

Table 2 details these totals under three headings: operations, manufacture, engineering-maintenance-support. Excluding manufacture, which gives a total considered more relevant by our American colleagues, the ratio drops to 699p/GWe and the rate of subcontracting to 38% (62% carried out by EdF).

At site level, (for instance, the case of Gruas, four 90OMWe units), for 1244 EdF staff members, non-EdF resources represent 570 FTE, making an outsourcing rate of 32% for total staff numbers of 504p/GWc.

For maintenance, the staffing ratio is 40% internal to 60% external (during outage, it is 15% EdF to 85% contractors).

The American data available to date shows a total of 59,400 staff members in the utilities (594p/GWe against 433p/GWe for EdF, that is +37%), even though the utilities have implemented significant staff reduction programmes (Duke Power decreased from 8800 persons for 7GWe in 1991 to 4313 in 2002 with a target of 3962 in 2006, that is 566p/GWe).

The Overall resource' level is probably higher than that of the EdF fleet, which can be seen through comparison of operating expenditure, but it seems that the outsourcing rate is lower in the USA than at EdF (Duke estimates it at 30%, Navigant Consulting at 25%).

However, the 'make-or-buy' policy is a very pragmatic approach, with streamlining factoring in the stability of the body of staff as much as direct economic efficiency. In this respect, the impressive cuts in overall staffing were often achieved firstly by a reduction in outside contractors. While necessary for introducing such deep-rooted changes over a long period, this approach is not at odds with alliances with the main contractors.

Based on over ten years' practical experience in some cases, alliances have become common in the USA. They are long-term contractual relationships (3-5 years, or more) covering fairly wide areas with clearly identified responsibilities and interfaces on both sides, in which risks and profits are shared on a 'winwin' basis according to formalised, accepted incentive systems. Thanks to their duration, these contracts foster additional investment in equipment as well as training. They promote joint forward planning of services (at least a year, against up to four months at EdF). It should also be noted that in the case of subcontracting, the prime contractor assumes overall responsibility for all the work to be carried out.


As a general conclusion to this benchmarking, lessons can be learned from the experience gained in the USA over the last ten years. In addition to the relations already well-established (EPRI, INPO, Framatome ANP and Westinghouse), EdF wishes to further its link with the NEI and some utilities.

In pragmatic terms and as has been highlighted for reduced outage duration, EdF's DPN (Nuclear Power Operations Division) has already taken on board a certain number of conclusions but also for all aspects of 'good practices' for which systematic analysis has been undertaken.

Above and beyond that, this work will be used in the 'Nuclear Initiative 2015' (IN 15), which has just been started within the Production-Engineering branch. This approach is aimed at preparing for and implementing changes in the nuclear field in order to face future challenges within a background marked by the following four points:

* Continued priority of nuclear safety.

* An increasing demand for competitiveness.

* A fleet which will include units reaching maturity, units at the end of their life and units starting to be operated (EPR) by 2015.

* Considerable renewal of skills as a result of numerous retirements.

Nine projects in IN 15 will review all aspects covering the responsibility of EdF in the nuclear field and identify areas for improvement and performance enhancement.

Laurent Stricker, Director of Nuclear Power Operations Division, Electricité de France Engineering and Production Branch, Site de Cap Ampère, 1 place Pleyel, 93282 Saint-Denis Cedex, France; Jacques Leclercq, Chairman of JAL Consulting, 21 boulevard Suchet, 75016 Paris, fronce

Copyright Wilmington Publishing Ltd. Dec 2004
Provided by ProQuest Information and Learning Company. All rights Reserved


Blogger Вячеслав said...

Is anibody who can give me this article. I need to see tables, which not included in this post. thanks.

10:03 PM  

Post a Comment

<< Home