In an industrial facility, the key-concept is reliability. It is all the more true in utilities such as electric power companies, since one must be able to trust the electricity provider 24/7, year in, year out. It is even crucial when it comes to nuclear power production, where one should expect high sustainability and total safety. In this industry, rigorous mastering of the production tool is thus a necessary condition for the technical and economic performance. The heart of this industrial model is engineering.
The performance of an industry reflects its technical and economic efficiency, but you can also assess it through the sector's capacity for anticipation and commitment. Such performance can only be attained through sufficient mastering of the production tools. What kind of mastering are we talking about? One could say that it comes from the concentration, into one single competence, of three elements: firstly, a capacity to design the production tools; then, an extensive knowledge of the suppliers and equipment that constitute these production tools; and, lastly but of prime importance, the return of experience (ROE) gained in operating these tools.
It is the interaction of these three elements that determines the quality of the industrial process as a whole. Operational ROE allows to improving the design and performance of the equipment. Drawing on this ROE, it is possible to optimize the design according to the industrial capacity of the sub-contractors and the existing technologies. Subsequently, it is possible to orient the industry through direct confrontation of the operators’ requirements with those expressed by the suppliers.
Gaining full control of the tools is the ambition of all industrial companies, whatever methods they develop to do it. A case often cited is that of Rolls Royce, famous not only for its luxury cars but also as the world‘s second aircraft engine manufacturer. On the assembly lines all RR engines are fitted with sensors that transmit engine data (oil pressure, etc.) continuously, as soon as the aircraft has taken off. These data are analyzed by the company’s engineering departments. Thus Rolls Royce can draw on an incomparable knowledge base of engine behavior over all their operational life. Michelin enlarges the ROE base of its tire systems with the same rigor.
The French nuclear power industry is no exception here. Benefitting from over 1,700 reactor.year operations and millions of engineer.hours, it has now become a case reference for industrial mastering, especially in the aftermath of Fukushima. This result is largely due to EDF’s integrated industrial model.
With some 220,000 employees and 2,500 contractors, the French nuclear power industry has an annual consolidated turn-over of 46 billion euros and an added value of close to 15 billion euros. It covers all aspects of nuclear power generation, ranging from R&D, forward planning and design, to waste treatment and disposal.
France has an original industrial organization – going back to the strategic choices made in the 1970s – in which the national operator EDF participates closely in reactor design (including all ancillary equipment) and construction phases. The fact that the owner-operator EDF is its own architect-engineer is a key factor.
Right from the early days of the construction of today’s working nuclear plants, EDF has therefore been able to build up an integrated engineering management, in charge of design, building, procurement, gathering and analyzing, as it does, the ROE. Approximately 5,000 engineers, supporting 20,000 operators, keep implementing a continuous improvement loop.
On the basis of the ROE in construction work and operating its earlier nuclear power plants, EDF has acquired a complete industrial mastering of its production tool and developed its own reactors. In parallel, the process led to gradual standardization of the French nuclear sites again with a systematic and continuous ‘improvement loop’. This has enabled the French nuclear power industry to benefit from two reliable advantages.
The first advantage stems from the constant improvement of safety, efficiency and performance level factors, to the extent that each event is duly analyzed and used to improve on the design, the process or the equipment components. This brings with it certain consequences: in France the construction costs are lower than elsewhere and operating costs are between 40 to 70% lower.
The second advantage lies in the possibility for EDF to implement an industrial policy. It is the role of the architect-engineer to define how the contracts are divided into batch lots, the competition policy to be applied and the certification of the national or foreign suppliers, according to the technical or strategic sensitivity. It is the architect-engineer who inspects the products ordered and supplied. It is also its responsibility to specify and certify the inspection tools used during manufacturing/assembly of system component parts. The architect-engineer forwards the precise needs of the operator to the suppliers, thereby providing the clearest possible description of the operational requirements and the contractual orientations. The relationship between architect-engineers and the equipment suppliers is a key part of industrial mastering. It presupposes that the operator has a full knowledge of the industrial context, allowing for the identification and qualification of each potential supplier.
It is essential to draw on return of experience and technological progress offered by the equipment suppliers, not only to improve today’s reactors but also to help design new models.
Those countries which today possess the most dynamic nuclear industries, France, Russia and China, have all adopted an integrated model for their nuclear programs from the scratch. They are, however, the exception, since most countries operating nuclear facilities – the front-runners being the USA and Japan – have chosen preferably the turn-key model.
There are, it transpires, several serious disadvantages to this option.
Firstly, it is not the plant operator but the constrictor who defines industrial policy, who negotiates with the suppliers, the operator simply footing the bills. This option leads to considerable overheads: the German nuclear program cost twice as much as the French equivalent, the Japanese program three times as much.
It is also the constructor who, with the national nuclear security agencies, defines the regulatory framework in which the country’s nuclear power stations will be operated.
The site operator is precluded from discussions with the security agencies and the suppliers during the design and construction phases and, indeed, is reduced to playing the role of third party paymaster. And yet it is the operator who is in the front line as soon as power production begins on the site. The industrial context and its engineering are shut out vis-a-vis the site operator, and consequently neither can benefit from any return on experience, albeit the latter is essential to assuring the highest levels of safety and security.
This is why it makes no sense to compare the nuclear industry with automotive manufacturing. If car owners are responsible for the way they drive cars, they are not required to have an in-depth knowledge of how the cars work to be able to drive them. And the consequences of a road accident, even fatal accidents, will never occur on the same scale as a nuclear accident. What happened in Fukushima dramatically reminded observers of the prime responsibility of the operator, in charge of the reactors and the site. The events also demonstrated implicitly that nuclear security is not only a question of applying technological solutions but, above all else, relates to the industrial organization and site operational quality.
The errors committed by the Japanese nuclear power authorities can be summarized in two perfectly symmetric ways. The Japanese operator did not have full mastering of its production units; the Japanese site contractor in no way benefitted from the operator’s ROE. In addition, the high number of operators made the sector fragile in that it segmented any possible ROE. In the early 2000s, a corrosion problem was recorded in a plant – a small crack at a pipe bend in the secondary cooling circuit. Four years later, the same weak-point led to a steam leak at another operator’s site, causing the death of four technicians. Why? Because the ROE from the first noted incident was not been integrated in the design of the second site circuits by the constructor. Having homogeneous materials used does not change matters: all of Germany’s nuclear plants are based on the same Konvoï model, but they are operated by regional companies coming under differing standards and security procedures dictated by the local Länder. This called for technical adaptations at each plant site installed.
In contradistinction, with an integrated model such as is the case in France, ROE is centralized and relies on a progressively enriched set of standards and specifications. Though there are currently seven different reactor models in France, a program of security assessments is conducted notably at each ten-year inspection, which allows improvements to be integrated on a regular basis, both in the equipment used and site management, for all French reactors. With the experience gained, the overall level of security rises progressively and homogenously. The currents estimate (by the Nuclear Safety Authority) is that the risk of having a reactor core melt-down has been lowered by a factor 10.
It is this continuous chain of improvements and integration of ROE that is lacking for turn-key models.
Naturally, some of the faults noted above can be corrected. The USA, for example in the 1990s, with a view to offsetting the fragmented operator companies, created the Institute of Nuclear Power Operation (INPO), responsible for security assessments, station staff training and sharing the ROE for the benefit of all American nuclear power plant operators. Japan today, after drawing conclusions from several years’ thoughts on operator responsibilities, is progressing towards the creation of a similar agency, for the purpose of federating the ROE from its ten national electric power operators. In the same vein, several other countries are considering how to use ROE in industrial systems that previously did not resort to this source of information. Hence the creation, for example, of the World Association of Nuclear Operators (WANO), which groups together the nuclear power operators from over 30 countries.
Will this be enough? The concept of return on experience is absolutely central in the nuclear industry and it may not prove sufficient just to share it. The real challenge is how to integrate ROE into a loop that aims at continuous improvement, providing the operator with the constructor’s intelligence and, reciprocally, the constructor with the operator’s experience.
The lessons we can draw from serious accidents occurring elsewhere can prove useful. The core meltdown that took place at Three Mile Island, in 1979, was considered by the constructor to be an infinitesimal risk. The cause identified later was an operator error. The latter was analyzed and immediately integrated by EDF who launched a change of equipment, installing sand-filters capable of filtering 99.9% of the cesium should a leak occur at any of the national reactor sites. This ROE was not integrated at Fukushima Dai-ichi.
Another example will illustrate the interactions between the operator and the reactor constructor. At Three Mile Island, as the meltdown occurred, all the alarm systems in the control room were activated simultaneously. One can readily understand the difficulty of rapidly identifying what was happening in the core vessel! A major progress lay in upgrading the alarm systems, as EDF did, reserving distinct security control panels for different levels of incident/accident. This allows the operator to identify more rapidly the original cause for the event, thanks to a reactor status approach; in this manner, the security control engineers have sufficient leeway to advise the site operators as to the best protocol to adopt.
Integrating these procedures in a computer aided control & command chain can help operators to avoid making human errors. But here too, the difference between integrated and turn-key models are instructive. Some nuclear plant constructors have started a race to implement computer-aided control & command systems. However, when we visit their power stations, the reality is more modest: lots of data and useful information is available and the computers integrate these and take the necessary corrective measures … but the set-back is that the procedures themselves remain, so to speak, in the paper tray. This is not astonishing per se. Integrating the procedures into the control paths, as EDF has done, is a huge task, and is in fact only possible if and only if the ROE is integrated during the reactor design phases.
The continuous improvement culture that characterizes the integrated industrial model does carry some constraints. Notably, it is assumed that the constructor preserves all his specific technical capacity, and this is not self-evident for Western utilities who manage large numbers of sites without large-scale growth perspectives or plant renewal plans.
Of course, implementing a continuous improvement concept leads to technical upgrades in the equipment at each reactor site. We should bear in mind that, apart from the reactor vessel and the concrete shield walls, all the equipment installed in a nuclear power station can be modified and/or replaced, when deemed opportune. But such a security-conducive culture is only meaningful if it is an integral part of a truly industrial approach. Engineering capacity cannot be preserved just by replacing certain equipment parts through time. For this reason, it is very important for electric power industrialists to be present in those countries where the nuclear option is being developed, i.e., in Asia today, notably in China.
Each year currently, China builds and installs the equivalent of 100 Gigawatts, i.e., the same as in the entire French nuclear power industry. Before the catastrophic events at Fukushima, the Chinese began building a nuclear power plant every month. EDF was quick to seize the opportunity, over thirty years ago, to participate in the launching of the Chinese nuclear power program from its beginning with an EDF team integrated to the China General Nuclear Power Corporation (CGNPC), to ensure the main architectural and construction contract control for the building of China’s first nuclear plant at Daya Bay. Ever since then, EDF has been involved in the development of the Chinese nuclear program, associating all of the French nuclear industry partners.
The very dynamics that comes with new projects provides a tremendous opportunity to build an “industrial spirit” for the French nuclear power industry. To attain this, we must be able to call on a true project engineering vision and, through orchestrating all the riches of the nuclear industry, develop its own industrial model round the world.
This challenge goes well beyond the mere cost control aspects. One must have a constant, full vision of what remains to be done; to progress from the target to identify those means needed to attain the objectives, and to be able – if ever a project derails – to identify the causes and to correct and control them.
Above all other considerations, however, a project is a human adventure relying on a shared desire mobilizing men and women to attain common targets, understood and accepted by all concerned. No matter what project I had the privilege to manage personally, from the building of the Daya Bay nuclear plant to the design and construction of “N4,” the most recent nuclear reactor to be commissioned in France at the end of the 1990s, step by step, I gradually acquired the conviction that a collective success story needs a blend of intelligent alchemy between a strong organizational structure and a tightly knit team of collaborators.