No technical system is perfect. In every production plant at any moment something may going wrong: a leaking coupling, a stuck valve, a bad electric contact, or whatever. Generally such failures can be ironed out without interruption of the production process or without endangering the personel. In a nuclear plant the health risks are much larger than in conventional plants. A small spill, only a nuisance in a conventional plant, may have serious consequences in a nuclear plant. For that reason the quality specifications for materials, control systems and personel in a nuclear plant and other nuclear facilities, such as reprocessing plants, are considerably higher than in non-nuclear plants.
High quality specifications mean a high degree of predictability of the properties and behavior of materials and structures. The higher the specifications the lower the tolerance for random occurrences, for impurities in the materials and for deviation from the dimensional specifications of the structures. High quality standards can be met by stringent control during the production process and by a large input of energy, most of it embedded in materials and specialized equipment. From the Second Law [more i41], it follows that the energy inputs exponentially increase with increasing quality specifications of a given amount of material or piece of equipment.
Bathtub hazard function
The risks for catastrophic breakdown of technical devices, including nuclear reactors, change as the devices age, much like the risks for death by accident and illness change as people get older. There are three distinct stages in the lifetime of a technical system or living organism:
• the break-in phase, also called the burn-in phase or the infant mortality phase,
• the middle life phase, also called the useful life,
• the wear-out phase.
The risk profile, the total failure rate as function of the time, for these three phases curves like a bathtub. Applied to technical devices only, the bathtub curve may be considered to be the sum of three types of failure rates.Obviously, the boundaries between the three life phases are not sharp.
• Early life ('infant mortality') failures, caused by bad design, defective manufacturing, material imperfections, faulty installation, unanticipated interactions, poor workmanship imperfect maintenance. and ineffective operation. The failure rate of this type decreases with time. The steepness of this curve depends on factors such as the amount of ‘pre-flight’ testing and the effectiveness of the quality control during manufacturing.
• A constant rate of random failures during working life, caused by accidents and random events. The height of this rate depends on, among other, the quality of the materials, of the design and the professionalism of the operators.
• Wear-out failures, caused by ageing, deterioration of materials, etcetera. The rate increases with time. Wear-out failures are typically the consequences of Second Law phenomena [more i39, i41].
The concepts behind the bathtub curve are playing an important part in space technology. The reliability and predictability of the behavior of each component of a spacecraft or launch vehicle has to be extremely high to achieve a specified reliability of the complex assembly as a whole: the spacecraft or launch vehicle. Extensive testing and screening procedures are applied to pass all components and assemblies through the break-in phase and to eliminate design flaws, manufacturing defects, etcetera. Functional flexibility by redundancy in the design of the spacecraft systems and very high quality standards minimalize the occurence of random failures and postpone the wear-out failures. Exhaustive screening and pre-flight testing and stringent quality control make spacecraft possible to function unattendedly for a decade or longer. The effort needed to achieve such a level of reliability is exceedingly large, a direct consequence of the Second law. Large efforts mean high input of energy, materials and human resources, and consequently high financial cost.
Bathtub curve and nuclear technology
In the commercial nuclear technology no 'pre-flight' testing occurs. A nuclear power plant is assembled at the location chosen by the utility which will operate the plant. Design flaws and manufacturing defects are uncovered during construction and during the first several years of operation of the nuclear power plant: the burn-in phase. Historical evidence indicates the burn-in phase of nuclear power plants to be several years. Major failures of nuclear reactors, including Three Mile Island 2 and Chernobyl, occurred during the burn-in phase.
Exactly the factors contributing to the burn-in phase failures are the cause of massive cost overruns of nuclear power plants and other large technological energy projects, as analyzed by the RAND Corporation . Recent examples of above mentioned habit of the nuclear industry, building before testing, are the troubled construction of the EPRs at Olkiluoto in Finland and at Flamanville in France, causing dramatic cost overruns and time delays.
Ageing processes of technical systems are consequences of the Second Law and are difficult to detect because they usually occur on the microscopic level of the inner structure of materials. The number of incidents and reportable events will increase. In addition, the aging process is leading to the gradual weakening of materials that could lead to catastrophic failures. Most notable among these processes is the embrittlement of the reactor pressure vessel. Failure of the pressure vessel of a PWR or BWR inevitably leads to a catastrophic release of radioactive material to the environment.
No human-made structure can be made absolutely fail-safe during tens of years. In the first place accidents and random events are impredictable by definition. The quality of the properties and the behavior of materials and structures predictably decline with time by ageing, cracking, wear, corrosion and other Second Law phenomena: the rate of wear-out failures predictably increases with time.
Inherently safe nuclear power is inherently impossible.
Figure 15-1. Bathtub hazard curve
The bathtub hazard curve is the sum of three types of failures rates: the early life failures, decreasing with time, the random failures, constant over time, and the wear out failures, increasing over time. The bathtub curve is valid for technical devices, including nuclear installations, as well as for living organisms.