Maintaining safety

A HUNDRED years of operation has resulted in excellent technical support for hydro power stations and consequent high availability (over 98%). Because availability requirements vary with the seasons and this affects inspection periods, the plants have been maintained for many years using a time-oriented maintenance strategy. In the past the maintenance of hydro-electric power plants was shaped by:

• A high technical level.

• Few failures.

• Relatively high maintenance costs.

The deregulation of the European electricity market, with the associated competition between power suppliers, has been the trigger for enormous pressure on costs in the electricity industry. The generation sector has been affected the most by this, since the competitive environment is very distinct, compared to the distribution side.

Maintenance costs for power generation plants have also come under heavy pressure and solutions had to be sought in order to make savings. The companies have reacted to this in two ways.

The first was organisational. In many sectors of the electricity industry a change from decentralised organisational structures have occurred, either to integrated forms (with co-operation between central and local working groups) or central organisations. Maintenance personnel were pooled across large company divisions, with the advantage of the more flexible use of existing staff.

This trend towards centralisation was usually accompanied by a significant reduction in the number of personnel. As a result the personnel's relation to a particular plant was impaired. Knowledge management (securing and distributing existing knowledge) is of particular importance especially during such developments.

The second change was in maintenance strategy. Originally the hydro sector used a 'time-related strategy' with fixed inspection or audit intervals, partly because the availability requirements were subject to extensive seasonal variation. But this means the question of technical necessity has not always been addressed and an approach that considers operating hours or one that optimises operating cycles allow operators to take advantage of the wear reserve.

What is more, as suggested by the name, in 'condition-oriented maintenance' many additional devices are installed to monitor the condition of the generating sets (vibration monitoring, oil analysers, automatic gap measurements, cavitation sensors etc). The results are important in making operating decisions.

An alternative method is 'damage-oriented maintenance', where repair follows damage with the appropriate preparation. It is not yet common but will become more important.

Each of these approaches has its justification, but which strategy is the correct one for the component considered in each case? After thorough analysis 'reliability centred maintenance' provides the answer to exactly this question.

The RCM method

RCM was developed in the aircraft industry and has been used in many sectors of industry. This method is particularly common where high safety requirements exist.

A quote from Albert Einstein best expresses the fundamental idea of this method: 'Recognising the problem is more important than finding the solution, as the accurate representation of the problem automatically leads to the correct decision.'

RCM is an industry-independent, systematic analysis technique for determining optimal maintenance strategy. RCM analysis requires the following steps:

• Specifying detail depth. The plant components for which the analysis is to be carried out are specified. The further down in the plant hierarchy one starts, the costlier and more difficult the analysis becomes, so it is advantageous to start on a less-detailed level. During the analysis process it will then become apparent whether a finer breakdown of the plant components is sensible.

• Which function does a certain component perform? The function of each component is ascertained. If one component has several functions and they can malfunction independently from each other, then they should also be listed independently. If the function is connected with a detectable measured value, this measured value should also be represented.

• Which malfunctions are possible? For each function of a component the possible malfunctions are listed, focusing particularly on those malfunctions that are highly probable or have already occurred, or which are expected to occur towards the end of the service life. A distinction between partial and complete loss of function can be useful.

• What is the cause of the malfunction? Possible causes are assigned to each malfunction. Human error must also be considered, if the effects are drastic.

• What is the effect of the malfunction? Experience shows it is useful to acquire the following data: the effect that shows the malfunction has occurred (eg automatic message, water leak); the measures to be taken if the malfunction has occurred (eg immediate emergency shutdown, replacement of parts); additional follow-up costs such as downtime, which entails a loss of generating capacity. In this connection a distinction between unplanned downtimes, downtimes that can be planned for at short notice, and downtimes that can be planned well in advance is useful.

• What can be done to prevent the malfunction? The assessment of the effect determines the maintenance strategy, which is specified together with a description of the measures using a decision tree. The decision tree can lead to: a planned condition-related measure; a planned overhaul; planned replacement; trouble-shooting measures; design modifications; a combination of measures; or no measure at all.

Applying RCM

The RCM method was applied at Energie AG hydro power plants. Energie AG is a leading infrastructure group of companies in the federal state of Upper Austria with activities in electricity, heat generation, gas, water, disposal and telecommunications. The enterprise's core business however is still the generation and distribution of electric power in Upper Austria. It operates 34 hydro power plants, 25 of them run-of-river and nine using storage. They are of small and medium size with a combined generation of around 1059GWh annually.

RCM was tested using the turbines of the Wagrain power station and the weir plant of the Plankenau power station.

The Wagrain storage power station came into operation in 1987 and is equipped with two Francis-spiral-case turbines with vertical shafts. The plant is particularly heavily affected by erosion and has had many improvements. Apart from the installation of a desander, design modifications such as the strengthening of wearing parts were carried out and the operating mode has been optimised. The turbine inspection interval could be extended from one year to three years.

The components considered in the analysis are shown in the diagram above.The analysis provided the following results:

• Condition-oriented instead of time-oriented turbine inspection could be employed, by opening the control valve in the crevice water discharge pipe to measure the wear of the split rings and carrying out regular measurements to determine the current efficiency level of the turbine.

• Technical measures that could increase service life included redesign of the split ring layout and coating individual wearing parts.

• Additional spare parts should be acquired to reduce the risk of long downtimes.

• Plant documentation should be improved.

The Plankenau weir plant has four waterways. Three waterways have rigidly connected wooden upper and lower bulkhead gates. One is a steel construction and is equipped with upper and lower bulkhead gates, which can be moved independently from each other. Three waterways are operated hydraulically, while one wooden weir is operated electrically by means of a spindle.

For the RCM analysis of the weir plant 25 components were considered and in the detailed analysis of the hydraulic unit a further 38 components were analysed. The greater depth of detail, compared to the Wagrain plant, was predominantly because the weir plant was examined especially in terms of its safety.

The analysis found that:

• Inspection measures that could provide early detection of faults were the waterway opening times and leakage losses in the hydraulic system.

• Installation of additional components would increase operational reliability.

• Environmentally friendly products could be used.

• Plant knowledge and documentation should be improved.

Wider findings

Selection of the personnel to carry out the RCM analysis is very important. A second aim of the technique is to counteract the loss of knowledge that follows personnel reductions. The systematic questions even promote easy knowledge exchange between the various organisational units. Therefore, personnel from the respective operational sector, maintenance personnel with inspection experience, and personnel from the design department (for new plants) should all participate.

An important factor for the success is the willingness to provide knowledge. In particular the attitude of the foremen concerned is vitally important, as the essential plant and maintenance knowledge is concentrated here.

Excel data sheets were used for data acquisition and documentation. For each component, an information and decision sheet are available in which the measures are described, and in which the interval and respective executant are specified. This documentation activity has proven to be very time consuming. A corresponding database application featuring data exchange capability with Energie software for the plant management was a useful solution.

The handling of the RCM decision tree has also proven to be complicated. Following the authors experiences and using as a basis a decision tree developed by Professor Sturm, a simplified decision tree was compiled. The process proved time consuming, particularly because of the absence of suitable EDP support and the use of the more complex decision tree.

It became clear relatively quickly that the systematic approach has many strengths that can be used, if a suitable EDP tool is available that connects as closely as possible to the existing operating and maintenance software.

On the basis of an MS-Access application programmed in-house he authors contacted Siemens, manufacturer of BFS++ and managed to arouse the product managers' interest. In the meantime a prototype 'strategy manager' is already available, which is subject to an intensive test series. In this EDP version a considerably more simplified decision tree developed in co-operation with Siemens and Professor Sturm was used.

The advantage of this solution is that the strategy identification module has been integrated into Energie's plant management software. Thus the existing plant data can be referred to, and other modules such as fault messages, operations scheduling or work orders can be connected almost seamlessly to the analysis module. The entire documentation of the analysis process is thus also effected within the same system.

In summary, the RCM method has a number of strengths. It is a systematic and purposeful procedure for determining maintenance strategy and it promotes a holistic view of a system where often a often small cause has a big effect. It includes fault analysis of safety-relevant components and compact documentation of the decision-making process and and it provides a method for securing the entire knowledge available within the company, improving co-operation between various divisions.