Service trials

Few would argue against real service trials being the first choice, although in practice the conditions and timescale required frequently rule it out completely. Simulated service trials using test rigs are more widely feasible and are often considered essential for critical components. 

Although a great number and diversity of product test rigs have been used, the cost and difficulties have restricted their application and it is fairly clear from accounts of failures that there are many instances where a service trial or rig test could have detected the problem before it occurred in the field. For example, in the case of radiator seals cited earlier the use of a relatively basic test rig would have found the deficiencies in the material chosen. 

A different use of service trials is in the procedure of simulated design life. The product is subjected to environmental exposures which equate, with the inevitable approximations, to a chosen design life and then performance is assessed by operating the product. The exposures normally have to involve accelerated procedures and may comprise several environmental agents applied simultaneously or sequentially. 

---

In another service trial on the Western Region of British Rail, sleepers of copper steel (Cu 0.5%) were found to lose up to 50% less weight by rusting than ordinary steel sleepers when laid in the open track. In agreement with what has been stated above, there was, however, little difference between the performances of the two types of sleepers when they were laid inside the Severn Tunnel. 

A common feature of both these methods is that the quantity of treatment chemical can be calculated from stoichiometric relationships in the reactions involved. This is not so with conventional inhibitor treatments. With these the concentration of inhibitive chemicals can only be determined on the basis of experimental laboratory studies, service trials and overall practical experience. 

Few would argue against real service trials being the first choice if the conditions and time scale do not completely rule them out. Even when this approach is not feasible before the launch of a product, it is highly desirable that real life trials are started as early as possible because, at the very least, they can warn of impending disasters in the field. 

Strictly, the useful life of a product can only be measured directly by service trials or tests on the complete product. Most assessments of lifetime of plastics are made by considering some measure of performance and specifying some limit or threshold value for the property, which is taken as the end point corresponding to where the material is no longer usable. 

Simulating service conditions avoids all the risks of using real service, and offers the possibility of moderate acceleration by simulating the worst conditions possible. The name implies that all factors present are considered, for example mechanical stress and the environment. However, the time scales will still be long and in many cases it is difficult, if not impossible, to simulate real conditions accurately. Clearly, simulated service trials are most attractive where the expected lifetimes are relatively modest and the conditions to be simulated are not too complicated. 

In the case of a new design it can be more expedient, and certainly effective, to subject prototypes to real service rather than to develop simulation tests. However, there are many cases when this is simply not sensible for time, cost or safety reasons. So, when real service trials have to be ruled out and prediction from laboratory material tests cannot be relied upon then there must be whole product testing. 

Controlled health services trials of effect of changes in institutional policies effecting the patient milieu, such as staffing, lighting, organized patient activities, medication protocols, etc., on the sleep of institutionalized dementia patients. 

However, because of these difficulties, candidate materials for an application should not be selected by an arbitrary method if a laboratory test is not feasible. The published data should be consulted taking into account the parameters of the application with a critical appraisal of the environmental conditions. An in-service trial can be undertaken and the problems associated with this approach are considered next. 

It cannot be denied that data taken from a service trial compares bearing materials in their true working environment, although a good performance in one application will not necessarily ensure the same in another, unless the mixture of wear mechanisms is identical. However, duty requirements and economic considerations do not always allow full in-service trials to be undertaken. 

The paper has outlined the problems which can arise when using wear data from laboratory tests or from service trials. The examples illustrate the potential error of wear rates and hence life predictions. It would be inappropriate to conclude this paper without some positive recommendations which may help to reduce this error. 

Wear is important in applications as diverse as tires, conveyor belting, footwear, and windscreen wiper blades, yet most of us would accept that the rubber industry still awaits a laboratory test or series of tests that can predict service performance with any confidence. Over the years, there certainly has been no shortage of small-scale tests for abrasion resistance and indeed no shortage of claims or reports of satisfactory correlation with the behavior of various products, but strong doubts remain, and some sectors, not least the tire industry, conclude that there is no substitute for a service trial or a simulated product test, for example, a road trial on tires held at a small slip angle to accelerate wear. 

Hudson, J. C. (1960) Final report of service trials of steel coal wagons (1939-1954) J. Iron and Steel Inst. 194 (Jan) 45-49. 

Development work and service trials are being carried out in a wide range of rubber products, e.g. milking inflations, pharmaceutical and food products, oil hose and seals, conveyor belting, shoes soles and adhesives. 

Generally, it must be concluded that in too many cases service trials were deemed not feasible, sufficient knowledge or experience was not available and accelerated testing was thought too expensive. 

Several full-scale bonded panels have been prepared and the results have proved successful in demonstrating the capabilities of structural adhesives. Assembly and service trials will be conducted within the next 15 months. 

---