Laboratory experiment as a basis of commercial plant design has been the subject of much analysis and thought and on which a large number of articles and papers have been published. We would be forced to choose between an unduly brief exposition if we wanted to cover it completely, or no exposition at all.
How much of this fundamental scientific attention has, as a matter of fact, been accorded to the subject, mainly by our sister profession, Chemical Engineering. The close similarity to chemical engineering problems that we meet, praticularly in hydrometallurgical operations, suggests that Chemical Engineering publications offer a rich source of information and discussions of methods useful to us. We have only to think of pressure leaching, ion exchange, liquid-liquid extraction, precipitation and so on, all problems in chemical engineering unit processes or of the new Freeport Sulphur Nickel plant in Cuba or the Sherritt Gordon plant and many others.
In the design of any commercial plant, problems arise which cannot be answered by laboratory tests, but their neglect would seriously jeopardize the success of the venture. We only need to think of accessiblity for ease of operation, supervision and maintenance, materials handling, storage as well as others which are primarily matters of experience.
A notorious example is the difficulty of predicting by laboratory tests what corrosion problems will be encountered in practice because in most cases the rate of corrosion is controlled by the fluid velocity, so that discrepancies between laboratory experiments and plant results can most certainly be expected.
We can, therefore, eliminate from the experimental program such that is readily available, and amply backed by previous experience albeit experience of others in such matters of crushing, grinding, settling, classification, agitation, filtering, transportation and a host of others. In fact, the probability is that in many cases, the equipment manufacturer, if he is informed of what is needed, can be of great help in simplifying and cheapening the development of the experimental program as well as by assisting in the design of the plant itself.
The words “average sample”, “representative sample” occur again and again. Much thought has been given to this subject but I submit that the nature of the raw material is itself one of the most important of the variables, and a study of the amplitude, the frequency and periodicity of variations in the raw material cannot be laminated simply by providing an average sample, no matter how carefully or at what expense it has been obtained.
The inevitable conclusion seems to be, therefore, that a broad area of engineering still remains an art rather than a science, not subject to exact mathematical analysis.