How Science Is Applied in Technology

Abstract

Unlike basic sciences, scientific research in advanced technologies aims to explain, predict, and (mathematically) describe not phenomena in nature, but phenomena in technological artefacts, thereby producing knowledge that is utilized in technological design. This article first explains why the covering‐law view of applying science is inadequate for characterizing this research practice. Instead, the covering‐law approach and causal explanation are integrated in this practice. Ludwig Prandtl’s approach to concrete fluid flows is used as an example of scientific research in the engineering sciences. A methodology of distinguishing between regions in space and/or phases in time that show distinct physical behaviours is specific to this research practice. Accordingly, two types of models specific to the engineering sciences are introduced. The diagrammatic model represents the causal explanation of physical behaviour in distinct spatial regions or time phases; the nomo‐mathematical model represents the phenomenon in terms of a set of mathematically formulated laws.

Acknowledgements

For constructive comments and suggestions on the content of this article, I would like to thank Henk Procee, James McAllister, Rens Bod, Susan Sterrett, Michael Heidelberger, and Margaret Morrison. The writing of this article, and the organization of the PSA workshop ‘Applying Science’ in which this article was presented, was supported by a grant from the Dutch National Science Foundation.

Notes

[1] My view on how science and technology were related became more refined of course, since it became clear that the natural world was much more recalcitrant than it had appeared in high school. In the real world, for instance, chemical reactions between components A and B produced the desired product, P, but also undesired ‘waste’, W. Therefore, the task of the engineer was to find physical conditions that would prevent the production of W and stimulate that of P; moreover, it required additional technologies to separate P from W and remainders of A and B. Nevertheless, finding optimal conditions and developing refined technologies could still be approached with scientific means, it seemed.

[2] The term ‘engineering science’ for indicating this research practice is problematic, particularly because actual research practices themselves are not very clear on this matter. Engineering sciences may be used in a rather narrow sense, indicating exclusive research into physical phenomena that occur in technological artefacts. But the term is also identified with systematic approaches in technology. In this latter denotation, the engineering sciences produce different types of technological knowledge.

[3] The industrial aim of reactive adsorption is sour‐gas purification. Toxic compounds such as hydrogen sulphide or ammonia are removed from waste gas by means of absorption into the washing liquid, where the toxic compounds react. The technological device (called a scrubber or absorber) consists of a column containing horizontal plates with holes in it. Washing liquid flows downwards through the plates, whereas the toxic gas rises upwards.

[4] From this exposé, it is clear that instruments also constitute an important relation between science and technology. de Solla Price (1984) termed them ‘instrumentalities’, which are the crafts and techniques (such as the laboratory methods) of the experimentalist and inventor. The advent of such instrumentalities has simultaneously opened up major new opportunities for scientific investigations and technological innovations. The mathematical method of science may also be regarded as a bridge between science and technology (see, for instance, Vincenti’s account of ‘theoretical tools’); this view may however be disputed (e.g. Staudenmaier 1985).

[5] This problem of hydrodynamics needs further explanation. Basic laws for describing flowing fluids are ‘conservation of momentum’ and Newton’s law of viscosity. However, when applying these laws to concrete flow phenomena, it is not possible to solve analytically the mathematical equation thus derived. Applying the principle of conservation of momentum to flow phenomena involves translating this principle to an inventory rate equation (in order to describe the dynamics of this conserved quantity). The rate equation for any conserved quantity ϕ takes the form (in words):

Rate of input of ϕ – rate of output of ϕ + rate of generation of ϕ = rate of accumulation of ϕ.

The application of this equation to flowing fluids produces partial differential equations, which are nonlinear, and therefore cannot be solved analytically. Therefore, the approach in classical hydrodynamics was to simplify this equation by assuming constant fluid density and constant viscosity, which is called a Newtonian fluid. These simplifications resulted in the Navier–Stokes equation. Unfortunately, when applied to concrete circumstances, i.e. filling out the Navier–Stokes equation at the specific boundary conditions of a system, those equations can still only be solved for a few very simple cases. A further simplification of the equations ignores the viscosity terms; in other words, it assumes that the viscosity of the fluid is zero. This simplification produces the Euler equation. This equation is only applicable to perfect fluids, i.e. Newtonian fluids with negligible viscosity. Mathematical solutions of this simplified equation are possible and agree well with observed behaviour of several kinds of flows but cannot appropriately describe flow past solid surfaces (e.g. Bird, Warren, and Lightfoot 2002). Since shear stress in the fluid is neglected, mathematical solutions of the Euler equations do not agree with the observed behaviour of flows in channels and pipes, and of forces on solid bodies caused by flow past them, etc.

[6] Cartwright’s account of models, as well that of earlier authors such as Hutten (1954), Nagel (1961), Achinstein (1964), and Hesse (1966), is related to how models are used in actual scientific practice.

[7] A proper distinction between observable phenomena and theoretical entities is complicated and not essential for my account of models of phenomena. Intuitively, a clear distinction is possible between directly observable phenomena such as flow patterns, and theoretical entities such as electrons. But when measurement techniques are available for determining the behaviour of theoretical entities such as viscous force or flow velocity, a distinction between observable phenomena and non‐observable (theoretical) entities becomes problematic.

[8] The model is termed nomo‐mathematical because the terms ‘mathematical model’ and ‘theoretical model’, which are used by several authors, are confusing. Theoretical explanation of a phenomenon may involve theoretical laws but also causal explanations in terms of theoretical entities. This confuses the required distinction between the causal story and the mathematical description. Another term in use for the mathematical description is the mathematical model. However, mathematical models have their origin in mathematics and do not necessarily have a physical meaning. The chosen term is an analogy after Hempel’s nomo‐logical model. In a ‘nomo‐mathematical’ model, physical laws are mathematically related instead of logically.

Additional information

Notes on contributors

Mieke Boon

Mieke Boon is at University of Twente, The Netherlands.