Skip to content
September 30, 2010 / or4green

Entropy and Green O.R.: Interview with Dr. Stefan Gößling-Reisemann

When optimizing with sustainability in mind, typical objectives include minimizing energy usage, material consumption and waste generation. There is another perspective, one that puts matter and energy on the same scale and recognizes the fact that from a practical point of view, while energy and matter are unlimited, their potential utility is not. The quantity associated with this perspective is entropy, and it is a key concept in the work of Stefan Gößling-Reisemann, assistant professor at the University of Bremen, Faculty of Production Engineering, in Bremen, Germany. I mentioned Dr. Bremen in an earlier post. He was kind enough to answer several questions I sent him.

greenOR> Life cycle assessment (LCA) tools typically measure the flows and consumption of matter and energy. But you argue that this is not a complete accounting since, for instance, the utility of a material can change without its energy changing or it being consumed. You use entropy production to approximate the loss of potential utility. It supplies a “common measure” across different types of consumption including “chemical reactions, physical transformations, heat transfer, and so on.”

S.G.-R.> LCA tools only measure the flows of matter and energy (not the consumption, i.e. they don’t tell you anything about their transformation). Of course, LCA really focuses on the environmental effects caused by theses flows. Consumption is not really an environmental effect, it is rather economic in nature.

The utility of a material changes exactly WHEN it is consumed, at least that is my understanding of consumption. When the utility changes, consumption takes place. This can be happening for a variety of reasons, physical or chemical transformation and dispersion is one of them, and a very common one. That is why I think entropy production is a good proxy measure for consumption.

greenOR> For readers who may not be familiar with it, how would you define thermodynamic entropy?

S.G.-R.> Microscopic disorder is a good first approximation to entropy. Unavailability of energy is also a valid, but far more abstract and rather indirect way of defining entropy. Entropy grows when thermodynamic gradients are dissipated, i.e. leveled out: diffusion, dissipation, mixing, heat flows across temperature gradients, transformation of chemical into heat energy, and so on.

greenOR> Why do we need to go beyond matter and energy consumption to include entropy production in life cycle assessment? Is it primarily to put the loss of utility of energy and matter on the same measurement scale, or are there transformations that lead to a decrease in utility that are not accounted for by matter and energy measures?

S.G.-R.> Putting energy and material utility on the same scale is important for a comprehensive analysis of consumption. Plus: in principle matter and energy are available without limit (conservation of energy law). What is limited, though, is there potential utility. If we want to address the finite nature of resources, we have to take this into account. As long as we “just” measure energy and matter in- and outflows (which is hard enough, believe me), we do not capture any decrease in utility. Another point is, that the “entropy disposal rate” of earth is also limited, and it mainly depends on the temperature of earth’s surface and atmosphere. This “garbage collection” has a pretty high capacity, but for all practical consideration it is fixed nevertheless. This gives us an absolute scale to compare our resource consumption to, something we cannot get from looking at material and energy flows alone.

greenOR> In creating a product from its raw materials, the potential utility of those materials diminishes. But the utility of a product is generated, where before there was none. One can imagine two processes generating the same entropy but creating two products of differing utility. Have you looked at utility from the product side, rather than the raw material side? This would seem to be much more difficult and subjective to measure.

S.G.-R.> Good question! The potential utility of the product can only be a subset of the potential utility of the raw materials (including the energy needed to produce it). But remember that I am only using “potential utility” defined as “the size of the set of possible products/services derivable from the materials at hand”. This is a very abstract measure and I do not assume that is measurable by any means. I am only interested in the differences between potential utility of raw materials, products and wastes. These differences can be approximated by entropy production. The economic concept of utility, as far as I understand, measures something very different. Although maybe it doesn’t. 😉

greenOR> Has your approach been taken up by any on the industry side? If not, what would be a likely scenario of application?

S.G.-R.>There are two optimization approaches using entropy or exergy, respectively, which have been quite successful in industry, I believe. One is called Entropy Generation Minimization (developed by A. Bejan) and the other is Exergy Analysis (Szargut, Kotas, Ayres, Finnveden, and others). My PhD thesis was done in collaboration with a major European copper smelter. However, to my knowledge the results were not used within the company for any decisions. They liked the analysis though, and used the flow sheets for educational purposes.

greenOR>> What do you plan to work on next? It would be interesting to see your approach applied to food.

S.G.-R.>Currently I am focusing on applying entropy analysis to (metallurgical) recycling processes. Progress is slow, mainly because entropy analysis requires a huge amount of data. An application to the food sector would certainly be interesting, but I fear the data problem would even grow exponentially: there is hardly any source on thermodynamic properties of biological materials and the production processes, from farm to fork, are really complex and hard to analyze on a thermodynamic basis.

%d bloggers like this: