What do we need to change to improve the industry’s ability to manage materials corrosion and to reduce the cost for it?
In March 2016 NACE International released the International Measures of Prevention, Application and Economics of Corrosion Technology (IMPACT) study, in which the global direct cost of corrosion was estimated to be US$2.5 trillion, equivalent to roughly 3.4 percent of the global Gross Domestic Product (GDP). Statistically, the cost of corrosion/GDP has stayed more or less constant (1-5 percent) for almost half a century based on many previous studies with the earliest studies dated back to the mid 1950-ies. Since IMF predicts an average GDP growth of 5 percent for the next five years for the emerging market and developing economies, the world’s cost for corrosion is likely to grow and new solutions to the corrosion problems are therefore highly at demand. Hence, there is a need for a step change in the way materials corrosion problems are managed and solved.
Corrosion, at its core, is the exchange and reaction of electrons, atoms, or molecules between a material and its environment and hence, it is highly predisposed to a mechanistic/analytical modelling approach. Yet, corrosion, despite its ubiquity and long history of study, as stated, essentially remains an unsolved problem that continue to cost the industrial and agrarian industries dearly. Corrosion bridges a number of topics; metallurgy, solid-state physics and physical, inorganic and organic chemistry which makes any sort of mechanistic/analytical modelling fiercely challenging and the resulting models not always sufficient, very practical and/or useful for the intended users of and responsible for the management of corrosion or for implementation in corrosion management strategies.
So, are there other possible and unexplored solution approaches available?
People have begun to realize that most of the world’s work is non-mathematical in nature. Only a small segment of activity has its kernel in the formulas used in mathematical based sciences. Even in sciences like chemistry and particularly in corrosion, most thinking is done by inference, not calculation. The same is true of biology, most of medicine and all of law. Most business management is done by symbolic inference, not calculation. In short, almost all the thinking that professionals do is done by reasoning, not calculating. As computing gets cheaper and professionals look to computer and data scientists to relieve their ever-growing information processing burden they will want to use methods that involve automated reasoning and symbolic knowledge. These methods, also called expert systems, have now demonstrated that a computer is capable of the same kind of intelligent behaviour as a physician making diagnosis, or a geologist deciding where to seek minerals. These expert systems do this in much the same way human experts do – by combining textbook knowledge, or factual knowledge with experimental knowledge and associated mechanistic/analytical modelling.
The human experts have acquired their expertise not only from factual knowledge (textbooks, lectures, etc.) but from experience, by doing things again and again, failing, succeeding, wasting time and effort, by getting a feel for a problem and learning when to go by the book and when to break the rules. They build up a repertory of working rules of thumb, also called ''heuristics'', that combined with the book knowledge, make them experts in solving complex problems.
The competitiveness of businesses in the market will depend on their intelligent utilisation of expert knowledge both now and in the future based on the awareness that the technology of artificial intelligence (AI) and knowledge engineering is now sufficiently matured and can critically contribute to competitiveness and complex problem solving.
Corrosion knowledge engineering
Corrosion knowledge engineering integrates corrosion knowledge into computer systems in order to solve complex corrosion problems normally requiring a high level of human corrosion expertise. Knowledge engineers and data scientists are employed to translate the information elicited from the corrosion expert into terms which cannot be easily communicated by the often highly technological corrosion expert. The knowledge engineers/data scientists interpret and organize the corrosion information on how to make corrosion management system decisions with the purpose of the job to work with a client who wants an expert system created for them or their business. The knowledge engineers/data scientists are involved with validation and verification and are required to carry out data collection and data entry, but they must use validation in order to ensure that the data they collect, and then enter into their systems, fall within the accepted boundaries of the application collecting the data. It is important that the knowledge engineer/data scientists incorporate validation procedures into the systems within the program code. After the knowledge-based corrosion management system is constructed, it may be maintained by the responsible corrosion expert.
Will the future corrosion management strategies include expert systems created for the industries and their businesses?
In summary, a step-change in the corrosion risk management strategies is needed and this may be achieved by implementing AI and machine learning techniques through the involvement of knowledge engineers and data/computer scientists working side-by-side with the corrosion engineers. However, these are fairly novel techniques and methods depending on which part of the industry that is considered and all relevant stake-holders should get together and iron out the way forward e.g. by conducting feasibility studies, building prototypes and implementation of international development projects. This should also be an excellent topic on corrosion prevention conference agendas and in corrosion societies technical committee work.