Interconnected systems for process capability and sustainability in the Industry 4.0
Basically, the goal of any industrial fabrication is to produce and provide highquality products to customers in a resourceefficient manner. The success of any industrial revolution stands or falls on the benchmarks of product quality and production efficiency, and it is therefore helpful to first gain some clarity on benchmarks and measurement methods.
Product quality by means of process capability
Process capability is about the ability of fabrication processes to deliver products in which the defined quality characteristics are within the specified limits with specified statistical certainty. Thus, there are two specifications, compliance with limits and statistical certainty. Given reasonable limits, matching the first is relatively simple. However, the methodology of the second may become arbitrarily complicated.
Process Capability Study
In industrial production, the proof of process capability has become established via the determination of the socalled process process capability index. In many areas, production is already carried out according to the 6sigma norm, i.e. the scatter of the quality characteristics does not exceed the limit values even with a sixfold standard deviation. This would correspond to a maximum rejection rate of 2 in every 1 billion parts produced. However, many fabrication processes are still considered 6sigma processes even for 1.5 sigma is allowed for slippage of the mean  strictly speaking, in that case, statistical confidence is reduced to 4.5 sigma, allowing 7 in 1 million produced parts being out of specification (OOS).
In surface technology, process capability investigations are uncommon. However, this is not because surface engineers are incapable. Rather, it is because the methodology of process capability studies is tailored to mechanical fabrication processes with easily determinable characteristic values, and in the very details, these methods turn out to be largely unusable for the chemical processes of surface technology.
On the other hand, surface technology cannot ignore the demand, reasonable in itself, that the quality characteristics should be guaranteed by its fabrication processes with a certain statistical certainty, e.g. a maximum rejection of 6 out of 100000 parts, i.e. real 4 Sigma?!
Surface technology must therefore develop its own methodology, and as we shall see, align processes accordingly. The challenging areas are:

Process domination

Measurement methodology

Statistical confidence
Process domination
Basically, a capable process must be controllable in the long run.
Controllable means that one can set the process parameters in such a way that the defined quality characteristics are kept within their limits. In surface technology, it is common practice to set process parameters such as temperature, pH, and concentrations, however, the problem lies not in the setting of parameters. The difficulty is rather that the final result depends on the interaction of very many parameters.
In the long run means that the time course of the process parameters must not show any interruptions. In surface technology, however, we have a serious problem with this, because every new batch of cleaning bath, pickling, activation, passivation, etc., represents such a serious interruption in the time course of the relevant process parameters. The problem is not that the process does not work in this way, the problem is that the statistical confidence with which it works cannot be determined, because the position and variation of the underlying distribution functions are always changing.
Consequently, this means that ALL discontinuous processes must be converted into continuous Never Dump processes in order to demonstrate the required process domination. Corresponding schemes for many of the common processes have already been worked out and put into practice in individual cases.
See: Simulations of chemical processes in surface technology
Measurement methodology
In surface technology, the result (for example a corrosion protection) depends on the interaction of quite a lot of parameters. In many cases, it is not possible to measure or analytically record and adjust all influencing variables. Continuously operating Never Dump processes, however, can be relatively easily transferred to a steady state and kept there, so that the relevant parameters hardly change in relation to each other, and the process can therefore be controlled on the basis of sum parameters. These sum parameters must be measured with suitable methods and the measurement data used in an integrated system on the one hand for process control and on the other hand for process capability verification.
Online measurement methods have already been developed for some processes (e.g. ZincOperator) and work is in progress on further methods for other processes or in other problem areas (e.g. methods for electrochemical corrosion protection determination) in surface technology. Sensors for a variety of measurement applications are available on the market that can be combined relatively easily with the ADC/DAC systems.
Statistical confidence
Quality characteristics must be determined somehow, and in practice usually a certain batch of manufactured products is tested for the characteristic. In rather rare cases there is a possibility of 100% testing, i.e. testing of every single product. Only in the case of 100% inspection there is no need to determine the statistical confidence. In all other cases, the statistical confidence must be derived from the variance of the measured values using mathematical/statistical methods.
In many publications it is assumed that the variance of the measured quality characteristics shall obey a normal distribution, so that the statistical confidence can be determined and the process capability can be proven. This is actually not quite correct. It is rather correct that only in the case of normally distributed characteristics, the statistical confidence simply results from the arithmetic mean and the standard deviation. However, it is also true that statistical confidence can be determined for other distributions, as long as the course of the underlying distributions is known.
For example, corrosion protection values and coating thicknesses are certainly not normally distributed, because they have no real upper limit, but they do have a fixed lower limit, namely zero. The distribution of corrosion protection values is therefore usually rightskewed, i.e. to the right of the median there is a very long runout and to the left of the median the runout is comparatively short. With today's computer technology, statistical confidence can nevertheless be determined from a skewed distribution function. It boils down to performing a curve fit of a suitable distribution function to an evolving histogram of the measured characteristics. To determine the probabilities, this function must be integrated within the specified limits.
In summary, the process capability in surface technology that has long been demanded could finally become a reality through the integration of automated measurement methods in interconnected systems.
Production efficiency due to Sustainability
The term sustainability has been used for all sorts of things, and before one can work with it in a meaningful way, one has to limit it strictly logically to the essential core. In the present context, sustainability is to be understood as a global measure of the economic use of limited resources. And in the ideal case of 100% sustainable production, resources will never be exhausted by appropriate working methods. As a rule, a single process or series of processes (e.g., electrogalvanizing of screws) plays only a small role in the global life cycle of all involved products, intermediates, raw materials, water, energy carriers and human resources. However, sustainability is the attribute of the life cycle of everything, and a single process cannot be sustainable as such, but can at best support sustainability to a greater or lesser extent.
So how can you tell if a process supports sustainability more or less, or perhaps not at all, without having to go into the unproductive and thankless field of recording product life cycles and their life cycle assessment?
The surface technology processes must be operated in their defined system boundaries in an open manner in terms of energy and in a quasiclosed manner in terms of materials. The bottom line is that entropy increases as little as possible in the defined system boundaries, or is even reduced through the sensible use of energy. The golden rule here is avoid mixing and use separators and regenerators . This is where chemistry can bring a unique potential that does not exist in mechanical fabrications. Entropy changes of chemical processes can be calculated, provided that the substances involved, the quantities of substances and their changes are known at any time.
With a networked system of automatic measurement methods, as mentioned at the end of the chapter Product quality by means of process capability, integrating the collection, control and regulation of material flows, the entropy differences could be directly determined and optimized against energy use and the use of soft resources (labor, utilization of installations), resulting in the best production efficiency with the best possible contribution to the sustainability of the industry as a whole.
Copyright © Cyclaero Ltda.  rj  20181016 11:17:04
