|Automation can significantly influence|
how a plant interfaces with the environment
Most of the industrial facilities I visit these days are automated. That automation largely came about due to a need to control processes better, and/or to remove manual operations and hence reduce the number of people required to operate individual process plants and production lines. We could easily start another post on the social merits of reducing headcounts through automation. Proponents will often argue that automation is done for both reasons, since machines are “more reliable than people”. Certainly automation allows for higher throughputs, but it must also be said that faulty field instruments can lead to dire process problems and large economic losses. Problems need to be detected and dealt with quickly to limit the impacts of mishaps, be these related to humans or machines.
It is however undeniable that if designed, commissioned and maintained correctly, automation (and here I mean hardware, software and field instruments) can deliver superior process consistency, which can lead to improved product quality, safer operations, increased throughput and lower operating costs. Automation cannot always do this on its own, since it has to work hand in hand with a well-designed and capable process. A poorly designed process plant can be improved by (but will not operate perfectly through) sound automation. Poor automation can however make a well-designed process plant operate well below design levels of capability.
Many (if not all) of the benefits of automation directly support sustainability. It is also possible, through the use of the appropriate software and databases, to extract performance data and then use it to produce powerful reports. This is exciting from a sustainability perspective, as we can use this information to drill down to the unit operation level to assess performance in terms of energy consumption, water use, emissions, process yields and the like. The data can be aggregated and fed into organisation-level sustainability reports. These capabilities can be extremely useful for linking strategic sustainability objectives to their process-level drivers
What I’d like to do in this post is to discuss the topic of automation from a plant perspective and hopefully demonstrate how automation can contribute to improved sustainability performance if properly employed. I will do this through a series of mini-case-studies, all hypothetical, but all of which would definitely not be out of place in the real world.
CASE STUDY 1: ELECTROPLATING PLANT DRAGOUT LOSSES
Electroplating, which is the deposition of metals onto a surface by reducing metal ions to solid metal using an electrical current, is carried out in a number of process baths, each containing different chemical solutions. The process baths may be used to prepare items before plating (e.g. de-greasing and etching), to carry out the plating operation (this could be in a number of steps and involve various different metals) and to remove all traces of chemicals after plating, through rinse steps, some further treatments etc. “ Dragout” is a term used in the plating industry to refer to the adherence of liquids to items that are being electroplated as these items are removed from process baths. The idea is to minimise the amount of adhering liquid (dragout), since this in itself represents a financial loss, but also because these chemicals can contaminate subsequent process baths, and negatively impact on their effectiveness.
There are a number of well-documented ways to reduce dragout losses. The best way (since it deals with the problem at the source) is to allow the liquid to drip back into the bath before the workpieces are moved to the next bath. This could be complemented with the use of water sprays to assist in the removal of adhering liquid. The workpieces could also be mechanically manipulated (shaken) to promote liquid removal. This can be particularly useful for workpieces with complex designs, and which have recesses that can withhold liquid. In small plating plants, all of these things can be controlled by an operator, who would physically hold the workpieces, using a jig or some other means. In large automated plants, the plating plant could be controlled by an automatic “robot” system, with motors (which could be fitted with VSD’s) used to lift and move the jigs on which the workpieces are mounted. The whole operation would be controlled by a PLC, and there would be setpoints that determine drip times, bath temperatures, process time in each bath, the rate at which jigs are lifted and virtually every parameter that can be physically controlled by equipment in the field. Provided the control system is designed to be flexible enough, the level of control can extend to an identification of the specific workpiece characteristics on each jig, which could then be correlated to drip times and process steps that are optimised to limit dragout losses for each individual workpiece type. All of this could be managed within the context of the maximisation of plant throughput, and since drip times drive capacity, different drip times could even be employed on the basis of the current level of plant utilisation. The data could then be extracted and analysed in conjunction with plating chemical usage rates for individual process baths, in order to allow for continuous improvement of process yields based on live feedback.
CASE STUDY 2: CONTROL OF A SCREW COMPRESSOR
Oil-filled rotary screw compressors are very common in industry, particularly those operating on a “load-unload” philosophy. In this control philosophy, the compressor operates between two pressure set-points, going onto load when a low-pressure is reached in the compressor discharge, and going off-load when a high pressure is reached. This is not a bad thing where the compressor concerned is running at close to design capacity. However, where such compressors are oversized, they spend a lot of time running off-load, often consuming up to 40% or more of the energy they would consume when on-load, but this time without producing compressed air. In this scenario, fitting the compressor with a variable speed drive and controlling the discharge pressure to a fixed setpoint is a more efficient mode of operation. The compressor’s air delivery and power consumption is linearly related to its speed, and by controlling the speed of the compressor (above some pre-determined minimum to protect the motor), the input of power will always be accompanied by the production of air. For very low-demand periods, the compressor can still go into an “off-load” mode, this time running at reduced speed, and hence drawing less power when doing so. All of this automation would be in a PLC, typically dedicated to the compressor, but able to communicate with other PLC’s if required.
CASE STUDY 3: PLANT-WIDE DEMAND CONTROLLER
Most manufacturing plants pay for electricity based on energy consumption and maximum demand, with the added complication of time-of-use tariffs in some cases. Energy consumption is measured in kWh, and this is charged for at different rates depending on when the energy is used for sites that pay time-of-use tariffs. Clearly, carrying out energy-intensive operations during periods of the day when rates are lowest reduces costs, though of course this does not necessarily reduce emissions. Demand charges are typically based on the highest peak (in kVA) achieved in a month, regardless of when that peak occurs. The peak generally does occur during periods of overall peak demand on a utility’s power supply, the point being that increased energy consumption during off-peak periods can result in access to cheaper (from a financial perspective) energy and reduce maximum demand at the same time.
Copyright ©2013, Craig van Wyk, all rights reserved