Automating the smart tasks

“A good automation project should detail all the machines, the sensors, the actuators, and the flows involved. Whoever you choose as your supplier, make sure the documentation before the delivery of the items is as exhaustive as possible. The documents should detail what is to be done and how. Nobody will give you engineering information for free, but in the contract you could specify a time after the down payment, to supply drawings, manuals and plans, before any item is physically delivered. That way, you will be able to spot any problems beforehand and demand changes.”

Prof. Gustavo Sosa
Industrial Mechanical Engineer
Licensed Grain Inspector
MBA Project Management
SOSA – Engineering Consultants
ing.gustavo.sosa@gmail.com
Generally, automation is considered something you do to simple tasks, to the most routine and stupid tasks you could find. Makes sense to find those things that are repetitive and simple, and install some gadget to accomplish them.

However, it is essential to remember that we deal with human factors. We get tired, our minds wander, we make silly mistakes. Even making decisions, the most basic human task, can be difficult. Some people will avoid responsibility at all costs.

There we find a market for automation. The whole operation of a grain mill can be automated if we take the time to design the process that each grain should go through. In our central system, we have the flow sheets for each grain and for each recipe. They might be hundreds, but once they are in the system, it is just a matter of choosing one. The critical human intervention here becomes the closing of the control loop, by entering the parameters of the raw material and those of the final product. All of which are obtained by lab analysis.

If we did an Ishikawa analysis of the grain industry, we would see that most of the problems in production are a consequence of human factors. The root cause might be either lack of training or a complete absence of methods. Or it could be the poor quality of the available workers, but analyzing that and its solutions would take us down a completely different route.

Automation is good because it almost erases the need for training and it also forces us to develop methods. You can’t automate something that doesn’t have a method. And you don’t need to train something that is just plugged in your system. Think of the time it took to train a human bank teller. And not only that, but also the time it took him to be as proficient as his colleagues. My wife used to work as a teller at a Casino, before becoming a school teacher, and it is always amazing to watch her count money. It is like a superpower. Work at a mill might not be so fast-paced as paying prizes at a Casino, and tasks are rougher in nature, but the difference in abilities between a novice and a veteran should be astounding too. But if the human work becomes simply entering the values from the lab into the recipe of the SCADA system, then there is no difference. You won’t need a veteran who knows exactly how many turns he needs on that knob, or how much to pull on that lever.

This is automating the smart, instead of automating the dumb. This where real breakthroughs are achieved.

It is not that you won’t need highly trained people. In fact, you will need much better people, but not for the critical routine activities. You will need better electricians, better mechanics, and at least one programmer, to maintain those systems. But they don’t even need to be on payroll. If you are near a large city, you could outsource maintenance to a service company, and pay them by the hour, according to some maintenance plan to be agreed on.

Even if you have to put the technicians on payroll, consider that the higher qualified workers are less likely to unionize. These are people who are confident they can just quit and find another job if they don’t like something, instead of going on strike. And Union problems in a small to medium sized company, in a third world country, may mean sure bankruptcy.

As I said in a previous article, you should also consider the quality of the technicians available in your area. If you own a flour mill in a remote area of Brazil, then you should either offer people very attractive salaries to more to your town, or completely forget about automation. I have a client who owns a large facility at just two hours from Montevideo (the country capital of Uruguay) and he can’t find a PLC programmer who will get there in less than one week. Also, the more components a system has, the more prone it is to fail.
Automation also means the process is repeatable and avoids subjective views. Grain sampling and truck weighing, left completely in the hands of human operators, invite corruption. Of course, these sensitive areas should be audited regularly, but it is better to leave little margin for illegal practices.

In consequence, you save a lot of money on labor by automating your processes.

However, you also improve the quality of the work. The new in-line sensors provided by major manufacturers can evaluate critical parameters in real time and feed that information to the SCADA, allowing the system to adjust automatically to those changes. That is something that couldn’t be done before, because the time required to take a sample, analyze in the lab, and then enter the data in the computer, meant hundreds or even thousands of kilograms processed out of specifications. It was better than letting the production run wild, but it was not the best solution either.

Special care must be taken to choose the milling equipment that includes actuators to control the gaps and distances that in turn control how grain is processed. An actuator is anything that transform a signal (in general, electrical or pneumatic) into a physical action. For example, a valve that controls how much fuel is burned by the dryer. Or a hydraulic cylinder or a screw that controls the pressure exerted by milling rolls. Retrofitting an old system with actuators is generally possible, specially when they are controlled by screw and levers, but not always.

It is pointless to invest money and time in an expensive SCADA system if later you use it only as a simulator and change all the variables manually on the plant.

A good automation project should detail all the machines, the sensors, the actuators, and the flows involved. Whoever you choose as your supplier, make sure the documentation before the delivery of the items is as exhaustive as possible. The documents should detail what is to be done and how. Nobody will give you engineering information for free, but in the contract you could specify a time (like two months) after the down payment, to supply drawings, manuals and plans, before any item is physically delivered. That way, you (or some consultant you hire) will be able to spot any problems beforehand and demand changes.

You have to consider the pros and cons according to your specific situation. But, if you are reading this article, chances are your facility is mid to large. It is very unlikely a small miller would take time to read about the latest developments in industry. So, it is very likely that you will benefit a lot from automation. Try and send an email to one of the many advertisers in our magazine. I am sure one of them will be able to help you.

Check Also

SUSTAINABLE FOOD SAFETY AND FOOD SUPPLY CHAIN

What makes food safety unique and different from other areas of public health is its …