BLOG

New technologies come to the milling industry

09 March 202010 min reading

“Workplace inefficiencies could be virtually eliminated by smart manufacturing, by allowing the continuous recording of all processes and machines, and implementing automatic actuators that regulate the variables of the processes. For example, grain mixtures can be adjusted, roll pressure increased, sifting time increased, etc. A network of sensors gives you real-time data, as much as you want, as long as each variable can be defined by a numerical variable. With the advancement of this technology, manual adjustments will have to disappear, being replaced by actuators that can control every aspect of the operation of the machines.”

Prof. Gustavo Sosa Industrial Mechanical Engineer Licensed Grain Inspector MBA Project Management SOSA – Engineering Consultants

Things like IoT (Internet of Things) and Smart Manufacturing have been in vogue for a certain time in large process industries like oil refining and car-making, but due to high cost and lack of trained professionals have failed to find their way into the milling industry.

The First Industrial Revolution was about the use of steam and rivers as power sources, and also about the discovery of cheap productions methods for steel. Both innovations compounded and launched the highest jump in the quality of life in the history of mankind.

The Second Industrial Revolution copied the mass production methods out of slaughterhouses and implemented them into automobile manufacturing and every other mass consumption item.

The Third Industrial Revolution started with computers, automating intellectual activities the same way the First and Second did with manual labor.

The Fourth Industrial Revolution is about the internet, massive amounts of information, and connecting every electronic item to each other, creating self-sufficient artificial organisms. Think that Skynet is coming and we are doomed.

Now that high-speed internet is widely available, and every teenager can take a MOOC course from a top foreign university without having to wait for their local technical college, these technologies of the Fourth Industrial Revolution are spreading.

All these innovations have the same common goal: to achieve strong customization under the conditions of mass production. It will be done using: self-optimization, self-configuration, self-diagnosis, cognition and intelligent support of workers.

The design principles that guide Industry 4.0 are: • Interconnection: Machine to machine, people to people, and machine to people. • Transparency: Huge volumes of information (collected from the system) are readily available across the whole system. • Technical assistance: The systems are designed to assist human beings in boring or dangerous tasks, but it is also devised so the human operator becomes a “consultant” of the system, providing orientation to the artificial intelligence. • Decentralization: The systems make decisions on their own. This also means a paradigm shift, because you can’t (from a practical point of view) provide independence to machines and at the same time micro-manage people. Organizations will become flatter and more flexible.

The keys of the future mill are the high adaptability and rapid design changes. The customization of small batches of flours according to grain quality and consumer demands is only possible with a highly automated mill.

A digital twin (software simulation that replicates a real process), fed with real-world variables, allows predicting the results of the real-world process. Most modern mills use a SCADA system to control what the factory is doing, to self-adjust the individual machines according to variations in the product (be it raw material, in process, or finished), and also to manually adjust parameters according to recipes. However, digital twins of process industries go one step ahead by creating what would be a fake SCADA that, after entering the parameters of the real world (even automatically) will tell you what happens at each stage of the process, to each variable, and the final results achieved. This isn’t something to be bought as a package, as it requires extensive programming to teach the twin how his brother behaves in every circumstance, but they are more accessible that you could imagine.

Even though the machines used in milling are fairly simple, if you consider the whole mill as a machine, with the different flows possible considered as robot operations, then the topic of robotics is applicable. It means that if one machine breaks down, the system can divert flows, change capacities, etc., in order to continue operation. This reactive capacity that usually required human intervention now is natural for any computerized system. It is even possible to program an Artificial Intelligence that learns from what happens in the mill and from our reactions to it. This is called Machine Learning. It means using patterns and inference (like a human would do) instead of specific instructions. This technology has been able to achieve incredible things, like develop a new language from scratch. If you use Google Translate, for example, it is based on machine learning. The AI behind it has developed a new language of its own as a bridge between all the languages it translates, just because it “felt easier”. The main focus of machine learning is optimization, according to the goals the user wishes. It is also closely related to Data Mining, which is the use of algorithms to discover patterns in the information provided to the computer. This could be used, for example, to discover the reasons behind quality fluctuations in the products or the causes of machinery failure.

Another breakthrough, that is tied to customization, is using blockchain technology to trace every processing operation, from farm to table. A blockchain is created by the farmer, using certificates provided by reliable authorities, declaring that the grain comes from a specific part of his land and was grown under some specific conditions. For example, it could say the grain is GMO, drought-resistant wheat, grown in parcel 114 of his land, or coordinates such and such, and has certain qualities (weight, density, humidity, etc.). This information is passed to the elevator, which segregates or mixes batches according to qualities and demands, and generates another blockchain, containing the information of the farmers involved, the processing involved for each batch (if it was dried, fumigated, stored for how long, etc.). Lastly, the miller does something similar when manufacturing the flour, generating a blockchain with all the information of his process. In this way, the final consumer could read the blockchain in his 1 kg bag of flour and know all the process it went through, from farm to supermarket. What blockchain will mean, essentially, is that a housewife will be able to scan a code in the package of flour at the supermarket and know its exact history.

This will also require more segregation of grains and flours but will provide us with more information that we ever thought possible. A whole book could be written explaining how blockchain works and how it currently applies and could apply to agriculture and milling.

Workplace inefficiencies could be virtually eliminated by smart manufacturing, by allowing the continuous recording of all processes and machines, and implementing automatic actuators that regulate the variables of the processes. For example, grain mixtures can be adjusted, roll pressure increased, sifting time increased, etc. A network of sensors gives you real-time data, as much as you want, as long as each variable can be defined by a numerical variable. With the advancement of this technology, manual adjustments will have to disappear, being replaced by actuators that can control every aspect of the operation of the machines. Forget about the old supervisor that knows the machine by the sound it makes, but be prepared to pay big bucks to your supplier’s field service technician.

From a human resources point of view, this will make the low skilled workers redundant, but will also require a lot of highly skilled workers, with expertise in mechatronics, that simply haven’t been trained yet. Our technical schools will be more important than ever because they will have to educate the legions of young men and women on skills that almost didn’t exist 40 years ago. It looks a bit scary, but the automobile launched a similar revolution at the beginning of the 20th century, and yet everyone managed to get his car fixed.

Programming is easy. You can learn it by yourself, at home, for free. But mechatronics, the fusion of mechanics, electronics, and programming at a small scale, requires specialized and expensive equipment. It demands learning by trial and error using materials that are expensive, and not just code. It is up to trade unions and business unions to promote new educational programs that will prepare the workers of the future. And by future, I mean next year, or maybe six months.

Big data is, basically, knowing as much as possible about everything, all the time, and creating models with that information. It is in vogue in everything related to social networks, dealing with the tracking of human behavior, in order to sell us more junk, but it can also be applied to grains. Imagine that you could track every water drop in a rain, you could discover patterns in it, and find out how the wind affects them, and the sun, and the nearby buildings. Of course, it is impossible and that is why engineers like myself deal only with approximates, and never on exact values. The uncertainties, compounded, lead to a huge margin of error. However, we do track (for example) batches of grain, and (as technology advances) we may track smaller and smaller packages. Then we will be able to find out which silo had the water leaks that got the grain rotten, which supervisor was sleeping and dried the grain too much, how long the truck takes to arrive in Porto Alegre, and much more. A consequence of this is that we will be able to implement Just In Time (JIT – Toyota) in many points of the grain value chain. If a miller can trust the numbers of the grain delivery time, he can buy only what’s needed at a specific time, and not 50.000 tons for the whole year. The money spent on inventory will fall dramatically, making millers more financially sound. The same applies to the output. Major distributors, supermarkets, importers, will be able to handle their supply chain better, resulting in savings for everyone.

Big data systems tend to be huge, not something that you can implement in an old desktop computer. This demand in resources has encouraged the growth of another characteristic of Industry 4.0, which is Cloud Computing. It means that, instead of processing the information in a local computer, to service the local processes, you do it on specialized servers that connect to all your machinery and other systems. Being Industry 4.0 doesn’t mean that you rent foreign servers. You could use what is called Edge Computing and achieve the same results. However, in any case, they have to be connected to the internet, and not just the local Profibus network. All your systems have to be connected (maybe not directly) to your servers through the internet. That is what makes them “cloud”. Generally, they really are rented servers, because the specialized data centers tend to be cheaper and more efficient than your own, but that is not mandatory.

Summarizing, Industry 4.0 is not a new set of technologies, but instead a new philosophy for the integration of technologies that have been around for decades. It means unlocking the whole potential of computers, instead of using them as humans. Have you ever seen someone use a pocket calculator to make the calculations in an Excel file instead of just using the Excel functions, using the spreadsheet only to make numbers look pretty? Well, for decades we have been that guy. Now we are finally committing to taking an Excel course.

Articles in Cover Story Category
25 January 201912 min reading

Change in Climate Variables and Impacts on Agriculture

“There is substantial evidence that trends in change in climate variables can negatively or positiv...

18 February 20209 min reading

Fit for the digital laboratory with a ‘Smart Workflow’

The precise measurement of ingredients and rheological parameters is a key prerequisite for quality...