Big data analytics set to power real time efficiency for sugar millers

The research team at the Sugar Milling Research Institute NPC, under the leadership of Research and Development Manager Steve Davis, and group leaders, Dr Kitty Foxon and Sarah Weyer, is beginning a journey to bring the power of big data analytics and artificial intelligence into sugar factories and already the possibilities are being eagerly embraced by the sugarcane industry.

If mention at the recent congress of the South African Sugar Technologists’ Association is an indicator of trends in sugar technology, then Industry 4.0, or the 4th Industrial revolution is clearly an important emerging topic, particularly in the improvement of performance by South Africa’s sugar milling sector.

At the 2017 Congress, a single commercial paper on the internet of things (IoT) was presented in the agricultural sessions, and in 2018, there were two research papers and one commercial paper in the factory sessions.

Keynote speaker Bronwyn Williams spoke on The business of disruption: the ripple effect of game changing technologies, and her presentation was populated almost entirely with examples of applications of the various pillars of Industry 4.0.

The Sugar Milling Research Institute NPC also had a poster at their stand promoting Sugar Factory 4.0.

But what is Industry 4.0?

Despite the enormous hype around the topic, there is no one definition for Industry 4.0, because it is not one thing. Rather, it is a parcel of activities, technologies and services that make use of the explosion in ready accessibility of some concepts that have been around for a long time. These include robotics, automation, artificial intelligence and 3-D printing.

What has allowed these disparate sciences and technologies to become a phenomenon is advances in connectivity and data transfer capabilities. Some clever people have realised that they can get immensely more efficiency and productivity out of processes and services by accessing and processing data as it is generated. For example, a few years ago, your GPS would map out a route from one point to another and calculate the journey time from a static algorithm. Now it scans the locality, differentiates between congestion and accidents in real time, and suggests alternative routes based on current traffic flows.

Why should sugar factories be interested?

Sugar factories have a lot of data; at any moment there are thousands of individual measurements of temperature, pressure, flow, brix, pH, conductivity, tank level, valve position and many other parameters being taken. Much of the data is used in local control loops that maintain conditions near the desired set point for specific sections of the factory. However, deciding on the optimum values of the set points often requires consideration of the entire factory operation and anticipation of future changes in cane quantity and quality, for example. The sheer complexity of the problem makes it very difficult for an individual to make the optimum choices, but the financial implications of making poor choices can often be great.

A small fraction of the available data is carefully composited and aggregated to calculate shift, daily, weekly and annual performance metrics that are used to measure efficiency, identify problems and motivate for investments and maintenance.

However, the scrutiny of the aggregated performance information takes place after the event. Remedies can be applied to persistent problems once they have been identified, but many problems have already caused efficiency or revenue losses before they have been identified.

What could this mean for the sugar industry?

The Industry 4.0 perspective challenges us to consider whether there may be potential for significant improvements in overall performance, if some of the masses of data could be processed in real time to automatically determine current performance efficiency and projected efficiency based on current conditions.

The same question can be posed for the entire sugarcane value chain: if there were a system that could assess the location and state of ready-to-crush cane, cane harvesting and transport resources, factory capacity, product distribution resources and market quality and quantity requirements, could rapid feedback loops be created to support cane supply and processing decisions to maximise the revenue fed back through the entire value chain?

The types of projects that might arise to tackle the factory and value chain questions posed here require data to be collected, cleaned and processed in near real-time, and information generated to be fed back quickly to the people and control systems that can make efficiency-enhancing changes. Data collection, transfer and processing capabilities need to be developed and installed to handle multiple data types across varying time-frames. For most Industry 4.0 projects of this type, the “brains” behind the data processing are big data processing algorithms that determine the complex relationships between measurements and associated efficiency metrics.

The use of big data analytics to target processing efficiency using existing data is the low hanging fruit for sugar factory 4.0. There is a bigger field of technologies and applications that could be harnessed, ranging from image recognition to robotics and automation. The possibilities are limited only by our imaginations.

RELATED ARTICLES: