Kagome, Campbell: IoT for... Internet of Tomatoes
- François-Xavier Branthôme
- Read in french
Tomato transparency and traceability
During tomato harvest time – February to April – in the state of Victoria, Australia’s premier tomato processor, Kagome, processes around 180 tonnes of tomatoes every hour, 24 hours a day, for 70 days straight (about 300 000 tonnes in total, editor’s note). More than 4 000 tonnes of tomatoes are processed each day and, by the end of the season, 40 000 tonnes of tomato paste and diced tomato products have been made. The plant supplies bulk diced tomato to companies like Simplot, as well as pastes and other bulk products to Yum! Brands (owners of Pizza Hut, KFC and Taco Bell).
Kagome-owned farms grow about half of the tomatoes, with contracted growers supplying the rest. But all harvesting is done by Kagome, using its own harvesting fleet, which with a few modifications can collect tomato crops at 150 tonnes per hour, as well as carrots in winter. Today, cultivating, harvesting and processing tomatoes is a matter of automation and it can be a logistic challenge to get the tomatoes from the fields to the factory in the most efficient and controlled way.
Quality control covers Kagome’s entire process, from tomato seed management and growing crops through to in-store displays. Looking for an automated identification solution for its Echuca fields, Kagome operates twelve harvesters loading tomatoes into more than three hundred huge bins, each with a capacity of fourteen tonnes. Once a bin is full of fresh tomatoes, it is unloaded at a bin pad waiting for one of twelve trucks to pick up and return to the factory. One trip from the fields to the Kagome factory takes approximately ninety minutes and each truck can load three bins – that is an average of around forty-two tonnes of tomatoes per truck. A truck arrives at the bin pad, the bins are then loaded and taken to the weighbridge close to the factory.
“…a 500% return on investment in the first season alone.”
Three years ago, there used to be long truck queues at the weighbridge and the truck drivers had to wait for 12 minutes until they could get out of the truck to have the tomatoes weighed. As part of Kagome’s quality control process, three samples from each bin had to be processed in the company laboratory because it was not obvious which tomatoes came from a Kagome farm. In addition, drivers had to prepare paper-work to document the harvesting process as well as the quantity and quality of the yield. The process increased the potential for human error in a paper-based quality control system that can result in contaminated products reaching the consumer, conceivably creating widespread foodborne illness.
It was a mammoth task to coordinate harvesting, hauling, grading and unloading operations through 500-odd phone calls each day, as well as keeping paper-records. So to ensure traceability, it was time for Kagome Australia to look for a paperless automated identification solution to be implemented at the weighbridge. Kagome wanted to know exactly where the tomatoes came from, when they were picked, which truck delivered them, when they arrived at the plant, what processing conditions were like, when they were packed and where they went … In other words, they wanted transparency from vine to dine.
Kagome approached Microsoft’s partner, Advance Computing, to digitally transform the entire process. Advance Computing devised an Internet of Things (IoT) solution that incorporated data from on-farm sensors, in-truck devices and technology installed in Kagome’s loading bay to ensure that the company could have a clear window on its operations. The system is based on Windows 10 IoT devices and a range of Microsoft Azure cloud technologies, reporting and analytics tools.
“The tomato market is highly commoditised and it’s very hard to take costs out,” Kagome field operations general manager Nick Raleigh explained. Kagome’s CEO, Jason Fritsch, says that the Azure-based solution paid for itself five times over in its first season.
How it all works
Using the IoT solution and the range of Microsoft Azure cloud technologies, Kagome is able to collect RFID information about its raw product at each step of its journey from the farm to the factory. All harvesters, tractors and collection bins have been fitted with RFID tags and GPS technology. Load-measuring cells on the trailers that hold the bins as they are loaded with produce give farmers visibility of the weight of the produce as it is being harvested. This enables operators to track the actions in the field, and to record the amount of produce collected.
The GPS-located IoT system means yield can be tracked against paddock boundaries and visualised in heat mapping of fields, which will be collected over time to show productivity rates and yield patterns. Kagome is working on using this information to help forecast production and plan for potential yield ahead of harvest.
A web app enables the Kagome logistics team and drivers to move the perishable product on a first-in, first-out schedule while maintaining an efficient, steady flow into the plant. “The logistics team (at the plant) can see the information on each bins in the paddock and advise drivers which ones to pick up first, so they aren’t sitting in paddock too long,” Advance Computing director Bryant Alford said. “They can also manage how much product is on farm and adjust the speed of harvest to suit how much is already in the factory.”
One hundred B-Double trucks deliver the tomatoes to the factory, where a weighbridge marries up the incoming bins with an accurate weight measure for the incoming delivery. When a bin crosses the weighbridge at the plant, and is tipped into the processing line, an IoT device reads when the bin was harvested and the GPS location it came from. The bins are then allocated to a product line, where the contents are processed into either tomato paste or diced tomatoes. Processed tomatoes are then fed into “pack-off” bins and barrels, which are also all RFID tagged, and then shipped to end users across the country and the globe.
Streamlining harvesting processes with RFID
RFID technology (radio frequency identification) is increasingly found in food tracing as technology improves and prices come down. Implementation is not uncommon in the case of large containers containing raw products and in the mixing of bulk materials. It offers companies a number of ways to streamline and manage their capacities, focusing particularly on the issues of traceability and process reliability.
The RFU63x read/write unit from SICK Australia was presented to Kagome in 2012. In January 2013, the company installed six RFU63x units, each equipped with three antennas for double stacked bins, at the weighbridge and discharge hill at the factory in Echuca. RFID tags were attached to the tomato bins, accompanying them right from the start of the harvesting process. The device entirely met the processor’s requirements set out for the paperless automated identification of tomatoes, helping prevent the typical errors made during inbound and outbound goods processes, for example incorrect quantity and quality data, or missing accounting entries.
RFID allows real time identification of where the tomatoes come from and helps save time at the weighbridge. Indeed, truck time at the weighbridge has been reduced from twelve minutes previously to two minutes, allowing the truck driver to take an extra trip per 12-hour shift. With a fleet of twelve trucks and one truck loading an average of 42 tonnes of tomatoes, this means a productivity gain of 504 total tonnes. Thanks to the increase in reliable real-time data made available by intelligent identification technology from SICK, Kagome gained the possibility of making better decisions, thereby increasing productivity and efficiency, while reducing costs.
Company executives, farm owners, factory staff, end customers and food safety authorities can all access data held in the Azure cloud platform via PC or web apps accessible on mobile devices.
The full supply chain visibility afforded by the IoT/Azure system enables Kagome to be confident in the provenance of its product. This level of traceability means Kagome can tell the whole history of any packet of its tomato paste – down to where and when the tomatoes were harvested.
IoT: How Campbell uses data to drive sustainability
The Campbell Soup Co., one of the longest-operating food brands in the U.S., maintains data on every load of tomatoes that comes into its processing plants, and has been collecting this information over the past five years. But the application of data science to farming and produce processing is fairly new, and Campbell is working on a project involving tomato-growing suppliers to channel its trove of valuable information into a big payoff: sustainability.
“It’s driving progress on the farm and creating systems that everyone can use more easily. That’s how we drive sustainability,” said Dan Sonke, director of sustainable agriculture at Campbell Soup.
Finding relevant data was not the issue. Campbell had access to five years of public and private information, including satellite imagery, soil conditions, weather station reports and river flows, along with the company’s own private internal metrics. The challenge for both Athena and Campbell was to create the proper technology that would clean-up and normalize datasets. “As Campbell Soup was collecting all of the data, the entire industry had never really been familiar with the structure of how to actually use data in any kind of meaningful, data science or analytical way,” Sypnieski explained.
Campbell decided to focus on tomato solids, as the sugar content of raw material conditions industrial yields and energy expenditure. The Campbell executive said that the company is trying to “tease out” what soils produce higher solids or irrigation practices that raise sugar levels through the data analytic tools provided by Athena. “This is extending the reach of the data outside the four walls of our factory into the farm,” Sonke concluded.
Watch the complete video interview below:
Sources: foodprocessing.com, Kagome.au, sickinsight-online.com, news.microsoft.com, siliconangle.com/blog