Bulk material handling sites face mounting pressure to increase throughput and grow revenues; in other words, get more out of existing equipment. Further complicating this challenge is the fact that these sites must manage fluctuating needs, depending on when different agriculture products, such as wheat, barley, beans, and canola, are harvested in that country.
Improving Bulk Material Handling with the Internet of Things
Growers deliver agri-commodities to the bulk handling site, typically by truck or train. Smartphones or RFID is used to track incoming deliveries at the storage site entrance. With this scanner data, all pertinent information on that delivery is immediately available, such as tonnage, grower, grade, commodity type, and quality. This process of incoming deliveries is known as 'receival' in grain storage operations. Likewise, smartphones or RFID is used to track departures including tonnage, grower, grade, commodity type, and quality. This activity is known as 'outturn.'
The Internet of Things (IoT) is the network of physical objects or "things" embedded with electronics, software, sensors, and network connectivity, which enables these objects to collect and exchange data.
By looking for patterns in receivals, outturns, and inventory at bulk handling sites, we can better understand a grower’s operations, market behavior, and bottlenecks in warehouse efficiency. Using predictive analytics software, the understanding of these patterns can be used to plan site operations, both during harvest season and throughout the year.
Leveraging Time Series Data
We can use historical data scanned from previous deliveries to perform analysis and gain insights. Because we are looking at time series data, when gathering the data, it is important to select sites that have activity data at regular intervals. Missing values are always a challenge in time series data. Some sites will not have frequent receivals and outturns, hence are not suitable. To overcome the issue of missing values, we can aggregate data across sites, commodities, and grades by month.
In time series analysis:
- The series of data is a sequence of observations (generally quantitative) taken at equally spaced time intervals.
- Adjacent observations are dependent.
- Forecasts are based on analysis of the dependence relationships.
- To forecast future values, time series analysis extrapolates observed dependence relationships.
Time series forecasts for future periods are based on analysis of dependence relationships, such as trend and seasonality, among available observations. Time series decomposition breaks observations into three components:
- Trend: The trend in observed data; whether observations have increasing, decreasing, or constant trend over time.
- Seasonality: A cyclical or repetitive pattern in observed data over time.
- Random: The part of the observations that are not explained by trend and seasonality.
Using statistical software, we can derive the separation of data into the three different components: trend, seasonality, and randomness.
Forecasting
We can break out the components of trend, seasonality, and randomness to determine patterns. We can perform this analysis on the following types of historical data:
- receivals
- outturns
- inventory
- movements within a storage site from one bin to another
- movements among different sites
This type of analysis enables sites to measure efficiency of operations and plan for upcoming harvest seasons.
Eka has developed advanced analytics solutions to answer forecasting questions such as:
- What will my future receivals be?
- What will my future outturns be?
- What will my future inventory be?
Eka's Commodity Analytics Cloud platform delivers industry-specific apps that cover P&L, procurement, risk, and supply chain. Unlike generic business intelligence tools, Eka's advanced analytics platform has been built with the specific needs of commodities companies in mind. Learn more in "The Benefits of Analytic Solutions Focused on Commodities."