Configuring Your AutoStore System: 5 Design Parameters
AutoStore is an automated, high density warehouse storage system designed and manufactured in Norway. The system is modular and uses standard components, which makes its configurability nearly endless. It can fit around building columns, fill L-shaped spaces, tunnel over walkways, and deliver to remote access ports across the facility. However, optimizing the configuration for efficiency, speed, and cost-effectiveness is an engineering challenge, but a challenge that Bastian Solutions has successfully surmounted well over a dozen times in North America.
AutoStore is made up of structural bins placed in an aluminum grid. The robots work across the top of the grid to dig bins out of the matrix and deliver them to ports for product processing. The ports are typically along the periphery of the AutoStore matrix, but it is also possible to place them above, below, or inside a tunnel within the AutoStore. The robots are connected wirelessly, and the AutoStore software will prioritize their activities to your order requirements, even assigning multiple robots to a single order to increase speed if necessary.
There are multiple parameters for configuring AutoStore that interact in give-and-take relationships. We run computer simulations using the actual AutoStore software in a virtual warehouse to optimize the layout and configuration prior to presenting final designs. Considering how easily scalable AutoStore is, it’s not uncommon to run these simulations again for growing businesses as they expand their warehousing operations. A sampling of the parameters to consider are the following: 1. Grid Height The most typical application for AutoStore maximizes cube space within a warehouse, which means the grid will be built to its maximum height of 25 feet. This correlates to 16 bins tall using AutoStore’s “standard” bins (most common) or 24 bins tall using the “short” bins. Sometimes there is a speed benefit to using a lower system height (i.e., short and fat), which will allow the robots to do less digging. However, clever slotting (how you arrange multiple SKUs in a single bin) and AutoStore’s “forecast” mode (discussed later on) can overcome this need. 2. Footprint As an AutoStore grows taller, it can fit the same number of bins into a smaller footprint. The smaller footprint will also allow less travel time from a stack of bins to a port. In small AutoStore systems, the travel time across the grid is insignificant in the picking operation, but in larger AutoStores systems, port arrangement can drastically impact travel time. As seen in the accompanying diagrams, putting a tunnel in an AutoStore system will cut down the average travel distance from a picked bin to a port. Those couple of seconds saved each pick carried throughout the day will add up to significant time savings, which means you can achieve throughput with fewer robots (and less cost). 3. Access Ports The ports are your operators’ way to get bins from the AutoStore. Two port types are the carousel port and the conveyor port. The carousel port has shorter cycle times (more picks per hour), but the conveyor port has the ability to interface with a conveyor system to transfer AutoStore bins to ancillary equipment in the facility. Choosing a type of port can be tailored to your needs. For example, we’ve designed a system where conveyor ports were used to feed a sorter and replenish inbound bins without an operator present. That same system used carousel ports for “hot picks” for small orders and expedited shipping. Choosing the number of ports will be based on results from computer simulations of the system to achieve desired throughput at the best cost.
4. Robots As the number of robots increases, the ability to pick faster generally increases. More robots helping each other means more bins being delivered to your ports. However, there is a limit to the number of robots that can be installed in a system. As the number grows, congestion increases, and eventually additional robots become counterproductive. Computer simulations will help determine the ideal number of robots for a system. We’ve seen small systems for parts delivery with only a couple robots as well as large systems for massive e-commerce operations with nearly 100. Another robot question that frequently comes up is how long it takes to pick a bin. A single robot can pick a bin and deliver it to its port within 30 seconds. If a bin has to be dug out from the bottom of the grid, the time can be 3 – 4 minutes. However, AutoStore has a unique ability to limit digging—forecast mode. 5. Forecast Mode Forecast Mode is a feature of the AutoStore software that allows robots to look ahead to future orders and pre-position them for later picking. They will focus on the picks currently in the worst locations (the bottom of the grid) and pull them to the top. One of the best ways to take advantage of this operation mode is to allow the robots to forecast while workers aren’t present. This will typically be at night for operations that aren’t running 24/7. By loading a day’s orders the night before, the robots can pre-position the picks for the next morning in order to fulfill the orders with fewer people in less time. All of the orders will be in the top few layers of the grid each day. If your operation is 24/7, you can still implement forecast mode throughout the day. As your workers rotate in and out and as orders fluctuate, you will have natural peaks and valleys in pick rates. During the valleys, a portion of your robots will automatically forecast picks for the other robots. For example, if you have 100 robots in a system that are all picking bins during peak periods, you may only be using 70 robots during lulls, which will leave 30 robots to pre-position bins for other robots. This cooperation between robots will increase average pick rate throughout the day, ensuring your operators have a steady stream of work.
In order to run an effective computer simulation to optimize the AutoStore configuration, we collect order data to tailor simulations to each operation. The first parameter we dissect is SKU velocity and distribution. AutoStore systems work best with a skewed distribution. We typically look for operations where 80% of order lines come from 20% of the warehouse’s SKUs, or an 80/20 distribution. The more heavily-skewed the distribution, the more benefit the AutoStore will provide. We use the SKU velocity and distribution to create a “bin depth profile,” which is a model of how often bins get picked from each layer in the grid. The more often bins on top are picked (as opposed to bins near the bottom), the fewer number of robots will be required to meet a given pick rate. AutoStore has two methods of forcing orders to the top to help make the bin depth profile more favorable—forecast mode and natural percolation. Forecast mode was already discussed, but natural percolation is a phenomenon that occurs through operation of the AutoStore system. Frequently-picked SKUs are replaced to the top of stacks after picking several times throughout the day, while infrequently-picked SKUs are in bins that naturally drift downward over time. This means that the more an AutoStore is used, the more efficient it becomes. It will even automatically adjust as demands change (seasonality, consumer preference, etc.).
AutoStore is a highly-configurable automated warehouse system that emphasizes storage density and flexibility. We’ve installed many of these systems, optimizing their configurations for each operation. With engineers throughout the world specializing in material handling equipment, Bastian Solutions is able to investigate the applicability of an AutoStore in nearly any situation in distribution, manufacturing, or e-commerce.
No comments have been posted to this Blog Post
Leave a Reply
Your email address will not be published.
Thank you for your comment.