Integrate a full automatic warehouse
In 2017 I got the oportunity to work on a short term initiative with high urgency
The customer bought a fully automated warehouse and started building it up in one of the business units. The promised capabilities of that warehouse were impressive.
- high density storage of full pallets as well as mixed pallets
- full pallet out feed for internal replenishment
- goods to person outfeed for customer order picking
- single piece picking for bulky flat packs including stacking of the ready picked customer order
Unfortunately, no one ever really investigated the impact of such a device on the different operational processes in the unit. As well the integration towards the customers ERP system was defined only at a very high level. Since the build up started already it was urgent to investigate further on those open points.
01
Volume analysis
We started with a volume analysis.
Which ranges are suitable for which capability and what is the expected sale for those items.
Based on that we were able to map articles and by that the expected outfeed volumes to the three different capabilities full pallet, G2P and autopick.
02
Pick priority
Next, we implemented in the customers ERP system an option to flag pick locations as
conventional, G2P or AutoPick. where it was needed to allow an article to be located on a conventional AND on one of the automatic locations, but not on both automatic locations at the same time.
With these flags we developed a crude initial logic to decide if a certain pick should be done from the automation or from the conventional.
After this step we had the decision point in place "when" to request a pick from the new automation
03
Event definition
Once we knew which picks to be requested from the automation we designed together with the automation vendor the events for the communication.
The first bunch was quite limited and allowed to
- request a pick
- confirm the pick request
- reject the pick request
- confirm the pick as done
Especially for the G2P part we put some energy as well in defining a sequence of the pallets delivered to the G2P outfeed. This was needed to ensure that we do not end up with too many different orders being halfway picked, since the out-feed area did not offer enough space for storing such orders.
04
Merging
The picking module of the ERP supported already the flow where different parts of the same order were picked by different users in different areas, so luckily we were able to re-use that functionality and with some configuration changes the automation was considered as separate user in a separate area.
05
Try out
After a few weeks the most important elements were in place and we were able to start testing the integration as well as the machine as such.
Especially the automation as such struggled a lot to achieve the requested performance.
Even after several rounds of improvements, optimisation of the picking robots and further improvements on the integration the needed throughput was never achieved.
After one year the decision was taken to keep the installation in the test unit, but to use it only for high density storage and to not invest in any additional installations.
06
Conclusion
Especially on expensive hardware installations it is important to do proper analysis and simulations upfront. Do not get "blended" by the fancy machines. A Porsche might be a great car, but not if you need a transporter for a carpenter.
Being aware of what is needed and what can be expected from the solution plus letting the Vendor know upfront what is expected is an important milestone during the preparation of such a project.