Simulation shows how bottles fill and stack and whether they'll tip over on a conveyor.
Manager of Package Analysis
Procter & Gamble
A meshed model of a Tide detergent bottle was created in Altair's Hypermesh. Blue areas are thinnest; red are thickest. P&G engineers use Abaqus FEA software in the Virtual Package Simulation system to assess effects of different loads and forces on bottles traveling along a packaging conveyor.
In this simulation, mouthwash bottles travel down a conveyor toward an impact gate. An important FEA task is to determine the effects of impact and bottle accumulation as they strike the gate at the right. Abaqus/Explicit performs a dynamic simulation of the travel and impact.
As mouthwash bottles accumulate at the gate, they tip back and forth against each other but do not fall over. Designs that tip over typically have their centers of gravity lowered to ensure better balance during travel and on impact.
To speed bottle design, our engineers developed a method of analyzing bottles to determine their structural performance. Our Virtual Package Simulation (VPS) system uses Hypermesh from Altair Engineering Inc., Troy, Mich., as a preprocessor and Abaqus FEA software, from Abaqus Inc., Pawtucket, R.I. Because nearly all Proctor & Gamble (P&G) bottles share common features, such as being filled from the top, having thin walls, and using plastic materials, VPS includes several predefined load cases. These spell out standard loads and boundary conditions, what bumps and impacts to expect, and proper analyses procedures.
The system is invaluable because P&G uses hundreds of different bottles, and every month we tweak old designs and evaluate new ones. Proposals include new weights and sizes that range from 4-oz bottles of Old Spice to 2-gallon bottles of Tide.
THE BOTTLE STARTS HERE
Bottle design proposals and revisions come from a number of sources, including marketing reports, consumer surveys, and ergonomics tests. Each time a bottle feature changes, the design must run a gauntlet of analyses to prove it will survive shipping, sitting on the store shelf, and use by consumers.
Bottles must also make it through the plant well before they ship. For example, empty and full bottles must travel at steady speeds down conveyors. They must hold their shape while being filled, and stay upright when they hit guide rails and gates on the conveyor.
In the past, analysts would test bottle prototypes on an oval test conveyor nicknamed the Racetrack. Results were reliable, but the tests were time consuming.
Costs limited the number of prototypes that could be tested for any one bottle, restricting innovation. A recent addition to VPS is a dynamic testing system called the Virtual Racetrack. It simulates conveyorbelt conditions with speeds, impacts, and bottle backups. To develop it, company analysts reviewed data from years of prototyping on the Racetrack and developed formulae based on that data. The Virtual Racetrack standardizes model meshing and analysis methods. Also, engineers can use it to study static loads as well as realtime analyses of bottle travel and impacts.
Typical design changes include lowering the center of gravity on empty and full bottles to improve travel on the conveyor belt, strengthening bottle walls to prevent deformation on impact, and modifying geometry to better withstand fill loads.
A typical analysis starts after importing the CAD model of a bottle through IGES into the mesher for preprocessing. Each design is meshed and given a thickness based on bottle specifications or on the results of a blow-molding simulation, also created in-house.
The mesher lets bottle designers select appropriate load cases to analyze with Abaqus/Explicit. Simulations can have, for example, top-loaded bottles bumping into gates while empty or filled, accumulating along the conveyor, and moving with an assigned speed. Engineers specify conveyor speed, gate heights, load values, and other production equipment. Finally, job-management software transfers all these inputs to Abaqus, which runs consecutive analyses of the meshed bottle designs. Full analysis generally takes 12 hr using six CPUs.
The Virtual Racetrack has produced a number of immediate benefits, such as shorter design cycles than previously possible. It also tests a wider range of proposals at lower costs. This lets our engineers evaluate more designs, thus promoting innovation.
In addition, VPS applies standardized meshes and analyses for greater uniformity of methods and results which, in turn, instills a high level of design confidence. Our team has analyzed hundreds of bottles using the Virtual Racetrack and VPS system. When designs get the goahead, the bottles perform as predicted. Engineers are now working to reduce the run time for the Virtual Racetrack from 12 hr to three.
Our company plans to license the Virtual Package Simulation system, along with the Virtual Racetrack and a blowmolding simulation tool. Of course, such systems will not replace the experience of seasoned analysts, but it will help them get accurate results more quickly.