Predicting bucket loads on a machine that did not exist yet

The challenge

A heavy-equipment OEM was developing a new, high-efficiency skid steer loader, and the program was still at concept stage: no physical prototypes existed and many design parameters were open. One of the earliest decisions on the table was sizing the hydraulic components, and that sizing depended on a number nobody had: the power required to operate the bucket attachment across the varying material types the machine would actually work. Every hydraulic sizing decision downstream would inherit whatever error that load estimate carried.

Crude force models produce numbers quickly but are not credible enough to size hardware against. Semi-empirical approximations can be made accurate, but only after exhaustive effort to develop and correlate them, and with no prototype there was nothing to correlate against.

The approach

DSIM coupled EDEM, a discrete element method (DEM) material simulation, with multibody dynamics in Adams Solver. DSIM built a multibody model of the skid steer loader carrying the driving motions of a typical operational scenario and ran it in co-simulation with EDEM: Adams Solver controlled all system dynamics while EDEM computed the granular material response and fed the resulting forces back onto the bucket. Machine and material pile solve together, so the load path from material to machine is computed. EDEM's GEMM database of material models let the full loading and unloading cycle be tested in multiple material piles with no guesswork on material parameters.

Where the bucket force pointed

The strongest evidence that the coupling is doing real physics is what the bucket force vector did through the dig cycle. No bucket-load formula was programmed anywhere in the model. Every load emerged from the physics-based DEM material interacting with the bucket geometry, and the behavior matches what is seen in physical test data with buckets.

Phases 3 and 4 are the tell. A static or semi-empirical load model gives you material weight; it does not give you the inertial overshoot during lift or the forward reaction under rearward acceleration unless someone thought to program those effects in. Here they fell out of the physics.

The result

The coupled simulation determined the power requirements to move different materials for an identical set of driving motions, evaluated for two materials at far ends of the spectrum: gravel (high density, medium cohesion) and wood chips (low density, zero cohesion). That bracketed the duty the hydraulic components had to cover and gave the OEM's engineers a physics-based envelope for sizing hydraulics while the design parameters were still open.