Gear Generation Grinding Force Analysis Based on Worm Profile

• 1
  Gear generation grinding force analysis starts with contact geometry
• 2
  Normal force, tangential force, and what each one tells you
• 3
  Where the heat goes and why force is only half the story
• 4
  When a force model stops being useful
• 5
  Which signal tells you which problem

A worm grinding wheel is a threaded cylinder dressed to the profile of the gear tooth it will cut. It is not a formed disc that sinks into one gap. It meshes with the work like a worm drive, and the contact line sweeps across the flank as both parts rotate. One wheel pass therefore touches several teeth at once, and each contact is short.

That geometry sets the force model. The instantaneous chip load is not a single number. It is the sum over all active contacts, each with its own local depth, sliding speed, and grit population. Analysts usually start from a normal force per grit, then integrate along the contact line. Module, pressure angle, helix angle, and wheel diameter all enter that integral.

Change the module and the contact length changes with it. A coarse module spreads load over a longer line but removes more material per pass. A fine module concentrates the same spindle power into a shorter arc. The force per unit width can rise even when the total tangential force looks modest on the dynamometer.

This is why two gears ground on the same machine can behave nothing alike. The model has to be rebuilt for each tooth form, not scaled from a previous job. Treat the wheel profile as an input, not a constant.

A grinding contact produces three useful components. The tangential force acts along the cutting direction and drives spindle power. The normal force pushes the wheel away from the work and causes deflection. The axial component follows the helix and loads the thrust bearing.

Spindle power tracks tangential force closely, so it is the easiest signal to log. Normal force is the one that matters for accuracy. It bends the wheel spindle, the work arbor, and the fixture. On a tall gear shaft, that deflection shows up directly as profile error and lead error.

For hardened steel at 58–62 HRC, specific normal forces commonly fall in the range of 5–20 N per mm of contact width for conservative passes. Push the depth of cut and the number climbs fast. The relationship is not linear, because grit dulling raises the effective friction coefficient as the wheel loads up.

Measure the components separately if you can. A three-component dynamometer under the work table is the usual setup. If only spindle power is available, you can still estimate the trend, but you lose the deflection information that explains the size scatter.

Most of the energy at the contact turns into heat. Depending on process parameters, roughly 60–90% of the generated grinding heat transfers into the workpiece. The rest leaves with the chips, the coolant, and the wheel. That split is not fixed. It moves with depth of cut, wheel speed, and coolant delivery.

Heat partition matters because it couples back into force. A hot workpiece expands, so the effective depth of cut increases and the normal force rises. Then more heat enters the part. This feedback loop is why a gear can grind clean for the first twenty teeth and burn at tooth sixty.

Burning is not a force problem, but it is a force-model boundary. Once the specific energy climbs past what the coolant can remove, tempering starts under the surface. The force signal may even drop slightly as the material softens, which makes power monitoring a poor burn detector on its own.

Use coolant pressure and flow as model inputs, not as shop-floor habits. Through-spindle delivery at 20–40 bar reaches the contact zone far better than a flood nozzle aimed at the outside of the wheel.

A force model is a planning tool, not a controller. It predicts trends and ranking, not absolute values. Grit size distribution, dressing sharpness, and wheel wear all drift during a shift, and none of them appear in a clean analytical model.

The model also assumes a rigid system. If the fixture has 0.02 mm of compliance, or the tailstock is not seated, the predicted normal force will not match the measured one. Check the setup before you blame the model.

Wheel wear is the biggest gap. A freshly dressed wheel cuts with sharp grits and low force. After a few hundred contacts the grits flatten, the force rises, and the heat partition shifts toward the workpiece. A model that ignores wear will underpredict force late in the cycle.

Use the model to choose starting parameters and to explain a trend. Use in-process signals and inspection data to close the loop. That combination is what keeps a gear grinding process stable across a production run.