Simulation of road slope with brake applications on horizontal ground

Abstract: During some operations the brake system on a truck is exposed to severe stress and hightemperatures. To ensure function and safety during these operations, Scania test thebrake system both test rigs and in actual trucks around the world. There are tests thatare regulated by law that the truck manufacture must pass to be allowed to sell its vehicleand there are Scania internal tests to validate performance and quality. One of thoseinternal Scania tests is a severe test with extreme temperatures, called the Rossfeld-test,performed at the Rossfeld slope in the German Alps. The idea is that the truck rolls downthe slope only accelerated by the downhill force where the wheel brakes will be appliedwhen the vehicle speed passes 35 km/h and released when going slower than 25 km/h,this is repeated until the end of the slope. No auxiliary brakes are used. The objective of this thesis is to investigate the possibility of using a Scania truck as a testbench to perform a simulation of a truck rolling down a descent. And by this investigatethe possibility of recreating a wheel brake test as the Rossfeld-test on a horizontal roadby using the engine as the source of acceleration instead of the gravity. In this thesis a model was developed in Matlab/Simulink which through the CAN-busnetwork in the truck could request engine torque and brake pressure. Given the roadtopography of the actual German slope, the corresponding downhill force was calculatedand requested of the engine and brake pressure was applied when the truck passed 35km/h and released again when reaching 25 km/h. The model was verified against roll-tests performed in known slopes with three, eightand sixteen percent gradient. Results from simulations with the truck showed a biggerinconsistency with respect to acceleration and gear selection when requesting a low torqueas with the simulation of a 3% slope. This was not the case when requesting a highertorque as when simulating a 8% or 16% slope. The results showed that at a highertorque request the acceleration was more consistent but the truck was still acceleratingslower than the corresponding roll-test and an extra twelve percent of torque was addedto mimic the acceleration of the roll-test. The first 3000 m of the Rossfeld-slope wassimulated by driving in a circle and the temperature measurements catches the differencesin temperature build up as in the result from a real Rossfeld-test performed in the Germanalp. This shows there is a potential in this type of simulation.