Thanks so much for your in-depth response, Jan!! The way that Jakob implemented it in the ElecHybrid model sounds perfect to me! I will try it out.
By the way, I found a research paper written back in 2011 by Maia et al [1] which claims that they implemented an EV model for SUMO which takes into account both motor-torque- and battery-current-limits to calculate the next timestep's velocity and acceleration. I will check if they published their source code. Regards, Chris [1]: R. Maia, M. Silva, R. Araújo and U. Nunes, "Electric vehicle simulator for energy consumption studies in electric mobility systems," Online: https://ieeexplore.ieee.org/document/5973655/ On Thursday, 15 October 2020 15:36 sumo wrote: > Dear Chris, > > as far as I can remember one of the problems with EV simulation is the > ordering of simulation steps: The simulator first moves a vehicle according > to the car following model and then uses the energy model to compute the > actual power that was needed to make that step -- and if that power exceeds > the rated power of the engine, you have no means to influence the already > executed vehicle move. This is a side effect of CFM parametrisation: the CFM > knows only about vehicle acceleration/deceleration capabilities, it makes no > use of engine power and vehicle mass. If you look at the code, even if the > parameter MAXIMUMPOWER is stored in the MSDevice_Battery, it is not used for > anything there yet. > > We have had this problem when designing the fork of EV model for our > MSDevice_ElecHybrid, and if I remember correctly, what my colleague Jakub did > at the end was to extend the MSVehiclie::executeMove() to limit the maximum > possible acceleration of the vehicle based on the MAXIMUMPOWER of the > ElecHybrid device. This will not affect the current speed and position, but > it will affect the next time step. It is not quite correct but given the > separate structure of the car following subsystem and the devices, there is > probably no other chance to do so (except of changing the set of CFM > parameters -- the current ones are de facto standard in the microsimulation > world). Contrary to some reports, our simulations show acceptably similar > results to those measured in real traffic. > > I am not sure that we have made some analysis of how the trolleybus model is > affected by all this (it could be that the final effect of this change is > quite small; this seems to be another candidate topic to appear on our TODO > list). > > So, to sum it up, if you feel adventurous, try to swap your Battery for > ElecHybrid. In principle, it should also work for opportunity-charged > electric vehicles (but we did not test it recently; I am sending a copy of > this to Jakub, he might correct me). > > Hope this helps, > > Jan > > On Wed, 14 Oct 2020, at 4:56 PM, Chris Abraham wrote: > > Hi everyone, > > > > It seems that the built in electric vehicle model does not limit the > > maximum power of the vehicle. The Sumo documentation indicates that > > there is a property called `maximumPower`. However, when I set it to a > > value, it seems that the simulation completely ignores it, because the > > EV still uses more than the value I specify. Additionally I would like > > this `maximumPower` to limit the regeneration power. > > > > Chris Abraham > > > > > > > > _______________________________________________ > > sumo-user mailing list > > [email protected] > > To unsubscribe from this list, visit > > https://www.eclipse.org/mailman/listinfo/sumo-user > > > _______________________________________________ > sumo-user mailing list > [email protected] > To unsubscribe from this list, visit > https://www.eclipse.org/mailman/listinfo/sumo-user > _______________________________________________ sumo-user mailing list [email protected] To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/sumo-user
