Elevate simulations start with the lifts empty at the home floor. Naturally, the first passengers will experience waiting times that are shorter than if all the cars were busy serving other passengers. In a short simulation, this "head start" can significantly bias the results, usually resulting in an optimistic estimate of performance. When you simulate for longer periods, the effect of this starting condition is much reduced.
If you are running a situation where traffic is heavier than the system can sustain, then the longer you run the simulation, the worse the queues will be, and thus the reported performance is worse. In this case, you might want to experiment with the inputs. You might change the inputs so that your system will not be overloaded (e.g. change the speed of the lifts or even the number of lifts). Or you might wish to study how long after the last simulated passenger has entered the system before the system returns to equilibrium.
Also note that with longer periods, there is more chance of there being variations in intensity within the period. For example, if you have a period with 10 persons per 5 minutes for 15 minutes, then in the first 5 minutes there may be 12 people, the second 7 people, and the third 11 people. If the system is running close to saturation, it may fall behind in the first 5 minutes, and then never catch up. As a consequence, three 5-minute periods may give better results than one 15-minute period.
So how long a time should you run a simulation? The simple answer is to run a simulation for a long enough time so that the results are independent of the starting conditions.
ISO 8100-32:2020 addresses this issue by recommending a two-hour simulation, omitting the first 15 minutes and last 5 minutes from the analysis. Elevate provides an ISO 8100-32:2020 template.
CIBSE Guide D addresses this issue by considering the transients you see in real buildings. Elevate provides a range of CIBSE templates.
Was this article helpful?
That’s Great!
Thank you for your feedback
Sorry! We couldn't be helpful
Thank you for your feedback
Feedback sent
We appreciate your effort and will try to fix the article