Issue 2/2010


03/02/10

Using Simulation for Lift System Design

Rory Smith

Since around 1998, a PC based lift simulation system has been available for general use. The system permitted one to model any type of building and any type of traffic. However, the results were quite different than what one would expect based upon the results of Up Peak Analysis. This difference in results has led to the conclusion that while simulation is a way to evaluate the performance of a lift system it cannot be used as a tool for designing an appropriate lift system. Simulation can be used for lift system design but a new methodology is needed . This methodology is explained.
Category: Issue 2/2010
Posted by: Editor

Traditionally lift systems have been designed using the 5 minute Up Peak Round Trip Time, (UPRTT), calculation (Strakosch, 1998). This calculation permits both the quality and quantity of elevator service to be evaluated.

Quality of service has traditionally been defi ned as waiting time while quantity of service is defined by handling capacity (Strakosch, 1998).

The 5 minute up peak traffic used in this method does not occur in real buildings. It is simply a method used to design a lift system (Barney, 2003). However, the method works in determining the number of lifts required in a building.

The assumption has always been that if one designs for the up peak then one will have sufficient lift resources to handle the other traffic conditions such as lunch, down peak and interfloor.

Barney validated this assumption by calculating that lifts have the following capacities during the other traffic conditions relative to Up Peak (Barney, 2003):

  1. Down Peak: 160 %.
  2. Interfloor: 140 %.
  3. Mid-day: 130 %.

Computer simulation of lift system performance is now available and it holds the potential to improve lift system designs beyond what was possible with the Up Peak Round Trip Method . However, an accepted methodology has yet to be found for applying simulation to design.

Simulation

Background

In the late 70s the use of simulation to evaluate lift systems was introduced by Barney and Dos Santos (Barney & Dos Santos , 1977).

Around 1998, a PC based simulation system was made available for general use. This system, ELEVATE was developed by Peters (Peters, 1998). The system permitted one to model any type of  building and any type of traffi c. However, the output was quite different from the results of an Up Peak Analysis.

Saturation

Simulation has the unique ability to identify the traffic handling capacity of a lift system under any type of traffic by identifying system saturation.

Saturation occurs when more passengers are requesting service than the lift system can handle. When this happens the length of queues in front of lift entrances increases as do the waiting times.

Saturation can be detected using ELEVATE for any traffic condition by increasing the lift traffic over time using what is known as a step profile.

Figure 1 is a graphic representation of a step profi le where every five minutes the traffic level is increased by 1 % of the building population. In this example, the traffic level starts at 5 % and ends at 15 %.



The results of simulation using the step profile can be plotted graphically in several ways. The length of queues of passengers waiting for service, as shown in Figure 2, is one method.



The queue lengths grow exponentially when the traffic level reaches around 10 %. At this point the system is saturated and cannot handle the quantity of passengers that require service. In order to make a more precise determination of the Maximum Handling Capacity (MHC) of this system, a 4 hour simulation of this traffic at the 10 % arrival rate was conducted.

Figure 3 depicts the queue lengths during this simulation.

What this graph demonstrates is that the queue lengths are not growing. Therefore, this system has not reached saturation.

In order to bracket saturation a similar 4 hour simulation was run at the 11 % traffic level. Figure 4 shows these results.



It is extremely important to consider that this building has a total of 792 occupants. Therefore, a 1 % increase in traffic indicates that an additional 7.92 people are requesting service in a 5 minute period. Figure 4 shows an oscillatory pattern. The system is attempting to handle 11 % of the population and it can almost handle the traffic but in the end the queue lengths increase. Figure 5 shows the same system attempting to handle 12 % of the population in 5 minutes. Notice how completely the system is in saturation as indicated by queue length growth.

After several trial and error exercises, it was determined that the queue lengths did not increase at the 10.5 % traffic level. Figure 6 demonstrates this.

One can better understand what is occurring by considering stability. Something that is stable when displaced from equilibrium quickly returns to the equilibrium position. Something that is unstable when disturbed departs from that position. The time required for a system to return to equilibrium defines its level of stability. The more stable system returns to equilibrium more quickly.

The following is a series of queue length plots with progressively increasing traffic level:

What one can observe here is a system that is stable prior to reaching saturation. It is stable in that the queue lengths are stable.

In the zone between totally unsaturated traffic and saturated traffic the queue length plots demonstrate an unstable condition.

Stability and control are opposites. A stable system is difficult to control while an unstable system can easily be controlled. This may create some opportunities for enhanced dispatching.

Evaluating system requirements

In a modern building the traffic levels are dependent upon building type, location, and client expectations.

Observations at a continuously monitored building in the USA (see Figure 8 ­ Logging Analysis below) indicate the low-rise lifts are handling about 8 % of the building population during up peak. However, this traffic is divided into 66 % of the passengers traveling up and 33 % traveling down.

Lunch at the same building is spread over a 2 hour period. This means lunch traffic is spread over 24 periods of 5 minutes each. If the traffic level was uniform, then, in each period, 4.17 % of the population would use the lifts. However, a typical peak level is around 8 %.

Interfloor traffic tends to be around 5 % and is related to smoking and coffee breaks.

Experience at this building indicates that 5 minute down peaks are around 8 % of building population.

The traffic patterns in this building appear to differ from those previously presented by Strakosch (Strakosch, 1998). Therefore, it is important to understand the traffic levels and patterns that are anticipated for a proposed building in order to determine the system requirements.

Methodology for using simulation for design

The following is a proposed methodology for using simulation to evaluate lift system performance in proposed buildings:

  1. Use traditional Up Peak Round Trip Time methods to determine the approximate lift needs of a proposed building.
  2. Determine with the buildings architect, consultant and owner, what types of traffic are anticipated.
  3. Determine what level of service is desired for the building.
  4. Simulate the anticipated traffi c patterns using the step profile method to determine the approximate saturation point.
  5. Determine the precise saturation point using queue length.
  6. Determine the saturation point for each type of anticipated traffic.
  7. Compare the systems capabilities to the required service levels taking into account the impact of one lift being out of service.
  8. If the systems capabilities differ signifi cantly from the desired capabilities, evaluate other system options.

Conclusion

Simulation has the capability to change the way lift systems are designed. In lieu of designing a system using the Up Peak Round Trip Time calculation that assures sufficient lift service regardless of which lift manufacturer is selected, simulation permits a lift system specific evaluation of performance. It is now possible to determine how the lift system will perform under any traffic scenario. One can now quantify how much, if any, reserve capacity a lift system has.

Finally, it may now be possible to apply advanced technology to lift systems and reduce the number of lifts in a building. Through simulation it is possible to verify that the system will deliver the required performance level.

References
Barney, G. (2003) Elevator Traffic Handbook. London: Spon Press

Barney, G. and Dos Santos, S. (1977) Lift Traffic Analysis Design and Control. London : Pereginus

Peters, Richard (1998). ELEVATE Simulation Software. Mobile: Elevator World

Strakosch, G. (1998) The Vertical Transportation Handbook. New York: Wiley


 

2/2010

^ Top