Pedestrian and Traffic Management for the Design of a Parking Facility

Thungela is a leading exporter of high quality, low-cost thermal coal in South Africa. It has several operations in South Africa, but the one used for this project was the Khwezela Colliery, which is an open case thermal coal mine and processing plant.

Problem

There are approximately 1,200 employees and 90% of them use their own vehicles to come to the mine. There is a shift system with around 300 people per shift. Half of these people work in the mining pit and, due to its distance from the entrance, they require a shuttle service to get there.

There were some challenges with the parking infrastructure that resulted in poor control and management of vehicles, people, and goods coming into the complex. So, a new design of a parking facility needed to be developed for all employees where they would park or be dropped-off in the designated areas. Then, if necessary, they could take the shuttle to their place of work.

Analyzing the pedestrian traffic through the facility was necessary so that the appropriate number of boom gates and turnstiles could be identified without creating a bottleneck at the entrance.

Confirming the parking lot's size was necessary to avoid under or over designing it.

At times, there was a problem with alcohol, so 100% alcohol testing on entry also needed to be implemented.

Finally, in order to incorporate the most recent technology, the current security facility had to be renovated too.

Solution

Creating an AnyLogic model and running a simulation was identified as the best way to design a parking facility.

Simulation using AnyLogic has many benefits including:

• Providing a low-cost method to test modifications and receive prompt feedback on their potential effects on the KPIs required to be evaluated.
• Relatively simple updating of parking and entrance layouts and measuring traffic and pedestrian density using the AnyLogic standard Road Traffic and Pedestrian Libraries.
• The smooth integration of the Traffic and Pedestrian Libraries with standard functionalities such as Events, Select Outputs, the Process Modeling Library, etc.
• Visually identifying the accumulation of pedestrians and vehicles using heat maps. This makes it easier to identify the flow of traffic and spot potential bottlenecks, which could be useful for determining where to do extra sensitivity analyses, such as adjusting the number of turnstiles or altering the pathways so that people can enter in different ways.
• The ability to export the model as a stand-alone application for client testing and analysis.

Single run

Initially, a single run with animation was conducted. The 3D views helped with verification and validation as well as providing a more visual comprehension of what was happening.

All information was imported into the AnyLogic database from an Excel scenario file which contained all the parameters. When the model finished running, the results were downloaded to txt files which could be opened in Excel. Then the data was viewed and analyzed.

In this single run, it was possible to determine the likely values of variables such as arrival rates or delay times for opening a turnstile.

However, the model designers did not have all necessary information, for instance, what would be the final delay when opening a turnstile. To overcome this, a sensitivity analysis could be performed.

Sensitivity analysis experiment

A sensitivity analysis is when a simulation model is run multiple times while varying one or more parameters and showing how the simulation output depends on these.

In the sensitivity analysis experiment run for Thungela, all parameters were analyzed simultaneously. The number of employees, shuttle capacity, and other variables could all be configured at once. A new experiment was generated and ran in the background for each set of parameters. The number of days was set for 30. The final results could then be downloaded in detail or as a summary.

There were three output files – employee waiting, shuttle waiting, and vehicle waiting. These could all be exported into Excel to compare and analyze different variations.

Below is the Thungela pedestrian and traffic simulation model for a new parking facility from this case study.

Results

The sensitivity analysis produced all the results for this design of a parking facility since this range of outcomes was required to determine the maximum and average times for vehicles and pedestrians entering and exiting at each gate.

Maximum waiting time at gate stop

Drop and go turnstile maximum waiting time

In the second result, the chart depicts the drop and go turnstile. This is where the employees are arriving on the outside. Then they have 5 seconds for alcohol testing and 5 seconds for the turnstile to open, completing the process. The maximum number of employees arriving at a shift simultaneously was about 60 (indicated by the green line in the middle).

With one turnstile, there would be a considerable amount of waiting time, with 2 turnstiles about 250 seconds, and so on. The project team could then make an informed decision about the maximum acceptable amount of waiting time for employees.

Parking lot turnstile maximum waiting time

The parking lot turnstile is displayed in the third set of results, and there is far less variation than there was in the previous results. This is because people park in different places and take different amounts of time to get to the turnstile compared with the drop and go where 60 employees could arrive at the same place at once. This insight was only recognized after running the simulation and looking at the statistics.

Maximum waiting times for vehicles at the entry boom gate

The maximum number of vehicles waiting at the entry boom gate was shown in the final set of results. The estimated number of vehicles arriving at the same time was between 30-60. As such, the maximum waiting time was 10 minutes even if there was a 10 second boom gate delay, which is extremely long. Five seconds is the more typical time for an entry boom gate delay.

The slides are available as a PDF.