Project Management

Gantt Chart

The Gantt chart went through three major revisions over the duration of this project. The first iteration had twenty-four tasks. In the second version, the Gantt chart’s tasks were broken up into five phases. Additional tasks were then added as the second semester progressed, which caused a significant variation in the final copy compared to the original version. For example, the original end date for this project was scheduled for 20 December 2018. However, the actual end date was 13 December 2018 with all necessary documentation required to be submitted by 7 December 2018.

In the second iteration of the Gantt chart, the timeline of the project was modified in order to match the criteria of the Capstone 2 course. As each task was required to be no longer than one-week, the original tasks had to be broken up into smaller chunks that would be able to be completed within the five-day timeline. An example of splitting tasks would be breaking up the build phase into electrical/mechanical/software/wireless communications sections for designing, prototyping, testing and final construction.

For future Gantt charts, the addition of a buffer at the end of every phase would limit modification during the project. In addition, more thought must be given to all tasks to ensure that the tasks have been broken down into their smallest possible parts. However, the limitation of the program’s schedule and not being able to truly follow the Gantt chart limited its usefulness as a project management tool in this course. In the first iteration of the Gannt-chart, the phases were broken down as follows:

Design phase 1 May 2017 – 16 June 2017.

Purchase phase 19 June – 8 September 2017

Build phase 5 September – 18 October 2018

Test Phase 19 October – 12 December 2018

Report Phase 13 – 14 December 2018

In the final version of the Gantt chart, the timeline was as follows:

Design phase 1 May 2017 – 13 June 2017

Purchase phase 13 June – 1 December 2017

Build phase 5 September – 16 October 2018

Test Phase 17 October – 16 November 2018

Report Phase 19 November – 6 December 2018

The majority of the phases were completed on time with one exception. The build phase was not officially completed until the end of the test phase due to difficulties with the wireless communication portion. While this task was important for the final completion of the project, the lack of wireless control did not prevent the test phase from starting. Even with this delay, the test phase ended on schedule.

Budget

The initial budget for the project was $1000. This value was only to include the material cost for the project. The final cost was $1126. The budget was over by $126 due to prototyping and spare parts costs. Furthermore, Justin Vaughan saved an estimated $2000 to $3000 by contacting local companies to sponsor the project. The following is a list of companies who donated components/materials to the project:

· Arva Inc - Drawing design support & paint

· Advantage Machine & Tool Inc - Scoreboard enclosure plasma cutting and bending

· Lor-Don Limited - Control center & flat bar

· John Zubrick Ltd. - Tapping pads

· Trojan Technologies – Terminal blocks and din rail

Lessons Learned

The first lesson learned during the budgeting phase relates to ordering parts from other countries. When budgeting for source material, do not order parts that have long lead times. It is recommended to order source material from local and reputable distributors. In addition, by ordering local, the local economy is supported.

The next lesson learned is that when planning the purchase of project materials, the highest priority items should be ordered first. The high priority items for the scoreboard were the LED drivers, Arduino Uno, Raspberry Pi, and the main power supply. Additionally, it is important to check for materials on sale or buy in bulk when the material is at high risk of breaking, such as an LED burning out during testing.

Finally, when purchasing materials, it is recommended to purchase approximately 10 percent more stock in order to account for damaged components. The original budget allocated for 30 percent more stock than required; however, a 10 percent increase would be sufficient.

Labour cost was high due to underestimating the length of time it took to complete tasks. According to the EVA, the total project labour cost was $13804 for 812 hours of work. The original budget had predicted approximately 200 hours; however, the time actually spent was 4 times the estimated hours. The lesson learned is when budgeting for cost estimation, breaking down tasks to the lowest level allows for more accurate predictions of task durations. Overall, the project’s labour was underestimated by 612 hours.

Earned Value Analysis

At the beginning of the project, the Schedule Performance Index (SPI) and Cost Performance Index (CPI) were above and below the upper and lower control limits (UCL & LCL) respectively. This phenomenon was occurring consistently for two weeks in a row even though the team was meeting every week to contribute to the project. On 16 October 2018, after consultation with Mr. Volkening, the upper and lower control limits for the SPI and CPI were diverged. This consistency has been attributed to systematic over and under estimation of time for tasks. After the limits were changed (SPI & CPI, UCL 1.3 & 1, LCL 1.1 & 0.8), the CPI remained within the UCL and LCL until the reporting period of 22 November 2018. At this time, the CPI dropped below the LCL due to an underestimation on the amount of resources required to construct the final design. The SPI was outside of the UCL for two reporting periods after the divergence due to delays in the wireless control tasks. This was solved before the testing phase was scheduled to be completed, which allowed the SPI to return within limits after 22 November 2018. The final SPI and CPI values were 103.85% and 98.14% respectively. This convergence is due to the final few tasks requiring the allocation of all of the team’s resources towards the final tasks to ensure completion. The EVA provided an interesting glimpse into the data behind a major project. While this project’s EVA was skewed above and below the initial UCL and LCL, the EVA’s power was to highlight areas where the group was falling short and allowed the team to allocate more resources at a task as needed. Without the EVA, the team would not be able track a team member’s progress to better complete the project.

Risk Matrix

The risk matrix was reviewed in September 2018 to mitigate high risks (Figure 26). Green tasks are low impact or low probability, while yellow tasks are medium impact or moderate probability. High risk (red) tasks were re-evaluated and adjusted to be yellow risks.