Critical Design Review

Critical Design Review

Submitted to

Professor Jolanta Janiszewska

Prepared by

Nick Kanel

Matt Musso

Marcus Pereira

Engineering 1182

The Ohio State University

21 April 2016

List of Tables and Figures

Table 1: Concept Screening Matrix designs 1 and 2

Table 2: Concept Scoring Matrix designs 1 and 2

Table 3: Numerical representation of Figure 6

Table 4: Numerical representation of Figure 7

Table 5: Breakdown of the main tasks Completed

Figure 1: Isometric view of AEV Design 1

Figure 2: Isometric view of AEV design 2

Figure 3: An image of the Airflow simulator

Figure 4: Propulsion Efficiency vs Advanced Ratio for 3-bladed propeller

Figure 5: Propulsion Efficiency vs Advanced Ratio for 2-bladed propeller

Figure 6: Power (W) vs Time (s) for initial AEV design

Figure 7: Power (W) vs Time (s) for the final design

Executive Summary

Statistics show that the population of humankind on earth has been increasing. As there are more people inhabiting the Earth, resources are depleting at a much faster rate than ever before. For this reason, it is becoming increasingly important to become more efficient what the resources that are still available. This is where the need for an Advanced Energy Vehicle comes in.

The goal of the Advanced Energy Vehicle is to be as energy efficient as possible while still completing the given task. The research that was done throughout the semester was constant testing of the coding through an AEV run on the track followed by analyzing run data of the Arduino afterwards to check out the energy usage. By comparing the data of the different runs, it is possible to figure out which run or design was the most energy efficient.

The three Performance tests lead to discoveries that would be used for the rest of the semester. An important outtake of the first performance test was that mass, reliability, and balance were three of the most important aspects in a successful AEV. Concept screening and scoring matrices were made which showed that the first AEV design was the one that would further improved and continued based off of the findings. The second performance test showed that the speed coming down the incline had to be lowered. The AEV was going too fast down the track and wasn’t stopping in the designated area due to the momentum carrying it. Also, it was decided that the AEV could be a bit more aerodynamic in order to increase airflow to the propeller, which would cause an increase in thrust. Finally, an important outtake of the final performance test was motor, Arduino, and battery were causing inconsistencies. Several parts were replaced which lead to smoother runs. All of these performance tests were a part of the decision process in deciding the Arduino used in the final run worked the best.

Results

Figure 1. An isometric view of Design 1. This design was chosen to be developed and became the final design used for the final battery of test runs.

The design pictured in Figure 1 became the final design used in testing. Originally, the design featured the same configuration of equipment (ACS, battery, support arm) on a t-shaped chassis with both motors mounted on the outboard portions of the chassis, as far back as possible. This configuration led to the center of gravity for the AEV being positioned too far aft and led to the front wheel on the support arm separating from the track during testing.

The aforementioned created two issues: the AEV was incredibly unbalanced, and the thrust created by the propellers were not directed in the most efficient direction, which led to part of the thrust force being directed along the positive z-axis. To combat these issues, the original design was revised to yield that pictured in Figure 1, but with the outboard, trapezoidal structures facing upwards (these upward facing structures were later made to face downwards, as it was observed that the propellers were striking the support dowels on the test track). This configuration succeeded in returning the center of gravity to within acceptable limitations, centering the AEV’s mass about the contact points of the support arm. In addition, this design allowed for uninterrupted airflow to the propellers, leading to greater efficiency from the propellers.

Figure 2. An isometric view of the second AEV design.

Design 2 (pictured in Figure 2) was heavily based off of Design 1 (Figure 1), with only slight adjustments being made to the positioning of the motor and amount of motors used on the design. The driving concept behind this design was weight reduction, allowing for a significantly lowered thrust requirement. Thus, the outboard trapezoidal structures were removed from the center chassis, and a single motor was mounted on the aft-most portion of the chassis.

The design ultimately failed as the repositioning of a single motor on the aft section, again caused the center of gravity to shift out of acceptable limits, causing the front wheel to separate from the track during testing. The design did, however, used less energy on a number of different test runs, affirming the hypothesis that a reduction in weight would result in an AEV that consumes less energy. Moreover, Design 2 did not provide a location sufficient for mounting the servo, unless the configuration of the support arm was altered to accommodate it. This likely would have led to more difficulties with maintenance, and no design changes were made.

Criteria	Reference AEV	Design 1	Design 2
Balance	0	+	–
Maintenance	0	0	0
Consistency	0	+	0
Compatibility with Cart	0	+	–
Efficiency	0	–	+
Speed	0	+	–

Sum +	0	4	1
Sum 0	6	1	2
Sum –	0	1	3
Total	0	3	-2
Develop?	N/A	Yes	No

Table 1. The Concept Screening Matrix for Design 1 and Design 2 as compared to the Sample AEV constructed in “Lab 1: Creative Design Thinking.”

Concept Scoring Matrix		Reference AEV		Design 1		Design 2
Criteria	Weight	Rating	Weighted Score	Rating	Weighted Score	Rating	Weighted Score
Balance	25%	3	0.75	5	1.25	1	0.25
Maintenance	5%	3	0.15	4	0.2	4	0.2
Consistency (Precision)	20%	N/A	0	5	1	2	0.4
Compatibility with Cart	15%	N/A	0	3	0.45	0	0
Efficiency	20%	N/A	0	3	0.6	4	0.8
Speed	15%	4	0.6	5	0.75	3	0.45

Total	—	—	1.5	—	4.25	—	2.1
Develop?	—	—	No	—	Yes	—	No

Table 2. The Concept Scoring Matrix comparing Design 1, Design 2, and the Sample AEV referenced in “Lab 1: Creative Design Thinking.”

In comparing Design 1 and Design 2, a significant difference is observed between the two designs as given by Table 1 and Table 2. The criteria included in each Matrix was decided on by popular opinion and discussion regarding the significance of the agreed upon criteria. As balance was seen to be a significant issue during the initial series of tests, it was included on both Matrices, and given a greater weight in the Concept Scoring Matrix. Subsequently, consistency (precision) and efficiency (amount of energy consumed) were also each given higher priority in the Scoring Matrix. Following these were compatibility with the cart and speed, then maintainance. Design 1 was superior in four of the six criteria, a fact due very much so to the positioning of the components on the AEV. This specific positioning allowed for the addition of the servo to the design without extensive redesigning, and a much more balanced Vehicle. Along with this, the presence of two motors as opposed to one allowed for a much greater speed. As noted earlier, Design 2 did have a superior power consumption, but this was not a difference substantial enough to outweigh the Design’s downfalls.

Each design costed roughly the same amount to create as the major components were all the same. Taking into account the differences between designs (number of motors, brackets, extra structural attachments) Design 2 costs $15.39 less than Design 1. This difference is entirely acceptable for the increase in performance seen by Design 1 when compared to Design 2.

As the design process progressed, the Arduino code was slowly altered so as to produce the most efficient vehicle possible. For example, during some downhill portions of the track, power supplied to the AEV’s motors was cut in an effort to reduce the overall amount of energy used by allowing it to simply coast along the flat portion.

In addition to alterations in code to increase efficiency, it was realized that the construction of an aerodynamic body around the propellers could potentially increase their efficiency. The construct was to be a 3-dimensional hyperboloid (Figure 3) around the propeller. In doing this the mass flow rate to the propeller is increased by effectively funneling air into the propellers. The hyperboloid structure would allow for the greatest possible efficiency in both directions of travel. Moreover, the body would have been only an eighth to a sixteenth of an inch in thickness, and would therefore not add a sufficient amount of mass to counteract the benefit of having the body on the AEV.

Also, early on in the design process, the determination was made that the larger three-bladed propellers would be more efficient in producing thrust than the other two-bladed designs that were available. These three-bladed design were operated at peak efficiency (between 36% and 44% supplied power) for a large portion of the test trials. Furthermore, even though Design 1’s single motor construction was determined to be more efficient, the improvement was considered to be insignificant enough to yield no alteration to the final design of the AEV

Figure 3. The hyperboloid portion of the aerodynamic body designed to be put on Design 1. In this picture, the colored contours represent surface pressure, the lines running primarily along the x-axis represent the airflow over the body, and the tick marks represent vorticity vectors.

As the Performance Tests progressed, many new challenges were encountered and dealt with as they presented themselves. Initially, the issues with the AEV’s center of gravity led to the redesigning of the AEV. Thereafter, the propellers began contacting the dowel supports on the test track, which in turn led to the Design that was used for the duration of the MCR. Following this, the only design changes that occurred were regarding the AEV code. As test trials were completed and new data became available, more alterations could be made to the code regarding the positioning of the AEV, as well as portions of the test run during which the power could be decreased. As the tests progressed further to include the unweighted cart, and finally the weighted cart, adjustments were made increasing the power supplied to the motors to account for the added weight. This also required a hook design be printed, and the servo be attached to the AEV in some fashion. The hook was hot glued to a servo attachment piece and screwed to the servo, which was attached to the AEV chassis with a zip tie. This design worked well to achieve its intended goal, however the margin for positional error during attachment to the cart was very small. This limited margin of error thus required more attention in fine tuning the position of the AEV.

Figure 4. Graph of motor efficiency vs. advanced ratio for the 3-bladed propeller, which was used on the final AEV design.

The graph above is of the Efficiency vs. Advanced ratio for the 3-bladed propeller. Through the System Analysis 1 lab, it was discovered that the most efficient energy levels of the electric motors was around 30-35% power. This was taken into account when programming the AEV, and it resulted in the average power being used to level out at around 33%, which is right in the middle of the most efficient power level. The propeller that was used was also determined to be the most efficient based on the number of blades and the surface area that is covered when it spun through the air.

Figure 5. Graph of efficiency vs. Propulsion Efficiency for the two-bladed propeller, which was only considered for the final AEV design.

The graph above is a representation of the propeller efficiency vs. advanced ratio for the two-bladed propeller. The two-bladed propeller seems to be most efficient when the power level is at about 30-35%, which is very similar to that of the 3-bladed propeller described in Figure 4. The problem with this propeller is that it does not provide as much thrust at certain power levels as the 3-bladed propeller, so this propeller would require a higher power percentage to propell the AEV forwards. Due to this fact, it can be concluded that the two-blade propellers would not be running at their most efficient power levels and thus should not be used on the final design.

First Prototype:

Figure 6. A graph of Power(W) vs. Time(s) for the first design without a weighted cart with phase breakdown.

Phase	Prototype 1 Power	Commands
1	23.07 J	motorSpeed(4,23);
2	29.85 J	reverse(4),motorSpeed(35);
3	26.55 J	motorSpeed(4,23);
4	29.95 J	reverse(4),motorSpeed(35);
5	38.86 J	motorSpeed(4,23);
6	27.14 J	reverse(4),motorSpeed(35);
7	43.83 J	motorSpeed(4,35);
8	29.04 J	reverse(4),motorSpeed(35);
9	8.38 J	motorSpeed(4,12);
10	25.21 J	motorSpeed(4,17);
11	30.44 J	motorSpeed(4,23);
12	35.87 J	reverse(4),motorSpeed(35);
13	24.76 J	motorSpeed(4,23);
14	28.85 J	reverse(4),motorSpeed(35);
Total	413.741 J

Table 3. Numerical representation of the phase breakdown for the Figure 6 graph with Arduino snippets.

In Table 3, it is apparent that the function that stops the AEV (reverse and then motorSpeed) uses a significant amount of energy. Due to errors that will be mentioned in the later part of this report, the total energy that the stop function uses is not correctly represented numerically in this table. In order to correct this error and show the total amount of energy used, a more advanced Arduino board is required in order to run the program on multiple threads.

The first AEV prototype was able to complete the track in a relatively timely manner of around 63 seconds. As one may notice, there seems to be a very clear pattern to this run because there are many constant levels of power consumption followed by a spike and then a lull. This can be compared to the second prototype, which also has updated code, to show that the first prototype uses more power on average. In order to get up the first incline and grab the cart, the first prototype required a constant power of 5.6 Watts. This power consumption is continuous across the entire run because the first prototype was not set up to carry a cart. It, instead, simply ran the track in full and was able to hit all the marks that were required by the MCR.

Final design:

Figure 7. A graph of Power(W) vs. Time(s) for the final design including the weighted cart with phase breakdown.

Phase	Final Design Power	Arduino Code
1	18.165 J	setBoth(24);
2	24.731 J	stop(getVehicleDirection(), true);
3	25.424 J	setBoth(29);
4	3.927 J	descend(13,29,1)
5	31.384 J	stop(getVehicleDirection(), true);
6	59.167 J	setBoth(29);
7	29.133 J	stop(getVehicleDirection(), true);
8	76.891 J	setBoth(49);
9	39.278 J	stop(getVehicleDirection(), true);
10	11.745 J	setBoth(16);
11	11.681 J	setBoth(18);
12	5.245 J	descend(15,29,0);
13	25.491 J	stop(getVehicleDirection(), false);
14	50.219 J	setBoth(29);
15	36.756 J	stop(getVehicleDirection(), true);
16	13.576 J	setBoth(29);
17	22.339 J	stop(getVehicleDirection(), true);
Total	485.133 J

Table 4. A numerical representation of the phase breakdown in the Figure 7 graph with Arduino snippets.

The third AEV prototype, which happened to be the final design, was able to complete the track successfully in around 78 seconds while pulling the cart the entire way. The programming strategy had developed from simply doing many of the same commands over and over again in order to complete the task into doing similar commands in an organized and controlled manner that would save energy in the long run. As an example of this, if one were to look at the original design and see how much power the AEV required to climb to the top of the hill, they would notice that it is a greater amount than is required by the final design. When running with the fully-loaded cart, the final design used only 485.133 J of energy in comparison to the unweighted first prototype using 414 J to complete the track. Because the mass that was placed on the cart was 45g and the AEV has a mass of 280g, it can be inferred that if the AEV only had to carry the 45g of additional weight, excluding the cart itself, the first prototype would use 480.536 J of energy. If the weight of a cart were added to the 45g, the energy consumption would surely be much more than that of the final design when running at full capacity.

For each of these graphs, it is important to note the weight that each design was pulling. The first prototype was pulling no weight at all, and it used 434.526 J of energy, but the final design pulling the weighted cart (which was a cart plus 0.045 kg) used 485.133 J of energy. When that is broken down into energy-per-kilogram, it turns out that the final design is likely to be somewhat more efficient than the original design. This is not testable, however, due to the mass of the cart itself being unknown, which prevents an accurate calculation of Joules/kg.

During the final testing, the AEV was observed to run the course much like it had with the unweighted cart and also without the cart at all. Because the code was written to be dependant on a distance traveled, the AEV was able to successfully adapt to the new situation without any code being changed. The only code that was changed to allow the AEV to run the course during the final run was the motor speed, which was increased from 25 to 30 due to the added weight of the cart and cargo aboard. During the final run, the AEV used 485.133J of energy and it ran for approximately 78 seconds before successfully coming to a halt at the end. It was discovered that the more charge the battery had, the more aggressively the AEV would run the course, which meant that each run’s accuracy and consistency was dependent on how much the battery was charged between runs. To accommodate for this, electrical calculations were done based on the voltage and amperage output of the battery charger and the total energy used by the AEV. It was discovered that the battery would need to charge for approximately one minute and forty seconds to regain the energy lost by the run. Once this was applied after each run, the AEV became much more consistent with its stopping positions and this lead to a final run being completed.

While the AEV was not as efficient as other AEV’s, it still showed improvement throughout the testing and development. It started out using 1467.167 J/kg, but was eventually reduced to using 1720.330 J/kg with the final design. This was with the cart, so even though there is more energy use per-kilogram, the AEV is still more efficient, as was mentioned earlier. This improvement was brought about by making the body slightly more aerodynamic, moving the propellers out of the way so that they were not obstructed, and reducing the instability which caused the AEV to rock back and forth and lose energy to friction.

There were a number of sources of error in this lab, and most of those stemmed from things that were easily mitigated and corrected such as battery life. The most important source of error comes from the Arduino being capable of only running a program on one thread. This means that when the program is running and it encounters a while loop that doesn’t end until a certain condition is met, the AEV logging functions are unable to log data until the while loop exits. This created some interesting patterns in the energy usage graphs, which are straight lines starting where the while loop started and ending where the while loop exited. Because of this, there are some parts of the program where it is impossible to log energy usage data correctly without having more than one thread running. It is not clear whether or not the Arduino was using more or less data than was logged by the logger because of this. Due to it being over such a short amount of time, however, it does not affect the results very much at all. This error is systematic because it was experienced by all versions of the AEV, so there is very little possibility of it changing the outcome and final conclusions of this lab. In order to overcome an error like this, an Arduino board capable of multi-threading is required so that the program can do logging on one thread and the execution of the AEV program on another. If the Arduino was set up like this, data could be logged without issue.

Conclusion

This lab was overall a success. Through trial and error, as well as comparisons between designs and results, a final design and program were created which resulted in a perfectly-running AEV. The vehicle stopped within every destination it was supposed to stop nearly perfectly, was able to grab the other vehicle during the run, and was able to safely transport the vehicle along the track. All the objectives were met as well as finishing by the time constraints.

One very important result that was discovered early on was that by centering the pieces on the AEV, it created a better balance and allowed it to stay comfortably on the track along the curved section. Another thing worth mentioning is that, as Table 2 identifies, the final design uses 485.133 Joules of energy for the entire run. As stated in the paragraph following Table 2, it can be inferred that the other designs would require much more energy to do the same job the final AEV design used. While the first design may have finished much faster than the final design by more than 10 seconds, that does not make it better. Safety of the passengers of the vehicle is very important. Design 1 often overshot the stopping portions and sometimes even derailed due to the lack of balance. For the safety of the passengers of this ‘would be’ vehicle, it would not have been safe to continue this design, even though it was faster. Something else discovered about the first design was that it was not the most efficient design because it only had one motor. The most efficient run speed for it was roughly 30-35%. Unfortunately, it had to run at approximately 50% in order to complete the track with the towed vehicle, making it not as efficient as it possibly could be.

Overall, this final design worked very well and the final run went very smoothly. It stopped nearly perfectly in every section, with the exception of the bonus stop. The only reason the bonus stop was not attempted was due to issues with parts and the time constraint preventing it from happening. Compared to the other AEV designs, this design is arguably the best. This design has much less mass than some of the AEV’s made by the other students. This allows for much less energy to be used due to the fact that the AEV is hauling around much less matter. Another reason that this AEV was arguably the best design was because it was one of very few, if any others, to use two 3 bladed propellers. The 3 bladed propellers were efficient and allowed for the AEV to stop quickly and speed up quickly, allowing for the run to be very accurate. One of the most important things to note as to why this design may be the best is that others still seemed to have troubles with their design – some even needed to add multiple claws to their hook in order to be able to hook the towed vehicle consistently. This AEV was able to get hook the towed vehicle consistently with no issue with no extra hook features added.

One thing that would benefit future classes would be to have more than enough battery chargers that worked well. Often groups had a charger that did not work and they would have to borrow one from another group. If all the battery chargers worked just fine, each group would end up saving time overall, because there would never be a wait to charge a battery for a run. Another thing that would be very beneficial would be for the students to only come in during their class time. At times, nearly half of another class worth of students were in the room doing their own test runs, which cut deeply into the allotted time period for this period to work. Other than that, the project went smoothly, and there are no further recommendations.

Appendix

Table 5. Breakdown of the main tasks that needed completing during this entire project including labs, programming, and more.