+ Reply to Thread
Page 2 of 22 FirstFirst 1 2 3 4 12 ... LastLast
Results 11 to 20 of 217

Thread: Any way to objectively choose engine?

  1. #11
    Originally posted by Adambomb:
    "A good plan, violently executed now, is better than a perfect plan next week."
    -General George S. Patton, Jr.
    That is unquestionably the best quote I have heard for the FSAE application. Spending weeks on a decision like this will only divide your team further and waste your time. There is no right or wrong answer as long as you have a reason for what you choose. So jump in!
    "Good judgement comes from experience. Experience comes from bad judgement."

  2. #12
    You'll find your decision becomes much easier if you just view racing as a series of drag strips connected together.



    Seriously though, this is what I've found is best to consider (even for just dealing with a single engine). What is your shifting strategy (which typically revolves around your drivers)? The desired torque curve for a 'Leave it in 2nd' attitude vs. a 'Maximize torque at the wheel at all times' attitude is drastically different. With gearing you can get any amount of low speed torque that you want. It just comes down how often you want to shift and how fast your drivers can do it. Design the car to your drivers.

  3. #13
    Senior Member
    Join Date
    Jun 2003
    Location
    Melbourne Australia
    Posts
    762
    Lap simulation, linked to pointscoring analysis. Can't go past it. You'll learn heaps.

    A perfect lapsim will take a year. A simple lapsim will take a day, if that. The perfect lapsim is a refinement tool, for when you are tuning your established concept around a known track. The simple lapsim will give enough info to form your concept, which is what you want at this stage.

    Try to answer the following questions:
    A 1kg or 1 lb difference in vehicle weight is worth ... points.
    A 1hp or 1kW difference in power is worth ... points.
    A 1% increase in braking deceleration is worth ... points.
    A 1% increase in forward acceleration is worth ... points.
    A 1% increase in lateral acceleration is worth ... points.
    From the above, come up with rules of thumb like "A 1% increase in cornering acceleration returns ... times as many points as a 1% increase in braking decel".
    Your answers don't need to be exact, just good enough to guide your overall understanding of the comp.

    With the above, you can then make top level decisions between concepts on a "points per time required" and "points per dollar" basis.

    By simple lapsim, think the following: track = straight lines and arcs, lateral grip limit determines corner speed, and start with constant accel and decel. Any track will do. Once you have that working, try to implement power-driven acceleration. Then include aero drag and rolling res if you like. An energy analysis can drive fuel economy comparisons between concepts. Grip limited vs. power limited accel is easy with a bit of thought. All of this can be done simply in Excel.

    Remember that the engine has a number of attributes that affect the overall vehicle - weight, power, packaging, economy. Choosing an engine based on straight line accel (for example, selecting only on the "power:weight" criteria) doesn't account for how the engine affects other parts of the track (i.e. when power attribute is "off", but weight, economy & packaging attributes are still definitely "on").

    As a team, spend a week performing and assessing the above. Get everyone to the same level of understanding. Choose the most appropriate solution for your resources, put a freeze on it and move on. Either engine choice can be competitive, but your understanding and team cohesion will be hugely enriched by working through the process.

    Decision matrices are great when they are populated with quality, (read, quantified and justified) data. But as said above, more often than not they are just formalized guesswork. Most I've seen have been awful and are only used to justify a pre-determined solution.

    Cheers,
    Geoff Pearson

    RMIT FSAE 02-04
    Monash FSAE 05
    RMIT FSAE 06-07

    Design it. Build it. Break it.

  4. #14
    Senior Member
    Join Date
    Sep 2002
    Location
    Perth, Western Australia
    Posts
    717
    I agree with Geoff that lap sims are invaluable. You do need to keep in mind that lap sims usually work with a 100% capability of the driver. This can be simply dealt with by applying overbraking (a very common problem) and percentage factors to overall cornering. When choosing fundamentals like this it is good to consider the "best" situation, as well as the "usual" situation.

    Lapsims bring quite a bit of objectivity to decisions within the team as Geoff has indicated. There are however a number of other factors that need to be considered in the general design process. Even with respect to the overall performance. A couople of questions include:

    How do you assess the points value of ergonomics? (This is very important for things like shifting systems)

    How do you assess the points value of reliability? This can be analysed in some depth as a risk assessment.

    Add into this mix the higher team level issues. An engine choice for instance locks you in for a number of years if you wish to make the most of your choice. What sort of testing facilities are available? Is one method of construction easier or more difficult for you in comparison to other teams? How will decisions on your car affect funding opportunities?

    There has been quite a lot of study gone into accurate concept selection. Decision matrix style devices are usually helpful, but surprisingly the simpler methods work just as effectively as more complex versions. It is much more important how they are implemented in a group setting. Group dynamics being what they are nearly always encourage certain individuals over others. The whole squeaky wheel analogy. It is important to have methods to neutralise egos and have true concerns about one way or another come out.

    I would also strongly recommend simple statistical models. Plot out choices of teams vs points scores (in all events including overall). They will show the historical content of how the different decisions finish in the real world. This is likely to show things like reliability concerns as well as highlighting the difference between fastest cars and overall winners. With this approach you may be the team to miss out on being the first to do _______ properly. However there are usually very few benefits to being the first group to do something right for the first time. Look into the history of things like MP3 players and ground effects race cars and you see that the "firsts" usually dont hold an advantage very long at all. Simple statistics are a great help for these decisions. Especially with 400+ cars a year being made.


    Finally I think the decision between engines is not a fundamental decision. Your choice between either will not define whether you have a winning car or not. The proof of this statement is in the results from the last few years. You should be focusing your debates on fundamental decisions. Other aspects of your car design will have a massive effect on your engine choice. Is your team capable of building a car well under 180kg for a single? Is your team capable of running a four cylinder car at around 3L for the endurance run?

    Answers to these questions (and many others) will help to inform your decision for which engine to run.

    Kev

  5. #15
    Originally posted by Big Bird:
    Lap simulation, linked to pointscoring analysis. Can't go past it. You'll learn heaps.

    A perfect lapsim will take a year. A simple lapsim will take a day, if that. The perfect lapsim is a refinement tool, for when you are tuning your established concept around a known track. The simple lapsim will give enough info to form your concept, which is what you want at this stage.

    Try to answer the following questions:
    A 1kg or 1 lb difference in vehicle weight is worth ... points.
    A 1hp or 1kW difference in power is worth ... points.
    A 1% increase in braking deceleration is worth ... points.
    A 1% increase in forward acceleration is worth ... points.
    A 1% increase in lateral acceleration is worth ... points.
    From the above, come up with rules of thumb like "A 1% increase in cornering acceleration returns ... times as many points as a 1% increase in braking decel".
    Your answers don't need to be exact, just good enough to guide your overall understanding of the comp.

    With the above, you can then make top level decisions between concepts on a "points per time required" and "points per dollar" basis.

    By simple lapsim, think the following: track = straight lines and arcs, lateral grip limit determines corner speed, and start with constant accel and decel. Any track will do. Once you have that working, try to implement power-driven acceleration. Then include aero drag and rolling res if you like. An energy analysis can drive fuel economy comparisons between concepts. Grip limited vs. power limited accel is easy with a bit of thought. All of this can be done simply in Excel.

    Remember that the engine has a number of attributes that affect the overall vehicle - weight, power, packaging, economy. Choosing an engine based on straight line accel (for example, selecting only on the "power:weight" criteria) doesn't account for how the engine affects other parts of the track (i.e. when power attribute is "off", but weight, economy & packaging attributes are still definitely "on").

    As a team, spend a week performing and assessing the above. Get everyone to the same level of understanding. Choose the most appropriate solution for your resources, put a freeze on it and move on. Either engine choice can be competitive, but your understanding and team cohesion will be hugely enriched by working through the process.

    Decision matrices are great when they are populated with quality, (read, quantified and justified) data. But as said above, more often than not they are just formalized guesswork. Most I've seen have been awful and are only used to justify a pre-determined solution.

    Cheers,
    Thanks mate.

    This is pretty much what I was looking for!

    I am still in the process of getting some data from my team, but can you guys give me ballpark estimates on what acceleration and deceleration speeds I should be using? Also what is the average and top speed?

    Thanks alot everyone. I am going to attempt to do the calculations on the lapsim. If you guys have any other resources that could help I'd greatly appreciate it

    Thanks again!

  6. #16
    The rulebook gives track curvature. You could start there to get acceleration and deceleration values.

    I think most schools set top speed between 70-80mph and take average track speed to be between 25-35mph.
    "Good judgement comes from experience. Experience comes from bad judgement."

  7. #17
    I would also strongly recommend simple statistical models. Plot out choices of teams vs points scores (in all events including overall). They will show the historical content of how the different decisions finish in the real world. This is likely to show things like reliability concerns as well as highlighting the difference between fastest cars and overall winners. With this approach you may be the team to miss out on being the first to do _______ properly. However there are usually very few benefits to being the first group to do something right for the first time. Look into the history of things like MP3 players and ground effects race cars and you see that the "firsts" usually dont hold an advantage very long at all. Simple statistics are a great help for these decisions. Especially with 400+ cars a year being made.
    Thanks Kev!

    I actually was going to do what you said above but I came to a brick wall when I tried to find out what other teams have been running. Apart from some intense guesswork from random pictures, it has proven near impossible to get an accurate idea as to what the setup was for each team.

    Is there a resource to see what others have been running? I was at the competition and when i asked questions from some teams, they either BS'ed me or tried to hide their design. COMPLETELY UNDERSTANDABLE!

    Any tips on how I can still do what you mentioned??

  8. #18
    Originally posted by thewoundedsoldier:
    The rulebook gives track curvature. You could start there to get acceleration and deceleration values.
    Sweet thanks I found it

    as reference if anyone needs it, this is from 2010 rulebook

    The following standard specifications will suggest the maximum speeds that will be encountered on
    the course.

    Average speeds should be 40 km/hr (25 mph) to 48 km/hr (30 mph).

    Straights: No longer than 60 m (200 feet) with hairpins at both ends (or) no longer than 45 m
    (150 feet) with wide turns on the ends.

    Constant Turns: 23 m (75 feet) to 45 m (148 feet) diameter.

    Hairpin Turns: Minimum of 9 m (29.5 feet) outside diameter (of the turn).

    Slaloms: Cones in a straight line with 7.62 m (25 feet) to 12.19 m (40 feet) spacing.

    Miscellaneous: Chicanes, multiple turns, decreasing radius turns, etc. The minimum track
    width will be 3.5 m (11.5 feet).

  9. #19
    Senior Member
    Join Date
    Jun 2003
    Location
    Melbourne Australia
    Posts
    762
    Originally posted by Kevin Hayward:
    I agree with Geoff that lap sims are invaluable. You do need to keep in mind that lap sims usually work with a 100% capability of the driver. This can be simply dealt with by applying overbraking (a very common problem) and percentage factors to overall cornering. When choosing fundamentals like this it is good to consider the "best" situation, as well as the "usual" situation.

    Lapsims bring quite a bit of objectivity to decisions within the team as Geoff has indicated. There are however a number of other factors that need to be considered in the general design process....
    Fully agreed Kev (and nice to see you popping up again on the boards). As when you write such things you wonder when to stop, and given my post above was done on the way to work I left out the whole "potential vs execution" thing. (I've already bored everyone with that point on another thread anyway).

    What I find valuable about writing such simulations is not so much finding out which concept offers the greatest performance potential - because with the quite complex FSAE rules most of the better concepts are within a few points of each other anyway. Rather it is the reality check and the sense of priority that the exercise offers. As I've said many times before, if you are finishing 300 points behind the winners, and you find your concept's potential is within 10 points of the winners, then you realize labouring over vehicle concepts ain't going to get you very far.

    The driving force behind RMIT's change to the 450 single concept wasn't a performance potential advantage - in fact all of the sims I've done indicate a UWA style car is worth around 10 points more than an RMIT style car. Rather, we wanted to address a problem we had completing the event. We thought about some alternative concepts, checked they were competitive enough (even if they weren't necessarily the top ranking option on the sim) and then chose an option that we thought best aligned with our resources and skill base.

    Cheers, and good luck Spetsnazos
    Geoff Pearson

    RMIT FSAE 02-04
    Monash FSAE 05
    RMIT FSAE 06-07

    Design it. Build it. Break it.

  10. #20
    Senior Member
    Join Date
    Sep 2002
    Location
    Perth, Western Australia
    Posts
    717
    Geoff,

    I remember some discussions we had in 2004 (great year for the Australian comp). I came away thinking that for all we had done that you guys had come up with the better concept. Unfortunately I think the beauty of that original composite single from RMIT is still missed. The fact that you could get similar performance by doing less was really impressive.

    At the time I ran through the numbers and couldn't see that a single had higher performance, and we had a known quantity with our concept (with a few years of continuous development mapped out). UWA has considered the single cylinder option quite carefully and come very close to switching a couple of times. Usually for the reasons of rear end simplicity.

    In 2004 I would have said you couldn't do much better than a well worked out 4 cylinder with a spaceframe. Not far from the truth given that Woolongong won 2004 Australia, and Cornell won 2005 US. Both spaceframes with turbo charged 4's.

    However rule changes (and development in singles) over the last few years have favoured both composite chassis construction and single cylinder engines (or rather fuel efficient cars). I don't think its been a huge points swing, but probably enough to put the favourite in the single and composite camp.

    I would run the numbers through the sim again, but I think my power curves for singles would be a gross underestimate of what some of the teams are squeezing out now. Seeing the GFR car powersliding at Silverstone was quite impressive.


    In all this conversation though your point about resources and skill bases is the most important. I am now involved as Faculty advisor at Edith Cowan University. It is quite a different environment to UWA, both with the type of resources and students available. I have no doubt that the team has the potential of becoming a top level team, but it wont be able to so by following a UWA (or RMIT) style car.

    Kev

+ Reply to Thread
Page 2 of 22 FirstFirst 1 2 3 4 12 ... LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts