<HTML>Hi Bill
I know your problem well. I use mathcad to get the initial valve timmings and cylander deminsions. MathCad is the thing to use though. Mush better then my old calculator. And the main reasion I have not taken the time to work on the calculator. It is also a bit limited as I was not planning on ever getting close to the critical point and so never implemented the region that covers the higher pressure range. Figuring compression is sometimes a problem as IFC-67 formulations upper temperature limit was around 1480 F. If I rember correctly I gave the temperature and pressure limits in the help file that is in the zip file along with the program.
Anyway. With MathCad I can automate calculations. I can have it calculate arrays of cycles for instance. I calculate an array of cycle for the high presure stage. And go through that array. Find the max volume at end of expansion. Then figure the steam usage of each cycle normalized to that volume. I calculate a factor for each cycle that would have all operatoring in a common volume. Thats about as clear as mud. OK. The HP stage is going to have some volume though with clearance and stroak variations it may be changing for each cycle, But the volume is a given. Fixed or set by peremeters that deturmin the cycle. The steam density will vary from cycle to cycle. What I need is all the cycles to be normalized so that I can figure the relative steam quanties going to the next stage. I am trying to obtain the next stage size so that it will always use all the steam from the previous stage.
Then I run the second stage. It gets normalezed as well. Only the volume factor is figured in as well. I have the exhaust steam properties from the previous stage at this point. I initially use the exhaust properties as inlet to the second stage. A second run is made in which the inlet is the exhaust of the previous stage mixed with the first stage inlet steam. The first stage inlet steam in this mix is what I call makeup steam. Sense the second stage is designed to as much or more steam then the previous and the more is variable accross the power range there is a different amount of makeup steam required on each cycle. So they are all figured again using the mixture is inlet. This is repeted several times to get stable mix relations. This is repeted again for the third stage.
After all that I have three arrays of cycles. 1 for each stage. If I remember roght my arrays contain 100 cycles. But be more or less as needed.
This is all done in one work sheet. Where I can set the inlet inlet steam properties and several parameters like number of cylanders in each stage etc. What I wind yp with is relative stage cylander sizes and initial inlet and exhaust timmings.
I then use thoes valve timmings for the initial simulation perameters. I then am hand tunning thoes by hand . The simulator is being redone in vissim. From the work I did in MathConex, I can say they will change quite a bit from the static analysis in MathCad. But it give a starting point. And really don't know how close my adjusted simulation values will match real engine either. But at least the static mathcad abalysis gives a worm fealing it will work.
Havn't had mush time to work on it. My home computer has been down for a while. Mother board went out. And the upgrade parts I bought (64 bit AMD processor) wont run win 98 any more and having problms getting WIN 2000 or XP to recognize my old highpoint IDE raid drives. The new board has a different rade chip and siftware. But I had a HighPoint card and it boots up fine in dose. But new OS wount upgrade it or even see it on XPs case. Will have to get a new SATA drive for the new system.
I have been looking at ideas for a test rig. Where the cylander can be tracked. I know there are several problems using a pressure transducer. For one thing the heat isolation as the norm that can not take the steam temperature being used. And that isolation will introduce some delays in sensing. And sense it a loop filled with liquid error will be introduced bt the liquids compressability and/or initia. Temperature is another problem. But I found a non contect sensor that might work at very low speed. It has a settling time of 0.150 seconds.
[
www.me-us.com]
At any rate a test rig might be the best way to go. And it would not have to be a complete enngine. I figure a single cylander set up could do the job. Kind of like the artical linked up on another thread on SACA site.</HTML>