Student Projects

Showing results for 
Search instead for 
Did you mean: 

University of Wisconsin Madison Formula Student

See Attached File for Full Report

Contact Information

University: University of Wisconsin Madison

Team Member(s):

Faculty Advisors:

Email Address:

Project Information

Title: Time Based In-Cylinder Pressure Data Acquisition

TheUniversity of Wisconsin - Madison


TheUW-Madison Formula SAE team developed a system to collect engine in-cylinderpressure data and post process the data to crank angle resolved pressure tracesusing time-based methods.  With use ofequipment from National Instruments the team was able to make the new,time-based method more accurate than conventional methods.  With National Instruments 9223 module and a studentdesigned post-processor the team of students was able to collect crank angleresolved data without the use of an external encoder. This setup can be adaptedto any engine and produce extremely accurate results. Future work will allowlive monitoring of signals and calculated values like IMEP.

            The UW-Madison Formula SAE team, inorder to increase powertrain output and efficiency developed a time-basedsystem of collecting in-cylinder pressure data. Data was collected usingLabview software and a NI 9223 Module. Once the data was collected it waspost-processed in a student designed Matlab algorithm which produced crankangle resolved data efficiently and accurately. The time-based data wascorrelated to engine crank angle by a 24 minus 1 encoder within the engine. Thepost processor interpolates the signal from the encoder to determine time vs.crank angle. It then interpolates crank angle time to sensor time to sensoroutput. Once the data has been collected it calculates useful parameters suchas IMEP and heat release. The following Labview program was used to collect thedata at various frequencies.

Figure1: Labview data collection program.

            Once the in-cylinderpressure data is collected it is then stored in a data file for postprocesseing. A student built matlab program takes the data file recognizes theencoder pulses and uses them to interpolate instantaneous engine rpm. Manydifferent types of rpm interpolation methods were evaluated. Three differentinterpolation methods; linear, piecewise cubic, and cubic spline, were testedfor computation time and error. Linear interpolation simply creates a linebetween two points on an interval. The constraining conditions are as follows.

The second interpolation method, piecewise cubic, fits athird degree polynomial to the for consecutive points [ Xi-1  Xi Xi+1  Xi+2 ].The polynomial is then evaluated between the points [ Xi Xi+1 ]. This then creates a curvethat is easy to compute but not differentiable. The polynomial is constrainedunder the following conditions.

            The finalinterpolation method that was tested was a cubic spline. A cubic spline is aunique cubic polynomial that at every point is differentiable in the first andsecond derivative. The cubic spline is considered in general to be the mostaccurate interpolation method. However because every point influences everycurve it can be very expensive to compute expecially with large data sets.  A cubic spline is constrained under thefollowing conditions.

            The programwas constructed implementing each interpolation method and the computationaltime of each was tested. The program was then run with 1000 cycles, a worstcase senario of test cycles to average. The test results follow.

Computation time (s)


Average (s)


Cubic Spline












Piecewise Cubic
























Table1: Computation time of different interpolationmethods.


Crank Angle







   Figure 2: Instantaneous RPM vsCrank Angle of 3 different interpolation methods

Linear interpolation is the clear winner followed closelybehind by piecewise cubic. Cubic spline, because of its complexity, takessignificantly longer. Before a method was chosen the error of each method wasinvestigated. Worst case linear interpolation is represented as follows where||F’’(t)|| is the infinity norm of the functions second derivative and h is thesize of the interval.

Piecewise cubic and cubic spline worst case error isrepresented as follows where ||F’’’’(t)|| is the infinity norm of the functionsfourth derivative.

Piecewisecubic and cubic spline have a clear advantage over linear interpolation with h4compared to h2. Using these error equations the different methodswhen using a 24 minus 1 encoder were compared to the conventional 360 degreeencoder. A sample set of data was used to determine the function norms.























360 Degree Encoder


Piecewise Cubic


Cubic Spline








Table 2: Computed RPM interpolation error.

From theerror estimates using a 24 minus 1 encoder with Piecewise Cubic interpolationcan be more accurate then interpolating between a 360 degree encoder. Howeverthese calculations do not include the error from interpolating the time basedICP data to a crank angle. To ensure that this error is negligible a fastenough sampling rate is required. Under normal practices the equipmentssampling rate needs only to be as fast as the encoder pulses, which on anengine operating at 9000 rpm is 54 kHz.

Butbecause it cannot be assumed that your time-based data points will coincidewith crank angle enough points must be sampled to minimize error. If forexample sampling at 54kHz at 9000rpm and all data was interpolated usingpiecewise cubic, after computing IMEP the resulting calculation could be off byas much as 30% off of actual. Therefore a higher sampling rate must be used.National Instruments produces 3 modules that matched our research the 9215,9222, and 9223.  The 9215 module has asampling rate of 100 kHz which could produce 31% error and therefore was notfast enough. The 9222 module had a 500 kHz sampling rate which could produce a.05% error. While the 9223, a significantly faster module at 1 MHz, couldproduce only .003% error. The conventional way to sample data could produce .005%error when calculating IMEP. Therefore if data is collected fast enough it ismore accurate than conventional methods. The following is a sample of the data collectedby the Formula SAE team.

Figure 3: Sample log-log PV diagram.

            Using National Instruments hardwareand software University of Wisconsin Formula SAE team was able to collect anduse in-cylinder pressure data without having to make drastic modifications totheir engine. This development can allow smaller teams who do not have themeans to collect data conventionally to collect valuable data about theirengine. Our goal is to produce a live time-based monitoring system which isessential for engine tuning applications.

Nominate Your Professor: (optional)

<insert nomination. Does your professor use LabVIEW or other National Instruments technology to make learning difficult concepts engaging, interesting, and fun? If so, nominate him/her as an outstanding educator by telling us who they are, what they teach, and how they make learning a better experience for you.>

NI Employee (retired)

Hello there,

Thank you so much for your project submission into the NI LabVIEW Student Design Competition. It's great to see your enthusiasm for NI LabVIEW! Make sure you share your project URL( with your peers and faculty so you can collect votes for your project and win. Collecting the most "likes" gives you the opportunity to win cash prizes for your project submission. If you or your friends have any questions about how to go about "voting" for your project, tell them to read this brief document (

I'm curious to know, what's your favorite part about using LabVIEW and how did you hear about the competition? Great work!!

Also - please add the following information to complete your submission:

- Team Members and contact information

- Problem and Solution

- Use of NI Technology

- Any photos, vidoes, or VI code.

Good Luck, Liz in Austin, TX.