ASE 4712 · Capstone I Fall 2026 Mississippi State University

Measuring what the aero package actually does.

Design, calibration, and experimental validation of an aerodynamic‑load measurement system for the Mississippi State Formula SAE vehicle — and for vehicles beyond it.

Team
David Stout
Carlos Medina Gutierrez
Faculty Advisors
Dr. Shreyas Narsipur
Dr. Vivek Khare
Measurands
Downforce & drag
Center of pressure
Platform
MSU Formula SAE
Car 133
Scroll
Problem Statement

We design an aero package we cannot measure.

MSU’s FSAE team does not currently have a system for measuring the real‑world aerodynamic effects of the vehicle’s aerodynamic package.

Every wing, undertray, and endplate on Car 133 is sized from CFD and hand calculation. None of it has ever been checked against a load the car actually produced on track, so the discrepancy between theoretical and computational prediction and real‑world physics stays invisible — and compounds into every design cycle that follows.

Knowing how the aerodynamic package behaves in the real world, and the error against prediction, lets the team make better predictions in future and ideally produce a vehicle that performs on track as intended.

Where the Gap Sits

One side of the comparison does not exist yet.

Predicted

CFD & hand calculation

  • Downforce, balance, wing loads
  • Iterated every design cycle
  • Never checked against track data
Measured

No capability today

  • No live, on‑track dynamic load data
  • Error between prediction and reality unknown
  • Downforce cannot be confidently traded against drag, weight, or cost

This project exists to build the right‑hand column.

Research Question

The question the system has to answer.

How accurately can an instrumented FSAE vehicle measure the downforce, drag, and center of pressure location under combined loading, and can the aerodynamic effects of two wing settings be reliably distinguished?

Statement of Work — signed with advisors
  • Downforcemeasured
  • Dragfrom coastdown
  • Center of pressure, xcpderived
  • Loading conditioncombined
  • Method of validationconfiguration change
  • Correlationagainst CFD
  • Target uncertaintyTBD

Validation is comparative: one component is re‑trimmed between two known configurations and the system must resolve the resulting change in load above its own noise floor.

Scope

What this semester covers — and what it does not.

In scope — this semester
  • Full‑vehicle aerodynamic mapping
  • Drag from vehicle deceleration between two speeds
  • Variation of multi‑element trim
  • Live load sensing, logging, and correction
  • Calibration against known loads
  • Correlation of validated measurements against CFD
  • Accuracy, repeatability, and error
Deferred
  • Tire and suspension load characterization
  • Full competition‑season test campaign
Why It Matters Beyond Formula SAE

A modular alternative to costly experimental testing.

Beyond Formula SAE, a modular aerodynamic‑load measurement system could support the development of prototype, aftermarket, and other dynamic systems under combined loading.

By directly measuring the forces and moments under real operating conditions, the system could help engineers validate computational predictions, compare design configurations, and improve system stability, safety, and efficiency without relying solely on costly experimental testing.

Modularity, maintainability, and scalability are treated as first‑class design requirements, not afterthoughts.

Project Objectives · Agreed with Advisors

Five objectives, in dependency order.

  1. 01

    Establish measurement requirements and practical constraints

    For evaluating aerodynamic loads on an FSAE wing — ranges, bandwidth, mounting envelope, and what the car and the test environment will actually tolerate.

  2. 02

    Develop an aerodynamic‑load measurement system

    Suitable for the selected loading conditions and the intended testing environment.

  3. 03

    Calibrate and independently validate the measurement system

    To characterize its accuracy, repeatability, and measurement uncertainty against a reference the system itself did not produce.

  4. 04

    Evaluate the system’s usefulness for FSAE aerodynamic validation

    Including its ability to locate center of pressure, identify changes in aerodynamic loading, and its limitations in practical use.

  5. 05

    Build a foundational, robust, and adaptable system

    Suitable for future testing and engineering applications beyond the immediate FSAE use case, with an emphasis on modularity, maintainability, and scalability.

System Function Formulation

Defined by its inputs, functions, outputs, and boundary.

Inputs
  • Aerodynamic load
  • Wing setting (slot gap, overlap)
  • Vehicle motion
  • Ambient conditions
System boundary
  1. F1SenseTransduce wing load into an electrical signal
  2. F2ConditionAmplify, filter, and reject vibration noise
  3. F3AcquireDigitize and time‑stamp at an adequate rate
  4. F4LogStore synchronized data from the vehicle ECU
  5. F5ReduceApply calibration; output load with uncertainty
Outputs
  • Calibrated wing load vs. time
  • Stated measurement uncertainty
  • Load delta between wing settings
  • Measured vs. predicted comparison

Excluded from the boundary: full‑vehicle aero balance, drag measurement, tire and suspension load paths. No sensor is named — selection is the outcome of the Objective 02 trade study, not an input to it.

Measurement Approach

An inverted wing, and the change we make to it.

Fig. A — Inverted section, live. Streamlines around the wing with the suction surface facing the road. Colour follows local speed: deep maroon where the flow stagnates at the leading edge, pale steel where it accelerates underneath. Re‑trimming this element is the change the instrumentation has to resolve.

What the system has to resolve

  • Downforce, Lmeasured on the element
  • Centre of pressure, xcpresolved moment
  • Drag, Dwhole vehicle

Downforce is measured directly in the element’s load path, and the centre‑of‑pressure station follows from the moment resolved about a known reference. Drag never appears at the element at all — it comes from the whole car decelerating, which is the next section.

Drag · Method

Drag comes from the car slowing down, not from a load cell.

There is no practical way to put a sensor in the drag load path of an FSAE car. So drag is inferred instead: the car is allowed to coast from one speed to another with the drivetrain disengaged, and the time it takes becomes the measurement. A longer coast means less drag, and the relationship is exact enough to give a drag coefficient from nothing more than two speeds, an elapsed time, and the vehicle mass.

Rolling resistance and driveline losses ride in the same deceleration. They are roughly speed‑independent while aerodynamic drag grows with the square of speed, so running coastdowns across a range of speeds is what lets the two be told apart. Getting that separation right is part of the test design, not an afterthought.

The drag polar

CD = CD0 + k CL2

Each coastdown at a different wing setting gives one pair of lift and drag coefficients. Plotted as CD against CL2, those points should fall on a straight line: the slope is the induced‑drag factor k, and the intercept where CL = 0 is the zero‑lift drag coefficient CD0 — the drag the car carries with no aerodynamic load at all.

C_D0 — intercept at C_L = 0 slope = k C_D C_L²

Fig. B — The fit, schematically. Axes are deliberately unnumbered: this shows the method, and it will be replaced with real coastdown points once the instrumentation is calibrated and the runs are done.

Validated measurements are then correlated against CFD — the simulation is the thing being checked, not the reference the measurement is judged by.

Team & Responsibilities

Two engineers, four roles each.

David Stout

Analysis · Test Design · Calibration · Software
  • Aerodynamic Load Analysis Lead
  • Experimental Test Design Lead
  • Sensor Calibration Lead
  • Software Development Lead

Carlos Medina Gutierrez

Modeling · Instrumentation · Vehicle Test · Manufacturing
  • Aerodynamic Modeling & Simulations Lead
  • Instrumentation Design Lead
  • Vehicle Test Engineer & Driver
  • Manufacturing Lead
Faculty Advisors

One covers what we measure, one covers how.

Dr. Shreyas Narsipur

Department of Aerospace Engineering

Applied aerodynamics · Flow sensing and control · Unmanned air vehicles · Flight dynamics and control · Bio‑inspired flight

Covers the aerodynamics of what we are measuring: defining wing loading, designing the test matrix, and interpreting the effect of a wing‑setting change.

Dr. Vivek Khare

Department of Aerospace Engineering

Space structures · Aero‑structural analysis · Structural health monitoring · Non‑destructive assessment of materials and structures

Covers how we measure it: load sensing through a structure, sensor placement, aero‑structural response, and calibration of an instrumented component.

Reporting
cadence
WeeklyIn‑person with Dr. Narsipur, through his point of contact Mehdi Ouhabi
Bi‑weeklyIn‑person with Dr. Khare — twice per month minimum
MonthlyWritten updates to the class and to both advisors, plus this website maintained as work progresses
Project Timeline · Fall 2026 – Spring 2027

Course deadlines are fixed. The work plan hangs off them.

Complete In progress Planned Deadline Project milestone Advisor meeting Today

Swipe sideways to see the full timeline →

Deadlines come from the ASE 4712 Fall 2026 schedule and the AIAA Region II 2027 Student Conference call for papers. Project work is the team’s current plan after the October 2026 scope change; bar shading shows status as the team reports it. Advisor cadence is the reporting agreement recorded in the Statement of Work.

Follow the Project

This site is the running record.

ProgressWork accomplished, objective by objective, with the evidence behind each claim. UpdatesAnnouncements, monthly status reports, presentations, and references. End GoalWhat a finished, trustworthy measurement system looks like — and how we will know.
Work Accomplished

Progress

Updated as work lands. Each objective carries only what has actually been completed — nothing is marked done ahead of the evidence.

Objective status

01Establish measurement requirements and practical constraints20% · In progress
02Develop an aerodynamic‑load measurement system20% · In progress
03Calibrate and independently validate the systemNot started
04Evaluate usefulness for FSAE aerodynamic validationNot started
05Build a foundational, robust, and adaptable system5% · In progress

Percentages are the team’s estimates as reported in the September monthly update and reviewed by both advisors on 1 Oct 2026.

Log

27 Oct 2026In review

Scope and direction briefing prepared for advisors

After the September update the team researched the calibration‑matrix approach alongside the dynamics model, weighed hardware, cost, and schedule, and prepared a short briefing on project scope and direction. It goes to advisor review on 2 Nov 2026.

The scope shown on the home page is unchanged until that review. Whatever is agreed will be posted here with a dated scope history, so the original plan stays on record next to the revised one.

30 Sep 2026Complete

Theoretical dynamics model and noise‑sensitivity study

First version of a model that locates the aerodynamic centre of pressure from on‑car measurements. Downforce comes from the tire normal forces, vehicle weight, and vertical acceleration; drag from a drag polar calibrated to the car; and the aerodynamic pitching moment from a moment balance about the centre of gravity that includes pitch, wheel, and engine inertia. The centre of pressure is reported where the resulting line of action meets the ground plane.

A synthetic study (1,500 trials per case, with illustrative rather than sensor‑specific noise) showed the main risk: at low downforce the ground‑plane location becomes unreliable. That matters because many of the car’s important handling cases are slow, low‑downforce corners.

  • Steady baseline (180 N downforce): 95th‑percentile position error about 0.06 m
  • Low downforce (4 N): 95th‑percentile position error about 33 m, so the location is withheld below a cutoff
  • Alternative raised by Dr. Khare: a 3 × 3 calibration matrix mapping sensor channels directly to lift, drag, and pitching moment
05 Sep 2026Complete

Advisor meeting — scope and reporting agreed

Met with both advisors. The Statement of Work was revised out of that meeting: the system is now described as an aerodynamic‑load measurement system rather than a wing‑load one, Objective 04 explicitly includes locating the centre of pressure, and the broader application is framed as dynamic systems under combined loading. Reporting cadence and points of contact were fixed at the same time.

  • Weekly in‑person meetings with Dr. Narsipur, through his point of contact Mehdi Ouhabi
  • Bi‑weekly in‑person meetings with Dr. Khare — twice per month minimum
  • Monthly written updates to the class and both advisors
05 Sep 2026Complete

System function formulation drafted

The system is now defined by its boundary rather than by its hardware: five functions from sensing through data reduction, with inputs, outputs, and explicit exclusions. No sensor is named — selection stays the outcome of the Objective 02 trade study.

04 Sep 2026Complete

Project definition locked

Statement of Work signed by both team members and both faculty advisors. Problem statement, research question, five objectives, and the individual responsibility split are now the agreed scope of the project.

  • Statement of Work — submitted
  • Proposal presentation — delivered
  • Project website — published (this site)
28 Aug 2026Complete

Team formed and advisors secured

Two‑person team confirmed. Dr. Shreyas Narsipur and Dr. Vivek Khare agreed to advise, giving the project coverage in both applied aerodynamics and experimental instrumentation.

Next up

Advisor review · 2 Nov

Scope and direction review

Walk the advisors through the October briefing and agree the scope, the near‑term experiments, and how success will be judged.

Course · October

Monthly update — October

Written status report and oral update to the class and both advisors.

Course · 20 Nov

End‑of‑semester write‑up

Problem definition, research to date, the agreed scope, and the plan through the spring.

Announcements · Reports · Presentations

Updates

Newest first. Every course deliverable is archived here once submitted, alongside the announcements that go with it.

01 Oct 2026Submitted

Monthly update — September

The September report covered the first theoretical dynamics model for locating the aerodynamic centre of pressure from tire normal forces and vehicle accelerations, and a synthetic noise‑sensitivity study showing that the location becomes unreliable at low downforce.

It also set out the problems ahead: choosing between the dynamics model and a 3 × 3 calibration‑matrix approach suggested by Dr. Khare, and calibrating the sensors with the car on stands, loaded through ropes, pulleys, and known weights. Reported progress: Objective 01 at 20%, Objective 02 at 20%, Objective 05 at 5%.

Reviewed and signed by Dr. Narsipur and Dr. Khare on 1 Oct 2026.

18 Sep 2026Submitted

Article Review #1

Submitted. The reviewed source and a short summary will be added to the references below.

05 Sep 2026Announcement

Advisor meetings held — communication methods and requirements established

The team met with both faculty advisors to agree how the project will be reported and how often. Standing meeting cadence is now fixed for the semester:

  • Dr. Narsipur — weekly. In‑person meetings through his point of contact, Mehdi Ouhabi.
  • Dr. Khare — bi‑weekly. In‑person meetings, twice per month minimum.
  • Monthly written updates to the class and to both advisors, plus this website maintained as work progresses.

The Statement of Work was revised following the meeting and the objectives, scope, and timeline on this site now reflect the updated document.

05 Sep 2026Delivered

Project proposal presentation

Proposal delivered covering the problem statement, refined research question and scope, the five objectives, the system function formulation, advisor coverage, individual responsibilities, and the semester timeline.

04 Sep 2026Submitted

Statement of Work & project website

Scope agreed and signed. The team website is published and will be maintained for the rest of the semester — progress and announcements post here first, and monthly status reports are archived below as they are submitted.

Deliverable schedule

DeliverableDueStatus
Statement of Work04 Sep 2026Submitted
Proposal presentation04 Sep 2026Delivered
Project website04 Sep 2026Published
Article Review #118 Sep 2026Submitted
Monthly Update — September30 Sep 2026Submitted
Article Review #202 Oct 2026Pending
Monthly Update — October30 Oct 2026In progress
End‑of‑semester write‑up20 Nov 2026Scheduled
AIAA Region II Student Conference — abstract05 Feb 2027Scheduled
AIAA Region II Student Conference — manuscript05 Mar 2027Scheduled
AIAA Region II Student Conference — presentation05–06 Apr 2027Scheduled

References

Sources reviewed for the literature survey and the two article reviews are listed here as they are read. Article review sources will be added once both reviews are complete.

  1. [1]Ulbrich & Reed (2019)Calibration and Data Analysis Recommendations for Three‑Component Moment Balances. NASA Ames Research Center, ARC‑E‑DAA‑TN66180. Cited in the September update for the calibration‑matrix approach.
  2. [2]Zhao, Zhao, Liang & Lin (2025)Research on Calibration Methods and Experiments for Six‑Component Force Sensors. Journal of Marine Science and Engineering 13(9), 1811. doi:10.3390/jmse13091811
  3. [3]Andrade Chavez, Traversaro & Pucci (2019)Six‑Axis Force Torque Sensor Model‑Based In Situ Calibration Method and Its Impact in Floating‑Based Robot Dynamic Performance. Sensors 19(24), 5521. doi:10.3390/s19245521
  4. [4]Faber et al. (2012)A force plate based method for the calibration of force/torque sensors. Journal of Biomechanics 45(7), 1332–1338. doi:10.1016/j.jbiomech.2012.01.024
Definition of Done

End Goal

A measurement system is only finished when someone else can trust a number it produced. These are the conditions we are building toward.

Acceptance criteria

  • It measures downforce and drag, and locates the centre of pressureDownforce and centre of pressure from the instrumented component, drag from vehicle deceleration, all under combined loading rather than one axis at a time.
  • Its accuracy is characterised, not assumedAccuracy, repeatability, and measurement uncertainty are quantified against an independent reference.Numeric targets: TBD with advisors
  • It resolves a real configuration changeThe change in aerodynamic loading between two trim configurations of the same component is distinguishable above the system’s own noise floor.
  • Its limits are documentedThe conditions under which the system stops being trustworthy are stated plainly, not discovered later.
  • Its numbers stand next to the CFDValidated measurements are correlated against the simulation, and the disagreement between them is quantified rather than argued about.
  • It outlives this projectFoundational, modular, maintainable, and scalable enough for the next FSAE team — and adaptable to dynamic systems that are not this car.

What success buys the team

Immediate

A measured value of aerodynamic load on Car 133, and a quantified error against the CFD the aero section already runs. The first real calibration of the team’s predictions.

Next season

Aero decisions made against measured data instead of simulation alone — wing trim, slot gaps, and undertray work checked on track rather than argued about.

Beyond FSAE

A transferable method for measuring loads and moments on prototype, aftermarket, and other dynamic systems under combined loading, without relying solely on costly experimental testing.

Scope across two semesters

Capstone I · Fall 2026

Build it and prove the numbers. Problem definition, literature and background, measurement requirements, sensor and instrumentation trade study, mount design and fabrication, DAQ and software development, and calibration with an uncertainty budget.

The intent is to finish the fall with most of manufacturing complete and most of calibration already behind us, closing on static validation against known loads.