A design-build-test project is used as means of providing an academic-based, industry-focusedexperiential learning opportunity for students in a senior-level aircraft structures course taught inthe Department of Aerospace Engineering at Mississippi State University. Initiated as a paperdesign project in 1998, the project has rapidly evolved into a comprehensive learning experiencewith prototype development and testing as its two major elements. This paper presents thedetails of this experiential learning activity as well as a formative assessment of its effectiveness.
I. Introduction
Prior to the 1950's, it was common for engineering programs to offer in their curricula suchcourses as sheet-metal fabrication, casting, and machine shop. With the advent of computers andmore emphasis on the theoretical side of engineering education, the courses on mechanical artswere gradually phased out with most of hands-on activities reduced and squeezed into thelaboratory courses. This shift in engineering education is mostly responsible for manyengineering graduates to have a very narrow understanding of the product development processthrough which a design concept is transformed into a physical product. Exceptions to this areprimarily those students who have engaged in Co-op learning opportunities in industry.
There has been a renewed interest over the past decade and a half in bringing meaningful hands-on or active experiential learning opportunities to traditional lecture-based courses. Some ofthese efforts have focused on manufacturing1 and construction2 while others have focused on abroader design experience.3 The improvement in learning by engaging in hands-on activities hasbeen well documented mainly as a result of the pioneering work by Kolb.4
As part of a broad undergraduate curriculum enhancement effort in the Department of AerospaceEngineering at Mississippi State University, three freshman/sophomore introductory courseswere developed and some of the existing junior and senior-level courses were modified toinclude meaningful experiential learning opportunities for the students. The senior-level ASE4623 (Aircraft Structures III) is one such course that was modified as a result of this initiative.
The experiential learning activity in this course started initially as a simple column design projectrequiring students to bend a flat rectangular sheet of aluminum into a composite cross-sectionalgeometry that would enable the resulting column to carry as much axial compression as possibleprior to failure. Although somewhat limited in scope, the column design-build-test projectprovided a great learning opportunity to the students who also tremendously enjoyed the hands-on activity.
Over the span of several years the column project gradually evolved into a more comprehensiveactivity requiring teams of two to three students to design, analyze, optimize, build, and test astiffened aluminum panel. This project has added a unique feature to an otherwise theoreticalengineering science course with the students putting to test what they have learned in class in aspirited team competition.
In the subsequent sections of this paper, the details of the panel design-build-test project, itslearning objectives as well as the assessment of its outcomes are discussed.
II. ASE 4623 - Aircraft Structures III
ASE 4623 is the third in the sequence of three aircraft structures courses in the aerospaceengineering (ASE) curriculum at Mississippi State University. While the first two, taught in thejunior year, are primarily focused on structural analysis of simple and built-up structures fordeflection and stress distribution, the senior-level ASE 4623 is mainly focused on structuralfailure and design as indicated in Table 1 below.
Table 1. Description of topics covered in ASE 4623Topic Number of lectures
General Principles in Design of Aerospace Structures 2Manufacturing and Cost Considerations in Design of Aircraft Structures 2Manufacturability Analysis and Manufacturing Methods 1Elastic and Inelastic Instability of Columns 4Theory of Plate Bending 3Buckling of Flat Sheets in Compression, Shear, Bending, and UnderCombined Stress Systems
4
Local Buckling Stress for Composite Shapes 1Crippling of Composite Shapes and Stiffened Panels in Compression 8Basic Principles of Design Optimization and its Application to AircraftStructures
7
Failure Analysis of Aircraft Wing and Fuselage Structures 4Combined Stresses. Theory of Yield and Ultimate Failure 3Membrane Stresses in Pressure Vessels 3
In compliance with the departmental goals as well as the ABET 2000 Criterion 3, the courselearning objectives include the following:
To convey the necessary knowledge leading to the students' understanding of
• various design methodologies/criteria and the impacts of early design decisions onmanufacturability, life cycle costs, and other aspects of aircraft structures.
• theories governing the elastic and inelastic buckling of columns.• theories governing the bending and buckling of thin sheets.• theories governing the compressive failure of thin-walled sections and stiffened panels.• basic principles of design optimization theory.• theories of yield and fracture failure.
• design, analyze, optimize, build, and test a built-up structure.• submit assignments as technical reports, and to properly explain the analysis procedure,
interpret the results, and suggest design changes to enhance the desired characteristics.• function in a team environment with a high degree of professionalism.
The student's final grade in the course is based on his/her performance on an average of sixindividual and team assignments (18%), design-build-test (DBT) project (15%), three in-classtests (42%), and the final exam (25%).
III. DBT Project Description
Stiffened panels for fuselage or wing structures as shown in Fig. 1 are typically designed basedon a combination of in-plane loads with longitudinal stiffeners (stringers) designed to support theskin in axial compression and tension.
Fig. 1 Example of a skin-stringer (stiffened panel) structure
In this project, students have to contend with factors affecting structural performance of stiffenedpanels including sheet thickness and engineering properties of the skin material as well as thegeometry, size, quantity and arrangement, and material properties of stringers. They also have toaddress manufacturability and production cost by considering the complexity of stringergeometry, the quantity of stringers, as well as the number of fasteners used for skin-stringerassembly. In aircraft structures, weight is always a major design consideration.
The DBT project incorporates more than 60% of the topics covered in ASE 4623 and aims atproviding the students a meaningful experiential learning experience while directly addressingthe majority of ABET 2000 Criterion 3 as identified in Table 2 below.
The project learning objectives include:
• Enhance the understanding of subjects studied in ASE 4623 through development ofviable panel design concepts and analysis of their performance characteristics.
• Apply the fundamental principles of optimization theory to an aerospace engineeringproblem with quantitative objective function and design constraints.
• Gain limited experience with sheet metal forming, hand tool operation, and manualassembly of mechanically fastened structures.
• Become better familiar with laboratory testing as means of design validation.• Improve teamwork and communication skills.
Table 2. ABET 2000 Criterion 3 - Program Outcomes and Assessment • an ability to apply knowledge of mathematics, science, and engineering.• an ability to design and conduct experiments, as well as to analyze and interpret data.• an ability to design a system, component, or process to meet desired needs.• an ability to function on multi-disciplinary teams.• an ability to identify, formulate, and solve engineering problems.• an understanding of professional and ethical responsibility.• an ability to communicate effectively.• the broad education necessary to understand the impact of engineering solutions in a
global and societal context.• a recognition of the need for, and an ability to engage in life-long learning.• a knowledge of contemporary issues.• an ability to use the techniques, skills, and modern engineering tools necessary for
engineering practice.
For this project, the students are divided into multiple teams of two and sometimes threestudents. The teams work independently of each other although they are not restrained fromcommunicating on topics of mutual interest.
III.1 Design Problem Statement
A 24 in. x 18 in. rectangular aluminum sheet is stiffened in the short direction by an unspecifiednumber of formed stringers with identical cross-sectional geometry. The sheet and stringers areall made of 0.032 in.-thick 2024-T3 aluminum (bare sheet), and are assembled using 1/8 in.-diameter 5052 aluminum cherry rivets.
Determine the shape, size, quantity, and spacing of stringers as well as the required rivet spacingfor a minimum-weight panel to support an ultimate axial compressive force of 15,000 lb withoutfailure for at least 3 seconds. Assume the panel has end fixity of 1.5 along the loaded edges andis free along the unloaded sides. The limit load is based on a factor of safety of 1.25.
Additional design considerations include the following:
• While local buckling of skin and/or stringer is permissible prior to panel failure, the inter-rivet buckling instability is not.
• Stringer design concepts must be manufacturable through manual operation of the sheet-bending press in the lab. Panel manufacturing cost is determined based on the number andcomplexity (number of corners) of stringers and the number of fasteners used.
• Each student will be given an 18 in. x 30 in. sheet to cut and fabricate the necessarynumber of stringers. Thus, the developable width of each stringer cross section is limitedto 30/Nst in., where Nst is the number of stringers used.
• The number of stringers is limited to the range of 2 ≤ Nst ≤ 7.• Stringer height cannot exceed 2.25 in.• No portion of the stringer near the free edge should extend beyond the edge of the sheet.• The maximum distance from each loaded edge to the nearest rivet should not be less than
3/8 in. or greater than 5/8 in.
III.2 Structural Analysis
Based on the topics discussed earlier in the course, the student teams begin the project bydeveloping several design concepts for the cross-sectional geometry of the stringers. Forselecting the appropriate shape, the students must consider the failure of thin-walled columns inthe form of crippling, buckling, or the combination of the two as described by the Johnson-Eulerformula5
f = cs − cs2
4 2Ec
Le
2
(1)
where Le/ is the effective slenderness ratio, Ec is the modulus of elasticity of the material in
compression, and cs is the crippling stress, which for a multi-corner section is obtained as
cs = 0.56 cygt2
A
Ec
cy
0.5
0.85
(2)
where t is the wall thickness, A is the cross-sectional area, cy is the compressive yield strength
of the material, and g is the shape parameter determined by dividing the cross section intomultiple angle elements. In the case of a multi-corner section, cs cannot exceed 0.8 cy.
The students in each team work together to develop a FORTRAN program to computerize thestructural analysis and the calculation of failure load for a stiffened panel under uniform axialcompression. The theory as well as the analysis method has previously been covered in thecourse, and the students generally have very little difficulty with this part of the project.
By the time the students take ASE 4623, they are reasonably proficient in at least oneprogramming language. However, to help those students who are not familiar with FORTRAN,a handout containing some examples is provided.
III.3 Design Optimization
The optimization problem is to find the optimal values of design variables that would minimizethe panel weight while satisfying all the design constraints. The objective function is the panelweight and since the skin and stringers are made of the same material, it is possible to express itin terms of the cross-sectional area or volume of the stiffened panel.
In addition, the stringers must be spaced uniformly in such a way that the web section betweentwo adjacent stringers would have the same buckling strength as the skin flange near the freeedges.
The continuous design variables include stringer flange and web dimensions as well as thefastener spacing. The number of continuous design variables, hence, depends on the geometricshape of the stringer and can vary from one team to another as indicated in Fig. 2.
The discrete design variables that are controlled outside the numerical optimization loop are thecross-sectional shape and the number of stringers. This means that each team would have tooptimize the panel for a given shape and number of stringers, and repeat the optimization processby changing the number of stringers as well as the cross-sectional shape. By examining theoptimization results for different combinations of shape and number of stringers, the studentsdetermine the minimum-weight or "optimal" panel design.
Fig. 2 Examples of stringer cross-sectional geometry and the corresponding design variables, X
Each design variable is typically limited by the lower and upper bounds or side constraints. Priorto deciding on appropriate values for side constraints especially the lower bound, each student isgiven an opportunity to practice forming various shapes using the sheet-forming press at thedepartment's laboratory. While becoming better familiar with the manual sheet-formingoperation, each team realizes the need to consider the manufacturing constraints and how theyshould be integrated into the design optimization analysis.
For design optimization, the students combine their previously developed and tested FORTRANprogram with a general-purpose design optimization program, DOT.6 To facilitate this task, thestudents are provided with a detailed handout explaining the procedure for coupling the twoprograms. In addition, a minimum of three class-lectures is devoted to detailed discussions on
analysis and optimization programming and the related issues. Furthermore, the students areprovided with a "shell" program, which contains the code needed to interact with theoptimization package. Hence, they are not required to write the entire program, but only the partthat deals with the structural analysis and the equations describing the objective function and thedesign constraints.
Following the completion of the design optimization task, each team identifies its optimal designby specifying the shape, size, and the number of stringers that according to the team’s analysisshould be the minimum-weight panel that can support the design ultimate load of 15,000 lb priorto failure.
Upon verification of the optimization results, the members of each team proceed to fabricatetheir own panels based on the dimensions obtained from the optimization solution.
III.4 Manufacturing and Testing
In this phase of the project, each student is required to manufacture a panel that closelyresembles the team’s optimum design in terms of configuration and dimensions.
The 18 in. x 24 in. aluminum sheets used for the panel fabrication are precut with each studentreceiving one sheet. To make the students aware of possible flaws in the manufacturing process,they are asked to take measurements of the panel skin and fabricated stringers and to record thisinformation on the measurement worksheets to be submitted with their project reports. Takingnote of such measurements is helpful to the students in reducing flaws in their fabricated panelswhile at the same time the recorded information can be used as means of measuring the quality ofpanels produced. Although for most students this is the first time that they are working with sheetmetal, they are very careful with their measurements and the bending operation of aluminum stripsinto stringers (see Fig. 3).
After the stringers are made, they are attached to the skin using the specified type of fasteners (i.e.,cherry rivets) at the spacing determined by the students in their design analysis. This is one taskthat really helps the students feel the impact of their design decisions. For example, depending onthe shape of the stringer, it is possible to have more than one option for attaching the stringers tothe skin using either one or two rows of fasteners per stringer (see Fig. 2). Some students givemore consideration to manufacturing ease and choose one row of fasteners while others hope tocapitalize on additional strength offered by using two rows of fasteners. The students make thisdecision during the design optimization process and are not allowed to make any changes to thedesign during the manufacturing phase.
All the drilling of holes and installation of fasteners will take its toll especially on those studentswho opted to use two rows of fasteners per stringer. The students use a hand drill to drill the holesand a manual pop-rivet gun to fasten each stringer to the skin. Special care is taken prior todrilling the holes to make sure the stringers are properly aligned both with respect to the loadededges of the panel as well as the neighboring stringers. The students receive ample warning abouthow the manufacturing quality can influence the load-carrying capability of the panels and evencontribute to their premature failure.
(a) (b)Fig. 3 Students marking an aluminum strip (a) prior to bending it into a stringer (b)
Following the skin-stringer assembly, each panel is prepared for testing by potting the loadededges in a fiber-filled polyester resin mixture. Once ready for testing, each panel is carefullyplaced in the hydraulic compression-testing machine and loaded gradually to determine itsresponse at various load levels and to record its failure load (see Fig. 4). During this phase of theproject, the students are instructed on the proper operation of the testing machine and the properplacement of the panel between the loading platens so to avoid placing the panels in bending.
(a) (b)
Fig. 4 A panel placed in the testing machine prior to loading (a) after failure (b)
III.5 Design Report and Presentation
Each team is required to submit a single written report with the percent contribution of eachmember clearly marked on the coversheet. A member selected as the team leader ensures that
team members are given proper credit for their contributions. The students are given sufficientinstructions on the format of the report and the information it must contain. Upon submission ofthe project reports, one or more members of each team will present a 10 to 15-minute PowerPointpresentation of their project including key observations and the lessons learned. At the end of eachpresentation, the students are asked a few questions related to the project, the assumptions theymade and elements that contributed to the discrepancy between the measured and predicted valuesof failure load in the case there is a considerable difference between the two numbers.
IV. Evaluation Method
Although this is a team project, the students are graded individually and it is possible formembers of the same team to receive different grades based on their contribution to the overallproject as well as the quality and measured performance of their fabricated panels in laboratorytesting. The evaluation criteria are stated in the project assignment and all the students know thegrading scheme as described below.
Report: The written project report makes up 50% of the project grade, and in it the studentsprovide a detailed description of the project; explain the design approach used; the failureconditions examined and accounted for in the design analysis and optimization; the designoptimization problem (identify the objective function, design constraints, and design variables);the reason for using a particular cross-sectional shape for the stringers; the size and quantity ofstringers and form of attachment; the comparison between predicted and measuredcharacteristics (dimensions, weight, strength, etc.); and most importantly, a discussion of lessonslearned.
Design Quality: Counting as 15% of the grade, the design quality was a major discriminatorbetween individual panel designs. It took into consideration the measured strength, measuredweight, and manufacturability using the design quality measure DQ defined as
DQ = ˜ S + 0.5 ˜ W + 0.5 ˜ M (3)
where ˜ S ≡ strength index, ˜ W ≡ weight index, and ˜ M ≡ manufacturability index, with thecalculation of each index shown in the next section.
Panel Fabrication and Testing: The panel fabrication and testing makes up 20% of the grade. Inthis case, the recorded data on the individual student’s worksheet are used to determine theuniformity and quality of stringers as well as the panel assembly bearing in mind that thestudents are not experienced technicians. A visual inspection of the fabricated panel is alsoconducted prior to assigning a grade for this part of the project.
Oral Presentation: This activity accounts for 15% of the project grade. Using a technicalpresentation format, representative(s) of each team provide a brief summary of the projectincluding the design approach, the failure analysis and optimization, the comparison betweenpredicted and measured panel characteristics, as well as the lessons learned.
The DBT panel project has been a permanent part of ASE 4623 course for the past four years.During this period a number of modifications were made including changes to the designultimate load as well as restrictions on stringer design configuration. We have also experimentedwith the way the students are teamed up as well as the number of students in each team. In thissection the results of the most recent DBT project performed in Spring-2002 semester arediscussed with focus on grading scheme and project assessment.
The eleven students in the course were grouped into five teams. The design concepts for stringergeometry as well as the skin-stringer attachment method proposed by each team are described inTable 3 below. In addition, the optimal dimensions for the selected stringer geometry are givenin columns 2 through 4 of Table 3.
Table 3. Design concept and optimal stringer dimensions
TeamSkin-Stringer
Concept X1, in X2, in X3, in
1 0.510 1.100 0.980
2 0.488 1.033 0.902
3 0.467 1.031 1.018
4 0.263 1.169 1.278
5 0.500 1.000 -
The measured strength, weight, part count, rivet count, and intricacy of each panel are shown inTable 4. For the 11 panels tested, the mean failure load is found to be 16,711 lb with a standarddeviation of 1663.7 lb (COV = 10%) while the mean panel weight is 2.41 lb with a standarddeviation of 0.093 lb. Thus, the average strength-to-weight ratio for the designed panel is 6,934.
Table 4. Measured characteristics of designed panelsQuality Measure
Team Member Strength, lb Weight, lb Part Count Rivet Count Complexity
The predicted and measured values for failure load and panel weight obtained by each team arecompared in Table 5 below. The mean, standard deviation (SD), as well as the differencebetween the measured and predicted mean values are shown in the table.
Table 5. Comparison of predicted and measured characteristicsFailure Load, lb Panel Weight, lb
Predicted Measured % Diff. Predicted Measured % Diff.Team Value Mean / SD in Mean Value Mean / SD in Mean
The discrepancy between the predicted and measured failure loads have historically been tracedto the following factors:
• Underestimation of effective skin width used in the calculation of panel failure load.• Underestimation of the stringer failure load due to a conservative modeling of boundary
condition along the loaded edges of the panel.• Non-uniform loading of the panel in the testing machine.
The grade distribution for design quality is shown in Table 6 below with the normalized indices˜ S , ˜ W , and ˜ M found as
˜ S = MinMS
MSmax,1
10 (4-a)
˜ W =10
W* 1 −MW
MWmax
(4-b)
˜ M =10
M* 1−PC
PCmax
+ 1 −
RC
RCmax
+ 1 −
SC
SCmax
(4-c)
where the factors used in each equation are defined as follows:
• Measured Strength, MS (MSmax = 15,000 lb, specified design ultimate load)• Measured Weight, MW (MWmax = 0.1 (30+24)(.032)18 = 3.1104 lb, the total amount of
material that could be used in panel construction)• Part Count, PC = number of stringers plus the sheet (PCmax = 8)• Rivet Count, RC = number of rivets used in panel assembly (RCmax = 7 (2)30 = 420, two
rows of rivets for each stringer at 30 rivets per row)• Stringer Complexity, SC = number of corners in each stringer (SCmax = 6)
Each normalized index can have a maximum value of 10. In Eq. (4-a) an absolute scale is usedwith no bonus for strength higher than the specified design ultimate load. In contrast, Eqs. (4-b)and (4-c) use a relative scale, rewarding the student with the highest score in each category anindex of 10. In this case, W* and M* represent the largest value for the term inside the bracketsearned by any individual student in class. The manufacturability index considers the combinedinfluence of part count, rivet count, and stringer complexity.
Table 6. Grade assignment for design quality, DQ as defined in Eq. (3)
The grade distribution for fabrication and testing is shown in Table 7 below. The grade forfabrication is based on the following factors:
• Stringer Uniformity, SU = sum of standard deviations on stringer dimensions as reported onthe individual student's worksheet (SUmax = 0.25)
• Placement of Middle Stringer, PMS = standard deviation of spacing of middle stringers asmeasured at the centerline of each panel (PMSmax = 3.0)
• Placement of Edge Stringers, PES = standard deviation of spacing of edge stringers asmeasured at the centerline of each panel (PESmax = 0.25)
• Rivet Uniformity, RU = uniformity of rivet spacing as observed in each panel without takingprecise measurement (RUmax = 1.0)
Therefore the grade for fabrication is found as
FG = 2.5 1 −SU
SUmax
+ 1 −
PMS
PMSmax
+ 1 −
PES
PESmax
+ RU
(5)
If the panel is actually placed in the testing machine and tested to failure, the maximum grade of10 is given for that activity. Thus, the grade shown in the last column of Table 7 is the sum ofFG found from Eq. (5) and the grade for testing.
The students generally enjoy the experience associated with this project. The main complaint isthe sourness they feel from the manual installation of more than hundred rivets. For the studentsthis is truly an eye-opening experience when they see that the panel they designed and fabricatedcan actually carry a compressive load that is 6000 to 7000 times the panel weight. The actualfabrication and testing of structures provides a learning experience that cannot be captured in anyother way.
The following is a sample of written comments by the students about the DBT project.
• The applying of concepts to real life applications was fantastic.
• My teammate and I learned many important things from the various phases entailed inthe stiffened panel project.
• I learned that the spacing of the stiffeners will make a difference in the failurecharacteristics of the stiffened panel.
• Building the panel proved challenging when the stage of riveting was reached.
• Troubles with riveting the stiffeners to the drilled panels brought to mind the necessity forkeeping track of the assembly process and thinking about the little details.
• The panels not being perfectly square (at the corners), and thus loading up unevenly,made me realize how much a little variation—such as a sixteenth of an inch—could affectmanufacturing results.
• I did enjoy the project because of the ‘hands on’ nature. I think the design load shouldbe lowered so a wider variation in stringer cross-sections could be made.
• The designer should consider the manufacturing process of his design. We had thatexperience through this project. P
This paper described an experiential learning project in a senior-level aircraft structures coursetaught in the Department of Aerospace Engineering at Mississippi State University. This multi-faceted project is used to provide the students a real-world experience with design andfabrication of lightweight built-up structures similar to those found in modern aircraft and to helpthem gain a deeper understanding of the various topics that are taught in the course. Thestudents have shown considerable interest in this hands-on activity and the feedback they provideyear after year indicates that this project is fulfilling the desired goals and learning objectives.
Bibliography
1. Ortmeyer, T. H., Cunningham, K, and Sathyamoorthy, M., "A Manufacturing Engineering ExperientialLearning Program," Proceedings of the 2000 ASEE Annual Conference & Exposition, St. Louis, MO, June 18-21, 2000.
2. Dennis, N. D., "Experiential Learning Exercised Through Project Based Instruction," Proceedings of the 2001ASEE Annual Conference & Exposition, Albuquerque, NM, June 24 - 27, 2001.
3. Tener, R. K., Winstead, M. T., and Smaglik, E. J., "Experiential Learning from Internships in ConstructionEngineering," Proceedings of the 2001 ASEE Annual Conference & Exposition, Albuquerque, NM, June 24 -27, 2001.
4. Kolb, D. A., Experiential Learning: Experience as the Source of Learning and Development, Prentice-Hall,Englewood-Cliffs, NJ, 1984.
5. Bruhn, E.F., Analysis and Design of Flight Vehicle Structures, Jacobs Publishing Co., 1973.6. DOT: Design Optimization Tools, Version 5.0, Vanderplaats Research and Development, Inc., 2000.
MASOUD RAIS-ROHANIMasoud Rais-Rohani is a Professor of Aerospace Engineering and Engineering Mechanics. He received his BS andMS degrees from Mississippi State University and his PhD from Virginia Tech. Prof. Rais-Rohani teaches coursesin aircraft structures, structural mechanics, and composite materials. His areas of research include structural andmultidisciplinary design optimization and structural reliability.
