Review of

B. van Lierop, R. Balster, P.W. Paap, Shell roof model, Report for course CT3280 Shell Roofs, Delft University of Technology, Department of Civil Engineering and Geosciences, 2024, online: phoogenboom.nl/B&B_schaal_report_9.pdf

by dr.ir. P.C.J. Hoogenboom, Delft, 22 February 2024

 1. Submitted on time (1 Feb. 2024).

 2. You do not need to include group number, student numbers or instructors on the cover. This is not important to readers. 

 3. The report structure is clear and complete. The report should have a preface. Summaries do not get a chapter number.

 4. The summary is complete. The lines could have been organised in 3 paragraphs; objective, approach, conclusion. In addition, do not end negatively. Few will read past the last sentence of this summary.

 5. The introduction is good, except for the third paragraph. The first sentence of the third paragraph is an objective but the objective is already stated in paragraph 2. Moreover, "... knowledge about ... wood and concrete." is personal and not interesting to the readers of the report. The second sentence of paragraph 3 is a summary and the third sentence of paragraph 3 is a conclusions, which should not be in the Introduction. If the third paragraph is deleted, the introduction is fine.

 6. Good design ideas. Beautiful too. The designs can be much thinner.

 7. Good story ... but the design and the construction method were copied from team 1. A large opening would have made the shell original.

 8. Nice drawings. All dimensions included. Is the shell really 20 mm thick? This is very strong.

 9. The building process is clear. Good photos too.

10. Chapter Hand calculation. Something is wrong in the equation: s should be s/t. There is a typing error too: 0.002 m should be 0.020 m. The team should have applied a checking process. Nonetheless, the conclusion is valid. 

11. More hand calculations could have been made. Compressive stress at the foundation. Tensile stress at the foundation. Buckling load. The teacher was available to assist.

12. Page 16. The formula is applied correctly. However, the concrete compressive stress of 20.4 N/mm2 is very high for home-made concrete.

13. Good linear finite element analysis. I assume that linear buckling is computed with the point load of 4.9 kN. So, the buckling load factor is 419/6 = 70, where 1/6 is the knockdown factor. Since 70 > 1 the shell will not buckle at a 4.9 kN point load.

14. Interesting nonlinear analysis. The stresses are much smaller than in the linear analysis. You would think that the load can be increased ... but it cannot because it also buckles at 4.9 kN. Is this what happened? It is a coincidence; that crushing and buckling happens at exactly the same load. So, the collapse load is 4.9 kN.

15. The analysis is clear and valid. There should be conclusions in the report (see comment 13 and 14). You cannot leave drawing conclusions to the readers.

16. You submitted 6.5 kN as collapse point load, which is much larger than 4.9 kN. I assume that new calculations were made after the experiment. Good that you did this ... better late than never.

17. The cylinder tests are clearly reported. The tensile strength is high. Often it is just 10% of the compressive strength. Was the formula applied properly? (Brazilian splitting test)

18. The point load stress formula is correct. The result of 11 kN is much larger than submitted. Uncertainties are shell thickness, stress concentration next to hole and failure mode. Good that this is addressed in the analysis of the test.

19. Load distribution system is clear. The test set-up photo gives much information. Show it bigger. (Also because I am in the picture, of course. By the way, thanks for saving the teachers toes that day. Is appreciated.)

20. What is the largest distributed load applied? How many bricks? This is an important fact. It is not in the report. I wrote it down at the time. It is 3.5 kN.

21. Excellent load-displacement graph.

22. Page 25. The formula does not work for point loads. This calculation is nonsense. The correct calculation is 5.4 / 2 = n 0.24 + 1, assuming 10 mm thickness => n = 7. So, a dome with a diameter of 7 x 0.7 = 5 m or more, because we could not test it with a higher distributed load.

23. The conclusion states that the applied distributed load is 3.2 kN. This number does not occur elsewhere in the report. The Chapter Conclusion is not for presenting new facts. 

24. The conclusions are interesting and mostly valid.

25. Good references to literature. The shell reader has the name "Notes on shell structures".