Review of A.S. Binnendijk, A.E. Mikkers, K.B. de Rooij, F. van den Berg, Shell Structures, Shell structure design Bend & Break, report for course CT3280 Shell Roofs, Delft University of Technology, Department of Civil Engineering and Geosciences, 2024, online: phoogenboom.nl/B&B_schaal_report_3.pdf by dr.ir. P.C.J. Hoogenboom, Delft, 19 January 2024 1. The reports structure is clear. The report should have an abstract. The report does not have page numbers, which is really inconvenient. Few language errors. Handed in on time (16 Jan. 2024). 2. The Preface explains what is in the chapters. The Contents explains what is in de chapters. The Introduction explains what is in de chapters. This is really too much. 3. Page 4. This is a short description of what happened. Like you did it for fun. Why would another engineer read this? In technical reports, we start with a problem or opportunity, we continue with an objective or research question, only then we describe what happened, finally we answer the research question. Thus, the author interprets the work. We do not leave interpretation to the reader. It is a technical report, not a novel. The introduction could start with "Shell roofs have been built all over the world. They are efficient in that they use little material and fit well in a natural or urban environment. However, the analysis of shell roofs is not easy. In this project we explored the ... The question we answered is: ..." 4. Page 5. This grid shell is beautiful. Who made these? Who made the photo's? Write this in the report. We give credit where credit is due. This how science works. Without it, nobody would want to be a scientist. Moreover, the copyright law says that we must (https://wetten.overheid.nl/BWBR0001886/2012-01-01). People have lost their job for violating copyright. Famous example: prof. Diekstra, Leiden University. 5. Page 6. Good job in playing around with the materials. Later, the slats could be curved less. This comes back in the discussion. 6. Page 8. The figure needs words, like wood slats 4 x 8 mm spacing 75 mm; galvanised steel mesh diameter 0.8 spacing 5 mm; plastic foil; concrete 5 mm; ... Then it looks professional. 7. Page 9. The thickness t in the formula is measured in the horizontal direction. The thickness perpendicular to the shell is even thinner. The stress in a normal section of the shell is higher. 8. Page 9. What stress is used to calculate the thickness of 0.159 mm? The calculations need to be checkable. It is best to write in one line, the formula with symbols, the formula with numbers substituted and the calculated result. This can be checked easily by a colleague. Before we had computers, we used to do it like this, with a pencil on paper. 9. Page 9. Normal concrete has a mass density of 2400 kg/m3. Something must be wrong in the calculation. 10. Page 10. "... 0.2 MPa ... is acceptable ..." Compared to what? What is the compressive strength of the concrete? 11. Page 11. Good linear buckling analysis. A knockdown factor should have been used. 465 / 6 = 77. (Still not governing.) 12. Page 12. Good nonlinear analysis. However, the elements are a bit large. Can they follow the buckling deformation accurately? 13. All three numerical analyses have been performed. The computations and the results are presented clearly. 14. Page 13. Do you really think that your concrete has a strength of 17 MPa? It is very high. Did you make test cylinders? Did we test these? Wat was the result? Something is mentioned in the discussion but this should be in a paragraph with data and photos. 15. Page 13. The reduction of the strength with a factor 0.2 is very large. Was this wise? Good that it comes back in Discussion. 16. Page 14. The building process is clear. There could have been more photos, for example of the formwork mesh, of the casting process and of details. 17. Page 16. Does the measured displacement (-0.7 mm at 200 kg) agree with the computed displacement (0.2 mm at 2.5 kN/m2, 2.5 kN/m2 x 0.5 m2 = 1.25 kN)? What caused the difference? Perhaps Young's modulus? Perhaps the deflection of the wood foundation? Perhaps a different thickness? 18. Page 17. The explanation that the dome top moved up makes sense. However, it did not move up in the numerical analysis. Perhaps the displacement gauge was mounted up-side-down. 19. Page 17. The formulas are correct. Nevertheless, a reference to the source should have been provided. 20. Page 17. "Filling in the values appropriate ..." Which values? This cannot be checked. This is not how we present hand calculations. 21. Good photo of the failed dome. 22. Page 19. Interesting discussion. I think you are right. In addition, the 1.3 MPa compressive strength should have been used in the calculations. It must have been used in predicting the point load failure. 23. Page 21. In the Conclusion is a research question! Very good. The same question should be stated in the Introduction. 24. Page 21. I agree with the distributed load conclusion. We can build the real shell, provided we make much better concrete. 310 kg / 0.5 m2 = 6.20 kN/m2. 10 cm concrete is 2.40 kN/m2. Snow is 1.00 kN/m2. 6.20 > 2.40 + 1.00. It will be save. 25. Page 21. Something could have been written about building the real size shell. Will the wet concrete stay on the formwork? 26. Page 21. I do not agree with the point load conclusion. The real shell fails at a point load of 2.35 kN x 20^2 = 940 kN. This is the weight of 940 large men standing on each other's shoulders. No need to be this strong. (To make this calculation perfectly valid, we should have loaded the roof with the point load and a distributed load of about 120 kg at the same time.) Received a new version of the report on 31 January 2024. Many of the above comments were incorperated.