software engineering

Professor Young Kwon teaches at the Naval Postgraduate School in Monterey, CA. I used this book in his class and found little difference when I compared it to the first edition. The first edition has some minor errors, but the problem sets for each chapter along with the Matlab programs provided are very much the same. The first chapter has some basic Matlab necessities and the second chapter covers weighted residual methods (strong / weak formulation, least squares, collocation, and Galerkin). There is a one paragraph summary of the Raleigh-Ritz method (derived from Calculus of Variations), which is more accurate but not as popular as Galerkin. I found a much better explanation of these weighted residual methods in an undergraduate numerical analysis book. The remainder of the book covered shape functions, isoparametric functions, numerical integration techniques, boundary and initial value problems, eigenvalue problems, beams, shell and plates, and trusses.

Caution:
- If you are planning to use this book as a reference or to educate yourself. STOP! This is a boring and merciless graduate-level book. At times we (study groups) would spend anywhere from half an hour to an hour trying to decipher the sample problems. Don't waste your time looking for the solutions manual; it is harder to follow than the sample problems.

- If you want to learn the core of FE and can program easily, this book would help -- but it won't be easy.

- If you want to learn FE for practical applications, spend your energy using a commercial or open-source FE package.

More general comments:
- If asked to describe this book, I would reply with "grueling".

- The Matlab code is far from being easy to understand. Maybe it's easy to follow the logic, but there is very little documentation to support the code. You will spend A LOT of time figuring out the variables alone, which has a lot to do with how well you understand the code.

- Someone with a non-engineer background would NOT be able to understand this book. Someone with a technical background and a strong math foundation may be able to. For example, this book contains a few topics from Theory of Elasticity. This theory is taught in advanced solid mechanics classes -- not even the undergraduate Mech-E's cover it. You should also have a good handle on matrix algebra, integral calculus, and basic ODE's.

Bottom Line:
I did not like how this book presents the FE Method. If you really need it...good luck!

This book carries on from the first edition, lots of new theories and programs have been added. The errors in the last edition(1996) have been corrected. This book is a must for people who want to understand the finite element method and for people who want to learn how to program it using 'Matlab'. Matlab is one of the most widely used programming tool.

I have the old edition but i find this new edition refreshing. The new theories and programs help in digesting the book better.

overall this is an indispensable book for people trying to learn to program the finite element method because it gives a clear understanding of how a finite element code works.

thanks,

kartik srinivas

This book is extremely useful, especially for those new to the finite element methods. Despite the fact that it puts a lot of emphasis on matlab, the code is easy to understand and translate into other powerful languages like fortran 90. It took me a few days to start writting my own FEM code after buying the book. I would strongly recommend it even for those with a non-engineering background.