Modeling

Routines for solving systems of algebraic and differential equations are first developed, with others to display a time-based 'table' of the simulation results. The second half of the book develops his DYFLO routines for single, staged or distributed unit operations, with many diagrams and examples. In the later chapters, partial differential equations and process control models are developed.

Dr. Franks has been referred to as the "Father of Simulation", and after building some of his many examples, I believe that the phrase is quite appropriate. I first bought this book in 1973, but did the most work with it after the availability of personal computers. It was amazing. The simulations changed the way I thought about unit operations, engineering problem solving, and process control simulations. This thought process shaped the modeling framework that I was developing (building, and tuning interactive process control simulations).

The source code in Modeling and Simulation is provided in FORTRAN IV, which was the current engineering language 'standard'. It is quite compact, again to run on the machines of that time. I found that it was relatively straight-forward (albeit that you will need some other reference books) to convert to other languages in order to use the interpreters/compilers of choice.

I rate this book 4 stars instead of 5 because : 1) it's too expensive and, 2) the FORTRAN IV code and it's structure is dated. The programs needs to be commented and rewritten both more linearly and in a current language such as DELPHI, VB or C++ Builder (or Visual C++). You can sort it out if you are an experienced programmer, but the programs need to be restructured. This book desperately needs to provide at least a CD with the source code.

A company called QMC appears to offer a Windows 95/98 simulation framework for many of Dr. Frank's examples, and many other models and simulations.

"Modeling" also includes a section overviewing the various simulation software packages available to modelers. Though developers like High-Performance Systems, Vensim, Pugh-Roberts, and PowerSim have made product enhancements to date, the sections from each company provide a great introduction to what is out there how each package can be applied.

The most valuable aspect of the book is probably in the case studies and methodological explorations of several authors. A number of key insights are offered as authors reflect upon the successes and shortcoming of the methods each chose to use to explore and develop models in a variety of business and public environments.

This is definitely a must have for any SD library.

IF you are NOT a computational chemist, and are looking for a source of information on NMR *techniques*, this book is not for you. Also, if you are looking for a concise technical description on how to run such calculations: again, this book is not for you, either!

What the entire set of papers distill down to is use of UNIX-based software packages such as Gaussian94 to generate geometry-optimized conformers, whose cartesians are then exported into the SP NMR mode of Gaussian94 to determine the NMR isotropic chemical shift tensors. There are numerous descriptions of the applicability of various methodologies to various chemical systems: polymers, crystals, olefins, etc..

The 23 chapters include such jewels as: "Effects of Hydrogen Bonding on 1H Chemical Shifts", "Ab Initio Calculations of 31P NMR Chemical Shielding Anisotropy Tensors in Phosphates: The Effect of Geometry on Shielding".... and last, but definitely not least, "Recent Advances in Nuclear Magnetic Shielding Theory and Computational Methods". [whew!]

My own (**VERY limited**) work in this area suggests that although Gaussian94 ab initio analysis does a good job in generating isotropic tensors which are predictive of chemical shift, computational accuracy may well break down on the introduction of multiple bonding. (Using HF theory on a 6-31G* basis set.)

If this all sounds Greek, it is to me, too...!