This document gives some background to my career and explains how I came to write the book.

My introduction to analysis modelling was in the early 1960s when I was an assistant lecturer at the University of Glasgow in Scotland. At that time my PhD research was concerned with the analysis of shear wall structures - mainly focused on modelling of shear walls with vertical rows of openings. This was a timely topic because many buildings with walls of this type were being built in the UK at that time and there was very little published work on the subject - certainly less than 10 published papers when I started. In the1970s I was given a bibliography with over 1000 published references on the analysis of shear wall buildings. Therefore I was fortunate to be in at the early stages of a subject for which there was significant potential for developing ideas.

In January 1963 I went to a course on Stress Analysis at the University of Swansea in Wales. Oleg Zienciewicz had just come back from a sabbatical visit to the University of Berkeley in California with ideas about a new concept in structural analysis - finite elements. Although such methods had been used in the UK aircraft industry for some years, it was Ray Clough and his colleagues at Berkeley who had started to apply it in civil engineering contexts. The course at Swansea was, effectively, the introduction of finite elements to civil engineering in the UK. It was clear that this had potential application to shear walls with openings.

There were some published textbooks on matrix structural analysis at that time but none of them showed how to implement the methods in the form of a computer program. We had do develop our own techniques. There was only one computer on the University of Glasgow campus with access only by paper tape - no terminal access. I could only get one run of a program per day. By modern standards this is viewed as a very unsatisfactory environment for software development, but it was great fun and it I was able to write programs for special frame analysis and for plane stress finite element analysis for application to the wall problem.

A major issue promoted in Modern Structural Analysis is that models should be carefully validated (Section 3.4). One thing I remember doing in this respect was to validate the experimental work that I did in order to assess its potential accuracy. I made a list of all the assumptions involved in the physical models and tried to assess the likely overall error in the measured values. This was my introduction to validation.

I also developed at that time the principle of using the simplest practical model (Section 2.2.1). I found that a model of a shear wall with openings with a relatively small number of frame elements gave as good, and in some circumstances better, results as a plane stress element model with a relatively large number of elements. This had particular significance because the cost of computer processing time could be significant.

After graduating with a PhD in 1966 I went to Canada to work in structural engineering practice for a year and then spent 18 months as a design research engineer with the Portland Cement Association in Skokie, Illinois continuing to investigate methods of analysis for tall buildings.

Over the years in academic posts at Glasgow, Paisley and latterly at the University of Strathclyde in Glasgow as Professor of Structural Engineering, my research interests in the design of buildings involved a range of issues including: monitoring of performance, fire engineering, design of masonry structures, computer aided design and structural analysis modelling. I kept in touch with structural analysis via consultancy work and teaching. I came to realise that conventional textbooks covered the mechanics but, like the texts on matrix analysis that I used in the 1960s they did not well explain how to do analysis i.e. how to work with analysis models. In 1990 Ellis Horwood published Analytical Modelling of Structural Systems which I had written to try to fill the gap. This provided a balance between theory and modelling issues. It did not however provide a comprehensive treatment of the modelling process.

By the way, I no longer use the term 'analytical modelling' because I found that some people use this to mean solving problems directly from the differential equations (i.e. a continuum rather than an element approach).

In the early 1990s a colleague and I at Strathclyde were awarded a research council grant to support an artist, Jamie McCullough, to work with us on an investigation into innovative design in structural engineering. Amongst other relevant projects, Jamie had designed and built a very interesting timber bridge at Eton College. He was an excellent sculptor with a very good understanding of structural behaviour and a good technical background ( he sadly died in 1999). The idea was that he would seek new structural forms and help us in a study of the fundamental principles of innovative design. We interviewed a number of leading designers and published a paper on our results. My work on computer aided design had led me to consideration of design process; the innovative design project strengthened my developing view that attention to process is a neglected area in engineering education. It was time to apply this to structural analysis.

Traditional engineering education tends to deal with determinate problems i.e. one for which there is a unique solution. The solution process in structural analysis is determinate but the modelling process is non-determinate (i.e. there can be more than one valid solution) and is closely related to the design process. In 1995 I had a paper published in the Structural Engineer on A strategy for use of computers in structural engineering which set out the basic modelling process as described in Chapter 3. Involvement in a study group of the Institution of Structural Engineers led to their 2002 publication Guidelines for the use of computers in structural engineering. Ideas from the work of this group are also included in Chapter 3.

This therefore is how the modelling process described in Modern structural analysis Modelling process and guidance developed. Also included is reclassified and extended information on modelling from Analytical modelling of structural systems.

So that is how I came to write Modern structural analysis. I think that it is a very important book. We need a new view of what are the most important issues in structural analysis. I hope that you will find that Modern structural analysis provides a good introduction to these issues.

Iain MacLeod   August 2005