Dan Grove
An independent consultant and trainer in statistical
methods, I specialise in the application of experimentation and statistical
modelling to engineering design and production.
I also write and deliver training to support a broad range of other
statistical work.
|
Contact me by phone on ( |
|
|
|
|
|
|
|
Statistical
Engineering
There are very many
ways in which statistical methods can contribute to successful engineering
design and manufacturing. Statistical
Engineering is the use of statistics to underpin engineering tasks and
processes. The term Statistical
Engineering was probably first used by Jim Morrison in the 1950's. Tim Davis of Ford Motor Company has made a
major contribution to developing the concept.
In collaboration with Tim Nicholls I have made a detailed tabulation of
the steps in a generic product creation process and
the statistical methods that are relevant to each step. We call this a Guided Tour to the uses of
statistics in a typical engineering-based manufacturing company.
Design for Six
Sigma
DfSS got off to a
slow start, perhaps because many companies and consultants thought it was just traditional
Six Sigma applied to a new area. It is
now clear that, at its best, DfSS is a sophisticated framework for bringing
analytical tools to bear on the whole product creation process, from
understanding customer requirements through robust design to controlling the
production or delivery process.
Statistical Engineering provides many of the core tools for DfSS.
Courses
and seminars
Whenever possible I aim to develop tailor-made
training for individual clients. However, the examples below illustrate a range
of possibilities.
|
Half day
seminar: |
An overview of
the tasks involved in the design/development of products and services. Shows how statistical methods contribute to
effective and efficient design processes.
Key themes are robustness, transfer functions, designed experiments
and the integration of statistical methods with analytical approaches based
on computer (CAE) models. |
|
One day
course: |
Introduces the concepts and methods needed
to develop transfer functions (response surfaces) from multi-factor
experiments, including experiments on computer models. Built around a sequence of hands-on
exercises. Graphical methods are emphasised and the teaching is supported by the use of
spreadsheets. I present the
same core material in a range of styles and at different mathematical levels,
for different groups (e.g. managers, design or production engineers, research
scientists, technicians). |
|
Three-day course: |
Follows the steps
of the product creation cycle, applied to a mass-produced article. Statistical methods are introduced on a
“just in time” basis, to facilitate the achievement of specific
engineering objectives at each step of the process, and applied to a case
study. Functions both as
a foundation course in statistics and as a practical guide to the use of
statistical methods in engineering.
The unique teaching style based on dynamic graphics quickly moves
participants from rudimentary statistical knowledge to understanding and
applying important methods such as multiple regression, response surface DoE,
Weibull analysis and ANOVA. |
I am an
Associate of WDP Consulting, who offer Six
Sigma Green Belt and Black Belt training as well as a wide range of services
for the development of people, processes and organisations.
Consulting
The most important contribution I can make as a
consultant is to help you choose the appropriate statistical method to support
different stages of the design/development/production process, and to achieve
different engineering objectives.
More detailed consultancy activity includes:
- guiding teams of engineers in planning experiments
- analysing experimental (and other) data
- designing experiments for non-standard situations
- detailed evaluation of proposed experiments generated in-house
- advice on the best way forward when an experiment cannot be run as
planned.
Development of internal specialists
Over the years I have helped numerous engineers to
become a "local expert" in the
application of statistical methods and other quality tools, and to become
trainers themselves. Such people are now
often known as Six Sigma Black Belts and Master Black Belts. Steps in the development process include:
- formal training
- train-the-trainer sessions and feedback on performance
- informal coaching and "master classes"
- tapering support for applications.
Qualifications and
experience
My academic qualifications are a BSc in Mathematics
and Statistics from
I was an academic for about 20 years, at
Current and previous clients of my training
development, training delivery or consultancy include Caparo Engineering, CGI International, Champion Spark Plugs, Ford Motor Company, Hygrade Foods, Jaguar Cars,
Johnson Controls, Lucas
Commercial Diesel Systems, Lubrizol, Oxford Instruments, Pall Aerospace, Perkins Technology, Pfizer Global R&D, Rolls Royce Aerospace, Sauer Sundstrand, Stanton Rubber, Visteon.
My support for
Ford’s Global Calibration Methodology Team played a major part in helping
them to deliver substantial process improvements worth millions of
dollars. The importance of this work was
recognised by the company when it invested heavily in software to support the
new processes. This software, parts of
which I specified, is now available commercially as a Matlab Toolbox called
Model-Based Calibration (MBC). The
toolbox includes novel methods for the design and analysis of experiments which
were developed by a research group at
Other aspects of my work
with Ford included helping to design an international training programme in
quality management and methods (EQUIP, now FTEP) for Ford engineers, and an innovative
course in Statistical Engineering now available globally as an eLearning
package. Groups of engineers that I supported
in their use of statistical methods have on three occasions received Customer-Driven
Quality awards from the company.
I have also helped to design or deliver
post-experience education in statistical methods and quality management for the
Software
I do not produce
software on a commercial basis but I have many useful spreadsheets in Microsoft
Excel for analysing experimental data. These are available free of charge.
I am often asked to
advise companies who are considering an investment in statistical software. As a regular user of software such as Design
Expert, JMP and Minitab I can compare their capabilities with the requirements
of particular companies and work groups.
Publications
This book "Engineering, Quality and Experimental
Design", by D.M.Grove
& T.P.Davis, was published by Longman in 1992
(ISBN 0-582-06687-5). It is
unfortunately out of print.
"Statistics
for Engine Optimisation", edited by Simon Edwards, Henry Wynn and myself,
was published in 2000 by Professional
Engineering Publishing (IMechE).
A list of my research papers is available on request.