System Dynamics in Water/Wastewater Mgmt and Sewer Re

[jm62004@Jetson.UH.EDU]

Dear John,

Sounds indeed very interesting.
>As background, it is worth pointing out that the Australian water
>industry has been undergoing dramatic change including restructuring
>and privatisation. As a result there is a lot of interest in industry
>regulation and comparative performance.

Raison d etre for the NSF grant in Houston too!

>The model developed aims to do two things- looks at performance issues
>and introduces SD to the industry.

>One problem encountered has been that many of the managers involved like
>to work at a very high level of detail, ie they are operationally
>focussed and are less interested in some of the longer term dynamics.

You hit the nail on the head - one has to do a selling job here as to how
SD could be useful. What I told the city officials is:

Just as in any large and complex system (such as the Houston sewer system,
or in the Australian utilities you mention) there are issues such as
"limits and constraints" (money, time, talents, political boundaries),
"indirect effects" (action in one area leads to diversion of resources and
therefore less actions (effects) in other areas), nonlin. dynamics
(physical such as the air-hammer phenomenon/socio-political too numerous to
mention), risk, delays, equity considerations (extremely rich River Oaks
inhabited by millionaires and welfare-recipients such as in Third Ward),
efficiency, and multiple objectives.

It is very difficult to understand such a complex system by one individual
with one brain - so we use tools such as SD, and continously interact with
different stakeholders for model refinement and input.

When I made the first presentation, because a lot of stocks and flows in
the model seem to be like a water/wastewater network, I had to repeatedly
tell people that THIS IS NOT A HYDRAULIC/ENGINEERING MODEL (especially
because there are so many engineers in the group, and they will jump on
the obvious engineering omissions). This is MORE than engg as it looks at
the bigger picture - but it is also LESS than engg because it ignores the
gory details. Here the story of the 6 blind men is apt - and the task of a
SD person is to portray himself/herself as a nonblind person looking at the
elephant *from a distance*, rather than the 6 blind (very smart) people
feeling the elephant up close. I also stress that SD is like an extended
way of doing back-of the envelop calculations - bigger picture, systems
view but not such gross (verbal - mental map level) generalizations as to
be totally useless.

>The key dynamics arise out of the interaction of the capex and
>maintenance and replacement sectors; the other sectors have been kept
>very simple.

we are just starting so our model is much smaller. how do you deal with
the paucity of data - the biggest problem for us seems to be knowledge of
the life-cycle costs of pipes - implies it is difficult to estimate
reliability, & to compare costs of preventive maintenance (the situation
hoped for) vs. costs of reactive maintenance (sitution now).

>
>>From a SD point of view, the problem is to keep the model simple enough
>to easily communicate, yet satisfy the level of disaggregation being
>asked for by the clients.

Agreed - again boils down to a selling job of why SD would be useful - I
personally think it is very useful in these situations.

>we are hoping to extend our work to include more behavioral features
>including political aspects as you have noted.

we are making a strong point of this one, as they (the City) have already
done a good job of the engg part and according to a veteran city official,
it is the understanding of the socio-political cost-benefit area that is
the most difficult part of their work. Luckily we have a political
scientist in our group who has good survey data we can use for parameters
etc..

>Hope these comments are of some interest. In the first instance our
>model and report is the property of our client under a research
>contract, but it will eventually reach the public domain.

I will be very interested to hear further comments (we can take this
discussion off the list if it gets more specific/technical). I will be
interested too when your work reaches public domain. Do you have a web site?

For what it is worth, we have a web address if youd like to have a look

http://www2.egr.uh.edu/~heller/Suwer/iruws.htm

Thanks and regards,

Jaideep
****************************************
jm62004@Jetson.UH.EDU
Jaideep Mukherjee, Ph. D.
Research Associate
Department of Industrial Engineering
University of Houston
4800 Calhoun Road
Houston, TX 77204-4812

Phone: 713 743 4181; Fax: 713 743 4190
****************************************

[John.Barton@BusEco.monash.edu.au]

Dear Jaideep,

I have just completed a demonstration model for the Australian water
industry which explores the manner in which asset management policies
affect productivity.

As background, it is worth pointing out that the Australian water
industry has been undergoing dramatic change including restructuring
and privatisation. As a result there is a lot of interest in industry
regulation and comparative performance.

The problem is that an individual utilities performance is very much
linked to its asset endowment and the operating conditions associated
with physical environments.

The model developed aims to do two things- looks at performance issues
and introduces SD to the industry.

One problem encountered has been that many of the managers involved like
to work at a very high level of detail, ie they are operationally
focussed and are less interested in some of the longer term dynamics.

Our model has six sectors- network demand, network supply, cash flows,
capital funding and borrowings, employees and asset maintenance and
replacement.

The key dynamics arise out of the interaction of the capex and
maintenance and replacement sectors; the other sectors have been kept
very simple.

Patterns of maintenance and replacement for Australian utilities follow
similar patterns with the principal feature being associated with a 30
year period of rapid infrastructure development in our cities from 1945
to 1975 (approximately). So many of these assets are nearly 50 years old
and a "bow wave" with high levels of maintenance and replacement costs
is starting to take effect.

>From a SD point of view, the problem is to keep the model simple enough
to easily communicate, yet satisfy the level of disaggregation being
asked for by the clients.

we are hoping to extend our work to include more behavioral features
including political aspects as you have noted.

Hope these comments are of some interest. In the first instance our
model and report is the property of our client under a research
contract, but it will eventually reach the public domain.

John Barton, Systems Thinking and Organisational Learning Group
Department of Business Management, Monash University, Melbourne,
Australia
Postal: PO Box 197, Caulfield East, Victoria, Australia. 3145 Tel:+61 3
9903 2672, Facs + 61 3 9903 2718, Email:
John.Barton@BusEco.monash.edu.au