|
Dynamic (on-line) Risk Analysis of
Wastewater Treatment Processes, as an
Operational Control Tool
ABSTRACT
During an extended cooperation between the
Central Wastewater treatment plant for
Næstved city in Denmark, ENVICARE, as
consulting engineers and the software
company 7-Technologies, a new operational
tool, “RISK-VIEW”, for wastewater treatment
plants, has been developed.
RISK-VIEW is an on-line risk analysis tool,
predicting future events on a wastewater
treatment plant, providing the operator with
information to avoid upcoming problems.
RISK-VIEW will on-line give a list where and
when problems will occur in the process.
RISK-VIEW is based on traditional fault tree
analysis theory. The innovation in the
method is that the fault tree is made
dynamic by combining the real-time
information from the SCADA system and the
traditional fault-tree analysis. This makes
it possible to estimate incidents in the
future and provide a lead-time to react and
solve problems before they result in
unplanned shutdowns. Further when combined
with prognosis data from i.e. the
maintenance system, production or weather
data to include future variations the
prediction of future problems and incidents
is even more powerful.
RISK-VIEW can predict the behaviour of the
treatment process giving the operator the
possibility of:
Identifying bottlenecks before they occur
Operating the plant at a known risk level
Optimising and prioritising maintenance
By using RISK-VIEW as an add-on to the SCADA
system, RISK-VIEW will reduce downtime,
eliminate outages and give the possibility
of a more stable operation. This to improve
the productivity targets and resulting
increased capacity.
RISK-VIEW also serves as a structured
container for operator experience. For each
possible operational consequence the
resulting remedial action has to be entered
into RISK-VIEW. As a result RISK-VIEW not
only tells you where the problem is but also
how it shall be rectified – well in advance.
RISK-VIEW is not only usable in the
wastewater treatment industri, but also in
industries as chemical, oil & gas, food &
beverage and the utilities.
KEYWORDS
Risk analysis - Wastewater treatment -
dynamic fault trees - on-line analysis
INTRODUCTION
This paper describes a new development
within the automation industry, the dynamic
fault tree method. By combining data from
the SCADA system, the maintenance system and
other available data with a traditional
fault tree, you get the dynamic fault tree.
During an extended cooperation between the
Central Wastewater treatment plant for
Næstved city in Denmark, ENVICARE, as
consulting engineers and the software
company 7-Technologies, a new operational
tool, “RISK-VIEW” has been developed.
RISK-VIEW is based on the dynamic fault tree
method.
This paper describes as follows:
Background - Data availability in the new
millennium
The traditional fault tree
The theory behind the dynamic fault tree
Benefits with the dynamic fault tree
The RISK-VIEW program
A case story from Slangerup Wastewater
treatment plant in Denmark
Conclusion
Background – Data Availability in the new
Millennium
Being a wastewater treatment plant or
actually any plant in this new millennium
all face the same challenge. We have access
to and are bombarded with an almost
unlimited amount of data and information
from the departments of production,
maintenance, sales or and information from
the Internet. Further to this all of us sit
on a large amount of experience, which we
utilise when trying to operate the plant in
an optimal manner.
Further everybody, being plants or human
beings, are pushed to and above their
capacity – we want to get more out of our
plants and staff – to increase profit.
Our challenge in this project was to find a
way to utilise the relevant data only and
combine this with the operational experience
and the knowledge from our maintenance
department of how and when our equipment
needs repair or will break down. This should
also be combined with knowledge of predicted
events, i.e. weather forecasts to give us
the possibility of reacting before an event
occurs.
Basically we set forth to invent a ‘crystal
ball’ for the industry.
Another challenge with the project was to
find a structured and easy way to collect,
use and document the experience of the
organisation. Today, where staff mobility is
much higher than what it used to be,
precious operational experience is lost with
staff leaving the company. We need to find a
way to contain this information and
experience.
A solution encompassing all this is the
dynamic fault tree method. Carsten Fjordside
of Næstved Central Wastewater Treatment
Plant in Næstved, Denmark, has invented this
method. Based on the mechanics behind this
idea we developed the program
RISK-VIEW
containing and utilising the method.
The Traditional Fault Tree
The traditional fault tree analysis is not a
new invention; consulting engineers for risk
analysis purposes have used it for a long
time.
The Fault-Tree Analysis is a statistical
method that traditionally is used for
determination of risks at a plant.
Figure 1
The fault-tree is a graphical representation
of the connection between different errors
and incidents in a process. Logical
operators to one or more so-called top
incidents connect basic incidents at the
bottom of the fault-tree. Top incidents
represent different errors already occurred
or actual errors in the process and thus a
risk level can be outlined.
Typical top incidents can be:
Stop of production
Safety related matters
Explosion
Toxic leakage
The basic incidents at the bottom of the
fault-tree are typically components or human
errors for which static error data are
available. Typical basic incidents are:
Pump failure
Failure in thermostats
Failure in electric switches
Valve failures
Operator is not reacting
The traditional fault-tree analysis is
static and shows the effect of basic
incidents based on an imaginary process
status.
The advantages of the fault tree analysis
are that it gives a fantastic overview over
a process. It is easy to evaluate the
consequence of the failure of a specific
component – or alternatively to evaluate
which events will result in a certain
consequence.
It gives the users a fantastic understanding
of how the process work, and makes it easy
to present potential problem areas.
The disadvantage with the fault tree method
is that it is static. No matter how good the
analysis is it will still be void shortly
after completion because things change.
The Theory Behind the Dynamic Fault Tree
The Dynamic Fault-Tree Analysis converts a
fault-tree from being a static analysis to
being a dynamic analysis. This is achieved
by combining the fault-tree with real-time
data, hence combining the traditional fault
tree with maintenance and operation data
from the SCADA system. The risk for an
incident can then be calculated based on the
actual status of the process. For example
the equipment not in operation is considered
when calculating the result of the
fault-tree analysis, which is something that
cannot not be done in a static analysis. The
dynamic fault-tree analysis makes it
possible to estimate future risks based on
the actual status – a crystal ball for the
process. To complete the dynamic fault-tree
analysis the physical mode of operation of
the process units is incorporated in the
fault-tree analysis thus making the
occurrence of the incident in time
conditional upon mode of operation compared
to the actual operation; for example the
overloading of a reservoir proportional to
the liquid in/out of the reservoir,
including the actual level of liquid in the
reservoir. Another element is the handling
of dynamic variable objectives for the
process based on a prognosis; e.g.
calculating the future distribution security
as a time function for water distribution
networks, district heating systems or gas
networks, all seen from a time-fluctuating
consumption prognosis.
The Dynamic Fault-Tree Analysis is not a new technology in the SCADA
system or the fault-tree system. The
innovation in the method is the fact that
SCADA system data and fault-tree analysis
can be combined. Among other things it
makes it possible to estimate incidents in
the future and provide a lead-time to react
and solve problems before they result in
unplanned shutdowns or worse.
Benefits of the Dynamic Fault Tree
The benefits of the dynamic fault tree are
all the same as for the static. But where
the static fault tree analysis provide an
excellent overview this is further enhanced
in the dynamic fault tree analysis as we are
now qualifying these based on actual values.
In the static analysis we build our
assumptions on i.e. number of pumps in
operation. In the dynamic analysis we do not
need to assume - we know how many pumps are
in operation – we know which pumps are not
available and which pumps are on stand-by.
Further we know from the SCADA system how
long each pump has been in operation, hence
also when it is most likely to fail. We
therefore know the present status and also
quite a lot about the future.
By the dynamic fault tree we therefore have
the ‘crystal ball’ effect.
The RISK-VIEW Program
The outcome of the project development was
the software product RISK-VIEW. RISK-VIEW is
developed as a product, which can be
combined with any type of SCADA system,
maintenance system or directly to the
Internet.
Figure 3
The engine of RISK-VIEW is the fault tree,
which is drawn and implemented using MS
Vision. The fault tree can encompass a unit,
an area of the plant or the entire plant,
depending on the level in the tree.
Figure 4
The main screen is shown in Figure 4. The
upper right hand corner shows the specific
risk for a consequence as result of time –
from now and onwards.
You can at any time on the left hand side
see where in the process (or where in the
fault tree) you are, and on the lower right
hand side see which consequences are most
significant for the given time. By moving
the mouse on the graph you can move the
time.
The red colour indicates the time where the
process enters a critical situation. By
right clicking on the mouse on a certain
consequence you will be able to obtain
information about how the problem seen is
alleviated, re figure 5.
Figure 5
By implementing the above in a proper manner
all knowledge and experience of operation of
the plant can be systematically stored.
A Case Story
The RISK-VIEW program has successfully been
installed at Slangerup Municipal Wastewater
Treatment Plant during the autumn of 2001.
The main operational constraints presently
foreseen in Slangerup are that they wish to
maintain a performance of 8500 P/E. The
plant is however only designed for 6500 P/E
and it is not expected that it will be
extended until 2003.
Further the Ammonium-, Nitrate- and COD
concentration shall be maintained within the
set requirements to avoid taxation.
By installing RISK-VIEW Slangerup is now
able to react on problem areas in advance,
hence avoiding exceeding the above
concentrations. Further than a proactive
troubleshooting RISK-VIEW have given
Slangerup an excellent knowledge transfer,
which has already proven fruitful in
training new staff.
The staff of Slangerup did during the
development of the fault trees obtain an
in-depth knowledge of their system, and did
find several areas where the design could be
optimised. This resulted in modification of
the biological treatment methods,
reengineering of the pumping functionality
and correction in the PLC programming.
RISK-VIEW is today
fully integrated in the control and
operation of Slangerup Municipal Wastewater
Treatment Plant. Having only been in
operation for 4 months it is difficult to
estimate the payback time of the investment,
but it is likely to be less than one year.
CONCLUSION
Using the dynamic fault tree method opens up
a entirely new way of operation of the
plant.
It is now possible to react in advance based
on information of the past, the present and
the future. Further it is possible to
structure the knowledge of the experienced
staff and store in an operative manner,
making it possible for new and inexperienced
staff to draw on the experience by use of a
mouse.
RISK-VIEW brings in a new way of thinking
and of controlling plants – and an entirely
new way of utilising the abundance of data
and information available in the new
millennium.
|
