HITERM: Project Description
|
Within the framework of HPCN Information Management and Decision
Support, HITERM aims at expanding the application of HPCN
to decision support
in new domains: the central focus is the interface between
technological risk management and the environment .
|
|
Using distributed parallel computing, the project aims at reaching
better-than real time performance for the simulation of accidental
release
of hazardous substances into the atmosphere, ground and surface water,
using state-of-the-art 3D simulation models.
|
|
This information will be used, in the framework of on-line decision
support and advisory systems for:
-
the adaptive routing of hazardous transports, integrating
environmental risk criteria with other road information;
-
the support of emergency management tasks (and related staff
training) for transportation accidents involving hazardous substances
and for hazardous installations, as foreseen by the amended
Post-Seveso Directive (82/501/EEC, 87/216/EEC, COM(94) 4).
|
|
In addition to connecting the HPC
simulations to various on-line data
sources (primarily environmental and hydro-meteorological monitoring),
the project will explore two additional important aspects of
HPCN based decision support applications, namely,
-
the on-line integration of uncertainty and error analysis,
based on Monte-Carlo methods again realized by parallel simulation;
-
methods for on-line interactive data interpretation and
visualization for dynamic, spatially distributed, and probabilistic
model results for effective user interface design, supporting direct
understanding.
|
|
The main Project Objectives of HITERM
are to design and develop:
-
HPCN methods and tools for time critical environmental
applications, that are related to technological risk and
emergency management;
-
a prototype system, based on client/server distributed parallel
computing, for
-
the on-line adaptive routing and safety analysis of
hazardous transports including dynamically updated environmental
risk criteria;
-
the simulation of accidents (release
scenarios of hazardous substances) for the support of emergency
control measures and staff training for emergency management,
both for transportation accidents and Seveso-class chemical
process and storage plants,
|
|
including concurrent sensitivity and uncertainty analysis
by Monte Carlo methods as an integrated part of the forecasting
and decision support scheme;
-
tools for the real-time visualization of these dynamic and
spatially distributed stochastic model results (concurrent
ensembles of solutions), to provide
-
efficient interactive decision support tools for these
applications, using a discrete multi-criteria optimization system,
rule based expert systems, or neural nets where appropriate;
HITERM will implement and test these tools and methods in
concrete case study applications related to the road transportation
of hazardous goods, and to the chemical process industry,
including the storage of hazardous chemicals.
|
|
HITERM is developing a new generation of interactive,
model-based decision support systems. It will explore the integration,
in a distributed client/server HPCN architecture with local,
distributed, and mobile multi-media clients, of:
-
Real-time data acquisition systems (eg., transport telematics systems,
satellite imagery, weather radar, stationary and possibly mobile
observation stations including hand-held data acquisition systems and
video input);
-
With (distributed parallel) high-performance computing resources for
(better than) real-time modeling and forecasting of transport and
dispersion models, plus discrete multi-criteria optimization,
rule-based expert systems and neural nets,
as remote advisory and decision support systems;
-
And multi-media clients, including local and networked X Windows
servers, and http clients including distributed, mobile clients
(including hand-held computers),
|
|
to provide real-time information and
decision support for complex
and demanding technological and environmental risk management
applications with considerable economic implications --
industrial and
transportation accidents addressed by the system can cause extremely
large economic damage: even a small reduction due to better emergency
planning and management would make a system like HITERM highly
profitable.
HITERM will link networks of information
resources and analytical capabilities with a range of clients,
including
mobile field personnel. As an initial form of use the exploitation of
the system for staff training exercises can be foreseen.
The decision support paradigm underlying HITERM is based on
interactive, multi-criteria selection from large sets of (HPCN generated)
alternatives, supported by (HPCN generated) dynamic visualization and
multi-media representation.
|
|
