1. Objectives
|
The main purpose of the project is to investigate possibilities of applying evolutionary computation techniques (extended by genetic memory structures) for a particular optimisation process, which aims at reduction of pollution in the area of the whole Poland. A new system would be capable to include all major sources of pollution (70 sources, 300 stacks) as well as the current weather conditions. The outcome of the system would indicate the recommended of produced energy for each particular pollution source.
|
2. Achievements Prior to the Current Report Period
- Detailed conception of the whole system
- Methodological assumptions for valorisation model (i.e. model for the sensitivity of different environments)
- Computational grid compatible with MSC-W EMEP (Meteorological Synthesing Centre-West Co-operative Programme for Monitoring and Evaluation of the Long Range Transmission of Air Pollutants in Europe).
- Data of population density in 20 regions of Poland (20 from 49) (in each grid unit).
- Test data of SO2 emission from all major sources of pollution (technical and geographical data).
- Preparation and analysis of the aerodynamical roughness coefficient (zo) matrix for the project modelling area (900 km x 750 km) in the established project computational grid covering that area.
- Preparation and analysis of the land-use cover (10 categories) for the territory of modelling area.
- Preparation of the population density data for the territory of Poland.
- Analysis of the limitations connected with the existing possibilities of optimisation system for ecological power control in Poland.
- Simple version of air pollution transport model was created.
- Calculation of the sulphur dioxide concentration background values for the project modelling area (900 km x 750 km) in the established project computational grid covering that area.
- Development, implementation and testing of SO2 dry deposition submodel as input for air pollution dispersion model.
- Preparation of the actual (1997) sulphur dioxide emission data from the public power generation sector sources.
- Analysis of possibilities of power control in Poland.
- Computer implementation of the air pollution transport model.
- Implementation of the wet deposition module.
- Implementation of chemical transformation SO2 to SO4 factor.
- Statistical estimation of predicted restrictions on the sulphur dioxide concentration background values considering day- and year-long periods.
- Analysis, software specification and functional project of EROS - system (including constraints).
- Simulated calculations using system GENOCOP
- Preparation of the computer tools for the visualisation package of CRIT2 project.
- Project and implementation of data archives in Access database format (for Win NT platform).
- Project and implementation of the EROS system's kernel (the code in C++ for WinNT platform) with main functions
- Preparation of the nominal sulphur dioxide emission data from the public power generation sector (power plants and cogeneration plants) sources, which are included in the control system of CRIT2 project.
- Selection as well as collection of the needed data for the additional constraints to be used in the optimisation system of the CRIT2 project.
|
3. Work and Achievements in Current Report Period
|
1. Modifications in the Regfor. New version, called Regfor+ was implemented.
2. Statistical procedures were finally created and described.
There are several statistical procedures that operate on the sten data. One can check if the series of the SO2 concentration measurements has statistically higher value than given sten.
This latter procedure is based on: µl and sl, i.e. mean and standard deviation calculated from logarithms of a given series of measurements of SO2 concentration.
Another procedure can be applied for stens and/or logarithms only. This procedure assesses if two series of concentrations differ significantly.
Third statistical procedure gives indication if there is an alarm concentration in a given grid element. Such an indication is very useful to establish a radical change in the SO2 emission.
3. Final version of the EROS system was implemented. The main purpose of the system is to reduce pollution in the area of the whole Poland. The outcome of the system would indicate the recommended level of produced energy for each particular pollution source. The resulting vector represents a list of recommended power control levels for which relative levels of pollution in all regions in the country are least and the pollution is distributed to protect regions with more sensitive environment (e.g., natural ecosystems, especially forests). Every day the vector can be different depending on current weather conditions. The system is also ready for continuous optimization of pollution sources in the country, day by day. A redundant genetic material was introduced into an individual, resulting in a new polyploid representation and a dominance function, which controlled transfer of information between chromosomes inside the individual.
EROS is a visual application dedicated for Windows 95/98/NT/2000 platform. It was implemented and compiled in Borland C++ Builder 4 environment. Major part of data used by the program is stored in a Microsoft Access relational database. Program realizes connections with database through an ODBC driver. System EROS integrates three main modules: (1) Regfor++, (2) evolutionary optimizer, and (3) visualization tools. Regfor++ computes forecasted transport of pollution depending on weather reports, weather forecast and emissions of sources during last two days. The second module finds optimal vector of produced energy levels for pollution sources. Third module is a visual application that shows maps illustrating disposition of stens and concentrations over area of Poland.
4. Several experiments with real data (meteorological real data and weather forecasts, and real emission data for chosen summer and winter periods) have been conducted. Initially real concentrations and stens have been calculated for each chosen experimental period. Then the optimization process has been applied and results were compared with initial results. From the performed experiments, the following conclusions can be drawn:
- according to our assumptions the amount of produced energy was equal in both cases,
- average emission and average concentration in the country decreased substantially (12-15%) after optimization process,
- results, obtained when applying real meteorological data instead to weather forecast, were practically identical (less than 1% differences).
|
4. Deviations from Intended Plan
5. Supporting material
List of Project:
-
-
- 5.1 Deliverables
- POLSOX-I - regional Eulerian air pollution model for the area of the extended CRIT2 grid (WUT-FEE, Poland); status: confidential
- REGFOR - Lagrangian air pollution model for the area of the CRIT2 grid (SRI PAS, Poland) ; status: confidential
- Dry deposition submodel (RIVM, Netherlands) ; status: restricted
- EROS - Evolutionary Real-Time Optimization System (ICS PAS, Poland) ; status: restricted
- 5.2 Published Papers:
-
- 1. Michalewicz M., Juda-Rezler K., Trojanowski K., Matuszewski A., Michalewicz Z., Trojanowski M., Evolutionary Real-Time Optimization System for Ecological Power Control, submitted to the Ninth International Symposium on Intelligent Information Systems, Bystra, June 12-16, 2000.
- 5.3 Practical Experiments
- -
-
- 5.4 Public Demonstrations
-
|
|