Neural Network Models for Air Quality Prediction: A Comparative - - PowerPoint PPT Presentation

Neural Network Models for Air Quality Prediction: A Comparative Study

S V Barai A.K.Dikshit Sameer Sharma Department of Civil Engineering IIT Kharagpur

Presentation Outline …….

Background on Neural Forecaster Objectives Developed Neural Air Quality Predictors Data Collection and Data Analysis Performance Evaluation Results and Discussion Future Projections Closing Remarks

Generic Air Quality Forecaster

Input Variables Mapping Output

Air Quality Parameter Model Air Quality Forecasts

Equation or Mathematical Model

Neural Air Quality Forecaster

Input Variables Mapping Output

Air Quality Parameter Model Air Quality Forecasts

Equation or Mathematical Model

Neural Networks

Objectives

Data Collection -multiple air quality parameters -

Containing daily average pollutant concentrations at a specific location.

Implementation of variations of neural network models

for predicting air quality.

Identification of suitable air quality prediction model(s)

for hourly (short-term) data and Yearly (long-term) data.

Exhaustive simulation Study using various models. Performance study of models.

Implementation of Neural Models

Recurrent Network Model (RNM) Change Point Detection Model (CPDM_RNM) Sequential Network Construction Model (SNCM) Self Organizing Feature Maps Model (SOFM)

Data Collection and Analysis

Case Study 1- Long Term

Seven parameters namely VOC (volatile organic

carbon), NOX (oxides of nitrogen), CO (carbon monoxide), SO2 (sulphur dioxide), PM10 (particulate matter with size less than 10 microns), PM2.5 (particulate matter with size less than 2.5 microns) and NH3 (ammonia).

All concentrations are in micrograms per cubic meter. Annual average data for 15 years from 1985 to 1999 Source:US EPA website www.epa.gov. Data for 115 counties of California State in USA has

been collected

Case Study 2- Short Term

Three parameters namely RPMA (Respiratory

Particulate Matter Average), SO2 (sulphur dioxide) and NO2 (nitrogen dioxide) is collected for Delhi State at nine locations.

Daily average concentrations for last two years from

3/7/2000 to 20/8/2001.

Tata Energy Research Institute web site www.teri.in Data for Ashram Chowk has been used for carrying

• ut simulation

Statistical Properties Case Study 1

Statistical Properties Case Study 2

Model Performance Evaluation

Error Evaluation

Percentage Error = (target – output) / target *100

Results and Discussion

Case Study 1 – Long Term

RNM Parameters

CPDM_RNM Parameters

SNCM Parameters

SOFM Model Parameters

Models Performance

VOC emissions - SOFM

Case Study 1 10 20 30 40 50 60 70 80 1994 1995 1996 1997 1998 1999 Year VOC (micg/m3) Target Output

SO2 emissions - SOFM

Case Study 1

50 100 150 200 250 300 1994 1995 1996 1997 1998 1999 Year SO2 (micg/m3) Target Output

PM10 emissions - SOFM

Case Study 1

10 20 30 40 50 60 70 1994 1995 1996 1997 1998 1999 Year PM10 (micg/m3) Target Output

Observations

The models in general could predict with some

accuracy.

Self-organizing Feature Map (SOFM) based model has

performed extremely well in comparison to other models.

Typical results of SOFM for VOC emissions, CO

emissions, and NOx emissions results high-light the very good performance of the model. The discrepancy

• bserved in the model prediction can be due to the

modeling of the problem.

Observations (Contd.)

Demonstrated an example of an annual average

emission (long-term) data prediction using various neural networks models for a very limited dataset. Models in general have performed reasonably well even with the limited historical data. It is expected that availability of more annual average emission data can improve the performance of models studied.

Case Study 2 – Short Term

RNM Parameters

CPDM_RNM Parameters

SNCM Parameters

SOFM Parameters

Models Performance

RPMA emissions - SOFM

Case Study 2

50 100 150 200 250 300 350 318 322 325 329 332 336 339 350 378 381 Sample Number RPMA (micg/m3) Target Output

SO2 emissions - SOFM

Case Study 2

2 4 6 8 10 12 14 16 18 325 332 336 339 343 346 360 364 378 381 Sample Number SO2 (micg/m3) Target Output

NO2 emissions - SOFM

Case Study 2

20 40 60 80 100 120 339 350 353 357 360 364 367 371 374 381 Sample Number NO2 (micg/m3) Target Output

Observations

Models could predict with modest accuracy. SOFM based model has performed extremely well in

comparison to other models.

Typical results of SOFM for RPMA emissions, SO2

emissions, and NO2 emissions showed reasonably good match between model predictions with target prediction.

Future Projections

Models can have as inputs data from multiple sources,

such as historical air quality measurements, meteorological data etc.

Models can have along with emissions data, episode

levels definition, historical measurements of surface and upper air meteorological data.

Models should be able to give predictions for the

following three different time windows: 1 day, 1 week and 1 month predictions.

Future Projections (Contd.)

Models should have an easy to use interface and

should be able to pre-sent the results in an understandable way to non-computer experts.

The model parameters and architecture of models in

this project were arrived with trial and error. One can arrive at optimal and better performance model after carrying out systematic studies on networks models and their parameters using optimization techniques such as Genetic

Closing Remarks

Study was carried out on air quality forecasting using

various neural network models: RNM, CPDM_RNM, SNCM and SOFM.

The study was focused at preliminary investigation of

single variable based time series prediction.

The investigation was carried out for long-term as well

as short-term air quality data set.

Self-Organizing Features Maps (SOFM) used for time

series prediction came up as the best tool for time series forecasting.

Closing Remarks (Contd.)

These were found to be very useful for large training

datasets.

The results shown here are indications that the neural

network techniques can be useful tool in the hands of practitioners of air quality management and prediction. In that case, practitioners need not know even about the development of the model.

The models studied in this study are easily

implemented, and they can deliver prediction in real time, unlike other modeling techniques. The models can very well easily deal with input noise and uncertainty

Pollution Free Solution

Acknowledgement

Images used in the presentations are obtained from various web-sources. All these web-sources are gratefully acknowledged.