![Sample our Built Environment journals, sign in here to start your access, Latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5926/TOC-Subject-Sample-2021-Built-Environment.jpg"})
ABSTRACT
One of the most important issues related to natural gas is unaccounted for gas. Residential customers constitute a significant percentage of unaccounted for gas. To estimate the amount of unaccounted for gas, it is necessary to compare the amount of consumption estimated by the model with the one recorded by the meter. Thus, the value estimated by the consumption model are of great importance. Initially, a consumption model is developed for each customer using consumption data for the first 12 months and the average monthly ambient outdoor temperature related to the same time period. The models are developed using artificial neural networks and particle swarm optimization algorithm. The estimates made by the models are then compared with the values recorded by the meters. This method is then implemented on some real data (as the study area). The results show the effectiveness of the proposed method.
5. Acknowledgments
The authors would like to thank the Natural Gas Company of Fars for their support by providing the data, with special thanks to Mr Amir Safavian and Ms Fatemeh Sadeghian.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Abbreviations
| UFG | = | Unaccounted For Gas |
| ANN | = | Artificial Neural Network |
| PSO | = | Particle Swarm Optimization |
| FPNGC | = | Fars Province Natural Gas Company |
| MAPE | = | Mean Absolute Percentage Error |
| SMAPE | = | Symmetric MAPE |
| NSMAPE | = | Normalized SMAPE |