Forecasting residual free chlorine levels in Karbala using artificial intelligence algorithms
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
The pollution of water from many sources creates a public health danger, necessitating disinfection in the treatment of drinking water. Consequently, chlorine is utilized as it is the primary disinfectant, making sure the secure conveyance of water through pipelines. The regulations specify that residual free chlorine in the water distribution system must range from 0.3 to 0.5 mg/L, consequently it is crucial to sustain appropriate chlorine levels in potable water. This study involved the collection of water samples from locations inside the DWDN of the Al-Hussein Central Neighborhood Water Project in Karbala. The points were P0, P1, P2, and P3, which correspond to the project point, the network's starting point, the network's midpoint, and the network's endpoint, respectively. The site was assessed twice a week for residual chlorine (RC), with physical and chemical laboratory testing, to collect network data for training an artificial intelligence (AI) model capable of predicting the remaining free chlorine amount in the water distribution system. Four deep learning and machine learning AI models which are (Random forest (RF), XGBoost, General regration neural network (GRNN), Pattren Recognition neural network (PRNN)) developed and evaluated to identify the most effective model with superior performance. The results showed error rates of 0.377, 0.072, 0.002, and 0.181, respectively. And R squared values were 0.796, 0.961, 0.988, and 0.902, respectively. The GRNN, identified as the optimal model, illustrated the lowest error rate and the highest R^2 compared to the other models. This research faces challenges with sampling, choosing of available points, the test process, and the selecting of suitable models; yet, satisfactory results were achieved.The model could simplify the management process and enhance decision-making about the residual free chlorine dosage with higher accuracy prior to any delays.
Fields of Study & Subjects
Water Systems and Optimization
Hydrological Forecasting Using AI
Water Treatment and Disinfection
Publication details
- Journal / source
- AIP conference proceedings
- Volume
- 3405
- Pages
- 020005–020005
- Publication date
- 1 January 2026
- Language
- EN