Development of soft computing prediction model for the influent physicochemical characteristics of sewage treatment plants / Mozafar Ansari

Mozafar , Ansari (2021) Development of soft computing prediction model for the influent physicochemical characteristics of sewage treatment plants / Mozafar Ansari. PhD thesis, Universiti Malaya.

[img] PDF (The Candidate's Agreement)
Restricted to Repository staff only

Download (214Kb)
    [img] PDF (Thesis PhD)
    Restricted to Repository staff only until 31 December 2023.

    Download (2580Kb)


      The role of sewage treatment plants (STPs) is reducing the sewage contaminates to a level that minimises the risk environmental disasters by treating the sewage to acceptable standards before being discharged into the receiving waters. Design and operation of these treatment plants depend on the influent conditions. Unlike industrial sewage treatment plants, there is not enough control over the quality and quantity of domestic influent. In this research, physical and chemical influent characteristics from 2011 to 2013, including flow rate, biochemical oxygen demand (BOD), chemical oxygen demand (COD), Ammoniacal Nitrogen (NH3-N), pH, oil and grease (OG) and suspended solids (SS) of three sewage treatment plants, STP A, STP B, and STP C, in Kuala Lumpur were evaluated. Sugeno fuzzy inference system (FIS) algorithm was used to model influent parameter, and the FIS parameters were adjusted by ANFIS, integrated Genetic algorithms, GA-FIS, and integrated particle swarm optimisation, PSO-FIS, algorithms. To find the best modelling results, the root-mean-square error (RMSE) and coefficient of determination (R2) were used as primary evaluation criteria, and relative error (RE) and Nash–Sutcliffe efficiency (NSC) were applied as secondary validation indices. The best algorithm for each parameter was selected based on these criteria. Moreover, the influent parameters were compared with design values that have been recommended by Malaysia National Water Services Commission. The prediction results showed that both integrated GA and PSO fuzzy algorithms performed almost at the same level and provided more accurate results for all parameters than the ANFIS model. Based on the influent assessment, the influent COD for all sewage treatment plants was higher than the design value. Beside COD, influent BOD of STP A had events that was higher than the recommended value. Moreover, the BOD, NH3-N, and OG of STP B have exceeded their design values, and in STP C, the OG was exceeded several times. These parameters were forecasted for one year to find the future condition of selected STPs. The results showed that STP B would be the STP with the highest number of exceeded parameter and STP A would have the minimum number of exceeded parameters. However, the number of influent COD events that was greater than the design value for STP C would be expected to be higher than other STPs. One of the methods that can resolve the condition of COD is reducing elements which can affect the BOD and COD value. Despite this, wasted cooking oil can be for other purposes such as biofuel. Therefore, it was recommended to collect wasted oil before entering the sewage system.

      Item Type: Thesis (PhD)
      Additional Information: Thesis (PhD) - Faculty of Engineering, Universiti Malaya, 2021.
      Uncontrolled Keywords: Sewage treatment plant; Influent; ANFIS; Genetic algorithm; Particle swarm optimisation
      Subjects: T Technology > TA Engineering (General). Civil engineering (General)
      Divisions: Faculty of Engineering
      Depositing User: Mr Mohd Safri Tahir
      Date Deposited: 22 Jun 2023 02:51
      Last Modified: 22 Jun 2023 02:51

      Actions (For repository staff only : Login required)

      View Item