Constructed Wetlands: Prediction of Performance with Case-based Reasoning (Part B)

Abstract

The aim of this research was to assess the treatment efficiencies for gully pot liquor of experimental vertical-flow constructed wetland filters containing Phragmites australis (Cav.) Trin. ex Steud. (common reed) and filter media of different adsorption capacities. Six out of 12 filters received inflow water spiked with metals. For 2 years, hydrated nickel and copper nitrate were added to sieved gully pot liquor to simulate contaminated primary treated storm runoff. The findings were analyzed and discussed in a previous paper (Part A). Case-based reasoning (CBR) methods were applied to predict 5 days at 20°C N-Allylthiourea biochemical oxygen demand (BOD) and suspended solids (SS), and to demonstrate an alternative method of analyzing water quality performance indicators. The CBR method was successful in predicting if outflow concentrations were either above or below the thresholds set for water-quality variables. Relatively small case bases of approximately 60 entries are sufficient to yield relatively high predictions of compliance of at least 90% for BOD. Biochemical oxygen demand and SS are expensive to estimate, and can be cost-effectively controlled by applying CBR with the input variables turbidity and conductivity.

Artificial Intelligence techniques: An introduction to their use for modelling environmental systems

“…"Constructed wetlands: performance prediction by case-based reasoning." [101] Lee et al [100] tested the efficiency of constructed wetland filters by analysing the quality of water fed through them. Half the filters received inflow contaminated with heavy metals.…”

“…Half the filters received inflow contaminated with heavy metals. CBR was applied [101] to improve water-quality monitoring and interpretation. Biochemical oxygen demand (BOD) and suspended solids (SS) concentrations are water-quality indicators commonly used for constructed wetlands.…”

“…Measuring them is expensive, time-consuming and labourintensive. To reduce costs, Lee et al [101] applied CBR to predict BOD and SS concentrations of treated samples. It stored past cases with up to six variables, turbidity, conductivity, redox potential, outflow water temperature, dissolved oxygen and pH, selected for their predictive potential for BOD and SS, cost-effectiveness and ease of measurement.…”

Constructed Wetlands: Treatment of Concentrated Storm Water Runoff (Part A)

“…Furthermore, case-based reasoning of the experimental data set has been applied by Lee et al (2005). The purpose of this follow-up paper is to show how case-based reasoning can be applied for water quality control purposes.…”