Assessing the Accuracy of Infiltration Measurements Using Finite Element Analysis
Infiltration is an important process in many green stormwater infrastructure (GSI) installations. Yet, roughly 30% of infiltration practices fail at initial installation. One major reason is a lack of knowledge on the infiltration rate of the practice, which is often assumed and not measured. To solve this, GSI design and construction quality control need to utilize infiltration measurements. There are many different methods that can be used to assess the infiltration potential of bioretention systems or similar infiltration based GSI. Ring infiltrometers, borehole methods, pilot infiltration tests, and full-scale flood tests all have different levels of effort and accuracy. The focus of this study is to identify and document the systemic accuracy of these methods to increase certainty in design and construction quality control. Potential sources of systemic error relate to the assumptions of the method, such as assuming vertical infiltration when lateral spreading is occurring. Soil heterogeneity in the field makes it difficult to differentiate natural variations from inaccuracy in testing methods. One way to avoid unknown heterogeneity of soil properties is to numerically establish a soil volume with known homogeneous properties. The numeric soil volume can be used to simulate the operation of different infiltration measurements to determine the accuracy of the method in estimating the given soil properties, such as Ksat. This was done with a transient finite element analysis (FEA) implementing Richard’s equation. The FEA simulates infiltration measurements for measuring Ksat, infiltration rate, and percolation rate. The recession rate from a full-scale bioretention system is also simulated. For each simulation, soil properties are input to the FEA and the simulation is run in accordance with field procedures for a given method. The FEA then outputs the same data that would be observed in the field. This simulated output is post-processed in accordance with the applicable method to generate a simulated Ksat, infiltration rate, or percolation rate. The simulated output is compared to the known Ksat input to the FEA to determine the accuracy of each method. A range of soil properties and antecedent soil moisture typical of bioretention systems were evaluated. The accuracy and level of effort of each method is useful in considering the applicability of each method for stormwater planning, design, construction quality control, and maintenance assessments. The accuracy can be incorporated into an appropriate design factor of safety for infiltration-based systems.