Critical Piezometric Levels for Dam Safety Monitoring

A case history is presented for re-establishing critical levels for piezometric monitoring at existing earth dams. Measurement of piezometric levels within the dams has been a major part of dam safety monitoring at the Blenheim-Gilboa Pumped Storage Project since early 70's when the dams were constructed. More than 40 hydraulic twin-tube piezometers were installed within and in the vicinity of the dam. The dam design engineer originally provided the operator the critical levels for piezometric monitoring prior to the reservoir operation. Over the years, it was found that some of the critical levels were exceeded by up to 12 feet, while others were under by more than 20 feet. Such significant difference between the established critical and monitored piezometric levels raised doubts about the reliability of the critical levels for dam safety monitoring. In addition, many existing piezometers were found to be malfunctioning and were subsequently abandoned. Some of the abandoned piezometers were replaced with new piezometers. The exceedances of critical levels and the installation of new piezometers prompted the establishment of new critical levels for both the existing and new piezometers.

New flow nets were developed to re-establish the critical piezometric levels. Several methods were evaluated for development of a flow net. These methods included computer seepage programs and manual graphing using assumed anisotropy, permeability and soil conditions. However, due to little data available on soil anisotropy, permeability and presence of multiple native and man-made soil materials, the reliability of the aforementioned methods would be highly dependent on the assumptions made. Therefore, these methods were not considered appropriate for the flow net development.

A trial of traditional hand graphing based on anisotropy presented in the original dam design did not provide satisfactory flow nets. A similar trial based on isotropic soil conditions also proved unsatisfactory. Utilization of a graphing program and existing piezometric readings was evaluated and considered not feasible due to limited data points available and presence of multiple soil types.

It was found that the best method to develop a representative flow net is traditional hand graphing using the maximum piezometric readings while proper consideration was given to anisotropy in the native soils. Four flow nets (each for an instrumented cross-section) were developed. The resultant flow nets satisfy essentially all boundary conditions and laws of seepage flows. Based on these flow nets, new critical levels were established and compared with historic piezometric readings and original critical levels. 11 pp., 2 references.