Large Diameter Steel Pipe Piles Running Under Self Weight In Soft Clay ' Predicted Vs. Observed Behavior ' Richmond San Rafael Bridge Seismic Retrofit

Dover, A. R.
Organization: Deep Foundations Institute
Pages: 11
Publication Date: Jan 1, 2006
The Richmond San Rafael Bridge (Figure 1) is a 4.3 mile long major toll bridge span across northern San Francisco Bay. Tutor-Saliba/Koch/Tidewater (TSKT) conducted a comprehensive seismic retrofit, which was at the time Caltrans? largest-ever construction contract at $484 million. The robust foundation retrofit design of 11 bridge piers included the installation of 26 new, large diameter (126 to 162 inch OD, wall thickness 1.5 to 2.25 inches) CIDH and CISS piles, with steel pipe piles installed to tip elevations of Elev. ?140 to -220 ft, in water depths up to 70 feet. The open-ended steel pipe piles had initial sections ranging from 71 to 172 feet long, with full lengths that extended to the bay bottom and then through 100 to 130 feet of marine sediments to top of bedrock. The marine soils of most interest consisted mainly of soft to firm Young Bay Mud (YBM), a silty, moderately-sensitive clay common to SF Bay. A key constructability issue was the prediction of pipe piles ?running? under self weight (198 to 625 kips in air) and pile driving hammer weight (IHC S-500 with total assembly weight of 322 kips). The term "pile running" is used to describe the penetration of the pile into the soil due to the self-weight of the pile only. This issue was critical because 1) there was limited head room under the bridge deck, and 2) if piles run too deep, the pile splicing on several piles would be problematic (the lowest desirable pile top elevation after running was +10 ft). While running measurements are sometimes made during marine/nearshore construction, there is little published data on predicted vs. actual running. This paper discusses the running prediction under self weight for 26 piles, and the methodology used in developing the predictions, which ranged from 15 to 94 ft of penetration below mudline. Pile running was recorded and compared with predicted values. The prediction methodology was quite good, and no significant constructability problems were encountered. The authors ?tweaked? the model to provide a slightly improved prediction method. The paper presents the model, predictions, observations, and discusses important lessons learned.
Full Article Download:
(281 kb)