This appendix documents several inland navigation studies that have performed analyses to determine the

distribution of velocities and impact angles for tows in the design of their approach walls. The purpose of

this appendix is to provide some basic design information on distributions for velocity and angles so that

designers can gain an understanding of the scope of what needs to be developed for their own navigation

study efforts. The data summaries are presented for the typical design parameters (velocity, angle, and

mass) used in design of approach walls at Olmsted Lock and Dam (L&D) (Ohio River), Winfield L&D

(Kanawha River), Kentucky L&D (Tennessee River), Marmet L&D (Kanawha River), London L&D

(Kanawha River), and Greenup L&D (Ohio River). A brief description of the approach walls that were

designed, plans, and hydraulic flow vectors from the navigation model, if available, are presented for each

project. However, many of the distributions presented in these examples are documented using a Beta

Subjective distribution. These distributions have been converted to a lognormal distribution with match-

ing statistical parameters (i.e., mean, standard deviation and percentiles) since the Beta Subjective distri-

bution is not always recommended for probabilistic analysis. Correlation coefficients for the velocity,

angle, and mass are shown for the Marmet L&D project.

(1) The Louisville District began construction of the first phase of the Olmsted Locks and Dam

project in 1993. The Olmsted Locks project began in 1996 and included the construction of two 366-m-

(1,200-ft-) long lock chambers. Toward the end of the construction contract for completion of the locks,

the contract to construct the approach walls began in 1999. The Olmsted approach walls project included

four floating guard walls and one fixed guide wall; the four guard walls are aligned between the dam and

the lock approaches. Figure C-1 shows the layout of the approach walls at Olmsted L&D. The Olmsted

Locks are aligned close to the Illinois shore; thus the approach angles for barge trains entering the locks

are not expected to be large. The walls were designed for the Louisville District following the method

described in ETL 1110-2-338. At the time of the design, there was not yet a set of locks at Olmsted at

which the behavior of arriving barge trains could be observed. Therefore, the design of the walls included

data from model testing on a 1:120-scale model at ERDC and the use of time-lapse videotape of the

approaches at both Smithland Locks and Uniontown (currently called J. T. Myers) Locks since the

characteristics of the barge traffic and the flow of the Ohio River at these projects were judged to be

similar to Olmsted.

(2) The results from the scale model were used primarily to determine the barge impact parameters

for the design of the approach walls. The videotape data from Smithland and J. T. Myers were used to

validate the approach and landing of tows and the currents in the scale models, and engineering judgment

was used to combine the results of these discrete studies in development of the design parameters at

Olmsted. Figure C-2 shows the velocities and flow vectors from the ERDC scale model, and Figure C-3

shows the time trace of the tow as it makes its approach to the locks under controlled landing scenario.

From the processing of the scale model experiments, the probability distribution for the impact angle is

shown in Figure C-4, the probability distribution for the longitudinal velocity *V*0*x *is shown in Figure C-5,

and probability distribution for lateral velocity *V*0*y *is in Figure C-6. The probability distribution for mass

of the tows was taken from downbound traffic data at Lock 52, which is 40 km (25 miles) downstream.

C-1