Inquire about this publication
Project: T192700 #0247

Part 1

This was part one of a three-part project to investigate ice loads on dams. This project consisted of three topics: extreme ice loads and the role of ice physical limits, ice load non-uniformity, and ice loads on small, low-head dams.

Extreme ice loads were evaluated using various methods, including (a) empirical assessment, (b) partially-probabilistic, deterministic analyses, and (c) fully-probabilistic analyses. It was concluded that, to obtain realistic results, a fully-probabilistic analysis is required. Physically-limited 100-year ice loads were established for the Barrett Chute, Arnprior, and Seven Sisters Generating Stations, as 188, 215, and 378 kN/m, respectively. These values are credible in relation to the field data that have been collected at these sites. It is noteworthy that the data vary among the dams, so in order to obtain load-relief benefit from physical limits, ice loads need to be defined for various regions and dams.

Ice load non-uniformity could cause ice loads on short wall sections to be higher than the average over a wide front or in the reservoir ice sheet. Field data from one site for one winter showed that load augmentation occurred. Ice load augmentation was investigated using various approaches. It was concluded that detailed analyses are required to obtain results that are realistic and not overly conservative. Although relevant field data have only been collected at one dam for one winter, useful insights and conclusions were obtained. The 100-year ice loads on a short wall section may be higher than far-field loads by about 5-10% for a 10 m long wall, and about 10-15% for a 5 m wall. This conclusion is valid for a wide range of far-field load magnitudes and sites.

Ice loads for low-head (<10m) dams are of concern for the stability of these structures. An opinion was provided regarding whether ice loads on small dams would be substantially different than those for larger dams. This work was aimed at assessing whether or not the ice loads evaluated in the previous sections are applicable for small dams, and whether special consideration is needed. The work was subjective and considered potential variations regarding the environmental conditions at dams and forebay operating regimes. With respect to either thermal or combined ice loads, I opine that ice loads at small dams would be similar to those for larger dams.


Ice loads, ice, extreme ice loads, ice load non-uniformity, dams


Part 2

This was Task 2 of a project investigating ice loads on dams. The objectives of Task 2 were to quantify the load-sharing that occurs as ice loads in the reservoir are distributed among piers and their adjacent structures (gates, stoplogs, spillway, etc.). A finite element (FE) model was targeted as the method for achieving this.

First, field data to define loads on piers and gates or stoplogs were reviewed and investigated in detail. It was decided to focus the project on the load-sharing between a pier and adjacent stoplogs, as the available field data were most extensive and reliable for this case.

Extensive FE model development was undertaken, starting with two-dimensional analyses and progressing to three-dimensional FE modelling. This provided significant improvement and the stoplog loads predicted by the FE analyses agreed with the field data within 10% for the most realistic parameter values.

Sixteen runs were then made with the 3D FE model to evaluate the ice loads on wooden and steel stoplogs for a range of cases. The FE results were first examined in relation to the following two factors, as they are fundamental for developing an overall design approach:

(a) The proportion of the far-field load taken by the pier – it was found that the pier generally took about 80% of the far-field load, with a range from 69% to 91%.

(b) The width-averaged line load across the stoplog span – the width-averaged line load was about 31 kN/m and there was relatively little variation in general. The only major exceptions were FE runs done with twice the far-field ice load – which resulted in twice the width-averaged line load across the stoplog span – and a longer pier (i.e., 3 m vs 2 m) – which led to a width-averaged line load of 55 kN/m.

Wooden and steel stoplogs were compared. For wooden stoplogs, the analyses provided confirmation of present design practices, as the FE results gave a width-averaged line load of about 31 kN/m within a relatively narrow range for a far-field ice load of 150 kN/m (10 kips/ft). Current design practices generally use an ice load of 30 kN/m (2 kips/ft) for wooden stoplogs.

For the steel stoplogs, the FE results indicated that present design practices are quite conservative. Current design practices generally use an ice load of 75 kN/m (5 kips/ft) for steel stoplogs. The FE analyses produced much lower width-averaged line loads on the steel stoplogs, as they spanned 33 to 42 kN/m over the analysis range.

This project should be continued by using the methodology developed here to (a) establish a design approach for determining ice loads on stoplogs and other hydraulic structures and (b) evaluate a wider range of cases (e.g., ice loads on stoplogs for a wider range of conditions, ice loads on gates, and ice loads on other hydraulic structures such as spillways).


Ice, piers, gates, stoplogs, spillways, dams


Part 3

This project investigated ice impact loads to make a preliminary assessment of whether ice impact loads are likely to be the governing ice loads for a dam. The work started with a literature review, including techniques used previously to measure ice impact loads on structures.

Ice impact loads were evaluated by referring to codes or guidelines and conducting impact analyses based on physics and mechanics, which are likely to be limited by factors such as the energy of the drifting ice floe.

The most important difference between ice impact loads and static ice loads is that impact loads tend to be concentrated over small loaded widths, whereas static ice loads are applied over larger widths.

For dam safety analyses, ice impact loads are unlikely to be the design case, as these design criteria are based on loads over a wide front. For local analyses, ice impacts could create localized failures in, e.g., the dam’s wall or hydraulic structures (pier, gates, spillways, stoplogs, etc.).

It is recommended that this work be followed up with preliminary structural analyses to assess whether the impact loads calculated here are of potential concern either globally or locally and, depending on the results, further refining the ice impact loads calculated here.


Ice, ice loads, ice impact, dams