WP 2: Benchmarking - Breach Formation

M. MORRIS HR Wallingford, UK
m.morris@hrwallingford.co.uk

M. HASSAN HR Wallingford, UK
mam@hrwallingford.co.uk

1. Introduction
2. Description
Table 1 Breach test descriptions
Table 2 Numerical modelling programme - Downloads

Submission Guidelines

The objectives of the numerical modelling are:

• To identify the best numerical approach / approaches for simulating breach formation through embankments
• To assess and quantify the level of uncertainty of current breach modelling techniques
• To incorporate knowledge gained from undertaking the field and Lab. tests into existing modelling tools.

A significant number of the numerical model testing will be undertaken as 'blind' tests to ensure complete objectivity. This means that numerical modellers will be asked to undertake their work and submit their results before the corresponding field or laboratory experiments take place or the results from these experiments are released. Modellers will then also be invited to submit further (revised) modelling results after receiving the field or lab test results ('Aware' testing). This permits the modeller to use modified data sets or models, as the modeller feels appropriate, for the particular test case in order to achieve the best numerical model prediction of the conditions. Details and justification of any variations to the modelling approach (as compared to the blind tests) should accompany any results submitted.

Participation in as many of the tests listed below is encouraged from as wider a group of participants as possible. However, it is recognised that the number of tests listed bellow is large and probably beyond the time availability of many participants. Modellers are therefore invited to undertake a number of core modelling tests and as many of the remaining tests as they are able.

Two tables are presented below:

Table 1 Breach test descriptions - provides a brief summary of each of the sets of data against which numerical modelling is being undertaken.

Table 2 Revised numerical modelling programme - Downloads - provides a summary of the numerical modelling programme, including the type of test, date of data release, date of results submission (see submission guidelines) and the possibility of download the data to run the tests.

Test Description	Test Conditions
Field Test #1	6m high cohesive embankment (25 % clay and less than 15% sand with D50 =0.01 mm).
Field Test #2	5m high non-cohesive embankment (less then 5 % fines with D50 = 5 mm).
Lab Test # 5	Replicate Field Test #2 in the Lab (scale 1:10).
Lab Test # 6	Same material grading as of Lab Test #5 but different face slope (1:2 instead of 1:1.7).
Lab Test # 7	Same material grading as of Lab Test #5 but different crest width (0.30 m instead of 0.2 m).
Lab Test # 9	Same geometry and material grading as Lab Test #5, allowing seepage before overtopping failure

Run Code	Run Description	Release Date	Submission Date	Comments
NF1b	Field Test #1 (Blind Core Run)	Mon 10/02/2003	Fri 07/03/2003	Re-release due to new survey and calibration of gates
NF1a	Field Test #1 (Aware Core Run)	Mon 10/03/2003	Fri 28/03/2003	Release of the field data
NF2b	Field Test #2 (Blind Core Run)	Mon 10/03/2003	Fri 28/03/2003	Re-release due to new survey and calibration of gates
NF2a	Field Test #2 (Aware Core Run)	Mon 31/03/2003	Fri 18/04/2003	Release of the field data
NL5b	Lab Test # 5 (Blind Core Run)	Mon 21/04/2003	Fri 16/05/2003	Direct replication to field event
NL5a	Lab Test # 5 (Aware Core Run)	Mon 19/05/2003	Fri 27/06/2003
NL6b	Lab Test # 6 (Blind Optional Run)	Mon 21/04/2003	Fri 16/05/2003	Assess effect of face slope
NL6a	Lab Test # 6 (Aware Optional Run)	Mon 19/05/2003	Fri 27/06/2003
NL7b	Lab Test # 7 (Blind Optional Run)	Mon 21/04/2003	Fri 16/05/2003	Assess effect of crest width
NL7a	Lab Test # 7 (Aware Optional Run)	Mon 19/05/2003	Fri 27/06/2003
NL9b	Lab Test # 9 (Blind Optional Run)	Mon 21/04/2003	Fri 16/05/2003	Assess effect of seepage
NL9a	Lab Test # 9 (Aware Optional Run)	Mon 19/05/2003	Fri 27/06/2003

• Two dates are provided; a release date and a submission date. We ask that each modeller submit results according to the given submission date for two reasons:
  
  
  • We shall not release field / lab data to a modeller until 'blind' testing has been undertaken. Once the results of the blind test have been submitted then the data for 'aware' test will be sent to the modeller.
    
    
  • We would prefer not to release data to each modeller at different times. It is more efficient to release data sets to all participants simultaneously and to maintain complete objectivity for the modelling programme. (If you participate in the modelling programme and find that you are unable to meet a given deadline, please contact us to discuss how this might be accommodated rather than ceasing participation).

• Please email results and questions to Mohamed Hassan

back to WP2

back to Benchmarks Page

back to Top