« Previous | Contents | Next »
Listen
6 Audit Task 'b'. Strength Model of main cable
The objective of this task is to consider whether the assessed strength of the main cable, based on the findings of the internal inspection, can be accepted as reasonable. The strength assessment has been based upon the guidelines given in NCHRP Report 534. The audit has therefore considered whether the methodology used to record the condition of the cables, and hence to assess the current strength of the main cables was in line with the NCHRP Guidelines and has also considered whether those guidelines are strictly applicable to the Forth Bridge main cables. This task has been broken down into five distinct sections:
- The scope of the internal investigation.
- The methodology of the site investigations.
- The findings and recording of the internal investigation.
- The findings and reporting of the associated testing.
- The calculation of cable strength from the inspection and testing.
It is considered that because FETA required the investigation to be undertaken using the NCHRP Guidelines, the appointment of Faber Maunsell, in conjunction with Weidlinger, to develop, manage and oversee the investigation was appropriate. (2.2)
6.1 The scope of the internal investigation
The initial scope for the first internal investigation of the main cable is considered to be in line with the guidelines. Six panels were selected based upon the conditions noted during a detailed visual inspection of the cables. The scope was flexible which permitted the number of panels to be increased to ten once the initial intrusive inspections had determined the general conditions of the cable.
It should be noted that the initial six panels were selected for the internal inspection based upon the conditions noted during a detailed visual inspection of the cables. However, once these panels were opened there was little correlation between the initial, external visual inspection and the condition of the wires within that panel. It therefore cannot be assumed that the panels selected will represent the panels with the worst condition wires. (3.1 - 3.4)
The audit has identified the following aspect where the scope of the inspection was not strictly in accordance with the Guidelines:
Removal of a cable band (4.16)
- Page 2-6 of the NCHRP Guidelines states that the parties should be prepared for the removal of one or more suspenders and cable bands, if conditions require it.
- Page 2-26 states that one or more cable bands should be removed wherever numerous broken wires are found.
- Page 2-31 implies that a cable band should have been removed to compare the level of deterioration in areas near the cable band with areas in the middle of the panel. The Guidelines expect this to be done whenever stage 4 or broken wires are present.
Faber Maunsell have stated that consideration was given to the removal of a cable band but the decision to do so was taken no further owing to the difficulty of the work for the perceived benefit 1. Removal of a cable band is an extremely complex procedure requiring a substantial period for planning and preparation involving long term lane closures and traffic disruption. The decision not to proceed is understood and would have been unlikely to affect conclusions drawn from this initial intrusive investigation. The conditions found in the cable will necessitate further inspections within a five year period and it is expected that the decision to remove of one or more cable bands will be reviewed at that time in the light of the findings. It should also be noted that the NCHRP recommendations are based around the typical US suspension bridge designs that have cable clamps on a 30 ft (9.14m) spacing. On Forth, the spacing is double this, effectively doubling the amount of cable available for inspection.
6.2 The methodology of the site investigations
Overall, the methodology adopted for the work of the site inspection team was considered to be very much in line with the NCHRP Guidelines.
There are a few aspects where the Guidelines would indicate a more detailed examination of the exposed areas of cable was required. The diameter of the Forth Bridge cables is approximately 24". Within the Guidelines, it is recommended that on larger cables (Diameter >24") additional wedge lines should be driven in the outer half of the cable radius where the wedges are at a maximum spacing, to avoid reducing the fraction of wires that are observed; particularly when many broken wires are found and Stage 4 is expected to be extensive. The number of additional wedge lines applied to the ten inspected panels was only two and in each instance it allowed a decrease in the recorded severity level of the wires. In view of the statements in the Guidelines and the conditions found, additional, intermediate wedge lines would have been justified. (4.7) However, it is considered that such additional work would be time consuming and unlikely to result in any particular difference in the overall results. It would be justified if there was a particular local anomaly that needed to be further investigated.
6.3 The findings and recording of the internal investigation
The general impression from the records provided (and supported from the site visits made by audit team members during the period that the inspections were carried out) is that a high level of confidence can be given to the quality, accuracy and consistency of the recorded data for the internal investigation. The NCHRP Guidelines provide detailed instructions over the manner in which the cable condition is to be recorded and these have been adhered to in general. Confidence is also achieved by Faber Maunsell using the same two inspectors throughout the investigation and reporting that comparisons of their findings were generally identical. Weidlinger also reviewed some areas and satisfied themselves over the accuracy and consistency of the recorded conditions. (3.11 - 3.12)
We know from our discussions with Weidlinger that the inspectors used a slightly modified Stage classification. Stage 4 wires are defined as having 25% or more of the surface corroded rather than 30%, as specified in the NCHRP Guidelines. This results in a larger percentage of Stage 4 wires and no cracks found in wire classified as Stage 3. It is our opinion that this adjustment in procedure is reasonable as this classification is subject to the judgement of the inspectors in any case. For the cable strength calculations, the larger number of Stage 4 wires will result in a larger strength loss when the NCHRP method is used to project the number of cracked wires in the Stage 4 population. Using Weidlinger's latest assumption that only Stage 4 wires in the outer 6 rings are cracked, the strength calculation will be little affected by this shading of corrosion stages.
Samples of wires for testing were only taken from the outer eleven layers, or "rings" of wire. This is not common practice 2 as it may not be representative of the larger population of wires throughout the cable cross-section. This becomes especially important when one considers the assumption being made in the strength calculations that only wires in the outer six rings are cracked.
It was reported to Ammann &Whitney that it wasn't possible to remove deeper samples. This is contrary to the audit team's experience, as we routinely remove samples from several inches into the cable. However, it is very difficult to repair the wires in these deeper locations, thus removing wires for sampling would have a detrimental effect on the main cable strength. Faber Maunsell followed a policy of avoiding any works which would further compromise the strength of the cables.
6.4 The findings and reporting of the associated testing
Eight 6m long sample wires were removed for testing from each of the ten inspected panels, for a total of eighty .These sample wires were cut into 457 mm (18 inch) long specimens and a total of 704 were tested to failure in accordance with the American Standard, ASTM A 586 and A 370. About 10% of these had strain measured to failure. The others had strain measured to about 2% elongation, after which the extensometer was removed and the test continued to failure. Only 17 specimens broke at less than the original, specified strength of 100 tons/in2, eight of which had cracks. A total of 17 specimens were found to have cracks. Three of the cracked specimens were from the same wire sample 1031, three more were from the same sample 363, two from sample 551 and two from sample 961. The total of 17 cracked specimens therefore came from 11 sample wires.
Uncracked wires appear to be of high quality with good ductility.
Seventeen specimens of Stage 1 and 2 wires were tested for zinc coating weight and were found generally within the specification.
Seventeen specimens were also subjected to Preece tests and it has been reported that only five were within expectations for new wire. It is interesting to note that even two Stage 1 specimens did not meet the standard. It might be inferred that the original wire did not meet the specifications for the zinc coating.
Four specimens had chemical analysis performed on them. All appeared to be generally within the original specifications, although two had low Carbon contents of 0.74 and 0.75%. Original specifications called for 0.75 to 0.80% Carbon. This range indicates good quality, uniform wire (based on these very limited number of samples). Weidlinger reported that there is apparently good correlation between tested wire strengths and carbon content (the stronger wires having the higher carbon) as expected.
No fatigue testing has been performed. Fatigue testing is not specified in the NCHRP Guidelines because it is recognised that fatigue is not normally an issue in the cable service loads. However, it is the audit team's experience that fatigue testing is useful in identifying incipient cracking in wires that perform well in static tests. This may be advisable in determining a more reliable estimate of the percentage of cracked wires in the Stage 3 and Stage 4 population. Faber Maunsell reported in the 5th January 2006 meeting that they would be considering this for the additional tests being programmed at Lehigh University.
6.5 The calculation of cable strength from the inspection and testing
The estimation of cable strength is a probabilistic exercise, due to the impossibility of knowing the properties of each individual wire in the cable. The models for strength estimation depend on data gathered during internal inspections and on laboratory testing of samples removed during those inspections. It therefore follows that a more extensive and detailed investigation will increase the level of confidence that can be given to the estimation of cable strength.
In general, the NCHRP Guidelines have been followed by Weidlinger, with some minor refinements and with some more significant departures and expansions of the degradation model calculations.
A small difference is that Weidlinger used a Normal Distribution for the strength of Stages 1 and 2 wires and a Weibull Distribution for Stages 3 and 4. NCHRP recommendations specify a Weibull Distribution for all stages, but Weidlinger's opinion is that this is more representative of the actual distributions. In our experience, this would affect the results in a very minor way and is not significant.
An important assumption in Weidlinger's latest calculations is that cracking has only occurred in Stage 4 wires located in the outer six rings of the cable. Based on the test data and an extrapolation of it, Weidlinger has assumed that 26% of these outer Stage 4 wires are cracked. This is different to the NCHRP guideline procedure which does not segregate Stage 4 wires based on their locations within the cable.
We received Weidlinger's calculations for both cases; one using the NCHRP assumption that all Stage 4 wires are potentially cracked, and the modified assumption that only Stage 4 wires in the outer six rings are potentially cracked. The more conservative NCHRP method results in a 2004 strength loss of 15.4%, while the modified assumption, which follows also follows a distinctly different mechanism, produces a 2004 strength loss of 6.7% (We have independently calculated the current cable strength using the modified assumption and Weidlinger's data, and have produced the same result) 3.
We are therefore reasonably assured that the mechanics of Weidlinger's calculations are correct. The results hinge primarily on what is assumed regarding the number of cracked and broken wires in the cable.
We believe the NCHRP's assumptions are supportable since the number of cracked wires is projected based on a broad interpretation of test data and observed conditions in the cable. In general, this calculation method is conservative and as reliable as the data (assuming sufficient data has been obtained - however much statistical rigour is brought to bear, the available data in this case is still a small number on which to depend).
Because some US bridges have been found to have extreme variations in conditions from panel to panel, we believe the only sure way to know we have found the worst condition is to open and inspect the entire length of the cables. This has never been done except in conjunction with a cable oiling and re-wrapping project, during which every panel is wedged for oiling, and also inspected as the work progresses.
Weidlinger's modified assumption that only the outer six Stage 4 wires contain cracks needs to be considered closely as it is a departure from the NCHRP Guidelines, especially considering the small number of samples taken and the fact that no samples were taken beyond the outer eleven wires. On the other hand, a very small number of cracked and broken wires were found, and these tended to be clustered within the outer few rings at the bottom of the cable (with a few at the sides). This is consistent with our experience on other bridges. We often feel justified in considering these parts of the cable as a separate population. We therefore do not find Weidlinger's assumption unreasonable.
Following the audit review meeting on 5th January 2006, Weidlinger confirmed that having reviewed all strength models they considered the final reported figure for the current strength loss of the cable should be at least 8% owing to inconsistencies in the cracking patterns in the higher strength predictive models (The 8% figure corresponds with their modified NCHRP nominal strength with stage 4 wires in the outer 6 rings potentially cracked). This is an appropriate and reasonable response to take in the light of the amount of data gathered; we are in agreement with their final figures.
« Previous | Contents | Next »