A Review of Structure Flaw Life in Mill Operation

Because of the size of mill component pieces, they are often not ideally perfect. This leads to flaw analyses: will the piece with the flaw survive the required life; is it ‘fit for purpose’? The mathematical flaw analyses for mill components are very conservative, and they make basic, theoretical, assumptions which often are only approximations. For example, the exact, detailed, geometry of a flaw may be approximated by some conservative upper bound description. These calculations are used to decide whether a piece may be accepted or should be scrapped, possibly leading to schedule delays, and monetary losses for all parties. Thus, it becomes very interesting to know just how conservative these calculations may be. Yet the people who make these calculations, rarely, if ever, follow this up.

To get an idea of the conservativeness of these flaw evaluation calculations, one needs to follow up on the life of the accepted flawed mill pieces. Often a flaw is calculated as being rejectable per fabrication specifications. However, the flawed piece is still accepted into use, based on ‘additional experience’, extra warranties, having back up pieces, etc. Tracing the operational life of such flawed pieces would provide significant information. However, mining project A&E companies, and their consultants, usually depart right after a mine starts operating. Follow up surveys are also complicated by changes of mine ownership, mine closures, equipment resale/reuse, etc. Thus, one has to make a concerted effort to keep track of such flawed pieces. Otherwise, each flaw calculation starts at ‘ground zero’ again.

I have chosen to track a few flawed pieces arising from my experience. Usually these are castings, however, I have included one interesting case involving shell plate.

Case 1 – 16.5’ diameter ball mill heads with integral trunnions

These are two ball mill heads, made of grey iron, originally for the Carr Forks project in Utah. Care was taken to eliminate shrinkage flaws in the head-to-trunnion knuckle region, but in doing so, the flaw area moved to the conical section of the heads. The castings were radiographed, and showed extensive flaws, such that they were immediately rejectable. However, even if the flaw areas were well beyond level 5 radiography, due to schedules, the heads were accepted for use, with the proviso that a new head casting be kept in stock. The Carr Forks project closed after just a few years of operation, and this mill moved to OK Tedi, in PNG. It has operated there ever since. As of the start of 2019, the mill heads have approximately a 35 year life, and continue with no problems. Using the early radiography rejection criteria, based solely on flaw area magnitude, these were the worst appearing castings in my experience, yet they have proved ‘fit for service’.

Case 2 – 32’ diameter SAG mill heads

The first 32’ mill heads for Mt. Isa had an extremely difficult time in manufacture, for Dominion Engineering. They were an Australian steel foundry’s first attempt at making ductile iron, and the project encountered an extremely high casting rejection rate. Many extra head segment castings were made, with little improvement. Finally, the best components were chosen for use, but these still contained flaws well over 1,000 sq.in. in area. These heads have now operated successfully, about 26 years. One head was replaced, but that was due to slurry erosion, in a completely different location.

Case 3 – 34’ diameter SAG mill heads

In the 1980s some European foundries were producing head segment castings, in ductile iron, to net thicknesses, without machining the head interior surfaces. When these surfaces were inspected, for the Kennecott project, these interior surfaces were found to contain large dross areas. This being one of the first times this phenomenon was encountered, by the Dominion engineers, the castings were put to use. Some of today’s specifications would have rejected all these head segments. Yet these 34’ SAG mill heads have now operated about 30 years, without any consequences. This is especially interesting since the A&E company, which accepted these heads, now has specifications which question/challenge any dross allowances.

Case 4 – 36’ diameter SAG mill heads

The Telfer project also had difficulty obtaining dross-free castings. Dross was found in the conical areas, and on bolting flanges. After much discussion, a ‘stress versus dross string length’ criterion was developed, and most castings were put into service. The bolting flange edges/extremities were cosmetically upgraded using titanium metal putty, which hardened to a suitable strength. Today, these castings have 15 years of operational life with no problems noted.

Case 5 – 40’ diameter SAG shell flange

During the manufacture, in Brazil, of the first 40’ SAG mill, for Cadia, it was discovered that some ingot forged flange plate contained compressed voids in a few spots. One of these spots appeared in a high stress area. This occurred because the ingots were not vacuum degassed, and on some, when the poor area was cropped/cut off, some voids remained, which then did not fuse completely under ingot volume reduction. Thus, flaws appeared as small areas of ‘pepper-like’ magnetic indications. After much discussion, and metallurgical speculation, Svedala made some destructive tests on samples, and demonstrated potential safety in use. This was accepted by the mine owner. Over the course of the 21 years of mill operation, to date, the high stress area was checked several times. At one point, 7 to 8 years into operation, it was noted that the area showed some slight change in the MT indications. Since it was possible that slurry access was occurring, with associated corrosion attack, a small weld buttering/overlay, of about 3 mm cover, was applied. Nothing further has been noted in the 21 years of operation.

Case 6 – 40’ diameter SAG mill heads

This is the most recent example, and actually demonstrates best the value (or loss) of monitoring history. Several Los Bronces head segments had some minor dross indications. However, the A&E company wanted no dross. This is ironic, as this was the same A&E that accepted the Kennecott heads (Case 3). However, the company did not keep performance history records. Also, as the company specifications were typographically cleaned and edited, over the intervening years, the requirement for investigating dross became for rejecting dross, with no consultation with the original group who wrote the specifications. During the ensuing Los Bronces discussions, academic consultants were all negative on the dross, due to theoretical calculations, but none had practical experience. As a compromise, to use the segments, they were cosmetically ‘refurbished’ with titanium metal putty, extended warranties were given, and inspection methods were established. To date, the mills have run for 10 years satisfactorily. As always, the A&E company left after start-up, and has no further history of these castings in operation. Thus, any future project flaw evaluations will continue to have no back-up from experience.

The above cases demonstrate the somewhat unchanging nature of ‘the state of the art’. They also demonstrate the conservativeness of current flaw evaluation methods, which, if not understood, lead to financial losses, and start-up delays. In evaluating flaws, one must differentiate between cosmetic appearance, and fitness for purpose. There are papers available in the mill design literature, that illustrate cosmetic inadequacies, especially of castings, but none, to my knowledge, that accurately tie such inadequacies to subsequent failures in operation.

Basis of Code Weld Fatigue Criteria

Weld fatigue criteria wasn’t developed through the same testing as normal material criteria, like various steels. The codes developed the criteria not by treating welds as only weld metal, but rather as a complete welded structure. Such a structure has a number of description properties: it contains random imperfections (flaws), it contains complex residual stresses, and it contains an unknown precise local geometry. Therefore, the weld criteria was developed for distinct overall geometries (butt weld, T weld, etc.), rather than for a simple metal.

The geometry aspect is often misunderstood, and thus deserves further discussion. The code starts with an as-welded configuration, be it butt weld, T weld, or other. Thus, it recognizes that, a priori, in the design stage, one cannot guarantee a local weld geometry, when fabricating a large structure. When welding un-machined parts, which will not be subsequently machined, on ALL surfaces, the code recognizes that even automatic weld processes will not deposit the same exact geometry over large runs. Each weld process, itself, may also provide different geometry to the final cover passes. To limit this variation somewhat, the code expects that good welding practices are included, such as typical weld inspection visual criteria. Implied also (from drawings and test data) is that all fillet welds, or reinforcements, are equal leg (45 degree). A further variation limit is imposed by expecting the welds to have been 100% ultrasonically inspected, and cleared to a reasonable acceptance criteria for flaws. That this is so, is evidenced by the code concentration on weld toe cracks (and expectation that weld throat will be sufficient in fillet welds).

To create a graphical fatigue criteria for such a range of weld structures, that criteria must be based on a uniform testing procedure, and on a defined, measured, stress. Such a stress could be the “far field”, or “nominal” stress. Indeed this P/A + Mc/I stress is the basic stress considered, for codes avoiding the complex geometries of offshore platform weld joints, and the pictorial definitions of joints, in those codes, usually present a loading arrow to show the far field load considered (in the testing). The weld fatigue criteria obtained from that testing, is then graphically presented, per overall weld joint configuration (butt weld, T weld, etc.). This is shown BS 7608 [1], and as the first case in IIW [2].

As weld joints become more complicated/diverse, and the structural analysis by finite element method (FEA) is used, a slightly more complicated stress, called the “hot spot stress”, may come to be used. This is the stress at the weld toe, found by extrapolation from the far field results. In BS 7608, this is the basis for the “T” class, and IIW contains a second set of graphical criteria for use with this stress. Note that the changes between the first and second cases in IIW (nominal stress versus hot spot stress) is nothing more than a recalibration of the stress plots against a differently calculated stress variable. The second case also admits a more complex loading pattern than the original experimental data used to construct the nominal stress plots. Nothing else has changed in the weld joint configuration.

Often mining specifications require fully ground welds, and here is where another confusion arises. The weld codes are predominantly concerned with weld toe discontinuities and cracking. Therefore, the weld grinding considered by the code is TOE grinding. It can be seen that this is true, since weld grinding is given the SAME stress allowable advantage (factor = 1.3), as any other weld toe treatment procedure, such as TIG dressing or hammer peening, for example, which certainly do not produce the same geometry as toe grinding. Therefore, a “ground weld”, according to code, can still have rough, as-welded, surface elsewhere (exception = butt welds), and does not remove all geometric stress concentration effects.

Because of the above, codes do NOT accept the FEA calculated peak stresses, on fully ground welds, to be used with the graphical fatigue curves such as BS 7608, or IIW first two cases. Peak stresses, correctly calculated, can only be used in IIW case three, the effective notch stress. Note that for this case there is only ONE stress allowable plot, which makes sense, since the correct calculation/use of peak stress eliminates the need for all varied overall joint geometries. These geometries are now already fully incorporated within the FEA calculation of the peak stress magnitude.

The above is intended to illustrate the difference between how weld fatigue criteria is obtained, from that of ordinary steels. It shows that one cannot change the weld stress calculations, without correspondingly adjusting the weld criteria graphs, as per IIW cases one, two, and three. The use of the criteria graphs must be consistent with the stress calculation. It should be noted that similar logic is employed in setting the current, relatively low, stress allowables for large grinding mill castings.

[1] BS 7608, Code for Practice for Fatigue Design and Assessment of Steel Structures, including amendment 1, Feb. 1995.

[2] IIW – 1823 – 07, Recommendations for Fatigue Design of Welded Joints and Components, Dec. 2008.

Cadia Ball Mill Gear “Anniversary”

A gear is aligned in the field, in steps. First a “satisfactory” geometric alignment is achieved, and then it is modified, in steps, relying on mesh temperature, and temperature distribution along the tooth face width. Any static alignment position determines the initial tooth interaction pressure pattern, which then self adjusts, in the running condition, depending on tooth deformation and temperature in the mesh. If this results in a running thermal equilibrium condition, which is within the manufacturer’s guidelines, the alignment is accepted. If a thermal equilibrium condition is not achieved, or if it is outside the manufacturer’s guidelines, a further pinion position adjustment is usually made.

There are other items which influence this process. The tooth profile may have been modified in anticipation of some deflection such as pinion “windup”. Lubricant viscosity and quantity will also affect the amount of deflection “tolerated”. Of course, the overall gear structure also influences tooth deflection. If we want to analyze mesh behavior, we must realize that each pinion position in the alignment process represents a starting point for a new analysis. The unsatisfactory earlier ones should produce a continually increasing thermal solution, whereas the final running position should produce a satisfactory thermal equilibrium.  A paper such as [1] merely scratches the surface of this complex analysis, and omits thermal considerations entirely. It is useful, however, in identifying the overall behavior of different gear structures. It was prepared as a study of the original Cadia ball mill gear.

When the Cadia ball mill was started up in 1997, the initial gear alignment attempt produced damage at one edge of the tooth face. In analyzing the the causes of this, it was noted that the lubricant being used was not among those recommended by the gear fabricator. A new “in spec.” gear lubricant was obtained, and a new alignment produced a satisfactory running gear set for seventeen years. Analytical studies funded at that time, by Cadia, however, put the blame on the unsymmetric geometry of the rim in the Y shaped gear [2].  This geometry was deemed a “design flaw”, regardless of the fact that the gear was running successfully using a more viscous, fabricator recommended, lubricant. The research in [1] was done to counter that argument. In the process, we studied the one item that was “novel” in the Cadia gear/pinion system – the smaller diameter to face width ratio of the pinion, for this range of motor power. As shown in the study, the substitution of a theoretical, larger diameter pinion, reduces the flexibility of the mesh, and thus may be expected to result in an easier process to achieve satisfactory alignment.

In the last quarter of 2014, the original Y gear was replaced with a spare, procured quite some time ago. The new gear was a perfectly symmetric T shape, and was fabricated with a forged rim, rather than being cast, as the original. During the initial alignment/start up, this gear also suffered damage, on the drive end of the tooth face. Since the lubricant was acceptable, alignment continued and achieved the now satisfactory running condition with this gear.

What can be inferred from this recent occurrence? With the complete change in the gear shape surely any hypothetical “design flaws” associated with the rim non-symmetry were removed. Yet similar damage occurred. This just demonstrates the difficulty of aligning a flexible mesh. In both cases there are pinion position starting points which lead to unsatisfactory results. In both cases, the diligent care of the field personnel, combined earlier with a higher viscosity lubricant, which acts as a safety factor against metal-to-metal tooth contact, solved the problem. In both cases, according to [1], a larger diameter pinion would have had a significant beneficial effect in easing the work.

[1] Fresko, M, et al, “Use of finite element analyses in understanding alignment and load distribution in large grinding mill gear and pinion stes”, SME Annual Meeting, Feb. 2004.

[2] Meimaris, C., Duncan, M., and Cox, L., “Failure Analysis of Ball Mill Gears”, SAG Conference 2001, Vancouver.

QA and Fitness for Purpose

Specifications for mining equipment should be practical. They should be written aimed at achieving a goal which delivers safe, reliable, equipment, but without an unrealistic burden on the supplier. On the other hand, a suitable safety factor should be included. It is this safety factor  that defines the range between the ideal, as specified, equipment, and equipment that deviates from specification, yet is still fit for service. This is not so easy to do. Mining equipment specifications have only developed since the early 1980’s, and have changed over time, always trending towards the more stringent. Yet there is plenty of equipment, of earlier vintage, that would not meet today’s specifications, but still had/has long successful operating lives. Sometimes that older equipment is resold, and operates at loads not planned in the original design, and still is successful.

Quality assurance is involved in finding whether a written technical specification has been met. If the specification has not been met, it does not automatically mean that the equipment is not fit to operate successfully for many years. The evaluation of fitness for purpose is rather difficult, especially when restricted solely to theoretical methods. Often experimental testing will reveal very interesting results. An example of this is described below, for mill head castings[1].

Large mill head castings very often have entrapped shrink areas. Specifications require theoretical evaluation of these for acceptability. Thus, the shrink volume is measured, and a crack is introduced into the theoretical calculations, of the size of the overall shrink dimensions. This forms the theoretical pass/fail test. However, a shrink volume is made up of many, unconnected, small shrink pits. Many years ago we gathered ductile iron plates, containing shrink patterns, from various foundries, for testing. The initial goal was to define how long a life these plates would have, when subjected to the stress amplitude defined as the limiting design value, in specifications. The results were completely unexpected. ALL the specimen reached a (mill) 20 year operating life without failure. Furthermore, when these specimen were examined by ultrasonics, x-ray, or magnetic particle, none showed any evidence of crack growth connecting the individual shrink pits. We then tested the specimen at higher levels of stress to determine the safety factors, which were found to be quite large.

To contrast this to evaluation by theory, we note that the theoretical evaluation calculations STARTED with introducing a crack, the full size of the determined shrink zone. The testing program showed that, at maximum limit design stress range, and 20 year life, NO cracking had yet even started, pit to pit, in the shrink zone. Thus, testing can show the conservativeness of the theory assumptions. Similarly, for welds, mill instrumentation has shown that full contour grinding results in a higher safety factor than the “rule of thumb” 30%, allocated by weld codes for all types of toe dressing, and that under overall oven stress relief, welds with higher compression than tension in the stress range, are safer than the weld code discussions indicate.

Cosmetic appearance of castings is certainly important, but one should also remember that without stresses there will be no failure. So when faced with a casting of poor cosmetic appearance (which certainly indicates flaws), one should still remember to ask the following:

1. What is the fatigue stress range in this area?
2. How does this casting compare with other cosmetically poor castings, which have operated, trouble-free, for many years?

If the answers to the above questions are favorable, i.e., low stresses, and no documented earlier failures, then the decision to accept/reject the casting should not be made strictly for cosmetic reasons.

We continue to be the only small, mill evaluation, consulting company that funds independent testing out of it’s own profits.

[1] Svalbonas, V., et al, "Mill Head Castings – Educating Opinions”, SME Annual Meeting, Feb. 2009.

Types of Grinding Mill Structural Failures

In the 1970’s and early 1980’s mills experienced “overall” structural failures. Over time, analysis and design methods have improved, and the vast majority of failures we see today are “local” failures on the mill structure. These are often caused by stress concentrations of the geometry, such as at flange endings, or other local discontinuities. Sometimes they are caused by occurrences of rare local load combinations, which may not have been considered in design. For example, if a mill is stopped by a brake, acting on a flange integral with a segmented head, on rare occasions the brake may grip the flange just at the end of the segment. In that case, the resulting brake forces will not distribute, circumferentially, to both sides of the brake calipers, but only to one side, due to the flange split line. This produces a different load pattern than the expected typical.

Recently there have been some mill failures, on the shell mounted mills, that deviate from the norm. Since the shell cans (without horizontal split flanges) are axisymmetric, it is expected that, on a circular cross-section, every point will see the same stress range. Thus, if we had a circumferential weld, the expectation of a failure crack starting, would nominally be the same for every point on the weld circumference. Indeed that WAS the occurrence on the early “overall” failures. Yet several more recent mills, of the shell mounted design, have experienced circumferential weld failures where the cracking was very severe (through the thickness) in a local arc area, but the rest of the weld circumference showed no signs of cracking [1]. Also, this cracking occurred in only one of a group of similar mills, but different sites showed the same failure patterns.

To explain this, one needs to consider what possible non-symmetry can occur (ruling out a rash of “drop charges” at the various installations):

1. Weld flaw distribution need not be axisymmetric, since QA would not differentiate flaws which just passed a specification from flaws which passed by an order of magnitude. But this would not be unique to shell mounted configurations.
2. Machining one side of a shell plate, like a riding ring, could create thickness non-symmetry, if the fabricated shape were not circular to start. This would be unique to shell mounted designs.
3. In welding without stress relief, the residual stresses do not have to be symmetric. This would be true for field or shop welding.
4. Some non-symmetric distortion is specific to field welding. This is usually measured, and corrected in the field, but with less ease and accuracy than in a machine shop.

Finally one needs to consider shoe bearing design. These bearings are designed for self adjustment. That being so, they can tolerate more rotating structure deformation, yet still operate satisfactorily. But in doing so, they introduce non-symmetric load variations, if a deformation, like a local flat spot, passes over the bearings. Also one needs to consider the effect of friction in these shoe bearing components, as they adjust.

The above is pure speculation, but it does show that there are some non-symmetric influences, and combinations, in shell mounted construction, that may not exist in trunion designs. Since all these are variable for each individual mill, they could explain both the non-symmetric nature of the failures, and why only a few, out of a batch, failed. The answer as to whether any of these speculated effects are really significant, awaits a suitably designed mill instrumentation program.

[1] Svalbonas, V., and Schultz, K., “The Need for Strain Gaging in Mill Design”, CONMINUTEK, April, 2013, Iquique.

Addendum (January 31, 2016):  The above discussion tacitly assumes the failure initiates in the weld. The weld, being the weakest point in the design, and at a location of highest stress, seems to be the likeliest failure initiation area.  However, a recent investigation into an unrelated, fabricated gear rim failure, initiated another line of conjecture.

The riding ring undergoes some significant residual stress, during the fabrication sequences of NDE, rolling, welding, and more NDE. Suppose that these residual stresses combine with undiscovered flaw areas in the riding ring plate. This would create an arbitrary weak point(s) in the riding ring.

View fullsize

The figure above (left and right) shows an identical failure pattern on two different, shell supported mills, at two different mine sites. As noted before, it is natural to assume that the problem started in the weaker weld zone, where it cracked completely through the structure. But then, if the crack starts in the central weld zone, how does one explain that the crack growth does not continue along the weaker circumferential weld, but instead travels into the stronger riding ring parent plate, at both ends? An alternative explanation is possible. The cracks could START at the flaw points in the rolled riding ring plate (perhaps enhanced by points 2 and 3 in the previous discussion), and then meet up, traveling along the weld. While the weld is still the weakest area, and thus cracks through, the problem initiation points are at riding ring plate flaws. This explains also the non symmetry of the failure, since initiation is not dependent on weld stresses.

How is the riding ring plate examined? This is usually done by straight beam UT, which would miss planar flaws oriented through the plate thickness. During rolling, the induced stresses would act on any of these planar flaws oriented perpendicular to the rolling direction. This would create a weak zone along the direction of the observed crack ends, in the figure. While this is by no means a complete explanation, it does point to a possible failure mode, in shell supported mills, which may not currently be considered. The low probability (but not zero) of such flaws, also explains why only a few mills fail.