Bit Whirl - A New Theory of PDC Bit Failure

Abstract

Summary. This paper presents the results of a study showing that the most harmful polycrystalline-diamond-compact (PDC) bit vibrations can be attributed to a phenomenon called "bit whirl." During whirl, the instantaneous center of rotation moves around the face of the bit, and the bit whirls backward around the hole. Cutters on a whirling bit can move sideways, backward, and much faster than those on a true rotating bit. The impact loads associated with this motion cause PDC cutters to chip, which, in turn, accelerates wear. Laboratory and field results show the detrimental effects of whirl on PDC bit rate of penetration (ROP) and life. Limits of PDC Bit Application Since their introduction in the early 1970's, PDC bits have almost completely replaced three-cone bits for use in relatively soft, nonabrasive formations. They sometimes replace three-cone bits in harder or slower drilling intervals, if the section is uniform. It is fair to say, however, that drillers do not normally consider selecting a PDC bit when drilling in harder formations or even in soft formations with infrequent hard streaks. The problem with running PDC bits in hard formations is not the ROP's; adequate ROP's are PDC bits in hard formations is not the ROP's; adequate ROP's are possible, at least for short periods of time. The problem is that bit possible, at least for short periods of time. The problem is that bit life is too short. Fig, 1 shows that a sharp PDC bit can achieve ROP's several times those of three-cone bits in Carthage marble. Carthage marble has an unconfined compressive strength of roughly 16,000 psi and is stronger than rocks usually drilled with PDC bits, yet a new PDC bit will typically drill two to three times faster than the best PDC bit will typically drill two to three times faster than the best three-cone bit. Field experience also usually supports the claim that bit wear, not initial ROP, limits PDC performance in harder rocks. For example, Fig. 2 shows a typical ROP comparison between a PDC bit and the best three-cone effort through a section of the Oswego limestone at the Catoosa test facility near Tulsa. The Oswego limestone has an unconfined compressive strength of about 18,000 psi. Winters and Onyia describe in detail the lithology at the test site. The bit initially drilled the limestone at a rate three to four times that of the three-cone drilling-rate baseline. After 20 ft, however, the ROP fell to nearly the three-cone baseline. Glowka and Stone and Zijsling also claim that wear, not initial ROP, limits the application of PDC bits in harder, more abrasive rocks. Unfortunately for the drilling industry, a good percentage of total rotating hours occurs in slower drilling situations where the ROP advantages of PDC bits cannot be consistently achieved. The degree to which PDC bits have penetrated the bit market demonstrates this fact. Even in the face of an aggressive program to apply PDC bits to operations, only 4% of the bits Amoco Production Co. purchased in North America during 1987 were PDC. Even though PDC bits represented just 4% of the total number of bits Amoco purchased, they accounted for some 15% of the total footage drilled, which shows that most PDC bit use was confined to softer intervals. PDC bits clearly have had a dramatic impact on the soft, faster-drilling formations, but they have yet to replace three-cone bits in most slower drilling situations (i.e., under 25 ft/hr). Moreover (to the extent that Amoco's North American drilling activity is representative of the entire industry), Fig. 3 shows that 60 to 80% of all rotating hours occur at rates less than 25 ft/hr. Most of the footage is drilled at high ROP's, but most of the time is spent drilling relatively slowly. This statistic is similar for Amoco's New Orleans and Denver regions, which shows that even in widely different geologic provinces (i.e., deltaic to midcontinental), most of the drilling is done at slow ROP's. If the ROP advantages of PDC bits could be consistently applied to harder intervals, the number of hours spent drilling each year would be greatly reduced. Sharp and Dull PDC Cutting Mechanisms While Fig. 1 shows that a new PDC bit will penetrate several times faster than a roller-cone bit in hard rock, other evidence shows that the performance of a slightly dull PDC bit can actually be much worse than that of a three-cone bit. The reason for this drastic decline in performance is that the bearing area increases as wear-flats grow, producing lower stresses in the rock. Feenstra, Glowka and producing lower stresses in the rock. Feenstra, Glowka and Stone, and Warren and Sinor discuss this point at great length. One observation by Zijsling that has not been as widely reported, however, is that an apparently dull PDC bit (i.e., one with significant wear-flats) can still drill almost as fast as a new bit. The Carthage limestone drilling results shown in Fig. 4 demonstrate one such case. The explanation for this phenomenon can be seen in Fig. 5, which shows a PDC cutter that has 0.3 in.2 of wear-flat. But because a diamond "lip" sticks up above the tungsten carbide, the tungsten-carbide substrate does not act as a bearing surface during drilling. The cutter's ROP performance is not significantly reduced because only the diamond table, not the entire tungsten-carbide wear-flat, acts as the contact area. Therefore, high contact stresses are still maintained. Lip size varies with the formation being drilled, the operating conditions, the ROP, and the location of the cutter on the bit. Because of the phenomenon of lips, wear-flat area by itself is not a reliable indicator of a PDC bit's drilling performance. The bit can have a significant wear-flat area but have diamond table lips and still drill almost as fast as a new bit. Alternatively, a bit can have a limited number of wear-flats but no diamond table lips and drill much more slowly than a new bit. Therefore, we define a "sharp" bit as one with diamond table lips on enough cutters to cut the bottom of the hole completely. A "dull" bit does not have enough lips on the cutters to cover the entire hole. The presence of lips is important for efficient drilling in hard rock; however, our laboratory and field experience indicates that they are not critical for drilling in soft rock. Bits that cannot drill the hard streaks (because of too much wear-flat area) can drill as fast as a sharp bit in the soft intervals between the hard streaks. Fig. 6, which shows the results of laboratory tests comparing the depth of cut per revolution vs. rotary speed for various operating conditions and bits, presents a way to explain qualitatively the different cutting mechanisms of sharp and dull PDC bits. During drilling with mud at high borehole pressure (with either three-cone, diamond, or dull PDC bits), a decline in penetration per revolution exists at a rotary speed below 600 rev/min. However, penetration per revolution does not decline with a roller-one bit at penetration per revolution does not decline with a roller-one bit at low borehole pressure or with a new PDC bit (at any borehole pressure). In the laboratory, sharp PDC bits (even with mud at high pressure). In the laboratory, sharp PDC bits (even with mud at high borehole pressure) do not drill less per revolution as rotary speed increases, which implies that sharp PDC bits drill by a different mechanism than dull PDC bits. Garnier and Lingen previously showed that bits with crushing cutting mechanisms (i.e., three-cone, diamond, or dull PDC bits) create enough pulverized rock flour to cause chip holddown. As the rotary speed increases, it becomes increasingly difficult to clean the bottom of the hole effectively between impacts. The differential pressure across the bottom of the borehole is not relieved, and drilling efficiency declines. Under low borehole pressure, however, differential pressure is insufficient to hold pressure, however, differential pressure is insufficient to hold down the chips, resulting in a more constant penetration per revolution as rotary speed increases. With clear water, the pressure is partially equalized and the penetration per revolution pressure is partially equalized and the penetration per revolution decreases, but not as much as with the mud. P. 275