Neural networks are increasingly becoming a useful and popular choice for process modeling. The success of neural networks in effectively modeling a certain problem depends on the topology of the neural network. Generating topologies manually relies on previous neural network experience and is tedious and difficult. Hence there is a rising need for a method that generates neural network topologies for different problems automatically. Current methods such as growing, pruning and using genetic algorithms for this task are very complicated and do not explore all the possible topologies. This paper presents a novel method of automatically generating neural networks using a graph grammar. The approach involves representing the neural network as a graph and defining graph transformation rules to generate the topologies. The approach is simple, efficient and has the ability to create topologies of varying complexity. Two example problems are presented to demonstrate the power of our approach.

Summary Since they were first introduced to the market decades ago, polycrystalline-diamond-compact (PDC) bits have undergone numerous technological improvements. One such example is depth-of-cut (DOC)-control technology. Introduced in the early 2000s, this technology emerged as an effective means to mitigate torsional vibrations by managing bit aggressiveness. Today, however, the drilling industry faces the task of drilling diverse sections of the well with a single bit. This is a significant challenge for fixed PDC bits, requiring a compromise that limits vibration mitigation in some sections of the well and/or top-end rate-of-penetration (ROP) performance in other sections. This paper presents an innovative PDC bit that can self-adjust its DOC-control characteristics to the constantly changing drilling environment and mitigate vibrations while delivering improved ROP. The self-adaptation is accomplished through a passive hydromechanical feedback mechanism encapsulated in self-contained cartridges that are installed inside the bit blades. The DOC-control elements mounted on the cartridges respond to the external loads through strategically designed rate-sensitive retraction and extension strokes. During unfavorable dynamic events, the elements engage with the formation and mitigate dysfunctions. During normal steady-state drilling, the elements gradually adjust their exposure to enable fast and efficient drilling. The operating principle of self-adjusting PDC bits is first demonstrated through laboratory drilling tests under confining pressure by use of full-scale prototype bits. The testing was expanded to a research well in the field to assess the ability to self-adapt and mitigate stick/slip vibrations. The field-test facility enables a controlled yet realistic environment with reduced uncertainty that is often not available in the field. The tests compare stick/slip tendencies of bits by building stability maps in weight-on-bit (WOB) and rev/min space by use of downhole measurements under similar operating conditions. Bits with the self-adjusting mechanism led to fewer instances of stick/slip than fixed PDC bits in multiple formation types. Self-adjusting bits significantly expanded the stable operating region and enabled operation at higher ROP. The ability of the DOC-control elements to continuously self-adjust their exposure overcomes several other limitations posed by the fixed nature of traditional DOC control. The technology does not require a fixed predesigned exposure; prevents overengagement because of its ability to retract; and eliminates iterative tuning of DOC control currently in practice. The technology is also anticipated to absorb impacts in interbedded formations and reduce damage from improper starting procedures. With many PDC-bit drilling applications being torque-limited, the technology opens up several possibilities to reduce drilling costs.

The last decade has seen a tremendous increase in the use of oilfield data to understand problems of well construction, evaluate drilling performance and derive guidance on future drilling. Despite the tremendous increase in investment and application of various data mining techniques, the outcomes in the well construction domain have been subpar when compared to that of other industries. Yet analysis of drilling data has resulted in useful predictions on the small scale–comparing a few wells where the huge number of variables that influence drilling (formation, rig, bottomhole assemblies [BHA], etc.) have little fluctuation. However, making actionable predictions on a much greater scale, and in the presence of significant variability, is substantially more challenging. The authors introduce a new platform for automated drilling analysis and optimization. The system starts with an enhanced in-bit measurement device that enables an improved view of downhole conditions at a higher density. This data is then tied to surface measurements through robust data validation and streamlined capture of variables. Finally, the platform ties state-of-the-art, physics-based models, which capture the physics of drilling from cutter-rock interface, to power delivery from the rig through the BHA to quantify the influence of each variable. This coupling of physics-based modeling and models based on drilling data is used to deliver key holistic insights–not just on speed of drilling, but also balance it against borehole quality, efficiency and health of the tools. Results of this new optimization platform, which was tested in the SCOOP and STACK plays of Central Oklahoma, are presented in this paper. The platform shows the power of combining physics and measured data to detect hidden insights, derive actions and enhance drilling performance.

A rigorous engineering and research effort combined with targeted field testing has delivered a new generation of PDC technology. This technology is intended to be used for the most technically challenging drilling applications across the globe.During the development of this technology, several new design features were successfully tested in Sultanate of Oman in traditional PDC applications. Once the new technologies were fully developed, an effort was made to test these PDC bits in historically non-PDC drillable applications.One of the first applications identified for the new technology was in Field-A, Sultanate of Oman. The 8 3/8-in. section at Field-A consists of abrasive sands and hard shales of the Haushi and Haima formation groups and is typically drilled using turbines and impregnated bits.Successful application of PDC bits in Field-A prompted a second application: Field-B. This application is drilled with the turbine/impregnated bit combination through the same formation groups. Penetration rates in Field-B are typically higher and run lengths longer.In both applications, testing started with the first bit out-of-the-shoe using a rotary assembly. The objective was to understand the capabilities of the new technology, then apply key learnings to the next design iterations.The authors will describe the technologies developed for the new PDC bits. These new technologies have been able to extend the typical PDC application range into harder abrasive rocks. Improvements in penetration rate of 25% to over 200% have been realized with run lengths competitive to impregnated bits resulting in substantial reduction in drilling cost. Additional cost reduction is achieved by replacing turbine with rotary. Thus far, savings of up to almost 200% have been realized in a single section compared with offsets and 11-day time savings compared to plan.

During the last decade, a tremendous amount of work has been published on using surface measurement of mechanical specific energy (MSE) for enhancing drilling efficiency and maximizing rate of penetration (ROP). With an increase in directional drilling, surface measurement of MSE doesn't correlate with downhole conditions due to frictional losses along the drill string and bottom hole assembly (BHA). Attempts to measure weight on bit (WOB) and torque on bit (TOB) downhole typically involve expensive tools that are deployed 50-100 feet behind the bit. The authors present a novel in-bit strain sensor that measures WOB and TOB in the bit while drilling. This new approach is completely non-invasive and does not change the BHA. Field testing has shown that in many circumstances only 50% of the surface indicated weight reaches the bit and drilling performance suffers including low ROP and higher vibration levels. This highlights the importance of managing the directional plan and using methods to improve weight transfer. This new in-bit measurement tool has been used to drill over 100,000ft in applications around the world and provides useful insights into drilling performance.