COTS hardware in oil well production and monitoring

No longer is a laptop with a four-port switch a viable engineering toolset for compliance and monitoring at the well site.

Optimal oil production and the necessary environmental safety have been tied to effectively monitoring down-hole parameters of producing wells. Current trends in technology allow for mapping a large array of parameters in order to optimise the equipment utilisation and anticipate or mitigate revenue-impacting issues. With the advent of advanced fibre-optic measurements, Raman Distributed Temperature Sensing (DTS) and Brillouin Distributed Temperature and Strain Sensing (DTTS), precision down-hole measurements are possible and practical.

The base system is comprised of a series of fibre-optic cables, usually double steel encased, with a reflective mirror attached to the end of each of the fibre cables. Light energy travels to the bottom of the hole and is reflected back to a photodiode acting as an optical receiver. Pressures, temperatures, flow rates and proximal changes in the hole can be monitored along the depth of the hole through Bragg, Raman and Brilliouin photopic (optical time domain reflectometer) modifications on the signal. An analog-to-digital convertor circuit creates a digital profile that is processed through a signal analyser. Data is passed via Ethernet to the server, which is used to process the data using several advanced Fast Fourier Transform (FFT) algorithms.

The resulting data, which is now easily interpreted as temperature, pressure, flow and position, is stored in a RAID array to monitor the well production performance as trend information. These servers and switches are typically powered through a UPS system that is supplied by a conventional on-site generator set.

A rugged 750 W UPS with a 10-port 10/100MHz Ethernet switch and a 2U server can be packaged in a 40 kg 4U transit case to provide the in-field data processing. Data can be stored onsite or uplinked to a central monitoring station by the extraction team.

Industry challenges

The unique problem with monitoring oil wells with this technology is that the level of computing necessary to process the data into meaningful parameters is daunting. The effort is not possible with a Toughbook or laptop-class appliance. Furthermore, standard servers are not made for this type of application. Rig sites are dirty, dusty, inclement places where the electrical power is typically unreliable. Fracking chemicals are highly caustic to electronics and cause premature failures in untreated commercial off-the-shelf (COTS) servers. The site monitoring schedule is typically a few weeks to a month at one site, then off to the next well via the back of a truck. Transportation to the next well site is often via unmaintained roadways. The equipment needs to be small enough to be manually carried and set up on a platform or well head but tough enough to be transported from site to site without being damaged. This situation defines a challenge for a new breed of rugged COTS computing, networking and integration. Using COTS allows for leading-edge computation at a reasonable cost, and integrating and hardening the package provides a highly reliable system in a small form factor. Performing at this level for the oil industry has been an unmet challenge.

Starting point: transit case enclosure

When selecting an enclosure, we must keep in mind a few considerations that may pay great dividends. First, it is tempting to oversize the transit case for future expansion. Unless it is certain additional or deeper components are required later, one should resist the temptation to pick a 6U case when a 4U will solve the problem. The case is a large portion of the weight, and weight is the focus in this application.

Consider selecting a case that uses a removable inner frame. While this seems like an unnecessary luxury, the first time maintenance is required, the feature will have paid for itself. Being able to get around the rack, perform wiring, assemble the system and perform troubleshooting is beneficial when the case is not in the way.

The style and durability of the transit case is also important. Consider a thermal formed or carbon fibre case that has been designed to be stacked, locked and dropped. Cases designed for the music industry are lightweight but not designed for this type of environment or service life. The material needs to be rigid enough as well as tough enough to carry the load without damaging the internal equipment or failing in an unanticipated incident.

Keep in mind the cases can be ‘dynamically tuned’ for the load by the manufacturer making isolation a variable by working with the transit case company. It is important that multi-axis isolation techniques are used for this class of equipment, and to be effective, the correct dampening characteristics need to be matched to the mass being protected. This isolation mass information is usually in the manufacturer’s specification sheets, but contact the case manufacturer for your unique situation.

Figure 1: A rugged portable data collection system. Source: Crystal Group.

The computing challenge

Gone are the days of industrial or military-grade screened parts used in the manufacture of rugged computers. The challenge today is to adapt what is provided for the server class computing industry to harsh environments like those found in well-head monitoring or oil exploration.

The key trends that have dominated the COTS packaging for more than 10 years are:

1. Regulatory pressure to substitute tin-lead solders for tin-silver-copper alternatives has altered the fatigue life curve for cyclic temperature-induced failure, which can create latent reliability issues and ‘tin whiskers’ (spontaneous dendritic growth of pure tin columns). While the consumer electronics industry is unencumbered by the change in these materials, the lower ductility solder fails under higher loads but fewer cycles for SnAgCu based solder¹.
2. Integrated circuit technology continues to advance, allowing for smaller features in the microprocessor silicon, which is in turn impacting the connection density on the printed circuit boards’ ball grid array size and pitch. Finer-pitched features fracture more readily under a set strain or deflection. Intel and AMD are integrating more of the circuitry on the microprocessor substrate (video, memory controller, interrupt control hub, platform controller hub, etc) making the silicon substrate larger. This is chiefly because of the technology improvement mentioned above. Larger silicon packages create a need for less deflection in the boards and therefore the chassis in order to maintain reliable performance.

These factors conspire to make adapting COTS architectures even more difficult, although not unmanageable. Creating a server package that can withstand shock or impact loading and vibration or cyclic deflection is a critical factor in designing a well monitoring system. The heart — and arguably the most sensitive aspect — of the integrated product is the compute capability. Numerous enhancements are required to utilise a COTS motherboard in this application.

The key to making a rugged system lies in constructing a chassis that has such extreme stiffness that deflections in the solder joints are entirely avoided. Standard sheet metal construction for the computer chassis provides insufficient rigidity. An alternative to this approach is a billet-machined chassis. This type of milled box construction does an excellent job of limiting deflection; however, there is a weight penalty. Since this is intended to be a portable application, weight is critical.

Another approach is weight reduction through the use of composite materials such as cross-weave carbon fibre laminates for the enclosures. Weight savings of 20–30% are readily achievable using these materials. Additionally, the stiffness of the enclosures rivals those of traditional material selection making a viable option at roughly a 15% cost premium.

The protection of the electronics and the need for a lightweight structure drive the design considerations for the computer. Staking components for vibration and shock, protecting the electronics from humidity intrusion and using stainless steel hardware are good measures to take in creating rugged, durable systems.

Many of these monitoring systems also require the support of virtualisation software to accommodate a larger number of sensor applications running on a single computer. It is difficult to find, and expensive to develop, an embedded computer architecture that is capable of accepting a server-grade virtualisation operating system (such as VMware or Windows with Hyper V) as most small embedded platforms are not server-grade hardware and lack the compute power needed for virtualisation.

An economical alternative is to use platforms that have already been certified to be compatible with your OS of choice. This tends to be a system level challenge. Included in the mix is the choice of a network switch and its management capability. The same challenges faced on the compute side of the system are seen in the network side.

Clean power is a key issue onsite

Oil production sites are not typically known for having clean power. This was never an issue until high-end computing entered the equation. A month’s worth of monitoring could be lost with a single power spike. The use of a UPS is essential in maintaining the integrity of the data collected at the well site.

Ruggedness and weight are significant factors in UPS selection for well sites. Ruggedness because of the environmental extremes and weight because these systems are expected to be portable. But these are not the only parameters that should be examined when selecting a UPS. Other concerns relate to the use of generators and the ability to provide galvanic isolation. The use of generators and the poor quality they can provide can result in the loss of data or a break in the communication link. Generators can cause frequency instability and notching of the sine wave input to any unprotected load resulting in damaged equipment and downtime onsite reducing the productivity of the well site.

Most commercial UPS systems do not provide the ruggedness to cope with the extreme temperatures or the environment and fail very quickly. Most commercial line-interactive UPSs do not provide galvanic isolation or provide frequency stability to the load and have to utilise the batteries to cope with the voltage fluctuations or the frequency instability. This results in battery failure or a dramatic reduction in battery life. The extreme temperatures also affect line-interactive UPS as the UPS and batteries are designed to operate in a typical ambient temperature of 20–24°C environment.

Figure 2: Double conversion UPS block diagram.

The solution is to utilise a rugged design with a double conversion online UPS with input isolation. Double conversion online UPSs provide frequency stability, voltage stabilisation and complete galvanic isolation to the load without utilising battery power. Batteries are only used when power totally fails or exceeds -15/+25% on the input voltage. This utilisation rate results in longer battery life, less maintenance cost, reduced site downtime and a more productive well site.

A true online UPS has the following characteristics:

• There is no transfer time or interruption of power to the load if a blackout or brownout occurs because the inverter is already online, supplying 100% of the load.
• The true online UPS provides full-time conditioned clean power because the UPS is creating a new clean sine wave after converting or rectifying from AC from the utility to DC and then back to AC. The double conversion online topology fully protects the computer load from all ongoing and often transparent power problems on the utility line.
• Backup AC generators have severely distorted waveforms when supplying non-linear loads such as computers as a result of their relatively high output impedance. All of the standby-type UPSs interpret the voltage distortion as bad power quality causing the UPS to go to battery and back cyclically as generator load changes. Eventually the battery is exhausted, which shuts down the load. Only a double-conversion online UPS will solve these compatibility problems.
• An online UPS can provide tightly regulated output voltage, usually +3%, to the load even if the input voltage varies widely. Many online UPS products can provide this tight regulation without any battery drain.
• An online UPS provides longer run time from its batteries when needed because the battery will not be partially drained during a brownout.
• Overall battery life will be longer for an online UPS compared to standby and standby with boost (line interactive) types.

Conclusion

The need for server-class computing on a well site is a recent trend. This trend has provided challenges for companies that need to carefully monitor production for the purposes of regulatory requirements in addition to production optimisation. The oilfield is a relatively inhospitable place for computers, power supplies and network switches. Selecting the right components and protecting them from damage is a key factor in a long and uneventful life. Chassis stiffness and circuit card protection as well as transit case integration and a double conversion UPS power conditioning make well monitoring a viable science in the industry.

*Jim Shaw is the Executive Vice President of Engineering at Crystal Group (www.cystalrugged.com) and oversees product development and product strategy. Jim has a 30-year history of designing and building rugged electronics for industrial and military applications.

Reference

1. Kostic AD 2011, Lead-free Electronics Reliability – an Update, The Aerospace Corporation, <<http://nepp.nasa.gov/whisker/reference/tech_papers/2011-kostic-Pb-free.pdf>>

Top image: ©stock.adobe.com/au/passarut