Comment

1 August 2006

Wireless instruments identify leak sources

Oil platforms, plants, and refineries are needlessly burning off valuable product

By Wallace Lueders and Richard Stanley

Undetected valve leakage costs refineries and petrochemical plants millions of dollars in product loss every year, diminishing their profitability and sometimes triggering burdensome regulatory reporting. 

The goal of zero emissions is lost.

Acoustic technology available today lets plant operators detect, mitigate, and even prevent volatile organic carbons leaks and pressure excursions that otherwise drive up costs, reduce profitability, endanger personnel, and cause environmental damage or blemish corporate reputations.

Effective monitoring of relief valves, isolation, and bypass valves up to this time has been a daunting task. The valves are often in inaccessible locations, and existing monitoring technology requires invasive sensors and installation of high cost wiring for power and data acquisition. 

Without continuous monitoring and diagnostic instrumentation, these valves are normally on a periodic inspection schedule with intervals coming from their operating history. This practice leads to expensive handling of valves that are in good conditions, while valves in need of repair lie fallow. 

A very small proportion of the total valve population accounts for more than 80% of product loss and emissions due to valve leakage. Recently introduced industrial wireless instrumentation enables new best practices for valve condition monitoring on a continuous basis. 

The added benefit of automatic time stamping of leakage events also improves EPA reporting accuracy and mitigates environmental emission penalties. By immediately identifying leaking valves, repair resources can focus appropriately, while reducing plant leakage losses by 80% to 90%.

Wireless instrument introduction

For some time, industrial applications have been successfully implemented using licensed radio frequencies at high cost and with significant administrative burden. With the development of many consumer products utilizing wireless communication technology, the availability of good quality, reasonably priced hardware has improved dramatically over the past few years. 

At the same time, the FCC has allocated increasing amounts of bandwidth for license free use at low transmission power. Miniaturization and portability have proven to be of high value and convenience to consumers and business users in office environments. Further advancements in license-free radio products may come at an even higher rated. 

Unfortunately, most of the devices that are now commercially available do not meet the requirements of industrial plants. The requirement of the application usually determines the particular wireless implementation and technology used. 

For example, in some applications, high data throughput is necessary, and in other cases, infrequent transfers and small amounts of data are all that need to transfer. Usually the cost or logistics dictating the use of wireless communication also means no power is available so the product will be battery powered.  

Here, we will concentrate on the type of wireless system suitable for industrial sensors integrated with smart signal conditioning or wireless instrumentation equivalent to advanced loop powered transmitters. 

Some of the important considerations for this type of application are:

• FCC approved devices for use in license free ISM bands (no end user certification required) 
• Battery powered field devices (programmable measurement instruments)
• Deterministic updates and response times 
• Small message sizes, usually just transmitting new sensor values and diagnostic data 
• Good RF signal penetration through typical industrial plant 
• Robust operation in the presence of external noise or interference 
• Secure communication protocol with interoperability to legacy systems

There is considerable interest in wireless communication of data in industrial processes. Wireless sensors, in particular, have the following advantages when compared to their wired counterparts: 

1. Wireless instruments cut installation costs up to 90% by eliminating field wiring. 
2. Wireless instruments offer inherent electrical isolation. They do not conduct electrical noise, electrical spikes, and electrical surges.
3. Wireless instruments are quick to install and easy to move. 
4. The communication path does not corrode or get cut, burned, shorted, or dug up.

Much of the new radio communication technology comes from military developments. These developments have led the U.S. Air Force to conclude that a properly designed wireless communications network will be more secure and reliable than its wired counterpart is.

To understand how to make a reliable industrial wireless sensor, look first at the wireless characteristics and understand that not all wireless is the same. Wireless equipment, can be divided into three distinct groups.

1. High power, high data rate, long range and licensed: commercial radio, TV, military, satellite, and SCADA 
2. Low power, high data rate, short range unlicensed for home and office: Wi-Fi, office data networking, cell phone (convenience with moderate reliability)
3. Micro-power, low data rate, medium range unlicensed for industry: Process measurement, equipment monitoring, control (high reliability and duty cycle)

It is important to understand the differences that exist in the field of electromagnetism or wireless. A wireless communication capability developed for one type application will most assuredly not apply successfully in a significantly different application. 

In particular, a wireless communication capability that will work with industrial sensors must bear up to the harsh, demanding conditions it will face.

Wireless valve monitoring

A variety of sensors suitable for valve monitoring has integrated into industrial wireless instruments with the characteristics described above. 

Pressure relief valves are critical safety devices that act to protect equipment and personnel in the event of dangerous pressure build up in process piping and vessels. Normally closed, these devices are typically subject to control by a spring force and have no auxiliary source of power. 

Therefore, as a result, these valves require a margin between the maximum allowable working pressure (MAWP) and the system operating pressure. Normal manufacturer’s recommendations are system pressure should not exceed 90% of MAWP. 

As the operating pressure ap-proaches the release point of the valve, the valve will simmer or weep much like a teakettle before the whistle blows. When the sealing surfaces are in good condition, emission and product loss is limited to release of excess pressure. However, as time goes on, a percentage of these valves will leak to atmosphere or waste collection systems that normally go to flare. 

These valves are usually on the top of vessels and often at inaccessible locations that require special equipment and safety procedures for in-situ inspections. The typical inspection approach is to schedule the removal and shop inspection of pressure relief valves on a time interval that has an historical or design basis.

Here is an example typical of a large refinery or chemical plant.

The plant has approximately 2,500 pressure relief valves, most of which discharge into flare systems. Company studies have indicated 5 million pounds per year of product is lost to undetected relief valve leakage. 

The value of these products ranges from cents to many dollars per pound. Of course, known leaking valves receive immediate attention. Also and as a precaution, the plant pulls all valves involved in any sort of overpressure event to verify they are not leaking as a result of seat damage during operation. 

In the absence of any event, a five-year inspection interval is the normal schedule to remove valves from service and test them in a specially equipped shop.

During the routine inspections, 30% of the valves turn out to be leaking to some degree, failing the acceptance criteria for installation. Even more disturbing, 10% of them leak so seriously that they are major sources of product loss and possible pollution. 

This equates to 50 valves, or 2% of the total population leaking to a large extent all of the time and are not detected. In contrast, the valves that get attention after an overpressure operation are less likely to leak than those considered operating normally are. Only 10% of these fail leak testing, and few have developed major leaks.

This example proves one thing: Today’s best practices are not effective in detecting and correcting pressure-relief valve leakage. The challenge is how to find the problem valves as soon as they start leaking.

Wireless sensors provide the path to a new set of best practices as far as pressure relief valve monitoring is concerned. This technology enables continuous inspection that can identify sources of leakage as soon as they occur. 

Because they are truly tether free, they can deploy at will with minimal infrastructure and can easily reconfigure as conditions change. There are several available wireless sensor types available for this application, and we will discuss each of them. 

The installation environment will dictate which will be most effective, and the options include acoustic or ultrasonic; temperature, pressure, and multi-input analog; and discrete input sensor packages. All of these packages are intrinsically safe, battery powered frequency hopping spread spectrum transceivers.

The most versatile of the sensors for pressure relief valve monitoring is the acoustic device, which uses a passive ultrasound sensor to detect valve leakage and pressure release operation. 

This sensor is noninvasive and mounts externally on the valve with brackets that allow installation during normal operations. This sensor is effective on most types of pressure relief valves and normally closed isolation valves. 

Because it is completely portable, it can move from valve to valve to optimize utilization of assets in a monitoring program. While this sensor is very effective at detecting leakage in most refinery and chemical plant relief-valve service conditions, there are some limitations.

This sensing technique is not effective for leak detection in low pressure applications below 30 PSI for quiet areas, and pressures should be 50 PSI or higher in normal areas. In addition, this sensor type is not effective for liquid leak detection, but it will work with all types of gas, vapor, and flashing liquids.

Another limitation is the sensor can be ineffective in areas where there is a high level of ambient ultrasound, such as near reducing stations or other piping where there are normally high levels of turbulent flow. In these cases, alternative sensor technologies are necessary for continuous monitoring.

Other monitoring sensors 

Another technique that performs well in certain environments is the use of surface mounted temperature sensors as the input to the wireless instrument. This approach is effective when process fluid temperatures are either considerably higher or lower than ambient temperature conditions. 

For example, steam leaks are detectable when outlet-piping temperature rises even if flow conditions render acoustic sensors inoperable. Likewise, refrigerated process material leakage or flow is detectable when outlet-piping temperature decreases.

These devices can employ any type of surface mount RTD or thermocouple and can add in during normal operations by attaching sensors using hose clamps or other mechanical attachment approaches. Because this sensor type does not rely on ultrasound created by turbulent flow, it can work with all fluid phases including liquids and is applicable also to isolation valves as well as pressure relief valves.

Wireless instruments with pressure sensor inputs provide another monitoring option. An example where pressure sensing is the preferred approach is valve monitoring at an interstate gas-pipeline compressor station. There is a high level of flow-induced vibration at certain relief valves, particularly those immediately downstream of a reducing valve leading into a metering run. Because the gas in the pipe is at or near ambient temperature, there are no extreme temperature changes in piping as a result of relief valve operation. The station is normally unmanned, so there is no audible detection of operation either.

To effectively deal with this set of conditions, the user decided to install wireless pressure sensors as a pressure switch on the valve outlet. This configuration effectively detects overpressure release events, but is not capable of detecting low levels of leakage. The pressure relief valve in this case was a soft seat pilot operated valve construction, so leakage was not a concern. 

Another pressure sensing approach is to install the wireless instrument at the valve inlet and report an alarm condition if the pressure exceeds the MAWP or set pressure of the valve. With this approach, the user obtains the additional benefit of real-time pressure monitoring at the valve inlet as well as the ability to calculate mass flow of release events.

Still another possible wireless instrument approach to relief valve monitoring is use of multi-input field units. These devices accept multiple analog and discrete inputs. The analog inputs can be either raw or conditioned transducer signals, and devices can be specified for either milliamp or voltage inputs. The discrete inputs can serve as an effective wireless rupture disc burst indicator. For applications where rupture discs are in place to isolate relief valves, these wireless instruments enable the installation of low cost disc burst monitors without the high cost of wiring them up to an alarm system.  

Site survey and system

Before any installation involving a wireless system, it is important to perform a site survey to determine if there are any potential RF communication problems. 

Perform a signal strength test using wireless instruments equivalent to what will install. The signal strength test should take place at the base radio and also at each of the proposed field unit locations. 

This can take place quickly and safely with the actual devices that will eventually serve for the continuous monitoring. If this test indicates possible interference from other devices, a spectrum analysis of the site to obtain a better understanding of the electromagnetic environment is necessary.

The spectrum analyzer test allows one to measure the RF received signal strength at a given location of signals on various frequencies. The measurements made during this test are important in determining what kinds of interference a product might experience in a given environment. 

If a certain frequency has a high level of noise being transmitted on it from another device, that frequency is avoidable by using an RF system. Mitigating this type of interference is possible by using Frequency Hopping Spread Spectrum devices.

An additional survey test is the data link test to measure potential error rates over each of the proposed wireless paths between the base radio and field units. It operates by sending a burst of packets between the base unit and sensor unit, and then allows the sensor unit to report statistics. This test could have many configurable options—the amount of data transmitted in the payload of the packet, the error correction and detection schemes to use, the number of packets to send in a burst, the pattern to use to mark the beginning of a packet, the power level to transmit the packets at, and the frequency to use.

ABOUT THE AUTHORS

Wallace Lueders (wlueders@accutechinstruments.com) has engineering and operations research degrees and an MBA. He works at Adaptive Instruments and has been in the control and instrumentation industry for 30 years. Richard Stanley (stanlra@bp.com) is a senior member of ISA, a registered PE, and technical specialist instrumentation and controls at BP. He has a chemical engineering degree.