NL2023080B1 - Leak detection system - Google Patents

Abstract

A leak detection system comprising a sensor for sensing a local acoustic emissions signal which is generated by a leak, which sensor is mounted on a nut by means of a nut cap which has a shape adapted to the nut.

Description

FIELD OF THE INVENTION

The present invention relates to a leak detection system.

BACKGROUND TO THE INVENTION

Leaks are often encountered during static pressure tests of complex pipe and valve systems. A large percentage of the leaks that occur during these pressure tests are associated with valve elements.

Leaks during pressure tests can be a substantial source of nonproductive time. The time spent troubleshooting leaking systems is often the majority of the nonproductive time whereas the time required to mitigate a leak is a lesser portion. This is especially true when the leak occurs in an enclosed part of the system and does not present a visible indication.

Other than visual inspection, there are currently limited options for identifying the source of leaks during tests. Primitive stethoscope devices including screwdrivers and steel rods are used in some cases to detect vibrations caused by a disturbance. Listening in the audible frequency range is limited as it requires personnel to be in the pressured zone, thus introducing additional safety risks. Audible noise is also not a reliable indicator as lab testing has demonstrated that most of the acoustic signal energy is in ultrasonic frequencies above 30 kHz. As a result of these limited diagnostic options, the use of random or best guess servicing of the individual valves is often employed as a system to eliminate the leaks contributing to a high variability in time required to complete routine pressure tests. A science-based system to identify the location of leaks from a safe area outside of the pressured zone is therefore expected to improve test times, accuracy of results, and reduce HSE exposure.

Although not widely used during static pressure testing there are some devices available using acoustic emissions for detecting fluid leaks in valves and piping systems. These fall into two groups of devices - hand held sensor devices and permanently installed acoustic sensors. The handheld acoustic devices have several problems precluding their use in this application. Foremost of which, includes the need to have an operator in the pressured zone manually checking individual valves in series. These instruments also have a relatively high cost for little benefit; they may not significantly reduce the time to locate the source of a leak, as it requires the operator to manually scan various points under pressure throughout the system.

There is a limited selection of dedicated acoustic sensors designed for leak detection available today and even fewer of those sensors are certified for use in classified hazardous areas. Some acoustic sensors are used on permanently installed piping for various low pressure processes, such as city water distribution and waste management. However, they are not designed for high pressure applications, complex piping that endures frequent repairs, or rated for use in hazardous areas. In addition, the available devices have numerous limitations, including sub-optimal mechanical mounting, limited frequency response and low or unknown sensitivity. These sensors often provide minimal data, typically only a single amplitude at each location, and have high costs, thus excluding their use from static pressure testing applications as well.

There is one systemic approach on the market that uses acoustics to detect disturbances - fiber optic distributed acoustic sensing (DAS) systems. The sensor in these system is the fiber optic cable itself. Due to its fragile, finicky nature, installation in extreme, class-rated environments requires substantial protection efforts which makes its use cost prohibitive. DAS systems are also better suited for permanent installations. The install and configuration process can be complex and would become cumbersome and time consuming during repairs, ultimately eliminating the intended benefit of a leak detection system. Therefore, while DAS systems are capable of pinpointing leaks, they are impractical for this application. Lack of viable technology options fueled the research and development of the acoustic emissions system disclosed.

SUMMARY OF THE INVENTION

In one aspect, there is provided a leak detection system comprising a sensor for sensing a local acoustic emissions signal which is generated by a leak, which sensor is mounted on a nut by means of a nut clamp which has a shape adapted to the nut.

The system may further comprise a human machine interface (HMI) which includes an output screen to display information about the acoustic emissions signal. The HMI may be connected to the sensor by means of a modular network comprising one interface module and one or more sensor signal processing modules (SSPM) which are connected to the HMR via the interface module. The interface module suitably comprises a power supply to supply power to each of the SSPMs that are connected to the interface module. The interface module may further comprise a B&B convertor to convert signals received from the SSMP into a signal that is suitable for the HMI.

The acoustic emissions are suitable frequency analyzed to infer information about imminent fluid leaks.

BRIEF DESCRIPTION OF THE DRAWINGS

The drawing figures depict one or more implementations in accord with the present teachings, by way of example only, not by way of limitation. In the figures, like reference numerals refer to the same or similar elements.

Fig. 1 shows a clamp mechanism for mounting the AE sensor on a valve assembly;

Fig. 2a shows a graph of a baseline frequency response;

Fig. 2b shows a graph of a frequency response of a leaking system;

Fig. 3 shows a graph of a frequency response of a leaking system in a low frequency band;

Fig. 4 schematically shows an AE sensor leak detection system;

Fig. 5 schematically shows the leak detection system extended to include more AE sensors; and

Fig. 6 shows an example of a human machine interface output screen.

These figures are schematic and not to scale. Identical reference numbers used in different figures refer to similar components. Within the context of the present specification, cross sections are always assumed to be perpendicular to the longitudinal direction.

DETAILED DESCRIPTION OF THE INVENTION

Disclosed is a system of acoustic sensors with local signal processing and a central Human Machine Interface (HMI) for use in all environments including hazardous areas. The system provides leak associated acoustic signal information in real-time. The displayed data allows the user to rapidly localize the source of a leak in the system under test.

The person skilled in the art will readily understand that, while the detailed description of the invention will be illustrated making reference to one or more embodiments, each having specific combinations of features and measures, many of those features and measures can be equally or similarly applied independently in other embodiments or combinations.

The presented implementation of the invention makes use of Acoustic Emissions (AE) that are created by leaks of pressurized fluids. Testing of a variety of high-pressure valves commonly used in these systems has shown that acoustic signals in the 30 kHz to 200 kHz range are detectable when small leaks are present. Sensors with high sensitivity and low noise in the ultrasonic band, nominally from 30 kHz to 200 kHz, are preferred for optimum performance.

Processing electronics for the sensors signal may suitably be embedded in the sensor body or in a module nearby the sensor, such that there is minimal possibility for the introduction of electrical noise between the sensor and the processing electronics. The digital signal output from the processing electronics is then transmitted using industrial standard hardwire or radio frequency (RF) link protocols.

The HMI can be a PC, industrial computer, or PLC with integral display. The display device must incorporate a suitable data interface for connecting to the signal from the sensor processing electronics module. The HMI device requires a program to convert the digital signal from the sensor into data for display as acoustic energy at each point sensed. The excess acoustic energy displayed for each sensor is separated into frequency bands. The HMI presentation for each sensed point can be numerical or graphical in nature or a combination of the two. The data displayed for each frequency band of each sensor is the excess energy above the reference point level sampled before the actual pressure test. The operator initiates the acquisition of reference values for each sensor through the HMI software. The HMI program stores average values of this reference data taken during a quiet period before pressure is applied to the system to be tested.

The system described solves various problems presented by application of existing market technologies. This system is a permanently installed solution that can be utilized as a system wide diagnostic tool. This can more efficiently monitor equipment in parallel as opposed to troubleshooting equipment at single points in a series fashion. Because of the local signal processing and networked communications connecting to a remote HMI, the need for operators to be in the hazardous area is eliminated. The hardware is better suited for the intended hazardous environment and install is more straightforward and less costly than fiber leak localizing applications. This design achieves reduced safety risks as well improved time and cost savings during pressure testing.

Although initially designed for use while pressure testing equipment in hazardous areas, it has wide potential applicability outside of pressure testing. Some examples of possible uses are: equipment condition monitoring, process leak monitoring, flow assurance, machine learning, stress strain testing, commissioning and start up verification, and preventative maintenance indication. This system could be utilized in various fields and industries including pipelines, subsea applications, chemical processing, electrical distribution, and residential utilities although that is not an exhaustive list. Further signal processing and algorithm design in conjunction with this technology could have wide benefits.

Example Acoustic Sensor System used for Leak Localization

An implementation of the described system was assembled for assessing its suitability to the offshore environment and its ability to help reduce average execution times of routine static pressure tests. The system was designed to comply with the hazardous area requirements of the near well zones on drill rigs. This was accomplished utilizing a combination of explosion-proof and intrinsically safe hardware.

A test system which is illustrated in Figs 4 and 5 was used to instrument twelve valves with AE sensors. The AE sensors used in this implementation were the Piezotron Acoustic Emission Sensor 8152C manufactured by Kistler with the following specifications: 50-400 kHz frequency range, -55°C to 165°C operating temperature range, and ATEX/CSA Zone 0 rating for use in hazardous areas. There are variations of AE sensors that may be more suitable to other applications. The valves instrumented in this test installation were selected because of their history of being failure prone and/or difficult to access and thus troubleshoot.

Months of AE spectral sensor data from a variety of valve types and manufacturers commonly used in these high-pressure systems was gathered in a test environment and analyzed prior to system design. It was determined that AE signals caused by valve leaks were consistently detectable at the valve bonnet nuts. The valve bonnet retaining nuts or cap screws provide a reliable mounting method that is mechanically secure.

A clamp mechanism 12 was designed and manufactured for use in this test system as shown in Figure 1. The AE sensor 14 is attached to the clamp mechanism 12, which in turn is adapted to snugly fit around a valve bonnet nut 24. The AE sensor 14 may be bolted or screwed onto the clamp mechanism 12 as schematically indicated at 23. The ease, consistency and robustness of the sensor mounting to the valves is contributing to the success of the system.

Initial lab acquired leak data showed a detectable shift in broadband frequency response between 30 kHz and 200 kHz. Figure 2a depicts the baseline frequency response and Fig. 2b shows to the frequency response resulting from a system with a leak under 500 psi (3447 kPa). The window between 30 and 200 kHz is indicated in Fig. 2b for reference. Both figures are formatted with amplitude data on the vertical axis and frequency data on the horizontal axis.

Further examination demonstrated that there was considerable variation of signal strength versus frequency depending on the valve type, leak size, and system location. For example, the signals in the lower frequency band propagate farther and therefore are more likely to indicate when an adjacent valve is leaking. Figure 3 depicts a system with a 44 ml/min leak under 500 psi (3447 kPa) that presented primarily in the low frequency band. To maximize sensitivity without significantly impacting cost and complexity, an algorithm was employed to compute the amplitudes for three frequency bands within the identified response range. The three frequency bands defined for this implementation are 25 kHz to 87 kHz, 88 kHz to 137 kHz, and 138 kHz to 188 kHz. The sensor processing module then transmits this AE data from each sensor divided into these three frequency bands for display on the remote HMI. This design allows for maximum sensitivity to pinpoint small leaks yet gives enough information to quickly pinpoint the source of larger leaks and indicate leaks on adjacent sensors (valves) as well.

Figure 4 schematically shows a dual sensor system design. A sensor signal processing module (SSPM) 2 is housed inside an enclosure 1, which may an explosion proof enclosure. The enclosure 1 further houses an intrinsically safe barrier

3. The housing may be provided with cable glands 4 suitable for passage of armored shielded cables 6 which can carry RS-422 signals. The SSPM is designed with a microcontroller that can acquire and process two sensor channels with a bandwidth of greater than 200 kHz in less than 200 milliseconds. The SSPM services requests from the HMI for the frequency banded sensor data. Each module responds to a request for data in less than 0.5 seconds. The message returned to the HMI contains the amplitude information for three frequency bands associated with the two sensors connected to the processing module. Each SSPM also includes two current source power supplies that power the AE sensors and a low noise broadband amplifier for each of the two AE sensors connected to it. A junction box 19, rated for hazardous areas, is placed close to the instrumented valves that serves to distribute wiring to two sensors via NPT connection ports 7. A multi-pair cable 9 connects the sensor through these connection ports to a terminal block 15. Suitable cable is available from Kistler, the manufacturer of the Acoustic Emissions sensors.

Figure 4 shows two different clamping mechanisms, items 11 and 12, to affix sensors 14 to the valve bonnet nuts. Clamping mechanism 12 utilizes an optional nut cover 13 which clamping mechanism 11 does not. Many different clamping mechanisms may be required depending on the desired implementation.

Several armored shielded cables 6 are employed to carry signal information. One of the shielded cables 6 extends between the enclosure 1 and a Nema junction box 5. The Nema junction box 5 houses a power supply 17 and a B&B convertor 18. The power supply 17 is suitably powered with 100-240V AC via line 8, and provides adapted power (e.g. 24 V) to the SSPM 2.

The test system was designed to use a single cable 6 for both communication and power; an isolated RS-422 communication device provides the interface from each of the six SSPM microcontrollers to the cable carrying the RS-422 signal to the remote HMI. B&B convertor 18 converts the RS-422 sensor data to RS-232 for communication back to the HMI.

As illustrated in Figure 5, a plurality of SSPM 2 housed inside enclosure 1, can be linked, for example using RS-422 in shielded cables 6. Each SSPM 2 is connected to its own junction box 19 and terminal block 15. In the assembled system shown in Figure 5, twelve AE sensors are provided. Each sensor was instrumented on an individual valve, but the sensors can be arranged in various configurations including multiple sensors on the same valve for redundancy. This communication design minimized the overall system cost by reducing the cabling requirements installation effort. Figure 5 provides an overview of this system architecture with six SSPMs processing signals two sensors each for a total of twelve sensors.

The HMI in this implementation is an industrial tablet computer. The HMI program displays the AE data from each of the twelve AE sensors in numerical format. The sensor data screen of the HMI updates every 3 seconds, with individual sensor module (2 sensors) updates every 0.5 seconds. The data for the three frequency bands of each sensor is updated during each screen refresh. The normal data not indicating a leak is displayed in green and the data above the leak threshold value displayed in red.

Figure 6 provides a screenshot of the basic HMI implemented in this test system. Color coding is applied to indicate where leaks may exist. A sensor (valve) with one frequency band in red is an adequate indication that a leak may exist at the valve or an adjacent valve. The more frequency bands are indicated in red, the higher the confidence that the leak is near the sensor. If more than one adjacent sensors are indicating with red level data, then assessment of amplitude of signals can easily narrow to the valve with highest probability of the leak.

Summarizing, this invention has the following unique system characteristics:

1. The use of valve bonnet retaining nut or cap screw for the attachment of the Acoustic Emissions sensor to the valves provides a reliable mounting surface regardless of valve geometry.

2. Three frequency bands of data for each sensor are evaluated for alarming at each sensor. This provides more information for use when determining leak location.

3. The use of high level programming language (Python) to produce the code embedded in the Sensor Processing Module allows efficient software development and flexibility in implementing algorithms.

4. The use of a single cable for communication and power from the HMI to all of the Sensor Processing Modules in series minimizes the cost of system cabling and installation.

As described above, the detection system can provide a visual display of acoustic signals from an array of sensors distributed across the pipe and valve system under static pressure test. The sensor and signal processing electronics may be optimized for detecting leak signals from high pressure valves. For best results, it is recommended that each pressure tested valve has a dedicated sensor, although a one to one mapping of acoustic sensors to pressure tested valves is not an absolute requirement to attain benefits from the system. The number and distribution of sensors needs to be sufficient such that the system is adequately covered for localization and the maintenance of high leak sensitivity. The described acoustic sensor system reduces time required to identify sources of leaks during pressure tests of complex pipe and valve system geometries in extreme environments.

Claims (6)

CONCLUSIONS A leak detection system comprising a sensor for locally receiving an acoustic signal which is produced by a leak, which sensor is mounted on a nut by means of a nut clamp which has a shape adapted to the nut. The leak detection system as defined in the preceding claim, further comprising a human machine interface which includes an output screen to display information about the acoustic signal. The leak detection system as defined in the preceding claim, wherein the human machine interface (HMI) is connected to the sensor by means of a modular network comprising one connection module and one or more sensor signal processing modules (SSPM) which are connected to the HMI via the connection module. The leak detection system as defined in the preceding claim, wherein the connection module comprises a power supply which supplies each of the SSPM connected to the connection module with current and / or voltage. The leak detection system as claimed in claim 3 or 4, wherein the connection module is further provided with a signal converter adapted to convert the signals supplied by the SSPM into a signal suitable for the HMI. The leak detection system as defined in at least one of the preceding claims, further comprising a frequency band analyzer for each acoustic signal. 1/4 SP2333 2/4