Leak detection for tube bundle systems

Tube bundle monitoring is a critical component of gas monitoring systems utilised in underground coal mines in Australia. Kilometres of tubing are installed from locations underground to an analyser room located on the surface of the mine. Samples are drawn to the surface under vacuum to allow for automated sampling and analysis by infrared/paramagnetic analysis or by Gas Chromatograph (GC) (Simtars 2015). This tubing is routinely tested for leakage to validate the integrity of the system. There is currently no standardised method prescribed for leak testing. This opportunity has resulted in the development of several different methods for leak testing (Forrester 2017). This project investigates the application of a leak testing method integrated with the mine's tube bundle system, which employs the process of purging the tubing with inert gas from the surface before drawing it back through an analyser under vacuum. This method allows for routine leak testing, as well as the determination of tube length and location of tube leakage. This method will enable the assessment of tube status during emergency response underground.

Mine site scoping visits and surveys were conducted to determine any constraints or considerations which would impact the capability or functionality. An integrated leak testing prototype was developed in collaboration with Deltamation, and control system software was developed to operate this system. Comprehensive testing of the integrated tube bundle leak testing system was conducted at Simtars with a variety of tube infrastructure and simulated leaks to provide proof of concept and determine the advantages and limitations of this method.

Overall, the testing of the prototype demonstrated the ability for the system to perform leak testing to an accuracy fulfilling the requirements for monthly leak testing by Australian Standards. The system was able to estimate tube length based on tube dimensions and flowrates of the purge phase with an average absolute error of 70m for 5/8” tube and 56m for ½” tube from the true distance of tubing. In the case of a leak, the estimated purge time for a non-leaking tube was sufficient to reveal the leaking tube, and the approximate location of tube leakage was estimated as an average of -55.6m absolute error from the actual leakage point. It was determined that the presence of tube infrastructure was not detrimental to the application of the leak detection prototype apart from non-return valves. The most significant factors for resistance of the tube were the length of the tubing and the accumulation of water. The integrated leak testing system did not function on blocked or kinked tubes, however, self-draining seal panel water traps from two different suppliers were not an obstacle to the purging of water from the lines.

This method of leak detection has the potential to be automated via programmable logic controller (PLC) with an existing or new tube bundle system. When compared to traditional integrity testing as a form of leak testing, this method offers logistical advantages including savings in time and labour, as well as additional benefits in terms of information on status of tubes and tube health. This method of leak testing grants unique advantages for the determination of tube status during a mine emergency where the integrity of the tubes may be compromised and inaccessible.