Finding Trapped Miners by Using a Prototype Seismic Recording System Made from Music-Recording Hardware

Abstract

The goal of this project was to use off-the-shelf music recording equipment to build and test a prototype seismic system to listen for people trapped in underground chambers (mines, caves, collapsed buildings). Previous workers found that an array of geophones is effective in locating trapped miners; displaying the data graphically, as well as playing it back into an audio device (headphones) at high speeds, was found to be effective for locating underground tapping. The desired system should record the data digitally to allow for further analysis, be capable of displaying the data graphically, allow for rudimentary analysis (bandpass filter, deconvolution), and allow the user to listen to the data at varying speeds. Although existing seismic reflection systems are adequate to record, display and analyze the data, they are relatively expensive and difficult to use and do not have an audio playback option. This makes it difficult for individual mines to have a system waiting on the shelf for an emergency. In contrast, music recording systems, like the one I used to construct the prototype system, can be purchased for about 20 percent of the cost of a seismic reflection system and are designed to be much easier to use. The prototype system makes use of an ~$3,000, 16-channel music recording system made by Presonus, Inc., of Baton Rouge, Louisiana. Other manufacturers make competitive systems that would serve equally well. Connecting the geophones to the recording system required the only custom part of this system - a connector that takes the output from the geophone cable and breaks it into 16 microphone inputs to be connected to the music recording system. The connector took about 1 day of technician time to build, using about $300 in off-the-shelf parts. Comparisons of the music recording system and a standard seismic reflection system (A 24-channel 'Geode' system manufactured by Geometrics, Inc., of San Jose, California) were carried out at two locations. Initial recordings of small hammer taps were carried out in a small field in Seattle, Washington; more elaborate tests were carried out at the San Juan Coal Mine in San Juan, New Mexico, in which miners underground were signaling. The comparisons demonstrate that the recordings made by the two systems are nearly identical, indicating that either system adequately records the data from the geophones. In either system the data can quickly be converted to a format (Society of Exploration Geophysicists 'Y' format; 'SEGY') to allow for filtering and other signal processing. With a modest software development effort, it is clear that either system could produce equivalent data products (SEGY data and audio data) within a few minutes of finishing the recording. The two systems both have significant advantages and drawbacks. With the seismograph, the tapping was distinctly visible when it occurred during a time window that was displayed. I have not identified or developed software for converting the resulting data to sound recordings that can be heard, but this limitation could be overcome with a trivial software development effort. The main drawbacks to the seismograph are that it does not allow for real-time listening, it is expensive to purchase, and it contains many features that are not utilized for this application. The music recording system is simple to use (it is designed for a general user, rather than a trained technician), allows for listening during recording, and has the advantage of using inexpensive, off-the-shelf components. It also allows for quick (within minutes) playback of the audio data at varying speeds. The data display by the software in the prototype system, however, is clearly inferior to the display on the seismograph. The music system also has the drawback of substantially oversampling the data by a factor of 24 (48,000 samples per second versus 2,000 samples per second) because the user interface only allows limited subsampling. This latte