Evaluation of an Advanced Proximity Detection System for Continuous Mining Machines

Researchers at the National Institute for Occupational Safety and Health (NIOSH) are advancing the emerging technology of electromagnetic proximity detection, which provides a promising means of protecting workers around any machinery that presents striking, pinning or entanglement hazards. This technology is particularly applicable to mobile underground mining equipment such as remote-control continuous mining machines, which offer perhaps the most difficult safety challenges in the mining industry. The operators of these machines must maintain constant vigilance to keep themselves and others near the machine safe. Tragically, striking and pinning accidents involving continuous mining machines occur every year causing severe injuries and claiming lives. Proximity detection technology has been effectively implemented for other types of equipment in underground and surface mining as well as in other industries. However, applying this technology to remote-control continuous mining machines presents uniquely difficult challenges. Due to visibility and space limitations, the machine operator must routinely work in very close proximity to the machine. In order to protect miners without preventing them from doing their jobs or causing nuisance alarms, NIOSH is now developing intelligent proximity detection technology. This technology accurately determines worker position relative to the machine and responds by intelligently issuing situation-specific alarms to warn the operator or disabling situation-specific machine functions to protect the operator from machine movements that could result in injury. In this paper, the authors review existing proximity warning technologies, describe ongoing NIOSH research on an intelligent proximity warning system, and summarize current test results. The NIOSH-developed intelligent system has the potential to have a significant impact on the mining industry by greatly advancing the state-of-the-art in proximity detection technology, leading to increased operator safety, and reducing the frequency of injuries and fatalities. Introduction Operating large mobile equipment such as a continuous mining machine (CMM), shown in Figure 1, is a hazardous job that workers perform in underground coal mining operations. Figure 1: Continuous Mining Machine Some of the conditions which make this hazardous are the potential for roof falls, the close proximity of large moving machinery, decreased visibility due to low lighting and high dust levels, and high noise levels. In addition to the task of cutting coal from the face, continuous mining machine operators must focus attention on their own position, the location of other crewmembers, and the proximity of the machine to the crew. There are unsafe areas that the remote-control continuous mining machine operators and other workers must avoid. Some areas are clearly defined, such as beyond supported top which is defined by the last row of bolts supporting the roof. Recent NIOSH research by Bartels et al. (2009) identified safe and unsafe zones for the operator near the continuous mining machine. Since the mining environment is dynamic, creating physical barriers to keep operators out of the unsafe zones is not feasible. In the past, operation of continuous mining machines was performed from the machine cab in a seated position. In the 1980s, new technology enabled the transition to remote-control of the mining machinery. By removing the operator from the machine cab, several safety hazards associated with having the operator near the coal face were alleviated. With remote-control capability, the operators are now free to position themselves for better safety and better visibility of the workplace. Typically, the operator positions themselves behind and to one side of the machine during cutting operations. During tramming, the operator walks near the rear of the machine in high coal seams where the machine is less of an obstruction. In low coal seams, the operator trams the continuous mining machine while walking or crawling in front of it. This difference is because the operator cannot see over the machine from the rear. Unfortunately, operators have the tendency to step beside a moving continuous mining machine for a better view during forward, reverse and turning movements while cutting coal or tramming. Bauer et al. (1994) reported that the practice of extended-cut mining has increased the operators’ tendency to position themselves in hazardous locations. Additionally, Steiner et al. (1994) stated that an unforeseen consequence of remote-control operation is that an operator can position themselves in dangerous or hazardous locations that could result in a fatality or injury from possible roof falls, mine wall breakouts, pinch-points or other vehicle traffic. Adding to the hazards of operating a continuous mining machine is the restricted workspace with reduced visibility. The mine work environment, such as in low coal seams shown in Figure 2, puts continuous mining machine operators and helpers in awkward work postures for a job consisting of tasks that require fast reactions to avoid being struck by moving equipment. Figure 2: Typical Mine Environment Furthermore, Lewis (1986) reported that restricted visibility due to the nature of the mine environment and low lighting conditions further complicates the tasks involved in operating mining equipment. The Mine Safety and Health Administration (MSHA) recommends a set of “red zones” that define dangerous areas near the continuous mining machine, and operators are supposed to avoid these areas. These zones help operators to understand and avoid potentially dangerous areas within the turning radius of the machine. While this concept has been around since the mid 1990’s, fatalities and injuries continue to occur with moving machinery underground. A survey of the 2002–2008 accident data from MSHA reveals that an average of 252 accidents occurs per year involving remotecontrolled continuous mining machines. Since 1984, there have been 33 fatalities involving workers being struck or pinned by these machines. This indicates that violations of the red zone recommendation occur frequently. Research (Bartels, 2009) shows the red zone guidelines address potentially hazardous situations, but ignore what the operators need to see and sometimes conflict with where the machine operators would like to position themselves in order to perform their job. A technological control to prevent the continuous mining machine from making hazardous motions with workers nearby would reduce the frequency of these accidents. A promising technology for this purpose is electromagnetic proximity detection, which utilizes magnetic fields to determine the proximity of workers to the machine. In this paper, we present an advanced, intelligent system utilizing electromagnetic proximity detection hardware along with novel and efficient software for determining the 2or 3-dimensional position of a worker and intelligently responding with alarms or disabling machine movement. The implementation of this intelligent system could greatly improve the safety of miners while also reducing the frequency of false alarms that are a problem for some currently available proximity detection systems. Background Remote-control operation has required the continuous mining machine operators to divide their attention and process more information simultaneously. Defining and prioritizing what cues and feedback the operator needs and determining what operators focus their attention on can then be used to develop safe, realistic operating procedures. The cues that operators use are primarily visual but will sometimes include auditory information to compensate when visual cues are blocked. Researchers can use this information when analyzing human-machine systems, it is important to examine the components of the mining machine operator and the machine within the work environment. The mining environment is a unique challenge due to its dynamic nature, many hazards, and operational information that must be continually monitored by the continuous mining machine operator. The ability to process and utilize feedback information, in particular the visual cues the operator uses, is an important component of the human-machine system. Safe and effective control of the system is dependent upon the worker properly sensing pertinent information and processing it to make the right decisions. Experienced miners have expanded their knowledge, skills, and abilities to perform safely and effectively. By identifying the specific cues used by these experienced operators, interventions and training methods can be designed to improve safety for all operators. Previous studies by Bartels et al. (2009) identified and defined visual attention locations (VALs) associated with remote operation of continuous mining machines. In this research, VALs are particular locations needed and visually used by the operator for machine control and operation. The operator needs to consider safe work positions, sounds, vibrations, and operator VALs such as machine orientation, operating characteristics and other visual cues within the work environment to perform their job effectively. These factors have to be accommodated concurrently when considering a safe operator location. The optimum work location for an operator may differ depending on the length of cut, visibility, roof condition, ventilation and avoidance of moving machinery. As part of this research project, the study gathered information on operator work positions and VALs needed when the operator trammed a continuous mining machine during the cutting phase or when moving to a new location. Analysis of the data defined the operators’ risk of injury relative to the operators’ task, equipment and workplace environment. The results showed that operators of continuous mining machines needed to maintain a 3-foot minimum distance for safety. In addition, the data indicated that a major contributing factor to continuous mining machine related injuries is operators positioning themselves in a hazardous position in order to see cues or VALs. Several types of proximity detection systems using various technologies have been developed (Ruff and Hession-Kunz 2001; Ruff and Holden 2003; Ruff 2004; Kloos, Guivant et al. 2006; Ruff 2006; Ruff 2007). Some of the technologies utilized in surface mining and in other industries include the Global Positioning System (GPS), and radar-, laseror ultrasonic-based distance sensors. Unfortunately, these technologies are ineffective in underground mines, where GPS is unavailable, and the constant close proximity of mine walls makes the use of the other sensors extremely difficult. Another possible solution is the use of Radio Frequency Identification (RFID) technology. Many industries commonly use RFID for tracking the movement of personnel, supplies and equipment. It is also currently in use in the mining industry for tracking the movements of people, equipment and supplies through the mine. These systems are capable of providing information on whether a tag worn by a person or mounted on a machine is within a set range of the transmitter. These systems typically operate in the very-high (VHF) or ultra-high (UHF) radio frequencies, and interference from signal reflections and line-of-sight requirements make applying these systems to continuous mining machines difficult. Another emerging technology that may be applicable to this problem is intelligent video systems utilizing either single-camera or stereovision and complex algorithms to identify and locate people and machines in the visual scene. However, application of this technology in the underground mining industry is likely to be very challenging due to poor lighting, dust, and the extreme difficulty in keeping the cameras clean. In a survey of companies implementing proximity detection technology internationally, NIOSH found that mines in South Africa and Australia are using electromagnetic-based systems in several underground coal mining operations. Partnerships between coal operators and the companies marketing the system are a common mechanism for developing these systems. The systems are used on continuous mining machines as well as other underground equipment such as shuttle cars, roof bolting machines, and feeder/breakers. The systems provide warning and danger zones around equipment. An operator receives visual and audible warnings upon entry into the warning zone and further approach into the danger zone causes the equipment to shut down. Currently, in the United States, the available systems for underground mining use either magnetic fields or radio frequency technology to alert miners when a machine is close to another machine or a person. On continuous mining machines, these systems will disable all machine movement if an operator moves too close. Recent interviews with the mining community (Kingsley-Westerman, 2010) indicate this action results in frequent nuisance alarms and shut downs. NIOSH researchers are developing technology that adds a measure of intelligence to these systems. The NIOSH intelligent system accurately determines miners’ positions relative to the machine and responds by disabling only the specific machine functions that could cause injury. Intelligent Proximity Detection System If a proximity detection system is implemented that completely disables machine movement when a person is located near the machine, nuisance alarms are likely to occur frequently. From the standpoint of machine operators, it may seem that the proximity detection system is preventing them from performing their job effectively or standing where they need to stand. For miners to accept the use of proximity detection technology, the technology must provide the necessary protection while minimizing the occurrence of nuisance alarms. NIOSH researchers have developed a solution to this issue in an intelligent proximity detection system. This system accurately determines the position of miners around the continuous mining machine using magnetic fields generated by multiple pulsed electromagnetic field generators. The signal strength from each of these generators is measured by an operator-worn Personal Alarm Device (PAD). Distances are estimated from the signal strengths and are used to triangulate the PAD position. Based on this triangulated position, the onboard controller disables specific machine functions such that unsafe actions are prevented but safe actions are allowed. In this way, the operator maintains the freedom to stand in close proximity to the machine and continue to mine without being hampered by nuisance alarms. The implementation of this technology, along with proper training, is likely to greatly enhance the acceptance of proximity detection by mining machine operators. An intelligent system of this sort requires a method for accurate position calculation. Therefore, an accurate mathematical model of the magnetic field shape is needed. NIOSH researchers have developed such a model for magnetic field generators that use an antenna with a ferrite rod core typical of proximity detection systems (Carr, Jobes, Li 2010). The shape of the magnetic field is very complex and irregular. Equation (1) defines the shape of the magnetic “shell” in polar coordinates as all points having the same magnetic flux density. ൅ ܾ ሻ 2ߠ ሺ ܿ݋ݏ · ܽ ߩ ൌ (1) In this equation, ρ is the radial coordinate measured from the center of the magnetic field generator and θ is the angular coordinate measured from the long axis of the magnetic field generator. The coefficients a and b are functions of the magnetic flux density as defined in (2) and (3).