CN106461617B - Method and apparatus for detecting structural faults - Google Patents

Abstract

A detection system for determining a condition of structural degradation is described that includes an acoustic sensor that receives acoustic emissions. The acoustic emission waves detected by the acoustic sensor are considered to be an impact. The system includes an analysis circuit that determines state a, state B, and state C. The impact activity is higher for state B than for state a, with rate B exceeding A f1 times. The impact activity is higher for state C than for state B, with rate C exceeding A f2 times. When it is determined that state C exists and the state C duration reaches the TM value, an alarm is activated; an alarm may also be activated when a certain threshold is reached, as determined from an analysis of the number of impacts in state C and the duration of state C.

Description

Method and apparatus for detecting structural faults

RELATED APPLICATIONS

This application is a partial continuation of U.S. patent application 12/151,077 filed on 3.5.2008, issued as a U.S. patent on 13.5.2014 (patent number 8,723,673). U.S. patent application 12/151,077 should be considered a continuation of U.S. patent application 60/927,523 filed on 5, 4, 2007. Where both of these U.S. patent applications (12/151,077 and 60/927,523) are referred to herein, they are incorporated by reference in their entirety.

Background

Technical Field

The invention relates to an acoustic sensing and warning device for a support structure member and a method of installing the same. More particularly, the present invention relates to a self-administered sensing and alarm system for continuously monitoring the integrity and degradation of infrastructure structures, particularly for integrated structures such as composite structures such as bridges and integrated structures such as underground mine anchor supported roofs.

Description of the Prior Art

Under the report of the Mine Safety and Health Administration (MSHA), underground coal mines in the United states have 1,500 and 2,000 roof fall accidents to be reported each year. Roof fall is the leading cause of accidental death in underground mines, resulting in the death of tens of people and the injury of thousands of people each year. According to MSHA reports, 70% of all unexpected deaths in underground mines are caused by roof fall. The average cost of such accidents is between 150 and 350 thousand dollars, and billions of dollars of lost production and repair and cleanup costs are incurred annually in the industry.

There are 590,000 highway bridges in the U.S. Many of these bridges are old and constitute a catastrophic risk to human life in view of undetected or inadequately discovered structural component failure or aging problems. Current methods for assessing bridge structural integrity rely primarily on visual inspection to identify damage only late in degeneration.

Referring to fig. 1(a), roof bolts 100 are typically spaced four feet apart to form an underground mine support roof. As shown in fig. 1(b), roof anchor 100 is secured to roof 105, and a condensable resin 110 is applied around the anchor to secure together the different strata 115a, 115b of roof 105 after tightening. It is less common to use expandable mechanical bolts for fixation. In both cases, support is provided by tightening the anchor heads through the fastening plate 120 between the mine roof 105 and the anchor heads 125 in close abutment with the mine roof, transferring a tensile load to each anchor head.

There are several prior art approaches used to study and strengthen the beam strength and improve roof support. However, as the mining operation continues, the formation 115 may begin to separate and create a tensile load on the anchor as shown in fig. 1 (c). A face of the formation may also move laterally, creating shear stress on the anchor. These forces can lead to three different conditions, causing roof fall: (1) the anchor fixing point may be weak, causing the anchor to loose and slip; (2) the anchor may lose its function and eventually break; and (3) the roof may crack or separate above the level secured by the anchor, known in the art as roof chipping. Statistically, the probability of roof fall occurring in a particular mine is 2.5 times per year.

Past efforts to predict roof fall have not produced a viable result. Various prior art techniques based on anchor load, tension or strain measurements are directed to stress versus strain relationships in materials. Generally, in fig. 2, the anchor bolts deform (i.e., elongate) when loaded or stressed (i.e., tensioned by the weight of overburden 130 (typically clay, rock, coal, or sand) over the mine passage). When a critical stress and corresponding strain is reached, the anchor then enters the yield zone and thereafter breaks, as shown in fig. 2. One prior art method is to measure stress or strain to detect impending anchor breakage. Field experience has shown that such methods are unreliable. Sometimes these methods predict that the anchor will break, but not; other times it is predicted that the anchor will not break, but the anchor will break.

The reason for this failure to predict anchor breakage is inherent in the variables being monitored. Not all nominally identical anchors (i.e., the same model of anchor) are actually identical. The materials used in making a particular batch of anchors are not all the same. Different anchors may vary. Manufacturing dimensional tolerances exacerbate these differences and unpredictability. Thus, in fact, the stress versus strain curve for a particular type of anchor exhibits a distribution as shown in region a of fig. 2. In addition, the stress and strain curves for a particular anchor can vary depending on the stresses that they were subjected to. For example, in the mine roof, if the anchor is subjected to pressure gradually over a long period of time, the profile will be different from the profile subjected to paroxysmal pressure at the same stage and different from the profile subjected to the same stress at different time periods.

Most of the prior art has focused on measuring load, strain or tension in an anchor, and several of the prior art techniques include generating a signal by a measuring device, propagating the signal in the anchor, and then detecting long-term changes in the signal, such as U.S. patent No. 4,149,446 issued to Spengler et al, 4/17 1979; U.S. patent No. 4,114,428 to Popenoe issued on 9/19 of 1978; us patent to Choi (patent No. 4,318,302) issued on 9.3.1982 and us patent to kibbelewhite (patent No. 5,205,176) issued on 27.4.1993. Furthermore, the strain produced by the externally applied stress of the anchor may vary along different portions of the anchor. While the strain in one location of the anchor is still in the safe zone, the strain in another location of the anchor may have reached the yield point. The installation of multiple strain sensors on a single anchor would make the system expensive and it would be simply impractical to measure strain along each location of the anchor. Furthermore, strain data may not necessarily provide the required information. Shear loads can also have a significant effect on mine roof support anchor failure and are not at all considered in load or strain measurement methods using load cells, pressure sensitive disks or strain gauges.

Another method of predicting roof fall is to measure roof subsidence, i.e., using a deformer to determine the degree, location and rate of movement of the soil or rock surrounding the excavation. In the mining industry, this method is widely used to collect support design information and form the basis of safety monitoring systems. The deformation measuring device is installed in the drill hole. In the mining industry, the smaller the borehole diameter, the lower its cost. The simplest strain gauge uses a stainless steel spring-based anchor head to which is attached a stainless steel wire to a catheter indicator visible through an orifice. The movement is indicated by a coloured reflective band on the indicator which is progressively covered as the movement progresses. In the mining industry, such a simple deformation determinator is referred to as a "visual display" because it is capable of displaying the roof movement in an intuitive manner. A large number of such devices need to be installed to cover the entire mine. The National Institute for Occupational Safety and Health (NIOSH) developed a roof monitoring safety system for measuring the movement of open roofs, such as andalusite mines. NIOSH acknowledges that the system is not applicable to predicting roof fall.

The method of using display on site and relying on the visual display device to predict is unreliable. The situation that the device does not send out early warning roof fall in advance occurs; sometimes the device gives an alert of impending roof fall, but the roof is never collapsed. The root cause of roof fall is the number of measurements, i.e. roof subsidence, which cannot be predicted by means of visual display. The instrument measures the degree of subsidence of the top plate at a particular location relative to a reference point. The reference point is at a fixed position of the instrument, based on the assumption that this position does not change (an unreliable assumption). In addition, the instrument does not measure the actual attenuation of the formation or the structure holding the formation together.

Sinking of the mine roof can generate longitudinal and transverse pressure, which can generate axial and shear force to the roof anchor. The combination of tensile and shear forces is sometimes sufficient to cause anchor breakage. Whether the anchor breaks depends on the material, structure and dimensions of the anchor, the anchoring resin, the surrounding rock mass and the angle between the anchor axes and the direction of the boundary between the strata. The measurement of the visual display does not take into account or evaluate any one of such factors. Thus, the degree of roof subsidence cannot account for the condition of the roof structure, and the instrument cannot reliably predict roof fall. Although the formation may move, the anchor structures supporting the formation may still be fully capable of securing the formation together. Conversely, although the roof subsidence may be relatively small, the separation between particular formations may have reached a critical value, or the anchor structures securing the roof together may have weakened to a critical level.

Other prior art detection methods are based on microseismic emission studies that deploy geophones in the mine roof area. The upper frequency limit of the geophone is between 4.5 and 14 Hz. Due to severe attenuation of high frequency pressure or sound waves, high frequencies in a wide range of hundreds or thousands of kilohertz cannot be detected. Such systems require the installation of geophones in the mine bore and move as mining progresses or the addition of geophones in new bores. In this system, four calculation parameters are combined and interpreted by the operator in order to determine whether a roof fall will occur immediately and where the roof fall will occur. To date, such seismograph studies have not reliably predicted roof fall. Two problems associated with this approach are the inability to detect high frequencies and the dependence of position determination on the speed of sound wave propagation in different directions. These velocities cannot be reliably predicted because the velocities depend on the heterogeneous structure of the formation.

In summary, the prior art lacks the ability to predict impending roof fall or other structural failure of a rigid support structure (e.g., a bridge) of an underground mine in order to take timely proactive steps to prevent failure and associated injury. There is a need for a system that can discover alarm conditions in order to take proactive steps to prevent failures in a timely manner.

Summary of the invention

The system disclosed herein places sensors at strategic locations of a particular infrastructure and alerts alarms when a structure reaches a weakened structural state that requires proactive measures to prevent infrastructure collapse. The system can be applied to underground mine roofs and highway bridges and other structures. For example, in an underground mine, it is preferable to place one sensor on each target roof anchor head, but it is also possible to place sensors on a smaller number of roof anchor heads. When the infrastructure or component degradation reaches a critical level, an alarm, such as an audible alarm, a visual indication, or a communication warning (e.g., a page or computer alert) is activated. This may be done through a direct electrical connection, an electronic signal propagated by the transmitter or sending an alarm signal to a maintenance office display through a communication network (e.g., the internet). Another approach is to store the alarm state in memory and retrieve it from memory on demand by a receiving device, such as a passive Radio Frequency Identification (RFID) system. It will be apparent to those skilled in the art that any conventional warning or notification system may be employed.

The sensors preferably employed in the system react to acoustic emissions or acoustic emission pressure waves from the target metal material, the fixing resin or the cover layer (i.e. the surrounding substrate), which propagate through the target metal material as acoustic wave conductors. Conventional acoustic emission sensors are typically manufactured from piezoelectric crystals, which are expensive to purchase and operate. Current systems prefer to use sensors made with piezoelectric films that are much less expensive.

The sensor can be placed on the bridge, and acoustic emission is detected through pre-buried anchors or steel cables. For example, the signal may be sent by a miniature transmitter to a central transceiver mounted on the bridge, which may send the signal to a computer in a maintenance office or other monitoring station over a communications network (e.g., the internet). To achieve low cost and high efficiency, the sensors and associated electronics may be powered by solar cells.

When materials such as metal, concrete or rock are subjected to pressure, the emitted sound waves are typically in the frequency range between tens of kilohertz and megahertz. These pressure waves come from atomic dislocations and microcracks as well as cracks that evolve into macrocracks. The selection and combination of predictable parameters (e.g., amplitude, frequency, power, duration, and incidence) can be used to determine the characteristics of these pressure waves and to identify the pressure waves. As a material approaches a critical zone (e.g., yield point), the incidence of acoustic emissions can increase substantially.

Many infrastructures contain metals embedded in other materials. As noted above, miners typically install roof bolts between five and twenty feet in length in a four foot square area. On prestressed concrete road bridges, reinforcing steel bars and steel cables are embedded in the concrete.

One embodiment of the system disclosed herein is applied to roof anchors used to secure multiple floors of mine roof together. Sensors attached to the roof bolt detect acoustic emissions from the bolt, resin or coatings around and above the bolt. The sensor module activates an alarm state when a critical level of structural defects is reached. As described in this application, the sound waves detected by the sensors that are considered significant events are referred to as hits. When the impact rate increases substantially, preferably by more than a preset threshold parameter (e.g., eight times), it is determined that an event threshold level or alarm condition is reached. This rate change indicates that the anchor, peripheral anchor region or overburden has reached a critical degradation point. The sensor converts the pressure waves into a voltage that is processed by electronic circuitry. In a preferred embodiment, the voltage waveform representing the pressure wave or shock is envelope detected, the number of envelopes per a particular time interval is recorded, and the sensor output value is calculated. The value is preferably the ratio between the number of impacts measured or detected for a particular time interval and a baseline reference value entered in memory. The baseline reference value is an average value taken over a preset period of time when the anchor is initially installed. If any such condition is detected, the impact rate is increased, the rate at which the preset threshold is exceeded, and an alarm condition is determined to exist.

The alarm indicator may be a visual indicator such as a Light Emitting Diode (LED) with a switch or a radio frequency transmitter that emits a warning signal. In fact, situations in which immediate action is required can be defined according to the law of change associated with the alarm indicator. For example, a bright LED may not be of concern because the weight of the ceiling may be supported by adjacent anchors. Similarly, two lit LEDs that are several LEDs apart may not be of concern. However, three adjacent lit LEDs may require immediate action to prevent roof fall.

When a radio frequency transmitter is used, the radio frequency transmitter may communicate with a mine communication network, such as sending a signal to an underground or ground maintenance personnel office, and displaying the change law at a video terminal. Algorithms that include preset thresholds for various parameters allow the computer to automatically determine whether immediate action is required.

Some salient features of this system are: (1) the location of the infrastructure weakened area is not determined by triangulation techniques and human judgment, but is determined directly by the sensor location indicated by the LED or video monitor; (2) rather than measuring with "typical" or "average" values that may miss an alarm condition or cause a false alarm, infrastructure weakening is measured using a method of self-reference measurement; and (3) the membrane sensor is very low cost compared to conventional sensors.

The above and other advantages and features of the novel invention will be more fully understood upon consideration of the presently preferred embodiments of the invention and the accompanying drawings.

Brief description of the drawings

Fig. 1A is a plan view of a prior art underground mine roof anchor layout.

FIG. 1B is a cross-sectional view of the roof anchor of FIG. 1A taken along lines I (B) -I (B).

FIG. 1C is a cross-sectional view of a roof anchor when subjected to a force according to the prior art.

FIG. 2 is a schematic view of a prior art roof bolt stress/strain curve.

Fig. 3A and 3B are top and side views of a roof anchor equipped with a sensor of the present invention.

FIG. 4 is a schematic electronic circuit diagram of one embodiment of the present invention.

FIG. 5 is a schematic diagram illustrating a decision tree for a first embodiment of the electronic circuit logic of the present invention.

FIG. 6 is a schematic diagram illustrating a decision tree for a second embodiment of the electronic circuit logic of the present invention.

FIG. 7 is a graph of long term sensor output calculations.

FIG. 8 is a graph of cumulative acoustic emission impact versus time for a system that detects state A, then state B, and then a second state A.

FIG. 9 is a graph of cumulative acoustic emission impact versus time for a system that detects state A, then state B, state C, and then a second state B.

Detailed description of the preferred embodiments

Referring now to fig. 3A, 3B and 4, one embodiment of system 47 shows sensor device 135 attached to anchor stud 125 of roof anchor 100. The sensor device 135 may include only the acoustic emission sensor 1, as well as a transmitter (not shown) in wireless communication with the rest of the system, a sensor wire 20 in wired electronic communication with the rest of the system, or may include the entire system in a separate housing. The acoustic emission sensor 1 may be fabricated using a piezoelectric film sensor (e.g., LDTO-028K/L sensor manufactured by Measurement Specialties, Inc.).

Referring now to fig. 4, the sensor 1 sends an output signal to the buffer amplifier 2 via the sensor line 20. Because the sensor has a high output impedance, a buffer stage is preferred. Buffering the output makes the signal less susceptible to noise. The buffered signal is transmitted to the amplifier 3 via line 21. The output signal from the amplifier 3 is transmitted via line 22 to the amplifier 4 where the signal is further amplified. The signal is then sent to the automatic gain control amplifier 6 via line 26. The gain input signal to the Automatic Gain Control (AGC) amplifier 6 via line 24 and line 25 is used to control the gain. The gain input signal 140 to the AGC amplifier 6 is transmitted from the microprocessor to a Field Effect Transistor (FET) gain adjustment circuit 5 via line 39 and line 40. The microprocessor 11 adjusts the gain of the AGC amplifier 6 in dependence on the envelope amplitude from the envelope detector circuit 7, as detected via line 31, as described below. The envelope detector circuit 7 comprises a half-wave rectifier 8, a buffer and low-pass filter 9 and a dc level shifter 10. The output from the FET gain adjustment circuit 5 is input to the amplifier 6 via line 26. The AGC amplifier 6 sends the output signal to the buffer via line 28 and to the low pass filter 9 via line 29. The buffer and low pass filter 9 sends the output signal to the dc level shifter 10 via line 30. A dc level shifter is used so that the microprocessor will read the envelope detector circuit 7 output only if the envelope detector circuit 7 output signal exceeds a certain threshold (e.g. 0.2 volts). This requirement is to avoid erroneously considering noise as a signal. Before entering the dc level shifter 10, the signal has an offset corresponding to half the battery voltage because of the use of a single battery power supply. The level shifter 10 adjusts the threshold level to a desired value.

The AGC amplifier 6 also sends its output signal to the buffer amplifier 12 via line 32. The output from the buffer amplifier 12 enters the zero crossing detector 13 through line 33. Each zero crossing of the output signal from the AGC amplifier 6 triggers a schmitt trigger buffer via line 34. Schmitt trigger buffer 14 is connected to digital counter 15 via line 35 to allow digital counter 15 to register the number of zero crossings of the signal from the output of AGC amplifier 6. The count of the digital counter 15 is transmitted to the microprocessor 11 via line 36. When the digital counter 15 reaches a predetermined value, a signal is output from the microprocessor 11 via line 37 to reset the digital counter 15.

As described above, the signal waveform output of the sensor 1 can be fully reproduced from the envelope and zero-crossing information in the microprocessor 11. Please refer to fig. 5 and the related description below to understand how to use such information.

To conserve battery power, the system may sometimes be set to an inactive or sleep mode (defined below). This is controlled by the microprocessor 11 via line 41 which is connected to the inactive mode switch 17. When the inactive mode switch 17 is closed, the battery voltage common collector Voltage (VCC) is used as the supply voltage peak-to-peak Voltage (VPP) supplied to the system 47. The positive and negative poles of the battery 18 are connected to the battery power supply starting circuit 19 by lines 44 and 45, respectively. This will prevent over-discharge of the battery 18 when the system 47 is not in use. When the system 47 is ready for use, the battery power supply startup circuit 19 is activated. When the battery power starting circuit 19 is started, the battery 18 output voltage is applied to the system 47 via line 46.

In operation, microprocessor 11 receives signal information from acoustic sensor 1. The received information includes an audio signal envelope, which is typically in the 600 μ s to 2ms range and within the audio signal violation. The microprocessor 11 determines from this information when an alarm condition is reached (as will be further explained in fig. 5). In addition, the microprocessor 11 performs power management functions to achieve maximum battery life. The system 47 will most of the time be set to the inactive mode. Depending on the acoustic emission activity detected in the last active mode, the system 47 will periodically resume function, take readings, determine if an alarm condition exists, and return to sleep mode (as will be set forth below in fig. 5 and 6). System 47 may include alternative embodiments that do not alter its primary function. For example, one amplifier may be used instead of the buffer amplifiers 2, 3, and 4. In order to obtain high gain and wide frequency band, three amplifiers are preferably used. Obtaining the same performance price with one amplifier is more expensive and uses more energy, causing battery drain more quickly. In another example, the output signal may be separated using a zero crossing count in the microprocessor 11, for example separating a 300kHz signal from a 700kHz signal. Similar information can be obtained by inserting two band pass filters in parallel before input to the envelope detector circuit 7. Bandpass filters can be used to reduce the amount of electromagnetic information that must be processed. For example, a first stage or low frequency range band pass filter between 100kHz and 300kHz and a second stage or high frequency range band pass filter between 350kHz and 700kHz may be applied to filter the voltage from the sensor output. The output signals from both frequency ranges indicate that the acoustic emissions are from the anchor itself. The output signal from only the low frequency range indicates that the anchoring has been weakened because the fixing resin has deteriorated to a critical level or the cover layer has been broken to a critical level.

Several different embodiments may be employed in accordance with the program being displayed via software and executed by the microprocessor 11.

Figure 5 shows a preferred method. When power to the system 47 is initially turned on, variables and defaults of the microprocessor 11 are enabled, including setting the values of DELTA T1, DELTA T, and threshold variable F. Before an acoustic emission strike count is detected, a time period DELTA T1 is left in order for the system under monitoring to be ready. For example, in a mine roof embodiment, when the roof bolts are initially installed on the roof, the roof bolts and resin structure require a period of time to secure on the rock formation, thereby creating an acoustic emission signal unique to this transitional stage. After a period DELTA T1, the number of hits during the time interval DELTA T is recorded and stored in memory as a variable REF 1. This establishes a baseline reference value for all further measurements. In a mine roof embodiment, a baseline reference value is established for each specific anchor in the roof structure. The counter is reset and the number of new hits per time period DELTA T is recorded. As shown in fig. 7, the count for each period DELTA T is compared to the value REF 1 to create a sensor output value curve that may reflect a rate ratio, frequency measurement, or other calculated value as noted elsewhere in this application. Fig. 7 is a graph of impact values versus time, particularly illustrating the behavior of the anchor as it is subjected to increasingly higher stresses. The slope of the curve between point a and point b changes significantly, in which the ratio of the number of impacts measured to the base value exceeds a threshold value 8 (which is set to REF 1). If the count is below the REF 1 value by eight times (the preferred multiple), the counter is reset and a new hit count is started. If the count exceeds the REF 1 value by eight times or more, an alarm condition is determined to exist, an alarm signal is transmitted to the user, and an alarm signal is sent to alarm/transmitter 16 via line 38 (as shown in fig. 4). This may be done by a visual indicator, such as a flashing LED. Alternatively, the device may transmit an alarm signal to a remote location via a transmitter familiar to those skilled in the art, where the alarm is sounded or displayed. When the monitored object reaches a very high attenuation level, it is determined as an alarm state and an alarm signal is generated. In the roof anchor example, the material yield point is reached.

In the present example, the alarm condition is determined by comparing the number of impacts recorded over a fixed time interval with a reference value. Another way to determine the alarm condition is to use a derivative of the cumulative number of impacts. In the safe state, the inclination of the cumulative impact curve is a graph close to a straight slope. When the material yield region is reached, the slope increases greatly and a new, higher, nearly straight slope pattern is reached quickly. The microprocessor 11 may thus calculate the cumulative impact number derivative and determine that an alarm level is present when the derivative increases (e.g. by a factor of five). To avoid errors due to small fluctuations in the slope, the derivative average over a short period of time can be calculated.

In another alternative embodiment, the frequency of the signal for each impact may be monitored. For example, a transition from 600kHz high frequency to 300kHz low frequency indicates that the anchor zone or formation surrounding the anchor has weakened to the point where an alarm condition can be determined to exist. This can be done by inserting two band pass filters in the system 47 or by monitoring the violation of the microprocessor 11 as described above. Another method is to monitor the impact rate and acoustic emission frequency and determine an alarm condition using an OR function when one of these two measurements indicates that a hazard zone has been reached.

Figure 6 shows an energy saving mode embodiment. The system 47 is fully powered only for certain time intervals when readings are taken. To prevent missing important events, the length of the inactive period may be adjusted and varied according to the value collected when the reading was last taken. The inactive period is set to a longer interval when the reading shows a better result. The inactive period is set to a shorter interval when the reading indicates a detected data input or signal height change. The routine shown in the flowchart of fig. 6 includes a second energy saving performance. In such a procedure, the alarm signal is terminated, rather than being continuously transmitted (e.g., LED flashing) after the alarm condition is detected. After one or several preset periods, the timer is set to wake up frequently. In the following cycle, when the decision block shows "if alarm has been triggered", the output will be "yes", the alarm signal (e.g. LED flash) is triggered again, the system returns to inactive mode, and returns to the "monitor timer" block. After an alarm is activated once, the periodic path will be much shorter than before the alarm state is displayed, thereby saving more energy.

Another feature of the embodiment shown in FIG. 6 is the block "process data and view acoustic emission type features". The program uses the signal envelope and out-of-limit information and compares it to previous readings, which will determine if the signal was indeed due to acoustic emissions. This step is particularly helpful if the environment is very noisy. Those skilled in the art can conveniently add features such as a low battery indicator and certain LED flashing patterns to indicate the status of the system.

Continuation of the information added in part in the patent application

Infrastructure 105 (e.g., underground mine roof, highway bridges, or other structures) may change in configuration from time to time. Such structural changes may result in pressure waves radiating in all directions from the location where the change occurred. As described above, this phenomenon is referred to as an acoustic emission event (AE event), and when an acoustic emission sensor detects an acoustic emission event, it is referred to as an acoustic emission bump. The stability of the infrastructure 105 may be evaluated based on the acoustic emission impact rate that occurs in the infrastructure 105 per unit time. The stability of the infrastructure 105 at different stages may be defined in terms of acoustic emission impact rate.

When the number of acoustic emissions impacts experienced by the infrastructure 105 is nearly constant, then there may be a period of stability, referred to as state a200, and thus indicative of normal, steady activity. The state a200 may also represent acoustic emission impingement rates in the anchor 100 and the surrounding formations 115a, 115 b. State a200 may be an acoustic activity period during which the acoustic emission sensor 1 detects normal background emissions, or may be a learning period of the system 47 during which the acoustic emission sensor 1 establishes an acoustic emission impact rate of a particular infrastructure 105 against a normal stable background. State a200 may be a DELTA T period following a DELTA 1 period (see above) that allows the system to settle after anchor 100 is installed. State a200 may have multiple hits within each particular time interval defined as rate a. The rate a may be different at different locations of the same infrastructure 105 and may be different at different sites.

In FIG. 8, the cumulative number of acoustic emissions impinging on a particular infrastructure 105 is represented by the y-axis and time is represented by the x-axis. According to the figure, infrastructure 105 is in state a200 from time 0 seconds to time 1000 seconds. State a200 is characterized by a relatively small slope S1 for rate a, and the slope is at steady state. The line may be an actual line or a line consisting of interpolated or averaged values across different data points. Rate a may be a background rate registered by detection system 47 or may represent a steady state of infrastructure 105 and anchors 100 and their surrounding formations 115a, 115 b. Data may be collected from the acoustic emission sensor 1 by the analysis circuitry of the system 47 or a signal may be collected from the acoustic emission sensor 1 by the detection circuitry of the system 47 and transmitted to the analysis circuitry so that the analysis circuitry determines that the infrastructure 105 is in state a 200. In one exemplary embodiment, the slope S1 may be 2 and the rate a may be 2 hits per second. In other exemplary embodiments, the rate a may be other rates, such as 2.5 impacts per second, 3 impacts per second, 3.5 impacts per second, 0-4 impacts per second, or up to 5 impacts per second.

Changes in the formations 115a, 115b or other portions of the infrastructure 105 may cause the acoustic emission impact rate to rise above the rate a or range of rates a defined for state a 200. The analysis circuitry then determines the infrastructure 105 to be in state B202. An increase in acoustic emission activity rate may indicate that the infrastructure 105 is becoming more unstable, and therefore the system 47 may alert the user of an unstable condition. When the number of acoustic emission events within a particular time interval increases by a factor of f1 (possibly within fL and fH ranges relative to the rate a of state a 200), system 47 may determine that state B202 is present. For example, in one embodiment, fL is 3 and fH is 6. Therefore, f1 may be in the range of 3-6 above rate A. In other embodiments, fL may be 4 and fH may be 9, such that f1 is in the range of 4-9. In FIG. 8, the system 47 determines that state B202 exists between 1000 seconds and 1250 seconds. The cumulative number of impacts during this period is 2000 and the rate f1 is 8. If system 47 sets fL to fH to range from 4-9, system 47 determines infrastructure 105 as being in state B202 when rate f1 is 8.

The slope of the line at state B202 is S2, which is higher than the slope S1 at state A200.

The line at state B202 may be an actual measurement over an equal time span (here, 1000 seconds to 1250 seconds), or may be formed by a mathematical procedure such as interpolation. The lines may be generated in the same way all the time along the curve, or different line generation methods may be employed according to other exemplary embodiments. When state B202 is determined to exist, a visual and/or audible alert may be generated in order to alert the operator infrastructure 105 that an unusually unstable state is occurring. This higher unsteady state may be stabilized by a return to a lower acoustic emission impact rate by the earth or normal procedures inside the infrastructure 105. If an operator is alerted while in state B202, it may be too early, or even inappropriate, because infrastructure 105 is not already at risk of collapsing or otherwise failing.

Fig. 8 shows the situation where the system exits state B202 at 1250 seconds and the system 47 begins to detect a lower acoustic emission impact rate from the infrastructure 105 or the anchor 100 or its surrounding formations 115a, 115B. The system 47 begins to determine the regression state a200 at this point in time, the line slope again becomes S1 (2 at S1), and the impact rate is 2 times per second. Although this rate is the same as state a200 from 0 to 1000 seconds before, it may be different from this period in other studies, provided it is below the fL threshold that distinguishes state a200 from state B202. As the infrastructure 105 returns to state a200 and is no longer in state B202, the generated alert may be terminated and the user may no longer be alerted to this higher unstable state.

Fig. 9 shows a second analysis of infrastructure 105 using system 47. From 0 to 750 seconds, system 47 determines that state a200 is present; from 750 to 1500 seconds, then state B202 is determined to be present. As described above, an alert may be issued when state B202 is currently determined to be present. However, infrastructure 105 does not return to state A200 after state B202 occurs, but rather enters an even higher acoustic emission impact active state, which may be referred to as a "catastrophic" state or state C204. If such a state C204 persists, it will result in a collapse or other failure of the infrastructure 105. When the acoustic emission impact rate is a multiple of f2 (greater than f1, greater than rate A), then state C204 is determined to be present. The acoustic emission impact rate ratio in state C204 may be a rate C, which is greater than A f2 times the rate. As noted above, rate C may be greater than fH. For example, if fL-fH ranges from 3 to 6, system 47 will determine that state C204 exists once rate C is greater than 6. The slope of the line in state C204 may be S3, which is greater than S2 and S1.

In other exemplary embodiments, the values of fH and fL may be different. In other embodiments, fL-fH may range from 0.5-2, 1-2, 2-3, 2-5, 3-5, 4-5, 5-10, 1-15, or 3.5-10.

The alert issued in state B202 may continue or stop once the infrastructure 105 enters the unstable state C204. However, when system 47 determines that state C204 is present, an alarm may be raised. As described above, the alert may be a more urgent notification to the user, with infrastructure instability being more severe and urgent than when the alert is issued. The system 47 may be configured to issue an alert immediately upon detection of state C204, or the system 47 may be configured to issue an alert not immediately upon detection of state C204, but rather upon some further determined event. These further events may be described as alarm initiating events occurring after system 47 determines presence state C204, i.e., rather than generating an alarm solely upon determining presence state C204.

The first alarm initiating event may be a time measurement, i.e., measuring the time that the infrastructure 105 is in state C204 after the initial determination that state C204 is present. One possible scenario is that some of the architectural adjustments experienced by the infrastructure 105 cause it to exit state C204 and re-enter state B202, or even state a 200. If this occurs, there is no reason to worry about infrastructure 105 collapsing or other structural failure, nor is there a reason to send out an alarm. False alarms, if raised, occupy resources and cause the user to prefer to ignore future alarms, since the user will think these are also false alarms. Once state C204 is first determined to be present, system 47 may monitor the duration of state C204. As shown in the graph of FIG. 9, state C204 starts at 1500 seconds and ends at 2000 seconds for 500 seconds. The system 47 may have a TM setting that, when reached, activates an alarm. In certain embodiments, the TM is 5 seconds. If state C204 is determined to last 5 seconds, an alarm is generated. If state C204 exits, the system 47 determines to revert to state B202 or state A200 before 5 seconds are reached, and the alarm is not activated. Although time is described as 5 seconds, according to various exemplary embodiments, time TM may be 10 seconds, 15 seconds, 20-100 seconds, 500 seconds, and up to 1000 seconds.

Once state C204 no longer exists, the infrastructure 105 returns to state B202 or state a200, and the counter measuring the time at state C204 may be reset. Thus, if TM is 5 seconds, state C204 exits for 2 seconds, and then enters state C204 again, then it takes 5 seconds to alarm after re-entering state C204. However, another configuration of system 47 in this regard is optional. For example, if TM is 5 seconds, and state C204 lasts 2 seconds before exiting, then the alarm can be generated only 3 seconds after re-entering state C204.

The reason why the alarm is not generated even when the state C204 is entered is that the infrastructure 105 may become unstable, but the partial structure thereof enters a new stable state after the movement, and therefore, there is no need to issue a warning alarm. However, if the unstable state continues for a certain period of time TM, it means that the stable state cannot be restored, and collapse may result. Unlike alerts, alerts are alerts that inform a user that a more serious condition exists. The alert may continue until an alarm is raised, or may be dismissed after exiting state B202 without any notification to the user until the alarm is activated. However, in most embodiments, when state C204 occurs, an alert is issued before an alarm is issued. Both the alert and alarm may be a number of times the sound emission hits, a sound of different intensity, or a different visual display (e.g., light or color). The alert may be distinct from the alert, but need not be distinct by degree or number, and may be a direct notification to the user that an alert or warning is being issued.

The system 47 may employ other arrangements that do not use the time measurement TM of the system at state C204 to determine whether an alarm should be raised. In such other alarm initiating events, a threshold is set and the system 47 looks at the data to determine if the threshold has been reached. The threshold is determined based on a multiple of the duration of state C204 and the number of impacts during the duration of state C204. For example, looking at FIG. 9, state C204 exists between 1500-2000 seconds, lasting 500 seconds total. The cumulative number of impacts at the beginning of state C204 is 2500, the cumulative number of impacts at the end of state C204 is 5500, and the total number of impacts during state C204 is 3000.

The analysis may determine the area under the line of state C204 and activate an alarm if the area equals the value A. If the area is less than the value A, no alarm is activated. Thus, the alarm generation is based on the duration of state C204 and the number of impacts during this period. The area under the state C204 line can be further calculated in one of two different ways. The area can be calculated as the area 206 of the triangle under the line of state C204, whose base is a horizontal line 210 parallel to the x-axis, starting at the beginning of state C204 (1500 seconds). Assuming that the value A is 1,000,000, once state C204 begins, system 47 begins to calculate area A. At 2000 seconds, area 206 is 0.5X 500X 3000 ═ 750,000, which is less than 1,000,000. State C204 terminates at 2000 seconds, at which time state B202 is determined to be present. Because area 206 does not reach value A in state C204, no alarm is generated. However, if the value a is 187,500, the value a will be reached at 1750 seconds, at which time the calculated area a 206 is 0.5X 250X 1500 ═ 1500

187,500. Once this value a is reached, an alarm is generated.

In state C204, the area under the line in the second alarm initiation event is calculated by: an area equal to area 206 is calculated, plus area 208. Area 206 is calculated as described above. The area 208 is calculated by multiplying the height of the line 210 on the y-axis by the time on the x-axis. In fig. 9, the area 208 is 2500X 500 ═ 1,250,000. Area 206 is 750,000 at 2000 seconds. Thus, in state C204, the total area under the line up to the x-axis is area 206+ area 208-750,000 +1,250,000-2,000,000. If the value a is 3,000,000, the system returns to state B202 in 2000 seconds, and no alarm is generated. If the value a is a small number, the sum of the areas 206+208 may be reached at some time between 1500 and 2000 seconds, and an alarm may be generated in accordance with the settings of this alternative.

In this second scheme, an alarm is generated when the area under the line reaches the value a. In state C204, the area under the line may be calculated as area 206, and in other example embodiments may be calculated as area 206+ 208. If the value A is not reached before state C204 exits, no alarm is generated.

The principle of this second alarm initiation event using area to determine whether to activate an alarm is based on the acoustic emission impact amplitude and the duration of this high acoustic emission impact rate. The higher the acoustic emission impact amplitude, the shorter the duration of state C204 required to determine that the infrastructure 105 is at a critical level of instability and an impending structural failure.

The system 47 can be configured to generate an alarm upon detection of a first alarm initiating event or upon detection of a second alarm initiating event. The second alarm initiation event may be configured in one of the two ways described above. Still other exemplary embodiments are those in which system 47 monitors for first and second alarm initiation events upon detection of state C204 and activates an alarm upon the first occurrence of either the first or second alarm initiation event. This scheme monitors two alarm activation events simultaneously, and alerts once one of them reaches its alarm state, even if the other activation event has not reached its alarm state. In other arrangements, all of which differ from the one described above, it is necessary to determine that all alarm initiating events have occurred in order to issue an alarm. Such a scenario may include the system 47 determining the presence state B202 and issuing an alert, but may not include such an alert in embodiments where the activation state C204 occurs once or twice the alarm initiation state.

The lines depicted in fig. 8 and 9 appear as straight lines and the transition from one state to the other appears as an abrupt change in direction, forming an acute angle. In other aspects, the lines may not be straight lines, may be jagged or curved, or may be of various shapes. Mechanisms are known to determine the area under a curve or line of different shape. Further, transitioning from one state to another (e.g., transitioning from state a200 to state B202) may not be a slightly curved curve or other shape that does not form an acute angle. The disclosed numbers are for exemplary purposes only and in other exemplary embodiments, other numbers are possible.

The newly disclosed subject matter may include all of the components and aspects previously described. For example, the system 47 may be a recently installed system in a mine or other infrastructure 105, or may be a retrofit system installed later. The time delay for learning background noise may be state a200, but state a200 may also be the state of system 47 after the background-aware/learning phase.

While the foregoing is directed to the presently preferred embodiment, it is to be expressly understood that the invention is not limited to this particular embodiment, but may be otherwise embodied and employed within the scope of the following claims.

Claims (16)

1. A detection system for determining structural degradation, comprising:

a sensor secured to the structure and detecting each of a plurality of acoustic emissions as an impact;

an analysis circuit that determines state A, state B, and state C;

the number of impacts of the state A in each specific time interval is called a speed A;

said state B having a higher impact activity than said state A, said state B having a number of impacts per a particular time interval referred to as a rate B, said rate B being higher than said rate A by a factor of f1, but said rate B not exceeding a factor of a rate A fH, wherein fL being less than said fH defines a threshold between said state A and said state B;

said state C having a higher impact activity than said state B, said state C having a number of impacts per a particular time interval referred to as a rate C, said rate C being f2 times higher than rate A, said rate C being greater than said fH, wherein said fH defines a threshold between said state B and said state C;

the alarm is activated when one of the following two occurs: (i) determining that the duration of the state C reaches a set value; and (ii) is determined to be said state C and has reached a threshold value based on an analysis of state C impact number and state C duration,

wherein the analysis is one of: (a) the cumulative number of impacts in state C reaches the threshold; (b) drawing a line of the number of impacts detected over time, the threshold being defined as an area, the area under the line at the state C being calculated as a derived area, and the derived area being equal to the threshold area.

2. The detection system of claim 1, wherein the set point is five seconds.

3. The detection system of claim 1, wherein said rate B exceeds said rate a by a factor of f1, said f1 being in a range between said fL and said fH, wherein said fL is 3 and said fH is 6.

4. A detection system according to claim 1 wherein said state a is a background state and said rate a is collected from said acoustic emission waves directly arising from the stresses to which said sensor is subjected in the plurality of target metal objects and their surrounding areas; the rate a is collected for a preset period of time after the target metal object is installed and before monitoring of the degradation of the target metal object and its surrounding area begins.

5. The detection system of claim 1, further comprising a detection circuit in electronic communication with the sensor that recognizes acoustic emissions, wherein the analysis circuit is in electronic communication with the detection circuit, receives input from the detection circuit, and the impact is an acoustic emission deemed to be a significant event, and the alarm is part of an alarm circuit in electronic communication with the analysis circuit.

6. The detection system of claim 1, further comprising a warning device that activates a warning when the analysis circuit determines that state B is present and does not activate a warning when state a is present immediately after state B is present.

7. A detection system according to claim 1, wherein an alarm is not activated when state C is determined to be present and state a or state B is determined to be present before the duration reaches a set value; the time is reset upon exiting state C.

8. The detection system of claim 1, further comprising a plurality of sensors affixed to a plurality of target metal objects located within the structure, each of the plurality of sensors affixed to a different location in the structure.

9. The detection system of claim 1, wherein the threshold is represented by an area having a value a, the area being derived from an analysis of the number of impacts of state C and the duration of state C; the resulting area is equal to 0.5 times the state C duration times the number of state C impacts; when the area obtained is equal to the value a, i.e. the threshold value is reached, an alarm is activated.

10. The detection system of claim 1, wherein the acoustic emission wave originates from at least one of the structure and a region proximate to the structure.

11. The detection system of claim 1, wherein the rate a is any one of 2 impacts per second, 2.5 impacts per second, 3 impacts per second, 3.5 impacts per second, 4 impacts per second, and up to 5 impacts per second.

12. The detection system of claim 1, wherein the range defined between fL and fH is in the range of 0.5 to 15.

13. The detection system of claim 12, wherein fL is 4 and fH is 9.

14. The detection system as set forth in claim 1, wherein the set value is any one of 10 seconds, 15 seconds, 20-100 seconds, 100-500 seconds, and up to 1000 seconds.

15. The detection system of claim 5, wherein the alarm comprises at least one of a visual indication and a sound.

16. A method of identifying degradation of one or its surrounding structure of a plurality of target metal objects, comprising the steps of:

mounting an acoustic sensor on a target metal object, receiving acoustic emission waves emitted directly from structures within and surrounding the target metal object under stress;

detecting each of the acoustic emissions as an impact;

analyzing the acoustic emission wave velocity to determine the presence of state a, state a having multiple acoustic emission wave strikes within each particular time interval, referred to as velocity a;

analyzing the acoustic emission wave velocity to determine the presence of state B, which has a higher impact activity than state a, and which has multiple impacts within each specified time interval, referred to as velocity B, which exceeds velocity A f1 times, but which must not exceed velocity a fH, wherein fL, which is less than fH, defines a threshold between state a and state B;

analyzing the acoustic emission wave velocity to determine the presence of state C, which has a higher impact activity than state B, and which has multiple impacts within each specified time interval, referred to as velocity C, which exceeds velocity A f2 times, and is greater than fH, wherein fH defines a threshold between state B and state C;

sending out an alarm when the existence state B is determined, and not sending out the alarm when the existence state A is determined;

an alarm is raised upon the first occurrence of one of two events:

1-determining the presence state C, the duration of which reaches a set value; and

2-determining that state C is present and determining that the threshold is reached based on an analysis of state C impact number and state C duration, wherein the analysis is one of: (a) the cumulative number of impacts in state C reaches the threshold; (b) drawing a line of the number of impacts detected over time, the threshold being defined as an area, the area under the line at the state C being calculated as a derived area, and the derived area being equal to the threshold area.

Images