Non-Intrusive Monitoring: Types, Applications, and Benefits

Introduction to Non-Intrusive Monitoring

Non-intrusive monitoring was developed by George Hart, Ed Kern, and Fred Schweppe in the 1980s at the Massachusetts Institute of Technology. It is commonly discussed in terms of non-intrusive load monitoring (NILM), a means of monitoring an electrical circuit that encompasses a number of appliances, all capable of turning on and off independently of one another. Rather than attaching a separate monitor to each appliance, non-intrusive monitoring uses electric meters to determine the different patterns of power use within a given home. Similarly, nonintrusive appliance load monitoring (NALM) operates through "a sophisticated analysis of the current and voltage waveforms of the total load; the NALM estimates the number and nature of the individual loads, their individual energy consumption, and other relevant statistics such as time-of-day variations" (Hart). Non-intrusive monitoring is particularly valuable because it can measure voltage and current without requiring access to individual components or appliances within a home or other entity under assessment. The resulting data can be extremely useful to public policymakers, energy auditors, consumers, and appliance manufacturers, providing an accurate snapshot of energy consumed over time (Hart). This can highlight deficiencies, irregularities, or issues within energy consumption patterns.

Within non-intrusive load monitoring, power meter readings offer a clear identification of the loads created by specific appliances, generating a reliable way to verify load sheds in homes and buildings (Bergman et al., 2011). A clear example of this is within a home during a two-hour period. Over those two hours, the activity might show energy use from only a heater and a refrigerator: the refrigerator turns on and off three times, the heater twice as often (Hart). This process can easily illuminate the total energy used by these appliances and their individual expenditures: "By also considering measurements of the total reactive power or harmonic current, along with the real power shown, changes in the resulting vector function of time would reveal even more information about the particular appliances" (Hart).

The key principle of non-intrusive monitoring is that one uses information available from the normal operation of an item, placing no extra requirements on the system (Thornton & Sanghera, 2011). This type of monitoring can be applied to all forms of energy, including sound — this is where acoustic and vibration non-intrusive monitoring becomes relevant. Acoustic and vibration monitoring can detect signals or vibrations caused by the movement of material. "This movement causes impacts and frictional contact with a containing face, for example the inside of a pipe. The sensor is fastened to the outside of the structure, and its high-frequency detection picks up these signals, which are often undetectable to the human ear" (pulsar-pm.com). This type of detection relies on tools that function effectively even in environments with a tremendous amount of machinery noise, because they use advanced technology to detect changes or disruptions in acoustic emissions from equipment or machinery during operation (pulsar-pm.com).

This non-intrusiveness is one of the defining attributes of the approach: materials and processes are examined while engaged in their normal functioning. With acoustic and vibration non-intrusive monitoring, a sensor examines acoustic discharges from machinery during normal operation, and these sensors are sensitive to the smallest changes in conditions. This type of monitoring is used in industrial scenarios such as mills and plants, in the assessment of automobiles and trucks to determine engine efficiency, and even in the field of medicine to evaluate the success of joint and limb replacements or therapies. The possibilities with these technologies are genuinely broad.

Several different types of non-intrusive monitoring have already been alluded to above. Non-intrusive appliance load monitoring gives energy auditors, homeowners, building owners, manufacturers, and other interested parties a clear picture of the combined energy that their appliances use as well as each appliance separately. It is a way of measuring energy output that yields a more detailed picture of which electrical loads are contributing to the largest or smallest consumption figures.

One of the most significant disadvantages of this type of monitoring involves real and immediate privacy concerns for the individual. These concerns arise most directly when the energy use of a home or other residential structure is being examined. Fundamentally, the way that individuals or families use energy reveals patterns of behavior. Patterns of behavior are directly related to privacy rights: individuals have a right to keep private what they do at home, when they are home, when they are showering, and other personal or potentially sensitive activities. The concern is compounded when individuals are unaware that their energy use is being monitored. It is a genuine issue of balance, as this type of monitoring provides valuable information to society at large, yet no one wishes to sacrifice the rights or needs of the individual.

Different Types of Non-Intrusive Monitoring

Another type of non-intrusive monitoring is biological in nature. This is seen in technologies such as non-invasive blood glucose monitors for diabetics, sensors to detect intoxicated drivers, and other tools focused on biological functioning. For example, non-invasive glucose blood level monitoring uses a sensor placed closely against the skin: "a special camera, called a Raman spectrometer, inside the sensor uses light to identify and analyze glucose molecules under the skin, via interstitial fluid. Each glucose molecule has a special 'signature' the sensor identifies, and from there, analyzes and extrapolates a glucose value, which is transmitted via Bluetooth to a handheld device, like an iPhone or Android, or to a computer" (Allison, 2011). A Raman spectrometer is a device that also measures vibrations within a system, illustrating how this tool represents both biological and vibrational non-intrusive monitoring. The significant advantage of this technology is its capacity to transmit results immediately to a phone, meaning that the user's family members or even paramedics can be alerted if glucose levels reach a dangerous point. This demonstrates how sensors are making life safer and more manageable for people with chronic illnesses. The non-intrusive nature of this monitoring is apparent: the diabetic no longer needs to prick their skin to monitor glucose levels, greatly improving quality of life.

The same principle applies to non-intrusive biological sensors capable of monitoring the level of intoxication in a driver. While this technology is still developing, the possibilities are substantial. According to Mothers Against Drunk Driving, one in three people will be involved in an alcohol-related crash in their lifetime, and almost every 90 seconds a person is killed in a drunk-driving accident: "In 2011, 9,878 people died in drunk driving crashes — one every 53 minutes" (Madd.org, 2012). Biological sensors could have a tremendous impact on eliminating or greatly reducing the number of these accidents. These systems are designed to assess the "biological condition of a driver and issue warnings during instances of drowsiness. Moreover, many researchers have reported that biological signals, such as brain waves, pulsation waves, and heart rate, are different between people who have and have not consumed alcohol. Currently, we are developing a noninvasive system to detect individuals driving under the influence of alcohol by measuring biological signals" (Murata et al., 2011). The advantages of making such technology a permanent feature in all automobiles are significant for human safety and for society as a whole. Some critics argue that such measures infringe on personal freedom, but the life-saving potential of the technology is difficult to dispute.

Acoustic and vibrational forms of non-intrusive monitoring, which measure emissions from machines and other entities in order to assess their functioning, also offer clear environmental benefits. For example, passive, non-intrusive acoustic monitoring can observe the valve movement of bivalve mollusks (Di Iorio et al., 2012). Though this may not sound significant at first, bivalve mollusks are vital ecological and economic components of coastal ecosystems: "The formation and size of growth increments delimited by striae are affected by environmental stressors. The mechanisms linking shell growth and striae deposition in relation to environmental variations are poorly understood but are likely associated with the animal's valve-movement behavior" (Di Iorio et al., 2012). Non-invasive acoustic monitoring can illuminate these issues in ways previously not possible. Similarly, non-invasive acoustic and vibrational monitoring has been used by physicians to better assess in vivo hip conditions and improve understanding of total hip arthroplasty (Glaser et al., 2010).

"Acoustic emissions are elastic waves generated by a rapid release of energy at a localized source. They are produced by events such as particle impact, gas evolution, boiling, phase transitions, and precipitation. Some processes produce emissions that can be heard. A lot more emit either outside the audible frequency or at too low an intensity to be heard" (McLenna, 1995, p. 338). Using non-intrusive acoustic monitoring provides a means of continuously and effectively monitoring an entire structure (Wu & Abe, 2003). Acoustic signal-based monitoring can assess individual components and mechanisms or entire structures, pinpointing abnormalities, failures, or warning signs requiring attention (Wu & Abe, 2003). It provides experts with valuable information that enables focused research and further evaluation, improves the timing of repairs or rehabilitation, and helps adjust maintenance budgets (Wu & Abe, 2003). Managers, scientists, policymakers, and members of society should view this form of technology as a management tool that can shed light on the needs of communities, individuals, and structures.

The possibilities for acoustic signal-based monitoring are genuinely expansive, as these systems can detect phenomena entirely beyond the capability of the human ear. A compelling example is the monitoring of live chick embryos. At present the technology is semi-invasive, but it is expected to become fully non-intrusive in the near future. The research demonstrates how acoustic signal monitoring can be used to observe and assess the development of chick embryos, with direct implications for feeding populations and expanding the poultry industry (Liang et al., 2011). In the study "Monitoring of live chick embryo based on acoustic and vibration signals with a new semi-invasive technology" (Liang et al., 2011), the researchers determine the health and rate of development of a chick embryo by making a small hole with a pin at the top of the egg, after which the heartbeat signal is amplified and tracked. A microphone installed over the hole helps distinguish a live embryo from one that is small or weak (Liang et al., 2011). This type of data can help researchers determine whether an embryo is alive, which is absolutely critical for the development of the poultry industry.

Another group of researchers applied acoustic monitoring to animal habitat research. The study "Target Classification and Localization in Habitat Monitoring" by Wang and colleagues seeks to determine whether acoustic signal monitoring through sensors can be used to recognize and locate animal calls (2003). This form of habitat monitoring has two objectives: "The first part is to determine whether observed animal calls are of the specified type using their spectrograms. Each type of animal call has its own characteristic spectrogram which is input to the system. The classification of an observed acoustic signal is determined by the maximum cross-correlation coefficient between its spectrogram and the specified characteristic spectrogram. The second part is to locate the calling animal when its call is recognized" (Wang et al., 2003). Given that the animal kingdom relies heavily on sound for communication, development, and habitation, acoustic-based non-intrusive monitoring can be incredibly useful. Since human beings have extremely limited hearing compared to many other species, acoustic-based monitoring can help scientists better understand the wide range of processes present in the animal world.

The research study "Acoustic Sensor Networks for Woodpecker Localization" by Wang and associates confirms that sensors offer some of the most valid, effective, and non-intrusive ways to study animals. Studying animal behavior necessarily relies on non-intrusive methods, as intrusive research disrupts natural behavior. In this experiment, data were gathered from woodpecker vocalizations using four microphones arranged as a square (Wang et al., 2005). "All four audio channels within the same acoustic array are finely synchronized within a few microseconds. We apply the approximate maximum likelihood (AML) method to synchronized audio channels of each acoustic array for estimating the direction-of-arrival (DOA) of woodpecker vocalizations" (Wang et al., 2005). Wang and colleagues found a fundamental connection between microphone spacing in acoustic arrays and the reliability of detecting woodpecker vocal tracks, even accounting for typical background noise (2005). Experiments such as these, which rely on acoustic signals monitored in a non-intrusive fashion, can illuminate a great deal about animal behavior and ecology.

Acoustic Signals

The biological sciences are not the only field that benefits from acoustic signal monitoring. Numerous industries can benefit as well, since the correct functioning of machinery is closely connected to the acoustic signals it produces. "Acoustic Emissions (AE) and surface vibration monitoring technologies are promising non-intrusive approaches to online monitoring of industrial tumbling mill process and machine condition" (Meech et al., 2005). This allows technicians to receive valuable feedback about whether their processes are running effectively, what malfunctions indicate, and what changes need to be made — all without dismantling equipment.

The potential benefits of acoustic-based monitoring extend further still. Some companies use acoustic technology to examine and assess the noise caused by sand grains striking pipe walls at the radius of a bend: "This noise can then be used to calculate sand production rates based on knowledge of background noise, flow rate, and particle size. Acoustic sensors provide an immediate response to any changes in sand production by detecting acoustic signals emitted when a sand particle hits the pipe wall. Benefits include easy installation, the immediate detection of sand production, and low power consumption" (emersonprocess.com). This demonstrates how the technology can yield highly practical results, saving both time and money.

Beyond immediate financial savings, acoustic-based non-intrusive monitoring can also assist with the maintenance of critical infrastructure. The study "On the opportunity to use non-intrusive acoustic emission recordings for monitoring uniform corrosion of carbon steel and austenitic stainless steel in acid and neutral solutions" addresses the issue of uniform corrosion of steels (Jaubert et al., 2005). The researchers acknowledge that this type of corrosion is widespread and is responsible for the destruction of many industrial components, leading to significant economic and environmental losses (Jaubert et al., 2005). While classic corrosion formulas can predict average corrosion rates for certain materials, they cannot provide instantaneous, real-time corrosion values. However, using acoustic signals in connection with the release of energy within an item generating a transient elastic wave propagation, corrosion information can be obtained quantitatively (Jaubert et al., 2005).

The researchers found that "acoustic emission measurements are in good correlation with aggressiveness of the corrosive media, and a semi-quantitative correlation is obtained between AE activity and corrosion rate for austenitic stainless steel. Whereas initial surface conditions greatly influence the acoustic activity, AE monitoring appears to be a rewarding technique for detecting corrosion rate evolutions during process modifications" (Jaubert et al., 2005). These findings carry significant implications, as they point to a more effective way of determining the real-time rate at which a material is corroding. This means that more precise preventative and maintenance measures can be implemented for public safety and efficiency, as many of these industrial materials support bridges, highways, trains, locks, gears, and other critical infrastructure. Acoustic-based non-intrusive monitoring provides this information without destroying or invasively accessing the structures being assessed.

Assisting with industrial machine maintenance is one of the most significant benefits that acoustic-based non-intrusive monitoring offers. "Acoustic emission, as a technique, is well established for monitoring the condition of rotating machinery, detecting cavitation in pumps, detecting defects in boilers during pressure testing, detecting gas leaking through relief valves, etc." (McLenna, 1995, p. 338). For those unfamiliar with intensive industrial machinery, it can be difficult to appreciate just how massive and intricate these structures are — systems of pumps, pulleys, and hydraulics, all enclosed and supported by steel frames, nuts, and bolts. Not only is it critical to identify problems as soon as they arise, but it is entirely impractical to disassemble these structures for diagnosis, which would waste time and cost thousands in lost revenue.

Another ideal application of acoustic-based non-intrusive monitoring is with fluid bed processes. These processes are excellent candidates for this type of monitoring because the primary form of transport within the mechanism is the upward movement of low-density areas or gas bubbles (McLenna, 1995). "An acoustic emission transducer will detect an increase in signal as a bubble passes due to the surge of accompanying particles. The regular flow of bubbles appears as a series of rhythmic pulses in the average intensity of the acoustic emission signal" (McLenna, 1995, p. 339). This form of monitoring is particularly well-suited to these processes because it can reveal a great deal of detail about what is occurring internally, offering promising diagnostic possibilities across a range of fields.

Some of the most notable benefits of acoustic-based non-intrusive monitoring arise in the assessment of diesel engines. Diesel engines power a wide variety of machines by generating drive power to overcome resistance loads through the combustion of fuel, converting it into the mechanical energy needed to power movement (Elamin et al.). Because diesel engines are so powerful, they are susceptible to degradation from incipient faults, which can also cause substantial economic losses to the user (Elamin et al., 2010). Acoustic monitoring is one of several methods of scrutinizing these engines to ensure they operate as effectively as possible.

Recent research continues to display the wealth of possibilities for acoustic monitoring of diesel engines. Fog and colleagues engaged in an experimental investigation into exhaust valve burn-through on a 4-cylinder, 500 mm bore, 2-stroke marine diesel engine with an output of approximately 10,000 BHP (Mba et al., 2006). The researchers monitored three distinct types of valve — from normal to those with a large leak — and the vibration and structure-borne stress waves were observed and recorded (Mba et al., 2006). The findings demonstrated that acoustic emissions provided more information about valve and injector-related mechanical events during the combustion process than time-series data from other sensors (Mba et al., 2006). This is significant for all engineers who work on diesel engine performance, as acoustic emissions offer more diagnostic clues about the issues that may arise. Recalling the work of Friis-Hausen and associates, two distinct failure modes were identified in marine diesel engines: exhaust valve leaks and defective injection (Mba et al., 2006). The exhaust valve is responsible for sealing the combustion chamber from its surroundings during compression — a function that guarantees maximum pressure in the cylinder during the combustion event and ensures maximum engine performance (Mba et al., 2006). Mba and associates found a successful classifier capable of distinguishing between different exhaust valve leak sizes on the basis of r.m.s. acoustic emission signals alone, significantly clarifying the repair process and reducing the guesswork involved in maintaining this complex system.

Acoustic-based signals via non-intrusive monitoring can also be used for the detection of engine misfire. The work of El-Ghamry and colleagues demonstrated the value of acoustic signals in assessing the strength of air-fuel mixtures in a 30.56-litre Perkins 4-stroke, 8-cylinder turbocharged gas engine (Mba et al., 2006). Findings by this same research team also showed that acoustic emissions could serve as indirect measurements of cylinder pressure (Mba et al., 2006). "The AE r.m.s. was correlated to the pressure in the time and frequency domain… El-Ghamry noted the advantage of employing cepstral analysis for the model, stating that it used the frequency content of the AE r.m.s. signal rather than the energy content, which gave the advantage over signals with low energy content" (Mba et al., 2006). These results are significant, as relying on the frequency of the acoustic signal means that researchers can still obtain a clear reading regardless of the given energy content.

The research article "A study of the tribological behaviour of piston ring/cylinder liner interaction in diesel engines using acoustic emission" by Douglas and colleagues examines the application of non-intrusive acoustic emission measurements in providing information about the relationship between piston rings and cylinder liners in a collection of diesel engines (2006). The researchers determined that this form of monitoring is effective and can open a new avenue of exploration into this critical dynamic of engine operation (Douglas et al., 2006). "AE generated during normal engine operation is known to consist of contributions from a number of different sources such as injector and valve activity. A recent finding has been the identification of AE signals associated with the ring/liner interface, which presents the opportunity for in-service monitoring" (Douglas et al., 2006). The paper explores and analyzes the range of potential acoustic emission dynamics, including asperity contact and lubricant flow (Douglas et al., 2006). The significance of this research lies in showing that acoustic-based non-intrusive monitoring is a promising method for scrutinizing the tribological dynamics of the piston ring/cylinder liner interface — achieving feats that other forms of monitoring simply cannot. Furthermore, the inherently non-intrusive nature of the method means that the delicate balance of engine performance is not compromised (Douglas et al., 2006).

This is indeed crucial, as any negative effects that could compromise the reliability or performance of the engine would be counterproductive. Moreover, the ease with which acoustic-based non-intrusive monitoring can be applied indicates that it could also serve as a successful tool "for the in situ appraisal of component condition, wear rates and lubricant performance and could significantly aid condition-based monitoring strategies" (Douglas et al., 2006). The success of acoustic-based monitoring in this context demonstrates its potential to be further improved and evolved for even more challenging applications in the future.

Conclusion

Non-intrusive monitoring in its many forms represents a powerful and expanding toolkit for researchers, engineers, policymakers, and medical professionals alike. From measuring household energy use to diagnosing diesel engine faults and monitoring the behavior of wildlife, these technologies offer reliable, detailed information without disrupting the systems under observation. Acoustic emission monitoring, in particular, demonstrates a clear advantage over vibration-based methods for complex environments such as diesel engines, offering superior signal-to-noise ratios and the ability to detect fine-grained faults. Biological sensors promise to transform quality of life for individuals with chronic conditions and could one day reduce drunk-driving fatalities at a societal scale. Environmental applications — from monitoring coastal ecosystems to tracking animal vocalizations — highlight the genuinely cross-disciplinary reach of these technologies. As research in this field continues to advance, non-intrusive monitoring will only grow in capability and importance, enabling safer infrastructure, healthier populations, and a better-understood natural world.