Bridge vibration energy harvesting for wireless IoT-based structural health monitoring systems: A review

1.1. Bridge health monitoring

Bridge health monitoring (BHM) is a fast-growing field now, with researchers contributing to the use of many technologies regarding health monitoring. The use of sensors, wireless sensor nodes (WSNs), on-board microcontrollers or processors managing the data, the data transmission protocol, the power management for the WSN, the collection of useful data and rejection of non-useful data, the use of artificial intelligence for damage detection, are all the contributing areas to the better utilization of resources. Information about bridge acceleration, frequency, strain, temperature and environment’s humidity, wind speed, etc., on the bridge site, is usually utilized to assess the response of the bridge. World around, researchers have put sensors on bridges for a wireless Internet of Things (IoT) based structural health monitoring (SHM) as shown in Figure 1. The monitoring system comprised a WSN (for sensing the bridge’s acceleration, strain, deformation, or velocity), a wireless gateway router (for receiving the information from WSNs and transmitting the information to the cloud server), and a cloud server (for storing the data and make it accessible for IoT). Once the data is available on the cloud server, it can be retrieved worldwide through the internet for analysis and interpretation. Some of the sensor-based systems mounted on bridges and reported in the literature are given in Table 1, with information such as the bridge location, the sensors mounted, the aim to achieve, the data transmission method, and the powering method of the sensor system. The information in the table indicates that for SHM, the accelerometer is the most implemented sensor on bridges. Moreover, the sensors for monitoring temperature, inclination, strain, and wind velocity are also deployed on bridge structures.

OPEN IN VIEWER

Figure 1.

A wireless IoT-based structural health monitoring system for a bridge.

OPEN IN VIEWER

Table 1.

Sensor-based monitoring systems mounted on various bridges for SHM.

Ref.	Bridges and their location	Monitoring parameter	Power source	Communication module	GPS/GSM/RF/Bluetooth	Sensor
Sumitro et al. (2001)	Akashi Kaikyo, Tatara, Japan	Load effects and responses	Solar	Fiber optic	GPS	Accelerometer, Velocity gage, Seismometer, displacement gage, thermometer, and GPS.
Kim et al. (2007)	Golden Gate, USA	Load effects	Batteries	Wireless	Multi-hop	Accelerometer SD-1221 L and ADXL-202E
Sazonov et al. (2009)	RT11 Bridge, Potsdam, NY	Temperature	Harvester+capacitor	Wireless	Data logger	TC1047 Microchip Technology
Jang et al. (2010)	Jindo Bridge, South Korea	Temperature, Strain, inclination, Wind velocity, acceleration	Solar/Batteries	Wireless	WSSN	Thermometer, strain gage, inclinometer, anemometer, and accelerometer.
Rodrigues et al. (2010)	Leziria Bridge, Portugal	Strain, Deflection Temperature	Batteries	Fiber optic	Data loggers	Fiber Bragg Grating
Magalhães et al. (2012)	Infante D Henrique. Bridge Portugal	Long term monitoring	Batteries	Wireless	Data loggers	Accelerometer, strain gage, clinometer, thermometer.
Chae et al. (2012)	Yongjong Grand Bridge Korea	Acceleration, Strain, Temperature, Wind velocity	Solar	Wireless	ZigBee CDMA	Temperature sensor, AC-310-002 and ES-U2 accelerometers, KM-100B strain gage, 05103V wind gage RTD,
Hu et al. (2013)	Zhengdian Bridge, China		Batteries	Wireless	Multi-Hop	Accelerometer, strain gage, temperature gage.
Flanigan et al. (2017)	Harahan Railway Bridge USA	Continuous monitoring	Solar	Wireless	ZigBee	GeoSpace Geo-11D4.5-380 VT, Hitec HBWF-35-125-6-10GP-TR Strain gage, silicon Design 2012-002 Uniaxial Accelerometer, silicon Design 2422-005 Triaxial accelerometer,

Different aspects of SHM of bridges were investigated by several researchers. The cable-supported bridges in Hong Kong were studied and reported (Wong, 2004) and the SHM procedures were discussed. An intelligent sensor network that utilized Mote sensing unit clusters (Basharat et al., 2005) was used. The Mote sensing units transmitted the signal to a gateway Mote, which transmitted the data to the base station. Furthermore, the locally collected data were compiled and transferred to the main computer for processing and decision-making. Video cameras were used for bridge health monitoring (Ko and Ni, 2005) in Hong Kong and China. The study was made by using diverse sensors on these bridges. Modeling, data analysis, and safety considerations were studied (Li and Ou, 2016) for cable-stayed bridges. The Bayesian theorem was used for optimal decision-making to make the decision-making dynamic for the health monitoring of bridges (Neves et al., 2019).

Worldwide several bridge structures are reported for health monitoring. Different bridges are equipped with various wireless sensors for health monitoring to achieve the desired objectives. Two bridges in Japan were studied and reported (Sumitro et al., 2001). A suspension bridge “Akashi Kaikyo” and the cable-stayed bridge “Tatara” were selected for analysis. Different gadgets were selected and used for monitoring the bridge response and condition, such as an accelerometer, anemometer, seismometer, velocity gage, GPS, tuned-mass damper displacement gage, bridge edge displacement gage, and thermometer. Efficient network parameters were recommended for the monitoring of the structure. The structural behavior of a suspension bridge “The Golden gate bridge in the USA” was studied (Kim et al., 2007). The deployment of a WSN was the focus of this research for the health monitoring of the bridge. The interest was to find its resonant frequencies, which were found as 0.11, 0.17, 0.22, and 0.27 Hz, and the acceleration levels varied between 5 and 10 mg. A bridge in the USA that was a cantilever truss structure (Catbas et al., 2008) was studied for the variation of temperature effect when the bridge was subjected to wind excitation, dead and live loads. The response of a suspension bridge with a single pedestrian as a periodic excitation cause was discussed (Kala et al., 2009), and a finite element model was used for its analysis. A modern wireless sensor network was installed and reported (Jang et al., 2010) for twin cable-stayed Jindo bridges in South Korea. A wireless smart sensor network was used that employed the Imote2 module. The study was performed for the vibration characteristics of the bridge’s deck, cable, and pylon. For the measurement of vibration and wind velocity, two types of sensors were used. A beam-type bridge “Komtur Bridge” was investigated (Neitzel et al., 2011) in Berlin, Germany. The frequency range for the bridge was found to be 2–2.6 Hz. A dynamic monitoring system was installed at a concrete arch bridge located in the city of Porto in Portugal (Magalhães et al., 2012). The effects on the bridge’s natural frequencies by operational and environmental factors were studied and a strategy was provided for the identification of structural anomalies. This bridge was analyzed for more than 2 years for modal analysis and frequency changes with seasonal temperature changes. A decrease in the natural frequency during morning traffic jams was observed because of the increase in mass of the bridge. As many as 45 sensors were implemented on a suspension bridge (Chae et al., 2012). Yongjong Grand Bridge in South Korea used a ZigBee network for short-range networking. A protocol for 2G and 3G communications for long-distance communication is Code Division Multiple Access (CDMA). The sensors like accelerometers, anemometers, strain gages, and temperature sensors were installed on the bridge. The power requirement of the setup was harvested using solar energy. The bridge vibrations were measured using an indirect method (Malekjafarian et al., 2015) when a vehicle containing vibration measurement instruments was moved over the bridge to record the vibration characteristics of the bridge. The passing of a vehicle on the bridge excited the bridge, and the same vehicle containing sensors measured the vibration characteristics of the bridge for health condition monitoring applications.

The bridge damage can be detected from the bridge bashing or concrete cracking, stiffness, damping, and mass of the bridge which changes the vibration characteristics of the bridge. A railway truss bridge on the Mississippi River was investigated (Flanigan et al., 2017) and a finite element model was prepared for this bridge and vibration frequencies were estimated. The results from modeling were compared with actual measurements. The measurements of parameters were carried out on a busy bridge (Bień and Kużawa, 2020), the bridge dynamic tests were classified, the testing methods were reviewed, and short-term tests and permanent dynamic monitoring were reviewed. These days, a large amount of data is available from bridges but analyzing the data for early warning is important. Two algorithms (Guo et al., 2020) were used to effectively have a classification of operating conditions. A novel probabilistic approach was adopted (Sarmadi et al., 2021) by the use of the “theory of extreme value” and “goodness-of-fit measure” for the estimation of an alarming threshold. A video camera-based technique using vision-based sensing was used (Xiao et al., 2020) and several computer vision algorithms were proposed with specific camera types for the dynamic response of structures.

Sometimes, emerging techniques, out-of-the-box methods, and innovative thinking are employed for health monitoring. Certain BHM research data can validate temperature change (OBrien et al., 2020) which causes the stiffness change. The IoT data acquisition method (Li and Hongyan, 2020) was used to monitor bridge health while machine learning and artificial intelligence (Muin et al., n.d) were proposed for BHM. The wavelet synchro-squeeze transform method was proposed (Mostafa et al., 2020) for the BHM of a railway bridge. When a truck passed a bridge, the bridge frequencies were measured (Cantero et al., 2016). When data from many smartphone accelerometers were analyzed for Harvard Bridge (Matarazzo et al., 2018), the first three modal frequencies were estimated. A vehicle-bridge interaction model (Zhang et al., 2018) was developed and two piezoelectric (PE) harvesters were used to match the vehicle frequency and bridge frequency coupling system, the numerical and experimental results were compared.

The fast-growing interest in bridge health monitoring shows that it is a field of focus and different techniques and technologies are finding their way into this area. Innovative ways of taking bridge health-related data are compiled and analyzed. The means for data acquisition and its transmission to the main computer are still developing. The techniques relying on bridge vibration are also getting new ideas to be implemented for health monitoring.

1.2. Energy harvesting for BHM systems

The sensors may be installed on bridges for the life of the structure, whereas, the batteries powering these sensors drain out after some time. Replacing or charging batteries in these sensors is a laborious task in cases where the bridges are located in difficult-to-reach or remote locations. Powering the sensors through cables is also not a favorable option since it reduces the flexibility of the monitoring system and restricts the number of sensors in the network. When a weak point on the bridge is spotted then the number of sensors at that location needs to be increased but in the case of cable-connected sensor nodes, new wiring may be required, which is difficult to implement in an already developed system. Moreover, the old bridges selected for monitoring may not have built-in ducted cables, therefore, the wireless sensor nodes are a better option in such cases. Recently, wireless sensor nodes (WSNs) are being utilized in IoT systems for bridge monitoring. However, when monitoring with such sensors, problems surface when the powering of sensors is considered. Usually, batteries are utilized for the operation of these sensors, which require frequent recharging and replacement. Additionally, it is the most viable option to make the WSNs autonomous or self-powered systems to harvest their energy for operation from the surroundings, which may extend the usefulness of the WSN for years of operation.

On and around the bridge structure, several viable energies are available, such as solar, thermal, wind, acoustic, and vibration. With the energy harvesting technology (Ahmad and Khan, 2021b; Khan and Ahmad, 2016; Khan, 2015; Khan and Iqbal, 2016; Khan and Khattak, 2016) the ambient energy can be easily converted into electrical power to operate the installed sensors. However, due to the abundant availability and high reliability, the vibration energy on the bridge is targeted for harvesting in most cases. The excitation can be sinusoidal or stochastic, linear or nonlinear. The performance of the harvester is subject to excitation by sinusoidal or stochastic excitation for linear or nonlinear harvesters (Green et al., 2013; Zhang et al., 2023). Similarly, the harvesters can be resonant, non-resonant, and multi-mode (Ahmad et al., 2022) while sometimes frequency up-conversion is done for better power output. A PE harvester using impact for frequency up-conversion (Yao et al., 2021) was proposed for the bridge and other low-frequency applications. Several bridges were studied for vibration analysis. A detailed report (Gaunt and Sutton, 1981) about bridges in the USA presented frequency and acceleration for 62 bridges that were spanning less than 30 m. A problem with the change of frequency during the day was identified (Liu et al., 2017) because of several factors, one of which was the temperature change during the day or the season. Another study identified the temperature variation in seasonal change (Magalhães et al., 2012). The temperature change caused a variation in frequency and was reported (Wahab and De Roeck, 1997) for the B15 bridge in Belgium. The change of frequency necessitated a broadband frequency device for energy harvesting that could harvest vibration energy over a frequency range of the bridge. A feasibility study of a PE harvester was performed on the Palma del Rio bridge (Spain) (Infantes et al., 2023) with the bridge open for traffic and closed for traffic. A study of energy harvesting from bridge vibration using moving successive vehicles was performed (Mousavi et al., 2023) for PE transduction.

Acceleration and frequency levels are determined for several bridges and reported in the literature. The information about these measurements is compiled in Table 2, where the reported bridges are arranged in order of increasing span. The developed bridges are normally of different types, like, steel girder, beam, truss, suspension, concrete arch, and cable-stayed structures. The vibration characteristics depend on the bridge type, length, traffic flow, and transportation type. The steel girder and beam type bridge length vary from 12.3 to 67 m, however, the truss, suspension, concrete arch, and cable-stayed bridges are built for longer distances that range from 155 to 1280 m. The resonant frequency of these bridge structures ranges from 0.4 to 30 Hz, however, the acceleration levels reported are in the range from 0.001 to 0.23 g. For long bridge structures (length of more than 100 m) the resonant frequency range (0.4–4.6 Hz) and acceleration levels (0.007–0.061g) are on the ultra-low side. These low acceleration and low frequency conditions of vibration are posing a challenge to energy harvesting.

OPEN IN VIEWER

Table 2.

Vibration frequency and acceleration levels of various bridge structures.

Country	Bridge Location	Bridge type	Span (m)	Frequency range (Hz)	Acceleration (g)	Ref.
USA	RT11 Postdam	Steel Girder	12.3	3.1		Sazonov et al. (2009)
Sweden	Pershagen Bridge	Beam	18.4	7.8	0.072	Cahill et al. (2018b)
Singapore	Pioneer	Beam	18.16	5.47–29.33	0.025	Brownjohn et al. (2003)
USA	SB-A1	Steel Girder	18.3	6.39	0.23	Gaunt and Sutton (1981)
USA	SB-C-1	Steel Girder	21.94	4.82	0.13	Gaunt and Sutton (1981)
USA	Ypsilanti	Beam	24	2–30	0.001–0.0035	Galchev et al. (2010)
France	North Highway	Beam	33	3.9–14.5	0.059	Peigney and Siegert (2013)
Sweden	Skidträsk Bridge	Beam	36	3.8–4.5	0.005	Cantero et al. (2016)
USA	Chulitana river	Beam	48	2–45	0.023	Huffman et al. (2015)
Belgium	B15	Beam	53	1.88–8.93		Wahab and De Roeck (1997)
Germany	Komtur	Beam	67	2–2.6		Neitzel et al. (2011)
USA	Harahan	Truss (Railway)	155	0.679–1.5		Flanigan et al. (2017)
Czech	Labe river	Suspension (footbridge)	159	1.367–2.286		Kala et al. (2009)
Portugal	Porto	Concrete arch	245	0.7–4.6		Magalhães et al. (2012)
South Korea	Jindo	Cable-stayed	344	0.4–1.9	0.007	Jang et al. (2010)
Korea	third Nongro	Beam		2.74–4.1	0.025	Kwon et al. (2013)
USA	Golden Gate	Truss	1280	0–1.5	0.061	Abdel-Ghaffar and Scanlan (1985)

With vibration-based energy harvesters (Khan et al., 2010; Khan et al., 2014; Khan, 2020; Khan and Ahmad, 2016; Kuang et al., 2021), the bridge excitation energy is transformed into useful electrical energy to power the WSNs (Gao et al., 2020). The power take-off mechanism is developed (Van Den Bergh, 2023) for the bridge structure and then the transduction mechanism is added. Out of many transduction methods, two mechanisms are adopted for this study, electromagnetic (EM) conversion (Khan et al., 2010; Sun et al., 2021) and piezoelectric (PE) transduction (Khan and Ali, 2019; Li et al., 2021; Wang et al., 2021b), which are the most trusted methods for successfully powering the WSNs. A PE vibration energy harvester for railway bridges was introduced as a promising solution (Sheng et al., 2022).

Worldwide, researchers have assessed bridges for energy harvesting from vibration due to traffic flow or in some cases because of high-speed winds. The literature discussed here shows in chronological order the research focus moving from a simple computational model to a finite element analysis and simulations. A computational model was developed for a single-span bridge vibration (Valluru, 2007) to assess its use in the implementation of a PE energy harvester. The cantilever-type harvesters on concrete bridges were investigated (Zhang et al., 2014b). Simulations for a single vehicle moving over the bridge and continuous vehicle flow over the bridge was conducted, and the conclusion was that rough road and shorter-span bridges can contribute to higher power output generation. A PE energy harvester was analyzed for BHM (Hedayetullah, 2010). A bridge model was established for moving load and analyzed for the PE energy harvester (Ali et al., 2011). Finite element analysis was conducted for a railway bridge and it was analyzed for PE energy harvesting (Song, 2019). A bridge was modeled for a wind-induced vibration (Wang et al., 2021a) to harvest vibration energy using a bistable device (Farhangdoust et al., 2019). A novel approach for computation was employed for electromechanical harvesting and sensing device for a real-time and offline parametric identification framework (Kefal et al., 2019). The finite element method (FEM) was utilized to assess the PE energy harvesting for a time-dependent moving load on the bridge (Bendine et al., 2019). An ultra-low frequency tunable X-shaped harvester (Li and Jing, 2019) was proposed for bridge vibration applications that utilized coupled modes because of nonlinearity and structural coupling effect. A railway bridge was analyzed for PE transduction from train-induced vibration using a modal decomposition approach (Romero et al., 2021).

The research then moved to experimentation on actual roadside and bridge structures to utilize the available knowledge. A PE energy harvesting cantilever beam was mounted for getting open-circuit voltage to power a sensor used for damage detection of a laboratory-scale bridge (Fitzgerald et al., 2019). A hybrid harvester using both PE and EM transductions was used (Cao et al., 2023), utilizing wind power for energy input on a canyon bridge. Thus, there are different techniques and approaches utilized for harvesting bridge vibration energy reported in the literature, even though sometimes the road or pavement energy harvesters can also be employed in bridge structures for power generation. An asphalt pavement PE harvester (Yang et al., 2017) was tested in the lab and on the roadside in China. A road-traffic harvester (Gareh et al., 2016) was proposed with PE transduction for such applications. A self-powered weight-in-motion system using vibration energy of pavement utilizing PE transduction is a solution to enforce weight regulation and keep the roads and bridges safe (Birgin et al., 2023).

The bridges are low-acceleration and low-frequency structures. The harvesters that can perform in low acceleration and low-frequency environments have the potential to be used on bridges to extract vibration energy and can be claimed as bridge energy harvesters (BEHs). Some researchers have reported harvesters for such conditions and these harvesters have been simulated in a lab environment, however, they have not been tested in a real bridge environment. whereas, some reported BEHs have been mounted and tested in a real bridge environment.

This work has compiled research performed on BEHs that extract energy from the bridge vibrations and are specifically developed for bridge health monitoring applications. The BEHs which can function in a low frequency and low acceleration environment of bridges and the harvesters that have been installed in a real bridge environment have been studied and discussed. In most of the conducted research, the bridge vibration energy is due to traffic flow and due to high-speed wind and are the targeted energy sources for the harvester’s excitation. The emphasis of this work is on the harvesters’ architecture, transduction mechanism, modeling information, and the input and output of the BEHs. Moreover, the reported BEHs are compared based on input vibrations and output parameters, such as output voltage, power, and power density. This report has collected the latest research reports on the subject and will be helpful for researchers to find all the research trends till the publication of this compilation. Furthermore, this comprehensive work will fill the vacuum that presently exists in this research area.

2.1. Piezoelectric BEHs tested in-lab environment only

In PE-BEH, the direct piezoelectric effect of the piezoelectric material is utilized and the vibration energy (mechanical energy) is converted into electrical energy. In the PE material, the dipoles are initially randomly oriented, however, when the material is subjected to a strain, these dipoles align such that the material is polarized and voltage is induced across the material. When a cyclic input (vibration) is provided to the material an AC output voltage is obtained as electrical energy. The PE energy harvesters get recognized for the bridge energy harvesting cases when there is a need for small electrical energy (for monitoring system). Mostly, in PE-BEHs, a PE patch is attached near the fixed end of a cantilever beam. The amount of ambient vibration energy extracted by the PE-BEHs depends on frequency and acceleration levels of input vibration, frequency band, and other factors, such as PE material type, the inertial mass of the harvester, resonant frequency, damping co-efficient (mechanical and electrical), electromechanical coupling coefficient, and the external electrical load. There are several types of PE materials available with different properties and features. Detailed information about PE energy harvesting techniques is comprehensively available in reference (Liang et al., 2021). A variety of PE materials are available for PE-based BEHs and the selection depends on high piezoelectric coupling, curie temperature, dielectric constant, and modulus of elasticity of the material. The main properties of the piezoelectric materials are shown in Table 3. Among all piezoelectric materials, lead zirconate titanate (PZT) and polyvinylidene difluoride (PVDF) are the two piezoelectric materials that are widely used in BEHs. PZT is preferred due to the high level of piezoelectric constant and ease of in-situ fabrication, however, mechanical flexibility is the major preference of PVDF polymer, although it has a much lower PE constant in comparison to PE ceramics. For PZT, the electromechanical coupling constant is 2.5 times higher than that of PVDF. Comparatively, the curie temperature of gallium orthophosphate, sodium bismuth titanate, and potassium bismuth titanate is on the high side and can be utilized in BEHs used in high-temperature environments.

OPEN IN VIEWER

Table 3.

The available PE materials for PE-BEHs.

Category	Material type	Curie temperature (°C)	Density (g/cm3)	Dielectric constantAt 1 kHz	Piezoelectric constant(pC/N)	Modulus of elasticity(GPa)
Natural crystal	Berlinite		2.64
	Lead titanate	240	6.7	270		112.8
	Quartz	550	2.6	3.75	−2.3 (d11)0.67 (d14)	76.5
	Topaz		3.6
	Aluminum oxide
Polymers	Polyvinylidene difluoride	205	1.78	12	6–10 (d31)6–10 (d32)13–22 (d33)	8.3
Artificial crystals	Gallium orthophosphate	970	3.57	10.3	4.5 (d11)1.9 (d14)
	Langasite			51.2	−5.94 (d11)4.99 (d14)
Artificial ceramics	Aluminum oxide	311	3.95	11.5		413
	Aluminum titanate	20	3.4
	Barium titanate	120	6.02	1550	160 (d33)	67
	Bismuth ferrite
	Lead lanthanum	315	8	11262	160 (d33)
	Lead magnesium niobate	170	7.85	5500		61
	Lead zirconate titanate (hard)	330	7.5	1250	−55 (d31)190 (d33)	63
	Lead zirconate titanate (soft)	250	7.5	1725	−170 (d33)425 (d33)	63
	Potassium bismuth titanate	380
	Sodium bismuth titanate	655			10
	Zinc oxide	243			11.67 (d33)−5.43 (d31)

A PE energy harvester was developed (Rhimi and Lajnef, 2012) that utilized tensile preload in an axial direction of the beam as shown in Figure 2. A cantilever-type piezoelectric bimorph was used with tip mass. The tip mass was extending to the sides where threads for the screw were cut. Two springs with two screws on each side of the cantilever beam were used and the beam was pre-stressed. The loading on the two springs was adjustable. The harvester was modeled and tested in-lab for harvesting energy from structures. A plot of experimentation showed that under a load resistance of 10 MΩ and at a frequency of 40 Hz the harvester produced 27 V/g when a prestress of 4 N was applied on a beam.

OPEN IN VIEWER

Figure 2.

The PE-BEH with springs for preloading. Reproduced with permission (Rhimi and Lajnef, 2012). Copyright 2012 ASCE Library.

A multi-impact BEH was produced (Zhang et al., 2014a) and is shown in Figure 3. The harvester consisted of a spring-mass system hanging vertically and had two PE cantilever beams at each side of the hanging mass. The mass hung with spring moved at a low frequency of excitation. It had a series of rollers that were used to excite the PE beams by hitting the projections from the beams. Thus, the low frequency of excitation was converted into a high frequency of vibration of the follower PE beams. The device was tested with simulated bridge vibrations. The experimentation showed that the harvester produced 2.7 mW (load resistance of 9.7 kΩ) of power at 2.8 Hz when an acceleration of 0.29g is applied.

OPEN IN VIEWER

Figure 3.

A multi-impact PE energy harvester (Zhang et al., 2014a). Reprinted under Creative Commons Attribution 4.0 International License, from IOP Publishing.

A PE-BEH (Figure 4) utilizing magnetic levitation for nonlinearity was proposed (Zhang et al., 2015) for wider bandwidth of bridge vibration. The low frequency of bridge vibration was converted to high frequency through a multi-impact four doubly clamped PE beams. The multi-impact configuration was established when the low-frequency lavation magnet mounted on a series of rollers passed through the bulging center. It displaces PE beams to vibrate at high frequency until the next roller hits the bulging center attached to the PE beam. The magnetic levitation introduces stiffness nonlinearity in the PE system. A nonlinear model was developed and simulated with a sinusoidal as well as the bridge’s vibration spectrum. The simulation results revealed better power generation from the harvester.

OPEN IN VIEWER

Figure 4.

A nonlinear multi-impact PE energy harvester for bridge vibration: (a) design, (b) with two beams removed, and (c) a front view with two beams removed (Zhang et al., 2015). Reprinted under Creative Commons Attribution 4.0 International License, from IOP Publishing.

A theoretical model was developed and experimentally validated for a PE cantilever beam harvester (shown in Figure 5) for bridge vibration application (Karimi et al., 2016). The bridge interaction model was derived using a moving concentrated mass and a distributed mass due to traveling vehicles. Moreover, the bridge modeling results were compared to the data available in the literature. A mathematical model for a single PE patch and tip mass was developed using the Euler-Bernoulli theorem. A low amplitude chirp excitation test was used for electro-elastic frequency response. The optimal load was computed and was also found from experiments. The acceleration signal derived from the model was used on a shaker to simulate the bridge excitations. At the resonant frequency of 13.5, the output voltage was 2.7 at a resistive load of 200 kΩ. The experimental results were in agreement with the theoretical findings.

OPEN IN VIEWER

Figure 5.

A single cantilever beam PE energy harvester tested in-lab. Reproduced with permission (Karimi et al., 2016). Copyright 2016 Elsevier.

For the low-frequency bridge vibration, a multi-mode microscale PE harvester was reported (Bhaskaran et al., 2017). The developed harvester is shown in Figure 6. The device consisted of an array of cantilever beams with resonant frequencies spaced close to each other so that the device was able to harvest a wide range of frequencies. The harvester consisted of six cantilever beams consisting of five energy harvesting beams and one sensor beam. The power output of the device was 2.283 mW/g and the integrated accelerometer had a sensitivity of 27.67 V/g. The device was modeled and simulated to assess its feasibility for use.

OPEN IN VIEWER

Figure 6.

A micro-scale multi-mode PE energy harvester. (Bhaskaran et al., 2017). Reprinted under Creative Commons Attribution 4.0 International License, from Hindawi.

An impact loading harvester for pavement deformation of the bridge was presented (Yesner et al., 2019) and a single harvester from an array is shown in Figure 7. A novel cymbal design of a harvester was used that utilized PZT-5X soft ceramic and hard steel caps. Four rows of four modules were used and were stacked one over another which made 64 transducers in parallel. A load of 227 kg at 5 Hz produced 2.1 mW of power at a resistive load of 330 kΩ.

OPEN IN VIEWER

Figure 7.

A novel flextensional bridge PE energy harvester. Reproduced with permission (Yesner et al., 2019). Copyright 2019 Elsevier.

Experimental validation of a PE energy harvester was performed (Cahill et al., 2018c) for a civil structure. The diagram of the harvester is shown in Figure 8. A laboratory-based study was performed on railway bridges. Acceleration response from the international train fleet was used as a source of excitation for the simulation of the devised prototype harvester. The harvester’s performance was investigated numerically and validated experimentally. A sinusoidal acceleration of 0.5g at a frequency of 13.4 Hz produced 0.183 V RMS voltage in one of the studied cases.

OPEN IN VIEWER

Figure 8.

The PE energy harvester is attached to the host structure (left) and side view of the harvester (right). Reproduced with permission (Cahill et al., 2018c). Copyright 2018 ASCE library.

An auxetic structure has a material with a negative Poisson’s ratio which is utilized in a bimorph PE cantilever beam and was investigated (Farhangdoust, 2020) for energy harvesting to power wireless sensors of the IoT network for the sunshine skyway bridge in south Florida USA (Farhangdoust et al., 2020). The BEH was to harvest energy from the stay cable of the bridge as shown in Figure 9. The vibration of the stay cable was measured using a contactless laser vibrometer. The vibration data was used in simulations for the output of the device. COMSOL multi-physics was used for modeling. The acceleration of 0.1g at a frequency of 1, 3, and 5 Hz produced a power output of 614, 2663, and 12184 µW respectively when the resistive loads were 28, 8, and 6 kΩ. The performance of an auxetic cantilever beam was computationally investigated for the bridge stay cable (Farhangdoust, 2020). The harmonic excitation simulation at 3 Hz frequency and 1 m/s² acceleration showed 427.22 µW power output which was compared to the output (170.17 µW) of the conventional beam under the same condition.

OPEN IN VIEWER

Figure 9.

The BEH with auxetic structure: (a) Top and side views of the auxetic cantilever (left) and conventional harvester (right) and (b) the cable-mounted harvester (Farhangdoust, 2020). Reprinted under Creative Commons Attribution 4.0 International License, from SPIE.

A vibration energy harvester designed for a railway bridge and using a PE transduction was presented (Hou et al., 2021) for Guangzhou Metro No.14 bridge as an example. The schematic of the device is shown in Figure 10. Train speeds of 120 and 80 km/h were used as excitation and the first 10 natural frequencies for the bridge were found to be in the range from 1.308 to 5.297 Hz. To harvest this energy a BEH consisting of six multilayer stacks of PZT-5H was reported. The PE elements were arranged in series with a steel spring in the floating concrete slab so that it had large stiffness, high bearing capacity, small height, and dimensions. The steel spring fulcrum force was amplified for the effective harvesting of vibration energy. A total of 144 harvesters were arranged in 36 blocks making it a multilayer PE-stack energy harvester that produced 31.4 kJ of energy. The cover plate thickness of 20, 15, and 10 mm was used and at the vehicle speed of 120 km/h, voltage levels of 199.6, 195.8, 181.5, and 102.6 V were recorded. The experiment was also repeated at 80 km/h train speed. The theoretical and finite element method results were in close agreement. For a single harvester, with a cover plate of 20 mm thickness, the output voltage, current and power output of 195.8 V, 5.6 mA, and 1.09 W respectively were reported. Thus, 1.07 J of output energy was available in the effective acting time of the train with a power output of 0.21 W from single a PE-BEH.

OPEN IN VIEWER

Figure 10.

A multilayer PE-stack energy harvester with labels. Reproduced with permission (Hou et al., 2021). Copyright 2021 Elsevier.

A monostable nonlinear PE energy harvester was developed for bridge vibration (Zhou et al., 2021). A cantilever beam with tip mounted magnet and two stationary magnets oriented above and below the beam (fixed to the frame) were used for attractive forces as shown in Figure 11. The beam carried a PE patch for energy harvesting. A bridge model was fabricated for testing the reported harvester. On this bridge model, at a load resistance of 0.9 MΩ, the test results for monostable and linear PE-BEH were compared at different rolling speeds of 1 kg mass. For a rolling speed of 2.5 m/s, the monostable configuration of the device resulted in a power density of 1980 W/m³, however, as a linear system it produced 1150 W/m³. Furthermore, a maximum voltage of 7.49 V is reported for the monostable harvester. The linear system configuration of the device has a natural frequency of 14.5 Hz; in the monostable case, the frequency varies and at the speed of 1.07 m/s, its frequency is 4.9 Hz. Moreover, in another bi-stable configuration of the harvester (Zhou et al., 2022) the power density of 680 W/m³ is reported at a rolling mass speed of 2.5 m/s.

OPEN IN VIEWER

Figure 11.

The schematic diagram of the BEH (Zhou et al., 2021). Reprinted under Creative Commons Attribution 4.0 International License, from SPIE.

The PE energy harvesters can produce higher voltage levels from low acceleration levels and are more suitable for low-power and high-voltage wireless sensor nodes and gadgets that are used in bridge health monitoring. For better performance, the resonant frequency of the harvester has to match the low excitation frequency of the bridge thus longer beam lengths are expected (to have lower stiffness) for the energy harvesters. The length of beams can be kept short if the beam material is selected to have a lower Young’s modulus of elasticity. Most of the PE devices reported as bridge energy harvesters have long cantilever beams to match the lower frequency of excitation but do not employ any new technique to increase the power output except by trying a different PE material. The device (Zhang et al., 2014a) utilized the frequency up-conversion technique to increase the harvester’s output. Not much attention is given to exploiting other ways to improve the performance of PE-BEHs, for example, the device reported by (Bhaskaran et al., 2017) has utilized a multi-mode technique for broadband frequency harvesting. The harvester (Yesner et al., 2019) has utilized multilayer stacks to increase the power output. The device report by (Hou et al., 2021) has magnified the input excitation by increased displacement for better response and multiple layers of PE material are employed for increased power output.

2.2. Electromagnetic BEHs tested in-lab

The performance of EM energy harvesters is well established. The EM harvester consists of a magnet and a conducting coil which in the presence of relative motion induces a current in the coil. The magnetic field is made to cut the conductor by moving the magnet or a coil. The output current is AC which needs to be rectified to DC for storage in batteries or uses in electronics. Mostly the coil is made of copper, however, other options are also available for the coil’s material as shown in Table 4. Usually, either planar or wound coil is used in EM-BEHs, the wound coil is preferred due to high voltage and power generation. However, the planar coil has the advantage to be integrated into the standard MEMS fabrication processes. The diamagnetic materials, such as gold, aluminum, and copper are preferred as coil materials since these offer less repulsion force in the magnetic field and attribute to high power generation. On the other hand, nickel has high yield strength and young modulus and can survive better in a cyclic vibration environment due to its high fatigue life. Moreover, the electrical resistance of copper and silver is relatively low and will attribute fewer power losses in the coil. The materials for the permanent magnets used in EM-BEHs are listed in Table 5. Because of high remnant magnetic flux density, neodymium (NdFeB) magnets are widely employed in BEHs, however, the lower curie temperature limits these to be used in a relatively high-temperature environment. Comparatively, Alnico and Samarium cobalt magnets have a better chance to perform well in harsh temperatures. Additionally, for long-lasting, embedded applications, Samarium cobalt, and Neodymium magnets are preferred due to their high coercivity. The suspension system (spring), magnet type, coil, and architecture are the main features of EM-BEHs. Under bridge’s excitation, the performance of EM-BEHs highly depends on the magnet (grade, type, shape and size), coil (planar, wound, size, number of turns, material, and architecture (moving coil, moving magnet, beam type, etc.). The bigger the size of the EM-BEH and the better its performance will be. The bridge structures are big and heavy. The effect of the mass of the harvester is negligible compared to the mass of the bridge thus the size limitation is not a problem in the bridge applications. Thus, a harvester is designed to deliver the required power with minimum size.

OPEN IN VIEWER

Table 4.

Coil materials used in EM-BEHs.

Materials	Electrical resistivity at 20°C (nΩ·m)	Thermal conductivity (W/mK)	Coefficient of thermal expansion (1/K)	Yield strength (MPa)	Young’s modulus (GPa)
Gold	24.4	318	14.3 × 10−6	207	80
Aluminum	28.2	237	23.0 × 10−6	170	70
Copper	16.78	401	16.5 × 10−6	70	128
Silver	15.87	429	18.9 × 10−6	55	83
Nickel	69.3	90.9	12.8 × 10−6	150–480	210

OPEN IN VIEWER

Table 5.

Main properties of permanent magnets used in EM-BEHs.

Magnet	Curie temperature (°C)	Coercivity (MA/m)	Remnant magnetic flux density (T)
Neodymium (NdFeB)	310	2	1–1.4
Cast alnico (AlNiCo)	860		1.2
Samarium cobalt (SmCo)	825	2	0.8–1.1
Sintered alnico (AlNiCo)	810	0.028	0.62
Sintered strontium-ferrite (SrFeO)	460	0.03	0.4

An electromagnetic (EM) energy harvester was proposed (Jung et al., 2011) for wind-induced excitation. The developed BEH is attached to the stay-cable as shown in Figure 12. A numerical simulation was done and the results were validated experimentally. The ambient acceleration was measured and used in numerical simulations and experimental validation. The device consisted of a moving coil attached to a spring having a relative motion with respect to the fixed magnet. The device produced 233.49 mW peak power, and 27.14 mW root mean square (RMS) power when the input acceleration was 0.0748g and the attached resistance was 23.65 Ω.

OPEN IN VIEWER

Figure 12.

Prototype of a moving-coil EM energy harvesting device’s drawing (left) and picture (right) (Jung et al., 2011). Reprinted under Creative Commons Attribution 4.0 International License, from IOP Publishing.

Two nonlinear harvesters, one with a repulsion magnet on each side, while spring force on the other side and a second harvester with two magnets on opposite sides were presented (Shi, 2013) and are shown in Figure 13. The wind-excited bridge structure was considered to harvest bridge vibration energy for power generation. The excitation from the wind was modeled and the two BEHs were simulated in the lab. The first was a resonant device having a magnetic mass suspended from a spring moving inside a copper coil. Under the coil, a second magnet was used for repulsion. The second device consisted of repulsion magnets at the top and bottom to keep the central magnet (movable) suspended in the air. An array of the first device was proposed to be mounted on a bridge as a tuned damper for power generation. The devices were modeled and simulated for power generation.

OPEN IN VIEWER

Figure 13.

Moving repulsion-magnet EM energy harvesting devices: (a) a Single repulsive magnet and (b) both sides having a repulsive magnet (Shi, 2013). Reprinted under Creative Commons Attribution 4.0 International License, from MIT Libraries.

A harvester utilizing two latching magnets holding a latex membrane for powering sensors on the bridge was developed (Khan and Ahmad, 2014) and is shown in Figure 14. A latex membrane was used to hold two magnets latched together above and below it. The magnets oscillate in wound coils in two Teflon spacers. Moreover, the top and bottom of the spacers were also closed at each end with planar coils. The device had a resonant frequency of 27 Hz and generated a voltage of 15.7 mV from a single planner coil and 11.05 mV from a single wound coil while the power output of 1.8 and 2.1 µW is reported during lab testing.

OPEN IN VIEWER

Figure 14.

Exploded view of the developed EM harvester. Reproduced with permission (Khan and Ahmad, 2014). Copyright 2014 IEEE.

An EM energy harvesting device for a seismically isolated bridge structure was reported (Lu et al., 2014) and is shown in Figure 15. The isolated bridge had a frequency below 1 Hz, a velocity of nearly 1 m/s, and a displacement of the order of 0.5 m. The displacement was used to drive a gear mechanism to drive a crank-piston mechanism having a permanent magnet as a piston moving in a copper coil. A mathematical model was developed for the device and was analyzed.

OPEN IN VIEWER

Figure 15.

The crank pin mechanism EM energy harvester (Lu et al., 2014). Reprinted under Creative Commons Attribution 4.0 International License, from SPIE.

The damping of vibration energy at the stay cable of the bridge was analyzed for EM energy harvesting (Shen and Zhu, 2015). The diagram of the reported harvester is shown in Figure 16. The energy from the wind excitation of the bridge was converted to electrical energy thus both the damping and harvesting objectives were achieved. The device converted the linear motion into rotary motion which was used to generate an electromotive force (emf) because of the relative motion between the magnet and coil. Simulations revealed that the device could harvest 82.5 mW at a wind speed of 9 m/s and 2396.8 mW at a wind speed of 15 m/s.

OPEN IN VIEWER

Figure 16.

The cable-stay vibration EM bridge vibration energy harvester (Shen and Zhu, 2015).

A dual mass damper used for harvesting the bridge’s vibration energy was designed and its parametric study was undertaken (Takeya et al., 2016). The line diagram of the BEH is shown in Figure 17. A linear EM harvester using a single mass-tuned damper for power generation is already established. In this study, a dual-tuned mass damper system was used and was named a dual-tuned mass generator. The device was modeled and simulated and a parametric study was performed to attain a better generation of power from bridge vibration.

OPEN IN VIEWER

Figure 17.

Dual mass tuned-damper EM energy harvester. Reproduced with permission (Takeya et al., 2016). Copyright 2016 Elsevier.

An EM harvester utilizing bridge vibration and the wind gusts on the bridge due to traffic flow was fabricated (Khan and Iqbal, 2016). The device consisted of a cantilever beam carrying a tip magnetic mass and having a copper coil under it and a thin rod-supported bluff-body for harvesting air movement. The device had two sources of excitation that are the base excitation and the wind gust. When Prototype-I (Figure 18(a)) is applied to 3.6 Hz resonant frequency and 0.4g acceleration, the harvester produced 290 mV RMS voltage and 354.5 µW maximum power (at 54.5 Ω resistive load). Additionally, the harvester produced 7.1 µW at a wind speed of 6 m/s. A second harvester was produced in which the coil was mounted on the tip of the second cantilever (Khan and Iqbal, 2018). The prototype-II (Figure 18(b)) had resonant frequencies of 7.6, 33, and 45 Hz. When excited at 7.6 Hz, and 0.6g acceleration, the harvester produced 430 mV maximum voltage and 2214.32 µW optimum power. Furthermore, at the wind speed of 6 m/s, it produced 22 mV voltage and 9.14 µW of power.

OPEN IN VIEWER

Figure 18.

Dual-energy EM bridge energy harvesters: (a) fixed coil and (b) moving coil (Khan and Iqbal, 2018). Reprinted under Creative Commons Attribution 4.0 International License, from Hindawi.

A multimode hybrid BEH was fabricated (Iqbal and Khan, 2018) and is shown in Figure 19. The device consisted of two cantilever beams, one carrying a copper coil and the second beam had a tip mass as a magnet placed just above it. The cantilever beam carrying the magnet also had a PE plate mounted at the fixed end. At the tip of the magnet beam, an extension was used to carry an airfoil to capture energy from the wind gusts of the traffic. The device had three resonant frequencies at 11, 38, and 43 Hz. When excited at the first mode of vibration 11.1 Hz, 0.6g acceleration, and 28 Ω matching impedance the EM portion generated 401 mV voltage and 2214.32 µW maximum power. The PE patch produced 4500 mV voltage and 155.7 µW at 0.4g acceleration across 130 kΩ load resistance. When excited by the wind speed of 6 m/s the EM portion produced 25 mV voltage and 9.1 µW power. At 6 m/s wind speed and 0.4g acceleration, the PE portion produced 6421 mV voltage.

OPEN IN VIEWER

Figure 19.

Multi-energy hybrid harvester. Reproduced with permission (Iqbal and Khan, 2018). Copyright 2018 Elsevier.

Tamar Bridge in southwest England was used to record vibration data for 2 days in real traffic conditions (Gaglione et al., 2018) using a direct and autoparametric vibration energy harvester. The selected bridge was a suspension bridge having a total length of 563 m with a main span of 335 m. The acquired data was used in the lab to estimate the power output and multi-hop data transmission protocol to be able to use the harvester in a real bridge environment. Two locations were identified on the bridge where excitation frequencies were 9.1 and 18.7 Hz. The maximum output from the harvester was estimated as 661 µW. The harvester and experimental setup are shown in Figure 20.

OPEN IN VIEWER

Figure 20.

Experimental setup for the direct and autoparametric energy harvester (Gaglione et al., 2018). Reprinted under Creative Commons Attribution 4.0 International License, from Aston University.

A train-track-bridge analysis was performed for the induced vibration from the vehicle movement. The finite element analysis (ANSYS) model of the Guangzhou Metro Line-14 bridge was established and the first 10 natural frequencies were acquired for simulation purposes. A harvester with a mass block supported on springs, having two magnets and a wound coil was reported (Hou et al., 2018) as shown in Figure 21. The wound coil was attached with a mass block and was moving while the two magnets were stationary. The device was able to harvest at a low frequency of 5.5 Hz and 3.71 m/s² acceleration. It produced a peak power of 35.3 W for a device volume of 2 × 10⁵ cm³ which corresponds to 176.5 µW/cm³ power density.

OPEN IN VIEWER

Figure 21.

The moving coil EM energy harvester. Reproduced with permission (Hou et al., 2018). Copyright 2018 Elsevier.

A multi-frequency energy harvester was analyzed (Hou et al., 2020). As shown in Figure 22, the harvesting device consisted of four modules arranged parallel to each other, each having a different frequency. Therefore, each one could be tuned to a different frequency peak of external vibration. The modules were comprised of lead-block mass mounted on springs and carried two coils. Permanent magnets were installed opposite the moving coils. An analytical model for the device was developed and simulations were performed on three types of devices (with varying sizes of Lead-block dimensions, mass, and magnets central spacing). The average power output for the three prototype harvesters was 0.91, 1.41, and 0.84 W.

OPEN IN VIEWER

Figure 22.

The produced harvesting device: (a) top view and (b) side views. Reproduced with permission (Hou et al., 2020). Copyright 2020 Elsevier.

An EM energy harvester shown in Figure 23 is suitable for harvesting bridge vibration energy was investigated (Amjadian et al., 2021). The work aimed to investigate that the scaling factor does not reduce the output of the harvester if the EM coupling is strong. The device consisted of a flexible cantilever beam having a rigid link supporting a rectangular copper coil. There were magnets fixed with the base on both sides of the coil in alternating poling directions. A uniform poling orientation was also used for comparison but it produced lower output. An FEA software COMSOL was used for simulation and an analytical model was reported, which were in close agreement. The simulation results showed that an excitation frequency of 4 Hz could generate a maximum output power of 1.716 W at 0.1g acceleration and 0.5 Ω load resistance.

OPEN IN VIEWER

Figure 23.

The schematic diagram of the harvester side view (left) and front view (right) (Amjadian et al., 2021). Reprinted under Creative Commons Attribution 4.0 International License, from SPIE.

A two-degree of freedom EM-BEH was developed (Masood Ahmad and Ullah Khan, 2021) and is shown in Figure 24. The overall length of the harvester’s cantilever beam was reduced by the folding beam approach. The outer beam was split into two pieces along its length and was spaced apart to accommodate the second portion of the beam inside it and extend back to the fixed end. This way the length of the cantilever beam was reduced by almost half. Both the tips were loaded with magnets to make the harvester a two-degree-of-freedom system. Each magnet had a wound copper coil under it that was mounted on an adjustable table to vary the space between the coil and the magnet. The harvester had two resonant frequencies of 4.4 and 5.5 Hz. It produced 0.592 V voltage, 2.337 mW of power from coil-1 and 0.161 V voltage, and 0.173 mW power output from coil-2 at the first resonant frequency of 4.4 Hz at 0.09g acceleration while connected to 150 Ω load resistance. The output at the second resonance of 5.5 Hz from coil-1 was 0.368 V voltage and 0.901 mW of power, while coil-2 produced 1.213 V voltage and 9.801 mW power at 0.09g acceleration and 150 Ω load resistance. The power density of the device was 32.13 µW/cm³.

OPEN IN VIEWER

Figure 24.

Two degrees of freedom EM energy harvester: (a) design and (b) the prototype. Reproduced with permission (Masood Ahmad and Ullah Khan, 2021). Copyright 2020 SAGE.

Most of the bridge vibration energy harvesters reported in the literature are EM transducers because of their higher power outputs in comparison to PE-BEHs that generate higher voltage levels. Although, the harvesters reported in this section are not tested in a real bridge environment, however, some of these are tested for similar input conditions in-lab. Few researchers have taken the real bridge environment data of vibration and used that in-lab for the harvester excitation. The results of such a harvester’s experimentation are expected to be close to the results of testing in a real bridge environment. The devices designed for railway bridges are bigger in size and have larger outputs. The earlier EM-BEHs were simple and bigger in size and the harvesters having the stiffness member was coil-spring then the coil was kept movable relative to the fixed magnet because of the smaller mass requirement for lower stiffness (Jung et al., 2011). The effect of nonlinear architecture is implemented to increase the effective frequency bandwidth of the harvester and (Shi, 2013) reported two such energy harvesters. Furthermore, the bridges are heavy structures and thus have huge energy during motion, so multiple transductions were also proposed (Khan and Ahmad, 2014) for increased output. The excitation displacement amplification was employed by (Lu et al., 2014) and (Shen and Zhu, 2015) by using a gear mechanism. Since the bridge site not only has vibration energy but can also have wind gusts, therefore, the availability of wind energy was also utilized by (Khan and Iqbal, 2016) and (Khan and Iqbal, 2018) along the bridge vibrations. Additionally, in some EM-BEHs (Iqbal and Khan, 2018) dual transduction mechanisms were employed to make the harvester hybrid for better performance. Furthermore, for high power output, huge railway BEHs (Hou et al., 2018) and (Hou et al., 2020) are also developed and reported. Employment of more magnets to increase the power output of the harvesters is also performed (Amjadian et al., 2021) for the same size and in some cases (Masood Ahmad and Ullah Khan, 2021) the overall size of the harvester is reduced (folded beam architecture) for the better power density achievement.

3.1. Piezoelectric BEHs tested on real bridge site

Several PE-BEHs are developed and tested on the real bridge vibration environment. In a such direct application, the PE-BEH is specifically designed and tuned for the resonant frequencies of the bridge structure. Moreover, it is also important to target a certain bridge location where adequate vibration amplitude is available for energy harvesting. Additionally, the chaotic bridge vibration with limited (narrow) frequency bandwidth also needs to be considered while designing a BEH.

A simple long cantilever-type PE energy harvester was tested on a bridge (Peigney and Siegert, 2013) and is shown in Figure 25. Two bimorph PE patches (Mide QP20 W) were attached to the lower and upper surfaces of a steel plate 40 × 220 × 0.8 mm³ at the clamped end. A 12 g mass was used to tune the beam to the desired frequency. A prestressed concrete simply supported bridge with five cross-braced girders, having a span of 33 m located in the north of France, was used for the case study. The bridge had resonant frequencies at 3.9, 4.3, and 14.5 Hz. The frequency of 14.5 Hz was targeted for harvesting. A resistive load of 100 kΩ was used at an acceleration of 0.02g and 14.4 Hz frequency to produce 0.19 mW of power in the lab. The resistance was replaced with an ALD EH300 circuit that provided a controlled voltage of 1.8–3.6 V to charge a storage capacitor. When the harvester was tested on a bridge its power output dropped to 0.029 mW, but the half-power frequency bandwidth was increased from 0.2 to 0.8 Hz. The behavior of the harvester was predicted using the mathematical model of Euler-Bernoulli and was experimentally validated.

OPEN IN VIEWER

Figure 25.

A simple cantilever beam PE bridge energy harvester with experimental setup in-lab (Peigney and Siegert, 2013). Reprinted under Creative Commons Attribution 4.0 International License, from IOP Publishing.

A micro fiber-based composite (MFC) PE cantilever beam was used for energy harvesting (Liu et al., 2015) as shown in Figure 26. Theoretical analysis of the device, modeling, and evaluation by simulations was done for the harvesting device with experimental validation. The device was fabricated and a power management circuit was implemented. Energy management and prediction algorithm were also proposed. The device produced peak to peak voltage of 6 V when 4–5 Hz of vibration was applied. The device was tested on a cable-stayed bridge of frequency 1.2785 Hz and produced 0.4 mW of average power.

OPEN IN VIEWER

Figure 26.

The PE energy harvester with MFC material. Reproduced with permission (Liu et al., 2015). Copyright 2015 IEEE.

A multi-modal PE energy harvester (Cahill et al., 2018a) was developed to harvest bridge vibration energy (Cahill et al., 2018b) and is shown in Figure 27. The aluminum substrate was used to bond the PE layer of PolyVinyliDene Fluoride (PVDF) to it. Six such beams were utilized, three beams to cover the frequency range from 6 to 8 Hz and three beams with the range from 15 to 20 Hz. The in-lab calibration of the device produced a maximum voltage output of 0.182 V at 8.5 Hz for the cantilever-2. The second range of beams had a maximum output voltage of 1.626 V at a frequency of 17.6 Hz. The harvesting device was installed to detect train passages on Pershagen Bridge in Sweden, which had a span of 46.6 m and a central section span of 18.4 m with a double-track rail for trains. The peak acceleration was observed to be 0.071g. The maximum output voltage of 0.064 V occurred at cantilever-1 in the first range and 0.099 V at cantilever-4 in the second range.

OPEN IN VIEWER

Figure 27.

A cantilever beam with PVDF material PE energy harvester from an array of beams with the top view (left), and isometric view (right). Reproduced with permission (Cahill et al., 2018b), Copyright 2018 Elsevier.

An auto-parametrically excited PE vibration energy harvester was introduced for bridge energy harvesting (Jia et al., 2015) and the work was extended (Gaglione et al., 2018) in another research. The same harvester was presented and analyzed with bi-stability in an earlier study (Jia and Seshia, 2013) shown in Figure 28. The study (Jia et al., 2018) was reproduced for bridge and other applications. The prototype harvester when driven at 23.5 Hz and RMS acceleration of 1g produced a peak power of 78.9 mW having a 4.5 Hz bandwidth. This multi-axis high power density and wide operational frequency bandwidth harvester produced nearly 1 mW average power output. The harvester was tested on the Forth Road suspension bridge which is 2.5 km long having a main span of 1006 m. The harvester was fixed to various locations on the bridge and the output was measured. In the active frequency range of 7–26 Hz the raw AC power produced was 1050 µW and when the conditioning circuit was used 315 µW power was available.

OPEN IN VIEWER

Figure 28.

The parametrically excited cantilever PE energy harvester: (a) design and (b) the prototype. Reproduced with permission (Jia and Seshia, 2013) Copyright 2013 IEEE.

All the PE harvesters reported for bridge testing were characterized for a frequency range of more than 10 Hz. The long cantilever-type BEH reported by (Peigney and Siegert, 2013) was the first PE harvester tested on a real bridge site. A multi-modal BEHs to capture energy in a wide band of frequencies of the bridge is also reported (Cahill et al., 2018b). Similarly, to capture energy from multi-directional excitation, a multi-axis BEH (Gaglione et al., 2018) is reported but tested at very high acceleration.

3.2. Electromagnetic BEHs tested on real bridge site

The EM vibration energy harvesters are also popular for bridge vibration because of their better operation on the bridge structure. In these BEHs, the energy transduction is simple and is very flexible to optimize the output of the harvester by varying the magnet’s size and magnet flux density, the number of turns of the coils, and the relative gap between the magnet and coil.

A harvester using a linear electromagnetic generator was used (Sazonov et al., 2009) to harvest energy to power a WSN. The device consisted of a coil and magnet suspended from a spring that was attached to a girder which moved due to traffic flow. When the mass attached was 0.09 kg and spring stiffness was kept at 34 N/m, the device had a resonant frequency of 3.1 Hz. The coil was having 10,000 turns of AWG 46 copper wire having a resistance of 67 kΩ and an inductance of 34 H. The maximum power output was 12.5 mW when a 10 mm displacement was given. The device was tested for 7 days on the RT11 bridge located in Postdam New York which was a steel girder bridge. The harvester was used to trickle charge a battery powering WSN to transmit 25–480 measurements per day on a low-traffic bridge.

A parametric frequency increase EM generator was introduced (Galchev et al., 2011) to capture energy from low-acceleration and low-frequency excitation. The harvester shown in Figure 29 consists of a tubular structure having an inertial mass supported by two copper springs. Two frequency up-conversion vibrating mass harvesters were used, one above and the other below the inertial mass. The harvesters had a latching magnet that attached with low-frequency inertial mass when it came near and was detached when it was moving away. This way it imparted energy to the high-frequency (152 Hz) harvester’s spring-mass system. The small latching magnet was attached to a large magnet separated by a spacer. The large magnet was a generating magnet that was moving inside a copper coil to produce output. The device was tested in-lab from an acceleration range of 0.055g to 1g with a 1.5 kΩ resistor for matching impedance. At an excitation acceleration of 0.055g and frequency of 2 Hz, the harvester produced a 2.3 µW average power and 57 µW peak power. The BEH was tested on two bridges; the Groove Street (GS) highway flyover that was having a composite structure with steel girders and a concrete deck in Ypsilanti Michigan USA while the second bridge was an NC suspension bridge in Valejo California. The peak acceleration levels changed between 10 and 35 mg for the GS bridge and 10–130 mg for the NC bridge. The BEH was attached at the bottom of the girder at the NC bridge and produced 0.5–0.75 µW of average power without any modification.

OPEN IN VIEWER

Figure 29.

EM energy harvester with inertial mass and moving magnets: (a) Schematic diagram and (b) Solid model (Galchev et al., 2011). Reprinted under Creative Commons Attribution 4.0 International License, from IOP Publishing.

A previously reported EM-BEH (Galchev et al., 2011) was modified for improvement of output and reliability (McCullagh et al., 2014). One of the changes was decreasing the number of turns on the coil which lowered its output impedance to 300 Ω while another alteration was the use of double magnets to confine and re-route the magnetic flux. The suspension spring material was also replaced to improve reliability and a new assembling technique was adopted. Electromagnetic coupling was improved with better positioning of magnets relative to the coil. These changes improved volumetric efficiency as well for the harvester. For the optimized harvester, an average power output of 12.5 µW was reported at an acceleration of 0.034g and a frequency of 2 Hz. The BEH was tested on New Carquinez Bridge California USA and it produced 1.6–5.02 µW of power output. A power management circuit was also used to collect and regulate the power supplied to sensors. The bridge was kept under observation for data collection for 13 months; the problems in data transmission were also reported during measurements.

An EM-BEH was developed (Kwon et al., 2013) for energy harvesting on the bridge. To enhance the output of the harvester, repulsively stacked magnets were used to increase the oscillation frequency and flux density of the magnetic field. Eight-ring neodymium (NdFeB) magnets were stacked into a steel shaft with seven steel cores separating them. The whole assembly was moved into two copper coils which were enclosed in a casing and the casing was supported by a coil spring as shown in Figure 30. The natural frequency of the device was 3.65 Hz and the optimal load resistance was 680 Ω. Modeling and simulations were performed and the results agreed with experimental measurements. The average power output of 0.156 mW was reported. A bridge (fifth Nongro) in Korea was targeted for device characterization. The response was observed for over 10 min at an acceleration level of 0.133g. The measured vertical frequency of 2.39 Hz while the torsional frequency of 3.17 Hz was reported. When the device was tuned to the first mode vibration of 2.39 Hz, the peak power output of 5.07 mW and the average power of 0.12 mW were obtained from simulations. The field test was conducted on another bridge the third Nongro Bridge in Pusan, Korea with a similar structural system. The measured peak acceleration was 0.1g and the natural frequencies were 2.74 and 4.1 Hz. The device’s natural frequency of 3.65 Hz was close to these frequencies so no tuning was required. The coil-1 produced a peak power of 0.75 mW, however, an average power of 0.059 mW was observed for 10 s when a loaded heavy truck was crossing over a bridge.

OPEN IN VIEWER

Figure 30.

EM energy harvester with repulsively stacked magnets with actual picture (left) and line diagram (right) (Kwon et al., 2013). Reprinted under Creative Commons Attribution 4.0 International License, from IOP Publishing.

An EM-BEH using a fixed planar coil between two rows of movable magnets arranged in a Halbach configuration was reported (Orfei et al., 2016) for bridge application as shown in Figure 31. The coil having dimensions of 10 × 8 × 4 mm was formed of 1300 turns of 0.04 mm diameter copper wire. The Halbach array consisted of 5 magnets on each side totaling 10 magnets having an overall size of 15 × 20 × 4 mm and were placed 7 mm away from each other. Two couples of magnets at the top and bottom kept the moving mass suspended. The resonant frequency of the device was 8 Hz. A voltage regulator LTC3588 was integrated with a harvester for low-loss, full-wave bridge rectification with a buck converter. A bridge over river Tevere near Rome in Italy was chosen for testing the harvester and the LoRa transceiver was powered with it. A 100 mF super capacitor at 3.3 V stored 330 mJ energy in 3.5 h on the bridge for further utilization. The transceiver utilized 124 mJ of energy in 870 ms which was almost half the energy stored in the super capacitor.

OPEN IN VIEWER

Figure 31.

EM energy harvester with a planar coil and two rows of moving magnets mounted on a vibration shaker with an accelerometer (Orfei et al., 2016).

An EM nonlinear harvester for bridge vibration energy (Yang et al., 2019) is shown in Figure 32. The harvester is unique since its resonant frequency can be tuned, works at a low frequency of excitation, and a high-efficiency power management circuit is developed for it. The harvester was deployed at Yeongjong Grand Bridge and Yeondae Bridge South Korea for fatigue crack detection. The harvester developed a 2.27 V peak voltage with an RMS value of 0.21 V at 0.018g acceleration and 3.05 Hz frequency. The harvester produced an average power of 114.7 µW. The device had a volume of 686.9 cm³. The BEH to be attached to a bridge had a linear spring to which the tubular device was attached. The top and bottom of the tube had magnets for repelling the central magnet which moved inside a copper coil to induce voltage. Overall this was a two-degree-of-freedom harvester.

OPEN IN VIEWER

Figure 32.

Nonlinear EM energy harvester with repulsive magnets (Yang et al., 2019).

An EM-BEH tested on the bridges excited by traffic flow was reported (Peigney and Siegert, 2020). The disassembled harvester is shown in Figure 33. The magnets at the tip of the cantilever beam were added to the inertial mass and were arranged as two rows of three magnets on each side thus totaling 12 magnets and enclosing a copper coil in between them. The device was for low-frequency applications and was tunable to a frequency of around 4 Hz. During a field test, the BEH produced 112 µW power at 3.3 kΩ load resistance. Detailed modeling was done with optimization for coil shape and geometry. The device was tested in-lab and then in a real bridge environment as well. At the Roberval highway bridge, the harvester produced 4 V voltage and 112 µW mean power at 3.3 kΩ load resistance. This test lasted for two and a half hours. Testing was also performed on the Vareze bridge which had a half-power bandwidth lesser than the Roberval bridge.

OPEN IN VIEWER

Figure 33.

Disassembled view the EM energy harvester with copper coil in center and assembly of six magnets on each side. Reproduced with permission (Peigney and Siegert, 2020). Copyright 2020 ASCE Library.

A dual resonator type harvester was used (Ahmad and Khan, 2021a) for in-lab characterization and then testing in a real bridge environment. The device utilized two cantilever beams of copper, one inside another for compact configuration. The two beams were coupled using a copper coil on one (outer) beam and a magnet on another (inner) beam. The two beams had two resonant frequencies of 3.2 and 4.2 Hz. The magnetic coupling increased the power output between the two resonant frequencies making it a broadband harvester. The harvester produced 2.7 V voltage and 13 mW of power output at a resonant frequency of 3.2 Hz, acceleration of 0.07g, and load resistance of 555 Ω. The output at the second resonant frequency of 4.2 Hz was 3.1 V voltage and 7.8 mW power at 0.07g acceleration and 1200 Ω optimal load resistance. A Cockcroft Walton multiplier-type rectifier circuit was used to convert the raw AC voltage to DC voltage and produced 42.1 V open-circuit voltage at first resonance and 13.2 V at resonance-2 at 0.05g acceleration and 100 kΩ load resistance. The power density of the device was 43.3 µW/cm³. The reported harvester is shown in Figure 34. The device was tested on two highway bridges in Peshawar, Pakistan. The device was tested on the Bagh-e-Naran bridge and produced 0.27 V voltage and peak power of 0.13 mW at a load resistance of 555 Ω and 0.024g peak acceleration. The device produced 0.32 V voltage and 0.18 mW of power at 0.017g acceleration at the flyover bridge on the ring road.

OPEN IN VIEWER

Figure 34.

The dual resonator type EM energy harvester: (a) design and (b) the prototype. Reproduced with permission (Ahmad and Khan, 2021a). Copyright 2021 ASCE Library.

Most of the EM-BEHs are of bigger sizes and mostly resonant-type devices. Even the frequency up-conversion harvester has to resonate the primary mass to the excitation frequency. Different types of EM-BEHs are tested directly over the bridge structure. The linear generator-type harvester (Sazonov et al., 2009) is developed to extract the energy from the bridge’s original movement. Frequency up-conversion techniques (Galchev et al., 2011) are also utilized to improve the performance of the harvester at the low vibration of the bridge structure. The magnetic flux density was increased using repulsively stacked magnets (Kwon et al., 2013) in the harvester to enhance power generation. Moreover, to increase the power density of the harvester (Orfei et al., 2016) multiple moving magnets stacked together in two rows around the planar coil are used for a strong magnetic field. On a similar approach, sometimes, very simple devices with a cantilever beam architecture (Peigney and Siegert, 2020) are also developed as BEHs. To increase the frequency bandwidth, in some EM-BEHs, both magnet-beam and coil-beam are kept movable (Ahmad and Khan, 2021a) to have a multi-mode operation.