Research and Development

Our specialized research and development team is committed to meeting the specific needs of our clients, leveraging nearly thirty years of accumulated expertise.

Design of a Robot for Inspection of Steel Tank Walls Using an Electromagnetic Acoustic Transducer (EMAT)

The goal of this project is to develop a scanner for inspecting and measuring the thickness of the walls of aboveground fuel storage tanks. It consists of an electromagnetic acoustic transducer (EMAT) mounted on a wheeled wall-climbing robot. The transducer itself contains a permanent magnet and a spiral coil interfaced with a portable pulser and a signal processing unit. It can generate and receive radially polarized ultrasonic waves that travel through the thickness of the steel wall. By measuring the time of flight, the wall thickness can be calculated. Thanks to its ability to move along weld seams, the robot can quickly scan a large portion of the tank wall.

The internal PowerBox H (PBH) specification is used to pulse the EMAT coil and measure the received signal. The pulse voltage is 1200 V peak-to-peak, and the frequency is 3000 kHz. The received signal undergoes a first stage of signal processing by the PBH. The receiver data is then sent in real-time via Ethernet to a laptop running a Python program, where the signal undergoes a second stage of signal processing and is displayed on the user interface. The Python program controls the robot and measures wall thickness in real-time using the time of flight as previously described. The signal peaks used to calculate the time of flight can either be manually defined by the user or automatically determined by an integrated peak-search algorithm in the software.

The robotic system uses differential drive and features four wheels driven by DC motors. The motor speed is controlled by pulse-width modulation (PWM) using H-bridge circuits, which are in turn controlled by an Arduino UNO microcontroller. The Arduino UNO also acquires distance information from an optical encoder. A 12V 100Ah battery is used to supply power. To adhere to and scan the vertical steel wall, the robot is equipped with magnets fixed to its chassis. The robot also has a camera installed for visual inspection.

Inspection of Steel Pipes Using Electromagnetic Acoustic Transducers (EMATs)

This project explores the use of electromagnetic acoustic transducers (EMATs) for the inspection of steel pipes. EMATs allow for rapid evaluation of the wall thickness of steel pipes prone to corrosion. This is achieved using horizontally guided shear ultrasonic waves that are sensitive to the pipe wall thickness. EMATs based on both the Lorentz force and magnetostriction mechanisms are being studied. The goal is to develop a pipeline wall scanner capable of detecting progressive thinning and pitting due to corrosion.

A prototype of a Lorentz force-based EMAT scanner has been built. It includes a Lorentz force EMAT transmitter and receiver, both interfaced with the Innerspec PowerBox H, which serves as both the transmitter/receiver and signal processing unit. The transducers consist of a periodic array of permanent magnets above a coil track and use the Lorentz force mechanism to generate and receive horizontally guided shear waves in steel. They are connected by a hinge with two degrees of freedom designed to accommodate different pipe diameters. Each transducer is equipped with a set of wheels for scanning a pipe axially. The transmitter generates horizontally guided shear acoustic waves in the pipe wall that propagate circumferentially.

The wavelength of the ultrasonic waves equals the period of the transducer’s magnetic array and determines the appropriate thickness range of steel that can be inspected, as well as the operating frequency. The frequency equals the excitation voltage applied to the transmitting coil. It is configured for the EMAT to operate using the SH1 dispersive wave mode, which is sensitive to thickness changes caused by defects. When the EMATs operate using the SH1 dispersive wave mode, a reduction in the steel thickness will cause a decrease in signal amplitude due to wave reflection or the conversion of the wave mode to SH0. By analyzing the signal characteristics at the receiver using a signal processing program developed in Python, the presence of a defect between the two transducers can be evaluated.

A magnetostrictive EMAT scanner prototype has also been developed. The transducers consist of a permanent magnet housing with a meander coil between its poles and use the magnetostriction mechanism to generate and receive horizontally guided shear waves in steel. As with Lorentz-based EMATs, the frequency is determined by the excitation voltage applied to the coil. However, the wavelength setting for operation differs. For magnetostrictive EMATs, it is equal to the period of the meander coil. The hinge mechanism and wheel set are similar to those used in the other prototype. The scanning direction is axial, and the wave propagation direction is circumferential. Therefore, the only difference between the two prototypes lies in the mechanisms of emission and reception.

Since the optimal transduction mechanism for the transmitter and receiver may differ, the hinges of the prototypes were adapted to be compatible. This allows the transmitter or receiver to be switched between the two prototypes to create a hybrid system of different types of EMATs. For different pipe or steel plate thicknesses, the operating point of the EMAT is selected relative to the SH1 cutoff point for that thickness. The ability to modify the operating wavelength was integrated into the EMAT design. Five wavelengths between 0.33 inches and 1 inch were intended to be used. This range was chosen to correspond to the operating region between the SH1 and SH2 cutoff frequencies for steel plates ranging from 3 mm to 11 mm in thickness.

Measurement of Steel Thickness through Insulation Using Pulsed Eddy Currents (PEC)

Pulsed eddy current (PEC) testing is used to measure the thickness of a steel specimen even in the presence of an insulating layer between the measuring probe and the steel surface. Thickness measurement is performed by inducing eddy currents in the steel specimen by pulsing a coil and monitoring the rate of decay of the eddy currents. The principle behind PEC thickness measurement is that the rate of decay of the eddy currents is inversely proportional to the square of the steel thickness. The ability to take measurements without removing insulation layers is a major advantage of this technology. The primary application of this project is the detection of wall thinning in pipelines due to corrosion.

A prototype and software interface had already been developed. The software allows for system calibration and control over pulse duration and the number of samples to average. Another feature is the ability to adjust the range in which the PEC current decay slope calculation is performed by the software. This allows the user to control the signal window based on background noise. The prototype can be used to measure the thickness of steel structures even in the presence of insulation and aluminum sheaths. The objectives of our work are to study the effects of insulation delamination and to determine wall thinning or detectable defects in terms of volume loss and ground surface area.

Development of an AC Magnetic Flux Leakage (MFL) System for the Inspection of Steel Tank Floors

AC Magnetic Flux Leakage (MFL) utilizes the skin effect to better estimate the remaining wall thickness of aboveground fuel storage tank floors. The steel plate is magnetized at different frequencies, and the resulting MFL signals, which depend on the skin depth corresponding to each frequency, can provide information about the depth of detected defects. Conventional MFL systems that use permanent magnets to produce a constant magnetic field are only used for defect detection and are not reliable for quantitatively assessing defect depth. The goal is to develop an AC MFL system with the added functionality of being able to estimate defect depth.

An electromagnet was designed and constructed to develop and test potential algorithms based on AC MFL. It is capable of producing magnetic fields with arbitrary waveforms at frequencies up to 10 Hz. For this project, the main interest lies in sinusoidal magnetic fields. The idea is to detect defects using the electromagnet in DC mode and then switch to sinusoidal magnetic fields for depth estimation. The present algorithm involves monitoring the drop in the MFL signal amplitude as the operating frequency increases. The hypothesis is that the rate at which the amplitude decreases with increasing frequency primarily depends on the defect depth. The magnetic field waveform is produced by the current in the electromagnet’s coils induced by the voltage applied via an H-bridge circuit. The applied voltage is controlled by pulse-width modulation using an Arduino Mega 2560 microcontroller.

In addition to the electromagnet, a sensor array and Python program were also developed to complete the MFL system. The sensor array consists of 24 Hall-effect sensors to detect and measure the MFL signal. The sensor output is read using an analog-to-digital converter in a Teensy 3.5 microcontroller. The Python program is used to acquire and filter the data. It also logs the data in a CSV file and plots it in 2D or 3D graphs. There is also a software filter with an adjustable cutoff frequency to reject signal noise caused by variations in scanning speed.

Use of 3D Magnetic Flux Leakage (MFL) for the Inspection of Steel Tank Floors

Magnetic Flux Leakage (MFL) is a commonly used non-destructive testing technique for inspecting the floors of fuel storage tanks. A portion of the steel plate is magnetically saturated, and in the presence of a defect, the magnetic field will leak in the direction normal to the surface. This magnetic field signal can be captured by a network of Hall-effect sensors. This project studies the measurement of MFL signals in three directions. Currently, MFL measured only in the direction normal to the inspected sample is used as a screening tool for defect detection. Additional information could potentially aid in the quantitative evaluation of detected defects. The goal is to develop an algorithm that combines data collected along the three axes to estimate the shape and size of the defect in all dimensions.

To meet the project’s needs, a 3D magnetic sensor data acquisition system was developed. The system had to be capable of detecting defects, saving detection results in 3D based on position, allowing the quantification of defect width and length, and providing the necessary elements to analyze the conditions under which an estimation of defect depth would be made. The first step of the process was to acquire and store the information. Since the required information was the amplitude and variation of the magnetic field along three orthogonal axes, it was necessary to design and build an appropriate system from both a hardware and software perspective.

To analyze the magnetic flux leakage, a fast, reliable data acquisition (DAQ) system with sufficiently high resolution was needed. The DAQ system had to include 3D magnetic sensors, a sensor interface software, firmware for the implementation of the multi-sensor architecture with a complex programmable logic device (CPLD), microcontroller operating software, and software for user interface, data display, and analysis. The hardware and software components of the DAQ system had to be designed and built to meet the general specifications mentioned earlier. The solutions chosen for each element of the DAQ process had to be compatible with the existing 16-sensor configuration, and with the extended configuration of 32 or 64 sensors if higher resolution needed to be implemented in the future. To ensure fast and flexible communication, an Ethernet port had to be implemented on the microcontroller. The system also had to be capable of displaying data from 16 sensors, with 3 axes per sensor, in real time. This totals 48 data acquisition channels with a rate of 100 samples per second per channel. The DAQ data had to be saved in real time as well.

The prototype development involved the design, fabrication, testing, and implementation of several components. This included the 16 3D magnetic field sensors, CPLD software, the microcontroller board with Ethernet and serial communication, microcontroller software, a display panel for 48 channels at 100 samples per second in real time with data acquisition into a CSV file. The developed prototype is based on the MLX90393 tri-axis Hall-effect digital sensors. The sensor range is 500 Gauss. A TEENSY 4.1 microcontroller was used for data acquisition. A LabVIEW program was also developed as the user interface for the system. The software can display separate plots of magnetic flux density on the three axes in real time for all sensors. The software is also used for signal processing and data saving. Signal processing primarily involves filtering that reduces the effect of the residual magnetization of the plates on the signal.

Detection of Fatigue Cracks Based on Nanomaterials

Fatigue is a widespread and serious concern in mechanical structures, such as airplanes, steel bridges, and gas tanks. The initiation and growth of fatigue cracks in components in service pose a major issue for human safety. In Canada, fatigue problems may be more significant and exacerbated due to the synergistic effect of mechanical loads, large temperature fluctuations, and chemical corrosion. On-site inspection using non-destructive testing (NDT) techniques involves significant human involvement and lacks continuous monitoring of the structure’s condition. In addition to intermittent deployment of NDT, it would be very useful to develop a sensitive fatigue crack sensor that allows continuous detection of fatigue cracks at an early stage across multiple sites simultaneously.

The project will focus on developing nanomaterial-based fatigue sensors that enable early detection and long-term monitoring of fatigue cracks. Several low-dimensional nanomaterials and their inclusion methods will be explored. The role of the nanocharge-material interfaces will be deciphered. The knowledge developed through these studies will be applied within the project to design high-sensitivity, high-reliability fatigue sensors.

A successful conclusion of this project will enable us to translate the results and designs created in the research project into fully operational nanomaterial-based fatigue sensors, with high sensitivity and reliability. This will allow us to be more competitive in the international structural health monitoring market. Additional benefits include training engineering students in the use of fatigue testing equipment and software.

As a completely new project, recent work has involved conducting a literature review on the subject in collaboration with the University of Toronto. This began by identifying currently available technologies for fatigue detection and strain detection in general. The following categories were found:

• Electrochemical sensor
• Piezoelectric ultrasonic sensor
• Antenna sensor
• Piezoelectric strain sensor
• Eddy current sensor
• Capacitive strain sensor
• Optical fiber ultrasonic sensor
• Piezoresistive sensor
• Acoustic emission sensor

Given the goal of a primarily passive structural health monitoring approach and the expertise of our collaborators in nanomaterials, the types of sensors of interest are piezoelectric strain sensors, capacitive strain sensors, and piezoresistive sensors. All three can monitor fatigue passively while requiring energy only during the measurement acquisition. Additionally, the piezoelectricity, piezoresistivity, or permittivity of the nanomaterials can be exploited. The use of the piezoresistive effect in carbon nanotubes (CNTs) for fatigue monitoring has already been experimentally demonstrated in the laboratory.