Structural health monitoring (SHM) is being widely evaluated by the aerospace industry as a method to improve the safety and reliability of aircraft structures and also reduce operational cost. Built-in sensor networks on an aircraft structure can provide crucial information regarding the condition, damage state and/or service environment of the structure. Among the various types of transducers used for SHM, piezoelectric materials are widely used because they can be employed as either actuators or sensors due to their piezoelectric effect and vice versa. This paper provides a brief overview of piezoelectric transducer-based SHM system technology developed for aircraft applications in the past two decades. The requirements for practical implementation and use of structural health monitoring systems in aircraft application are then introduced. State-of-the-art techniques for solving some practical issues, such as sensor network integration, scalability to large structures, reliability and effect of environmental conditions, robust damage detection and quantification are discussed. Development trend of SHM technology is also discussed.

Composite materials have been widely used in many industries due to their excellent mechanical properties. It is difficult to analyze the integrity and durability of composite structures because of their own characteristics and the complexity of load and environments. Structural health monitoring (SHM) based on built-in sensor networks has been widely evaluated as a method to improve the safety and reliability of composite structures and reduce the operational cost. With the rapid development of machine learning, a large number of machine learning algorithms have been applied in many disciplines, and also are being applied in the field of SHM to avoid the limitations resulting from the need of physical models. In this paper, the damage monitoring technologies often used for composite structures are briefly outlined, and the applications of machine learning in damage monitoring of composite structures are concisely reviewed. Then, challenges and solutions for quantitative damage monitoring of composite structures based on machine learning are discussed, focusing on the complete acquisition of monitoring data, deep analysis of the correlation between sensor signal eigenvalues and composite structure states, and quantitative intelligent identification of composite delamination damage. Finally, the development trend of machine learning-based SHM for composite structures is discussed.

Lamb wave-based damage detection is one of the most promising structural health monitoring (SHM) technologies for aircraft structures. In this paper, a Lamb wave-based deep transfer learning network is developed for multi-level damage classification of plate-type structures. A one-dimensional convolutional neural network (1D-CNN) is employed to deeply mine the damage characteristics of complex Lamb wave signals with multiple modes and multiple boundary reflections. The concept of multi-level damage classification is carried out to get different results for different engineers, and a multi-task cascaded 1D-CNN architecture is established for three levels of damage classifications, which is corresponding to different SHM levels, i.e. identifying the damage presence, location, and severity, respectively. In the multi-task cascaded architecture, a fine-tune transfer learning concept is adopted to share partial structures and weight values among different classification models, which could greatly improve the efficiency of the model calculation. In the multi-level damage classification model, the one-dimensional Lamb wave scattering signals with different damage locations and sizes are used as the input without any preprocessing steps, while the classifications of the damage presence, location and size are designated as output of different levels. An experiment has been conducted to verify the proposed multi-level damage classification model. The experimental results show that the training time of this model is reduced by 35% and the accuracy of the proposed model is greater than 99%, which verifies the effectiveness and reliability of the proposed technique.

A flexible microfluidic super-capacitive pressure sensor is developed to measure the surface pressure of a complex structure. The innovative sensor contains a filter paper filled with ionic liquid, and coated with two indium tin oxide polyethylene terephthalate (ITO-PET) films on the top and bottom, respectively. When external pressure is applied on the top ITO-PET film of the sensor mounted on the surface of an aircraft, the capacitance between the two ITO-PET films will change because of the deformation of the top ITO-PET film. The external pressure will be determined based on the change of the capacitance. Compared to the traditional pressure sensor, the developed sensor provides a high sensitivity of up to 178.5 nF/KPa and rapid dynamic responses for pressure measurement. Meanwhile, experiments are also conducted to study the influence of the thickness of the sensing film, sensing area, temperature, and humidity.

Structural health monitoring techniques based on vibration parameters have been used to assess the internal delamination damage of fiber-reinforced polymer composites. Recently, machine learning algorithms have been adopted to solve the inverse problem of predicting delamination parameters of the delamination from natural frequency shifts. In this article, a delamination detection methodology is proposed based on the changes in multiple modes of frequencies to assess the interface, location, and size of delamination in fiber-reinforced polymer composites. Three types of machine learning algorithms including back propagation neural network, extreme learning machine, and support vector machine algorithm were adopted as inverse algorithms for assessment of the delamination parameters, with a special focus on the interface prediction. A theoretical model of fiber-reinforced polymer beam with delamination under vibration was constructed to learn how the frequencies are affected by the delaminations (“forward problem”) and to generate a database of “frequency shifts versus delamination parameters” to be used in machine learning algorithms for delamination prediction (“inverse problem”). Multiple carbon/epoxy fiber-reinforced polymer beam specimens were manufactured and measured by a laser scanning Doppler vibrometer to extract the modal frequencies. Numerical and experimental verification results have shown that support vector machine has the best prediction performance among the three machine learning algorithms, with high prediction accuracy and only requiring a small number of samples. For predicting the interface of delamination which is a discrete variable, support vector machine classification has observed better prediction accuracy and requiring less running time than regression. This study is one of the first to prove the applicability of support vector machine for structural health monitoring of delamination damage in fiber-reinforced polymer composites and has the potential to improve the prediction capability of machine learning algorithms. Another significant outcome of the study is that the interface of delamination has been predicted accurately with support vector machine.