Session 5 Structural Health Monitoring

Thursday, 12.11.2020, 08:45-10:40 o'clock

Introduction

08:45 - Intro: Conference & Day, Walter Lang

Presentations:


Liv Rittmeier: Investigation of in-plane wave motion in a Fibre Metal Laminate using PVDF foil sensors

09:00-09:20 o'clock

(Paper ID: 1129)

Structural Health Monitoring (SHM) via Guided Ultrasonic Waves (GUW) has been the subject of research for years and has already been described in detail for purely isotropic and purely anisotropic materials. Fibre Metal Laminates (FML) combine isotropic metal layers with anisotropic fibre composite layers to a laminate. They are mainly used in the aerospace industry to reduce mass while optimizing stiffness and strength properties. On the other hand, the combination of different materials leads to high differences in the acoustic impedance at the interfaces, which in turn leads to the question to what extent these materials can be monitored via GUW.

This research investigates the applicability of polyvinylidene fluoride (PVDF) foil sensors for the detection of GUW in the interior of an Aluminium Glass Fibre laminate.

Preliminary analytic investigations based on new adaptions of the partial-wave method gave insight into general GUW characteristics along fibre-metal interfaces and within an FML plate waveguide. Analytic and transient finite element simulation results were used to predict displacement fields. Composition and dimension of fabricated FML plates as well as excitation frequency were chosen based on these predictions, particularly maximizing dominance of the in-plane displacement components.

Preliminary experimental investigations confirmed the general sensitivity of reversibly bonded PVDF foil sensors as well as PVDF foil sensors embedded in GRP to GUW.

Further experiments based on self-fabricated Aluminium Glass Fibre laminates with embedded PVDF foil sensors proved their applicability to GUW based SHM. Contacting of commercial PVDF foil sensors with crimped contacts was optimized enabling the embedding of sensor and cabling into the laminate with a high degree of structural conformity. The position as well as the number of PVDF foil sensors were chosen minimizing local accumulations of material in the laminate. A cold-curing epoxy resin was used to prevent the temperature-sensitive sensors from being destroyed or depolarized during production. In a first run, the functionality of the embedded sensors was tested after production and a brief explanation was given of how a full-surface connection to an epoxy matrix inside the laminate influences the sensitivity and dynamics of the sensor and what needs to be taken into account for proper measurement calibration. After the design and application of a PZT actuator on the laminate, GUW were excited with particular attention to the frequencies in which a particularly dominating in-plane displacement field was assumed. Laser-Doppler Vibrometry (LDV) was used for reference measurements.

Paper PDF

Daniel Gräbner: Condition Monitoring of O-Ring Seals with Integrated Strain Gauges and Finite Element Analysis Assisted Signal Evaluation

09:20-09:40 o'clock

(Paper ID: 1191)

Smart factories are supposed to provide an overview on the current status of installed production machines. In order to collect data, intelligent machine parts have to be designed. Two major and widely underestimated issues in the use of those parts is the interpretation of the generated data as well as the deduction of consequences from the interpretation. The main reason for this is likely the requirement of very advanced and comprehensive knowledge of the intelligent machine part as well as the whole system that the part is installed in. In this work, we demonstrate a method that allows the interpretation of signals acquired during monitoring of O-ring seals with integrated strain gauges and the determination of the remaining sealing capability during its lifetime.

Sealing and sensing are two well-established technologies. Recently, we have demonstrated the possibility to monitor the decay of O-ring seals by combining those two technologies. Strain gauges have been integrated into rubber seals during compression molding and are capable of generating a signal that corresponds with the decay of the seals over its lifetime. However, the interpretation of the signal has not been addressed yet. In this work, we present a Finite Element Analysis-based interpretation of the signal that enables us to correlate the measurement signal with the remaining sealing potential of the gasket.

Degradation in elastomer materials is a very complex process. The two major molecular degradation mechanisms are polymer chain scission and recombination. Both mechanisms are heavily influenced by environmental conditions, e.g. temperature, presence of corrosive media, and by the shape and dimensions of the part. Extensive material models, which take into account all those parameters, are not available. However, the macroscopically observable result of degradation can be modelled as stress relaxation. In our work, the degradation of a seal is simulated using a basic stress relaxation model. The effect of degradation on the signal of an integrated strain gauge is calculated from simulation and is compared to experimental data. Then the stress relaxation parameters of the model are adjusted to make simulation and experimental data match with reduced error. This way, global stress relaxation parameters of the elastomer material are determined in the process and allow a close monitoring of the remaining elastic restoring potential of the rubber material. Using those parameters, the simulation can predict the remaining sealing capability of the seal.

Paper PDF

Osman Altun: Identification of dynamic loads on structural component with artificial neural networks

09:40-10:00 o'clock

(Paper ID: 1130)

Enhancing structural components by implementing sensors offers great potential regarding condition monitoring for lifetime analysis, predictive maintenance and automatic adaptation to environmental conditions. This article describes an approach to determining the operational forces applied to the front suspension arm of a car using strain gauges. Since suspension arms are components with free-form surfaces, an analytical calculation of applied forces by means of measured strains is not feasible. Hence, artificial neural networks are applied to approximate the functional relationship. The results reveal how artificial neural networks can be applied to identify load conditions on structural components and, therefore, deliver essential data for condition monitoring.

Paper PDF

Philip Johannes Steinbild: Strain-based monitoring system for ski poles with low impact on their total mass and inertia

10:00-10:20 o'clock

(Paper ID: 1152)

This paper focuses on the development of a fully integrated monitoring system for measuring mechanical loads and motions experienced by a cross-country ski pole. A specially developed miniaturised electronic system allows the integration of the monitoring system into the grip of the ski pole while increasing the mass of the pole only slightly. The digitalized measured data is send wirelessly to a smartphone and can be made available for coaches and scientists in real time. This technology can also be used in other applications involving strain measurements where the need for low mass and miniaturised electronics is given.

Paper PDF

Markus Kreutz: Investigation of icing causes on wind turbine rotor blades using machine learning models, minimalistic input data and a full-factorial design

10:20-10:40 o'clock

(Paper ID: 1111)

Ice formation on rotor blades of wind turbines cause significant downtimes in cold regions. Existing ice sensors can detect, but cannot predict ice. Therefore, our research aims to develop an ice prediction system based on historical and forecasted data. In a first step, the detection process is analyzed in this research paper. Two machine learning models are trained using minimalistic input parameters. A full-factorial experiment design is performed for the models, using the f1-score as the response variable. The most significant input parameter was the external temperature for both models.

Paper PDF

Discussion and Comments