Advances in Asset Management and Condition Monitoring

Inhaltsverzeichnis

1. Frontmatter

2. Chapter 1. Comparison of Amplitude to Real and Imaginary Features of the poly-Coherent Composite Bispectrum (pCCB) Components in Machine Diagnosis

   Kenisuomo C. Luwei, Jyoti K. Sinha, Akilu Yunusa-Kaltungo

   Abstract

   Earlier studies have successfully demonstrated the use of the poly-coherent composite bispectrum (pCCB) in the faults identification in rotating machines. However, only amplitudes of the pCCB components were used in the earlier studies. Since the pCCB components are complex numbers (both amplitudes and phases). Hence, the real and imaginary features of the pCCB components are also explored in the fault identification of rotating machines in the current study. The observations from the present study through the experimental rig are presented and compared with the earlier observations using the amplitudes of the pCCB components. It shows that the real and imaginary of the pCCB components shows improvements in fault identification and classification along with a good representation of machine behavior, compared with the magnitude only of pCCB components.

3. Chapter 2. Asset Management Simulation and Optimisation of Railway Bridges

   Dwayne Nielsen

   Abstract

   Bridges deteriorate with age and use, and eventually become unsafe without adequate maintenance. Infrastructure managers are being challenged to find economical solutions to address the issues caused by increased asset degradation, while simultaneously reducing maintenance costs on ageing infrastructure.

   The asset management approach adopted in this study suggests and allocates strategic level maintenance budgets to individual asset components based on asset importance, deterioration and cost of various maintenance options. The aim of the approach is to identify the most efficient combination of future maintenance actions for each bridge in the network based on a limited network budget. Asset management simulations are performed to identify the effects of changing current and future maintenance budgets, application of different maintenance actions and their impact on future maintenance requirements.

   Maintenance plans are created using a two-phase approach. The first phase applies an optimisation algorithm that combines dynamic programming with Monte Carlo Simulations to consider uncertain input values. The second phase uses the binary integer programming method to allocate detailed maintenance plans within budget constraints.

   A sample set of 17 bridges from an Australian railway network were analysed utilising this approach. The results suggest a maximum and minimum range for strategic level maintenance budgets for each year in a 100 year plan and a detailed maintenance plan for the next 5 years.

4. Chapter 3. Study on the Ultrasonic Attenuation Characteristic Due to Crack in a Two-Dimensional Isotropic Plate

   Xiaojun Zhou, Huifang Xiao

   Abstract

   In this paper, scattering and attenuation of ultrasonic waves by embedded horizontal crack of different depth is studied using a two-dimensional isotropic plate. The finite element analysis (FEA) with absorbing boundary condition is employed to simulate the experimental pulse-echo mode to obtain ultrasonic data. The presence of a dead zone directly under the surface in the pulse-echo test is recovered. Effect of crack depth on the scattering of elastic waves and the resulted attenuation behavior is characterized by the defined energy attenuation coefficient. The variation of energy attenuation coefficient with crack depth is approximated by a second order polynomial function. The maximum attenuation occurs when the crack is located at central plate in the thickness direction and the attenuation decreases as the crack approaches surface.

5. Chapter 4. Mimosa Strong Medicine for Maintenance

   Riku Salokangas, Erkki Jantunen, Martin Larrañaga, Petri Kaarmila

   Abstract

   The paper describes how the use of Mimosa open source data model supports the development of a low-cost condition monitoring system that is capable to carry out automatic diagnosis and prognosis. Mimosa follows the ISO 13374 definitions (condition monitoring) and links well with the ISO 17359 (diagnosis) and ISO 13381 (prognosis). The Mimosa data model defines all the necessary ontology for the automatic system. As a use case the paper describes the installation of the Mimosa data model in a Raspberry where MariaDB is used as the database engine. A low-cost accelerometer has been installed to a Raspberry thus enabling the collection of vibration data from rolling element bearings of a conveyor. In addition, a low-cost system that uses Arduino is presented for data collection in future use cases. The necessary signal analysis functions are programmed with Python which offers a wide collection of useful functions. The paper summarises the key role of Mimosa in building and using this kind of automatic monitoring systems.

6. Chapter 5. Fault Diagnosis of Motor Broken Bar Using Current and Vibration Fusion Signal

   Xiaoyun Gong, Yongjie Jing, Wenliao Du, Hongchao Wang, Baowei Zhao

   Abstract

   Motor broken bar is a common fault for the asynchronous motor. An intelligent diagnosis method for motor broken bar fault is presented. The intelligent diagnosis method combining Ensemble Empirical Mode Decomposition (EEMD) and Support Vector Machine (SVM) is used to identify the fault type of the motor broken bar. EEMD is used to extract the frequency character from motor vibration signal and current signal. By comparing the stability of the frequency band energy characteristics between the vibration signal and the current signal under different working conditions, it is concluded that the results of vibration signal is better than the current signal. At the same time, the quad-classifier kernel function parameters are optimized using the grid selection method. A multi-information fusion method based on current and vibration signal is designed. It is effective to identify broken bar fault from motor multi-faults. It can resolve the difficulty of multi-component fault feature extraction from multi-faults with broken bar fault.

7. Chapter 6. Mixed Kernel Functions for Multivariate Statistical Monitoring of Nonlinear Processes

   Karl Ezra S. Pilario, Mahmood Shafiee

   Abstract

   Machine learning techniques have now become pervasive in the field of process condition monitoring. In particular, kernel methods are those that use kernel functions to allow for the efficient nonlinear analysis of process data by projecting them onto high-dimensional spaces. A widely used kernel machine in multivariate process monitoring is kernel principal components analysis (KPCA). Many choices of kernel functions were used in previous KPCA studies. However, the use of single kernels alone was recently shown to give only limited expressive ability. In this work, we explored the impact of combining various kernel functions to the performance of KPCA for condition monitoring. Fault detection performance is defined by percent correct detection of faulty states and non-detection of normal states. Optimal kernel parameters were obtained using the genetic algorithm (GA). Visualizations of the boundary between normal and faulty states are provided for demonstration in a chemical process case study. This work can inform the development of mixed kernels for nonlinear process monitoring, not only in KPCA, but also in other kernel machines.

8. Chapter 7. Compound Faults Separation Based on Intrinsic Characteristic-Scale Decomposition and Sparse Component Analysis

   Yansong Hao, Huaqing Wang, Liuyang Song, Lingli Cui

   Abstract

   The frequent occurrence of rotating machinery faults seriously affects the operation of the equipment and the production of the enterprise. And it is worth mentioning that there are few single faults in the rotating equipment failure while the multiple faults are the norm, which undoubtedly increases the difficulty for fault diagnosis. Therefore, it is of great essentiality to address compound faults of rotating machinery. A novel compound faults separation method based on intrinsic characteristic-scale decomposition (ICD) was suggested to detect multi-faults in the case of underdetermined blind source separation (UBSS) when traditional diagnosis techniques fail. To achieve UBSS, ICD is utilized to decompose a single observation signal into multiple product components (PCs). Then, sparse representation is used to improve the signal sparsity, guaranteeing the normal operation of the sparse component analysis (SCA) algorithm. In addition, because the compound fault features cannot only be extracted by ICD, the spares-promoted PCs are arranged in the SCA processing to separate the multiple signal. Simulations and experiments based on the proposed method was successfully verified. Meanwhile, the Empirical mode decomposition (EMD)-based independent component analysis (ICA) is utilized as a contrast to verify the effectiveness of suggested method. The results suggest that the proposed method can deal with the multiple signal separation of roller bearing.

9. Chapter 8. Detection Method of Contact-Type Failure Based on Nonlinear Wave Modulation Utilizing Ultrasonic Vibration Driven by Self-excitation

   Takashi Tanaka, Yasunori Oura, Syuya Maeda

   Abstract

   This paper introduces the concept of new detection method of contact-type failure based on nonlinear wave modulation utilizing ultrasonic vibration driven by self-excitation. It is difficult to detect the contact-type failure in standard inspection utilizing linear ultrasonic vibration. The constructional element of infrastructure is excited by environmental disturbance or forced excitation. In this situation, the contact condition of the failure fluctuates in synchronization with vibration. As this result, local stiffness in the vicinity of failure location fluctuates at the vibration frequency. Thereby, the amplitude and phase of ultrasonic vibration are modulated caused by local stiffness fluctuation (nonlinear wave modulation). This phenomenon can be expressed by a linear time-varying system caused by fluctuation of the natural frequency. In this paper, the new detection method of contact-type failure utilizing self-excited ultrasonic vibration is introduced. Firstly, the concept of detection method of contact-type damage based on nonlinear wave modulation is proposed. It is indicated from time history response analysis using single degree-of-freedom model of nonlinear wave modulation that the fluctuation amplitude of the amplitude and phase of ultrasonic as failure index varies depending on viscous damping. Secondly, the self-excited method utilizing local feedback control and characteristics are introduced. This method realizes oscillator which start to oscillate at the natural frequency automatically. Thus, the frequency of oscillation signal controlled by local feedback control is fluctuated in synchronization with fluctuation of natural frequency when nonlinear wave modulation occurred. Lastly, it is proved that the fluctuation amplitude of frequency of oscillation signal is the failure index independent of viscous damping.

10. Chapter 9. Online Condition Monitoring of Engines by a Deep Analysis of the Electrical Conductivity and Relative Permittivity Changes of the Lubricant

    Manfred R. Mauntz, Jörn Peuser

    Abstract

    The requirements of renewable energy for large industrial gearboxes as installed in wind turbines on and off shore rise. The same applies for efficient gas and diesel engines. A larger flexibility is required of these devices such as maximum operational reliability and a long lifetime. Thus, the requirements for oil and oil condition monitoring grow correspondingly. This presentation provides information about a novel online oil condition monitoring system that gives a solution to the mentioned priorities in the energy sector. The different mechanisms of oil parameter variation in gearboxes and engines are addressed; the data interpretation has to be redefined to the dominating effect. From this, the very sensitive measurement of conductivity κ, relative permittivity ε_r and temperature, T, enables the detection of small changes in the conductivity and dielectric constant of the corresponding oil composition. Therefore, the sensor system effectively controls the proper operation conditions of engines and gearboxes. 24/7 monitoring of the asset during operation enables specific preventive and condition-based maintenance, which is independent of rigid inspection intervals.

11. Chapter 10. Are We Ready for Industry 4.0?

    Abdu Shaalan, David Baglee, Michael Knowles

    Abstract

    A significant number of manufacturing organisations are showing interest in Industry 4.0 due to the support it can provide for processing and visualising manufacturing data in real-time. Industry 4.0 techniques can be used to provide an assessment of machine condition by detecting and processing internal and external data of critical machine components. Currently, a few Small and Medium Enterprises (SME’s) still use ageing and non-computer numerical control, manufacturing assets are operated and maintained without the use of digital technologies to monitor and report operating problems before they occur. Which in return, creates a significant barrier to the implementation of Industry 4.0 applications. In order to facilitate the implementation of Industry 4.0, on ageing, manual manufacturing assets, certain technologies associated with the third industrial revolution, including electronics and information technology, should be examined.

    This paper presents the implementation process of an automation system for monitoring and control of a hydraulic press by firstly examining the required electronics and information systems for processing data and secondly by defining the needed tools and techniques associated with Industry 4.0 applications and the related implementation barriers.

12. Chapter 11. Operational Modal Analysis in the Presence of Pulse Train and Harmonics Based on SSI

    Fulong Liu, Jiongqi Wang, Miaoshuo Li, Fengshou Gu, Andrew D. Ball

    Abstract

    Operational Modal Analysis (OMA) is a popular and effective method to identify the dynamic characteristics of a structure for Condition Monitoring (CM). It is well known that most of OMA methods are under the assumption that the excitation loads are stationary white noise. In practice however, this is not true, the excitation with pulse train and harmonic loads are common for mechanical systems with rotation parts, such as wind turbine and vehicle tested on the roller rig. In order to investigate the effects of pulse train and harmonic loads on the OMA, a quarter vehicle model was developed to simulate a Y25 bogie tested on the roller rig. Moreover, Correlation signal Subset based Stochastic Subspace Identification (CoS-SSI) was employed as the OMA technique in this study. The simulation results indicated that pulse train excitation has no effects on the OMA, whereas harmonic loads have significant effects. On the one hand, harmonic loads will result in false modes, on the other hand, the harmonic frequency will overwhelm the true modes of tested systems when the harmonic frequency is close to system resonance frequency. Therefore, cepstrum editing process was introduced in detail, and employed to filter out the harmonic effects before the OMA process. It has been proved that cepstrum editing is an easy but powerful approach to address the challenge of OMA in the presence of harmonics.

13. Chapter 12. The Development of a Maintenance Gap Analysis Tool for Use Within the Automotive Supply Chain: A Case Study Perspective

    Derek Dixon, Kenneth Robson, David Baglee

    Abstract

    Automotive manufacture contributed £82 billion to the UK Economy in 2017. In addition, the production of vehicles within the UK continues to rise, with 1.65 million vehicles produced in 2017. Lean production methods, compounded by synchronous delivery to an Original Equipment Manufacturer (OEM), ensure membership of the automotive supply chain is challenging. To meet this challenge, participants in manufacturing operations within any business must operate both effectively and efficiently. Interestingly, despite the apparent success of the industry, research has revealed that disjointed maintenance practice within the supply chain is evident and augmenting a difficult production environment. This research gathered empirical data from four case study partners who operate at Tier 1 within the automotive supply chain. The findings demonstrate the majority of research participants operate with an underperforming maintenance department due to a number of barriers and constraints. A worrying consequence of poor maintenance execution and an unsupportive maintenance culture has also emerged. To mitigate the risk of poor maintenance performance, manufacturers are retaining excessive safety stock to ensure delivery targets are met. As well as establishing constraints preventing maintenance performance, areas of best practice have been highlighted, with both characteristics integrated into a Gap Analysis Tool. This paper will discuss the development and subsequent testing of the Gap Analysis Tool with one case study participant and the potential impact for a manufacturer within the automotive supply chain.

14. Chapter 13. Feature Subset Selection Using Sparse Principal Component Analysis and Multiclass Fault Classification Using Selected Features

    Biswajit Sahoo, A. R. Mohanty

    Abstract

    Feature selection is one of the most important aspects of condition monitoring. The condition or health of a machinery is manifested in the form of changes to its signal features. Traditional approaches for feature selection involve collecting an exhaustive list of features from different domains, all of which may not be significant. The objective then is to find a subset of features that reveal the changes in a significant way. The usual approach to select important features is by applying thresholding to loading scores of variables obtained from principal component analysis (PCA). However, there is not enough study as to the effectiveness of selected features in classifying multiple faults. In this paper, we show that by applying sparse principal component analysis (SPCA) to wavelet packet features, we obtain a subset of features that are interpretable and classification accuracy obtained in multiclass classification by using only those features is encouraging. This classification accuracy is compared to that of features obtained from conventional PCA by taking top four features with maximum absolute loading scores. It is observed that, in the particular application concerned, SPCA based features perform at least as good as PCA based features. The method has been applied to a real-world bearing data set that is widely used in condition monitoring.

15. Chapter 14. Vibration-Based Detection of Wheel Flat on a High-Speed Train

    Ruichen Wang, David Crosbee, Adam Beven, Zhiwei Wang, Dong Zhen

    Abstract

    Wheel flat is not only commonly unavoidable surface damage in railway wheels, it can result in possible damage and deterioration incurring high risk of running safety and high maintenance costs. Wheel flat is therefore necessary to be detected at an early stage to minimise safety hazard and maintenance work. This study explores the capacity of the vibration-based detection for high-speed train wheel flatness. A more realistic vehicle-track coupling dynamic model (a dynamic model of vehicle systems of 94 degrees of freedom with wheel flat) considering the dynamic factors of traction transmission, gear transmission and the track geometry irregularities, is established to calculate the dynamic responses of axlebox. In this paper, the proposed method is focus on processing the axle box vertical vibration caused by wheel flat in conventional time and frequency domain, as well as the envelope analysis with a band pass filter. Results demonstrate that the wheel flat can be successfully detected in a more realistic vehicle model, provide an efficient way to the wheel flat detection.

16. Chapter 15. Piezoelectric Energy Harvesting System to Detect Winding Deformation in Power Transformers

    Guillermo Robles, Mariano Febbo, Sebastián P. Machado, Belén García

    Abstract

    One common use of energy harvesting systems is the installation on applications where the access to conventional sources of energy is difficult due to availability, space constraints, environmental hazards or sealed equipment. In this work, we propose an alternative use of an energy harvesting system based on a piezoelectric that takes the vibration of a transformer tank due to winding deformations and hence helps to monitor the condition of the equipment. The system consists on a cantilever piezoelectric beam with a mass tuned to the resonant frequency of the vibration. The output of the piezoelectric is connected to a quadrupler, a low-drop regulator and a capacitive storage. The harvested voltage is planned to supply a low power microprocessor that detects changes in the vibration measurements to determine an abnormal behavior of the transformer. This work introduces the causes of abnormal vibration of transformers, describes the installation of the piezoelectric on a model that generates the same acceleration as the vibration of a transformer tank and studies the capability of charging capacitors to determine the feasibility of the method.

17. Chapter 16. Digital Asset Management: New Opportunities from High Dimensional Data—A New Zealand Perspective

    Marianne Cherrington, Zhongyu (Joan) Lu, Qiang Xu, Fadi Thabtah, David Airehrour, Samaneh Madanian

    Abstract

    As a testing ground for technology, the New Zealand microcosm is a stand out. In an age where many industries are vulnerable to disruptive innovation, a sustainable, integrated and entrepreneurial vision of digitally enabled asset management is vital. New Zealand is like a bridge between maturing Western economies and emerging Asian markets; with environmental values, a culture of sustainability and kaitiakitanga (stewardship), innovations that thrive in the diverse New Zealand paradigm are worth emulating. In our connected world, 5G and the internet of things are poised to explode the volume, velocity and variety of data. Survival may depend on an organisation’s ability to action knowledge from noisy, high dimensional data using machine learning. This is not as easy as it sounds; many challenges and barriers exist. This research proposes that by studying the emerging digital asset management trends in an educated and tech-savvy nation such as New Zealand, a clearer vision of how to allocate scarce resources can result. New value and exhilaration within organisations can be created so as to realise a myriad of sustainable opportunities.

18. Chapter 17. Introducing a Field Service Platform

    Maike Müller, Dirk Stegelmeyer, Rakesh Mishra

    Abstract

    The introduction of platforms disrupted many consumer industries. Prominent examples include Amazon, Airbnb, and Uber, who upended their respective industries. Eventually, platforms will also enter capital goods industries and their downstream service businesses. A part of these downstream businesses in capital goods industries are field services. In this conceptual paper, we show how the current field service delivery, which is characterized by static alliances, can be replaced with dynamic platform ecosystems, which are enabled through remote monitoring technology and mobile collaborative augmented reality. We propose a field service platform ecosystem framework by applying platform research fundamentals to the industrial field service businesses.