The maintenance cost of wind turbines needs to be minimized in order to keep their competitiveness and, therefore, effective maintenance strategies are important. The remote location of wind farms has led to an opportunistic maintenance strategy where maintenance actions are postponed until they can be handled simultaneously, once the optimal opportunity has arrived. For this reason, early fault detection and identification are important, but should not lead to a situation where false alarms occur on a regular basis. The goal of the study presented in this paper was to detect and identify wind turbine bearing faults by using fault-specific features extracted from vibration signals. Automatic identification was achieved by training models by using these features as an input for a one-class support vector machine. Detection models with different sensitivity were trained in parallel by changing the model tuning parameters. Efforts were also made to find a procedure for selecting the model tuning parameters by first defining the criticality of the system and using it when estimating how accurate the detection model should be. Method was able to detect the fault earlier than using traditional methods without any false alarms. Optimal combination of features and model tuning parameters was not achieved, which could identify the fault location without using any additional techniques.

This paper delivers a study involving the inspection of artificial surface cracks with depths ranging from 0.25-2.5 mm from the surface and with a crack angle of 30°, which is a typical angle for surface cracks in railheads. The inspections were conducted using three different techniques: phased array ultrasonics, single-element ultrasonics and alternating current potential drop (ACPD). For the ultrasonic techniques, the study focused on employing either longitudinal or shear wave signals. In the railway industry, shallow surface cracks in railheads are caused by rolling contact fatigue (RCF). In this study, artificial defects were made, allowing the authors to explore the extent to which the ultrasonic measurement techniques can detect such defects. The negative effect of a dead zone near to the surface in the ultrasonic tests was reduced by using a wedge attachment. A discussion on the extent to which the techniques can be used in field tests was also provided. The most important result is that shallow cracks ranging from 0.25-2.5 mm were successfully characterised with acceptable accuracy. The 2.5 mm-deep crack can be measured with an accuracy of 0.8% using a 20 MHz single-element probe and with an accuracy of 3.5% using a 5 MHz phased array (64 elements, 0.6 mm pitch). The characterisations were performed using a filtering method that was developed in this study.

1675605

Rail issues such as corrugation, rolling contact fatigue, noise and wear have been increasing with the increase in railway traffic. The application of top-of-rail friction modifiers (TOR-FMs) is claimed by their manufacturers in the railway industry to be a well-established technique for resolving the above-mentioned issues. There are various methods for applying friction modifiers at the wheel–rail interface, among which stationary wayside systems are recommended by TOR-FM manufacturers when a distance of a few kilometres is to be covered. TOR-FM manufacturers also claim that by using wayside equipment, the TOR-FM can be spread over a minimum distance of 3 km, over which it maintains a coefficient of friction of µ = 0.35 ± 0.05. To determine the carry distance of TOR-FMs, some researchers use tribometers to measure the coefficients of friction. However, moisture and deposits from the environment and trains can alter the top-of-rail friction and give a misleading indication of the presence of a friction modifier. Therefore, the coefficient of friction itself is not a clear indicator of the presence of TOR-FMs. In the present study, cotton swabs dipped in a mixture of alcohol and ester were used to collect surface deposits (a third body) from both the wheel and rail at various distances from the point of application. Subsequently, the third body collected on the cotton swab was analysed using an energy dispersive X-ray analysis. The results have shown that the maximum carry distance of TOR-FMs on the top of the rail is limited to 70 m when using a TOR-FM from one manufacturer and to 450 m when using a TOR-FM from another manufacturer. The carry distance on the contact band of the wheel is limited to 100 m and 340 m. The friction modifier on the edges of the contact band was detected over a distance of up to 3 km; however, this will not minimise the damage or friction at the wheel–rail interface.

Using ideation methods in an optimal way has a great potential to increase the number of ideas which a team can contribute during the conceptual phase of product development in industry. Previous studies on ideation methods have been mainly statistical studies in a laboratory setting. In the present study, however, the aim was to develop tailor-made ideation methods in a specific context, among actors on a deregulated railway market, through close interaction between researchers and engineers in a real-world context. Considering previous relevant research on ideation methods, associative memory models and the findings from tests of established ideation methods performed in the same group, a number of preliminary design principles were formulated and implemented in an ideation method that combines individual phases of rotational and gallery viewing with phases of verbal group interaction. The method was tested and refined in a cross-functional inter-organizational group comprising participants from different actors in the railway sector. Besides its provision of qualitative and quantitative test results, the present study has considered the opinions of the participants in detail, which can give important insights into the factors determining whether such a method will be implemented in industry. The participants found the method to be more useful and to generate more ideas that could be used in practice compared to the established ideation methods which they had tried. The learning derived from the specific case was formalized into a number of design principles for ideation methods to be used in cross-functional inter-organizational groups.

Rolling contact fatigue is a major problem connected with railway tracks, especially in curves, since it leads to highermaintenance costs. By optimising the top-of-rail friction, the wear and cracks on the top of the rail can eventually bereduced without causing very long braking distances. There are several research articles available on crack prediction,but most of the research is focused either on rail without a friction modifier or on wheels with and without frictioncontrol. In the present study, in order to predict the formation of surface-initiated rolling contact fatigue, a range offriction coefficients with different Kalker’s reduction factors has been assumed. Kalker’s reduction factor takes care ofthe basic tendency of creepage as a function of the traction forces at lower creepage. The assumed range covers possiblefriction values from those for non-lubricated rail to those for rail with a minimum measured friction control on the top ofthe rail using a friction modifier. A fatigue index model based on the shakedown theory was used to predict thegeneration of surface-initiated rolling contact fatigue. Simulations were performed using multi-body simulation, forwhich inputs were taken from the Iron Ore line in the north of Sweden. The effect of friction control was studiedfor different curve radii, ranging from 200 m to 3000 m, and for different axle loads from 30 to 40 tonnes at a constanttrain speed of 60 km/h. One example of a result is that a maximum friction coefficient (m) of 0.2 with a Kalker’s reductionfactor of 15% is needed in the case of trains with a heavy axle load to avoid crack formation.

Identification of component faults using automated condition monitoring methods has a huge potential to improve the prediction of machine failures. The ongoing development of the Internet of Things (IoT) will support and benefit feature selection and improve preventative maintenance decision making. However, there may be problems with the selection of features that best describe a specific fault and remain valid even when the operation mode is changing (for example different levels of load). In this study, features were extracted from vibration signals using wavelet analysis; a feature subset was selected using the random forest ensemble technique. Three different datasets were created where the load of the system was changing while the rotational speed remained the same. The tests were repeated five times by first recording the nominal condition and then introducing four faults: angular misalignment; offset misalignment; partially broken gear tooth failure; and macro-pitting of the gear. To improve previous feature selection techniques, a method is proposed where, before training a classifier, the most promising features are compared at different degrees of torsional load. The results indicate that the proposed method of using random forests to select top variables can help to choose good features that may not have been considered in manual feature selection or in individual load zones.

The aim of this study is to develop a novel two-level multi-objective genetic algorithm (GA) to optimize time series forecasting data for fans used in road tunnels by the Swedish Transport Administration (Trafikverket). The first level is for the process of forecasting time series cost data, while the second level evaluates the forecasting. The first level implements either a multi-objective GA based on the ARIMA model or based on the dynamic regression model. The second level utilises a multi-objective GA based on different forecasting error rates to identify a proper forecasting. Our method is compared with the ARIMA model only. The results show the drawbacks of time series forecasting using the ARIMA model. In addition, the results of the two-level model show the drawbacks of forecasting using a multi-objective GA based on the dynamic regression model. A multi-objective GA based on the ARIMA model produces better forecasting results. In the second level, five forecasting accuracy functions help in selecting the best forecasting. Selecting a proper methodology for forecasting is based on the averages of the forecasted data, the historical data, the actual data and the polynomial trends. The forecasted data can be used for life cycle cost (LCC) analysis.

This study implements an AutoregressiveIntegrated Moving Average (ARIMA) model to forecast totalcost of a face drilling rig used in the Swedish mining industry.The ARIMA model shows different forecasting abilities usingdifferent values of ARIMA parameters (p, d, q). However,better estimation for the ARIMA parameters is required foraccurate forecasting. Artificial intelligence, such as multiobjective genetic algorithm based on the ARIMA model, couldprovide other possibilities for estimating the parameters. Timeseries forecasting is widely used for production control,production planning, optimizing industrial processes andeconomic planning. Therefore, the forecasted total cost data ofthe face drilling rig can be used for life cycle cost analysis toestimate the optimal replacement time of this rig.

The conceptual design phase is a critical step, since it influences the subsequent steps during product development with regard to cost, quality and performance. Previous research has focused on cross-functional teams within an organization. However, many product development projects benefit from the participation of members from different organizations, not least during the conceptual design phase of technical products, where it is essential to consider different aspects of the product-to-be. Therefore, we conducted an in-depth case study of a cross-functional inter-organizational group testing ideation methods in a real-life setting within a development project in the railway sector. The group comprised participants from an infrastructure manager, a supplier, a maintenance contractor and research bodies. The tested ideation methods were Method 635, the gallery method and the SIL method. The participants found working in a cross-functional inter-organizational group to be beneficial both during the group-analysis of the topics and during the generation of ideas on how to address the ideation topic. Applying the ideation methods to the ideation topics facilitated the sharing of information between participants, and the diversity of the group manifested itself in several ways during ideation. Overall, the gallery method was most popular, and the SIL method was least popular among the participants. 

Rolling contact fatigue (RCF) and wear, two major deterioration processes, limit the lifetime of rails. These deterioration processes are even more severe on the curves of tracks used by heavy haul trains. Because wear is a material removing process, it can suppress the formation of RCF (also known as surface initiated cracks). In railways, cracks have a higher risk of instigating a catastrophic failure than wear; hence, it is comparatively better to have wear than to have cracks. By controlling the top-of-rail friction, both of these deteriorating processes can be reduced to enhance the lifetime of rails. In order to achieve these possible advantages, the infrastructure manager of the Swedish railway is planning to implement a top-of-rail friction control technology on the iron ore line in northern Sweden wherein RCF is a major problem on the curves. The present study uses a damage index model in a multi-body simulation software and predicts the probability of RCF formation with suppressing effect of wear for different friction control values. The effect of friction control is simulated on curve radii ranging from 200 to 3,000 m and axle loads ranging from 30 to 40 t at a constant train speed of 60 km/h. Findings show that on a very sharp circular curve, radius < 300 m, RCF can be eliminated without friction control due to the high wear rate. On moderate curves, 300 < radius < 1,000 m, a friction coefficient (µ) of, at most, 0.3 with a Kalker's coefficient of, at most, 30% is required to avoid RCF