Main elements of an e-Maintenance project Dynamite.

Contexts in source publication

Context 1

... his doctoral thesis [ 6 ] Karim also presents a summary of e-Maintenance related papers and goes through the definitions given to eMaintenance. Figure 1 shows the main elements of an EU IP project DYNAMITE (Dynamic Decisions in Maintenance) in which many key elements of e-Maintenance such as Smart ...

Context 2

... key point is to maintain or upgrade equipment as late as possible, but as early as necessary (ref. Figure 1). The higher the capital asset of the equipment and the more disastrous the potential consequences of a failure are, the higher the necessity for condition monitoring. ...

Context 3

... means a constantly high level of product quality and a reduction of maintenance costs. Such an overload situation is shown in Figure 10. The heavy torsional chatter which occurs during one rolling pass can potentially mark the surface of a strip or plate. ...

Context 4

... last example of practice of torque monitoring is shown in Figure 11. he effect of torque load distribution between the upper an lower roll when performing strip lubrication at the first three finishing stands of a hot strip mill. ...

Context 5

... is thus recommended to use permanently installed system for continuous, online monitoring (ref. Figure 12). Mobile solutions (ref. ...

Context 6

... solutions (ref. Figure 13) requiring data collection by personal are not an appropriate solution on rolling mill main drives, but may be a suitable, cost effective solution for equipment with more steady operating conditions. By experience gearbox toothings and bearings of rolling mill main drives are the most wear affected components which can be diagnosed by vibration analysis. ...

Context 7

... analysis is typically performed in frequency domain (ref. Figure 14). Characteristic frequencies, i.e. structure borne noise and mill vibrations from rotating components will increase if resonances are excited or in case of defects . ...

Context 8

... "amplified" vibrations are the root cause of chatter marks on product or on rolls (ref. Figure 15) and may even result in strip ruptures. The focus of the measurements and analysis at the tandem mill was thus to identify the critical frequencies, excitation mechanisms, to monitor their amplitude behaviour and correlate the results with the mill and process set-up. ...

Context 9

... protect electrical motors against internal winding shortcuts and consequently significant loss of production, temperature monitoring within rotor windings is recommended. Figure 16 shows an example from practise. Eight PT100 standard industrial temperature sensors were placed among the windings and connected to a multi-channel telemetry system. ...

Context 10

... result of online rotor temperature monitoring is a real-time temperature trend line and warning signals shown to the mill operator (ref. Figure 17), which allows to adjust motor load and production cycle according to the motor heat capability. ...

Context 11

... Figure 21) 9. The load collectives can be extrapolated to future load situations and resulting fatigue safety factor can be assessed. ...

Context 12

... architecture, based on the clonal selection principle, network pruning, representing the death of non-stimulated cells (apoptosis), boolean connection strengths, competitive network, with unsupervised learning based on a mutation mechanism ( Fig. ...

Context 13

... order derivatives increase the number of signal alternatives. Figure 10a shows the raw acceleration signal x (2) obtained from an accelerometer with a time length of 0.5 s and a sampling rate of 131072 Hz. All the time signals used in this paper were acquired with these settings. ...

Context 14

... could be seen in the corresponding amplitude spectra that frequency components up to 40 kHz are present. Figure 10e shows the signal x (2) in the case of unbalance, generated by a disk with an unbalance of 3.6 kgmm. The signal has a clear sine-shape with an additional noise level, which is comparable to the level in Figure 10a [27]. ...

Context 15

... the other signals in Figure 10 are derived from these acceleration signals. In Figures 10b and 10f both the time signals x (4) are very similar. ...

Context 16

... and 10f both the time signals x (4) are very similar. The reason is obvious, because these signals only differ in the low frequent unbalance component at 30 Hz. On the other hand, unbalance could be clearly identified in the x (1) signal, which is generated by the integration of x (2) . The velocity increases considerably in the case of unbalance (Figs. 10c and 10g). According to the ISO 2372 standard, the vibration rms value of approx. 30 mm/s is far away from any acceptable level. Figure 11 shows time signals from the two other classes: outer race fault and a combination of this with the unbalance. While the signals x (4) and x (2) from the outer race fault have a typical structure with ...

Context 17

... mm/s is far away from any acceptable level. Figure 11 shows time signals from the two other classes: outer race fault and a combination of this with the unbalance. While the signals x (4) and x (2) from the outer race fault have a typical structure with relative constant peaks (Figs. ...

Context 18

... to the ISO 2372 standard, the vibration rms value of approx. 30 mm/s is far away from any acceptable level. Figure 11 shows time signals from the two other classes: outer race fault and a combination of this with the unbalance. While the signals x (4) and x (2) from the outer race fault have a typical structure with relative constant peaks (Figs. 11a and 11b), the combinations of outer race and unbalance are less structured (Figs. 11e and 11f). The reason is the time dependent load that occurs during the contact between the roller and the damaged area caused by unbalance. This effect leads to slightly higher peaks than in the pure outer race fault situation. Fig. 10. Time signals and ...

Context 19

... from any acceptable level. Figure 11 shows time signals from the two other classes: outer race fault and a combination of this with the unbalance. While the signals x (4) and x (2) from the outer race fault have a typical structure with relative constant peaks (Figs. 11a and 11b), the combinations of outer race and unbalance are less structured (Figs. 11e and 11f). The reason is the time dependent load that occurs during the contact between the roller and the damaged area caused by unbalance. This effect leads to slightly higher peaks than in the pure outer race fault situation. Fig. 10. Time signals and frequency spectra for x (4) , x (2) , x (1) and x (0) : intact system (left), and unbalance ...

Context 20

... relative constant peaks (Figs. 11a and 11b), the combinations of outer race and unbalance are less structured (Figs. 11e and 11f). The reason is the time dependent load that occurs during the contact between the roller and the damaged area caused by unbalance. This effect leads to slightly higher peaks than in the pure outer race fault situation. Fig. 10. Time signals and frequency spectra for x (4) , x (2) , x (1) and x (0) : intact system (left), and unbalance fault (right) [27]. : outer race fault (left), and outer race fault and unbalance (right) ...

Context 21

... race and unbalance (Fig. 11f). The selected noise level masks the typical fault structure in the x (2) signal completely (Fig. 12). The signal x (4) is adequate to indicate the fault while the signals x (1) and x (0) highlight the unbalance. Fig. 12. Time signals and frequency spectra for x (4) , x (2) , x (1) and x (0) in a complex fault situation including an ...

Context 22

... race and unbalance (Fig. 11f). The selected noise level masks the typical fault structure in the x (2) signal completely (Fig. 12). The signal x (4) is adequate to indicate the fault while the signals x (1) and x (0) highlight the unbalance. Fig. 12. Time signals and frequency spectra for x (4) , x (2) , x (1) and x (0) in a complex fault situation including an outer race fault, unbalance and a high noise level ...

Context 23

... race and unbalance (Fig. 11f). The selected noise level masks the typical fault structure in the x (2) signal completely (Fig. 12). The signal x (4) is adequate to indicate the fault while the signals x (1) and x (0) highlight the unbalance. Fig. 12. Time signals and frequency spectra for x (4) , x (2) , x (1) and x (0) in a complex fault situation including an outer race fault, unbalance and a high noise level ...

Context 24

... roller bearing fault detection the fourth moment provides a reasonable compromise between insensitive low order moments and the very sensitive high orders. Figure 13 demonstrates the influence of the order of derivatives on two features, crest factor and kurtosis, for all 5 classes. Fig. 13. ...

Context 25

... for example, from the source of vibration to the sensor. For roller bearing fault detection the fourth moment provides a reasonable compromise between insensitive low order moments and the very sensitive high orders. Figure 13 demonstrates the influence of the order of derivatives on two features, crest factor and kurtosis, for all 5 classes. Fig. 13. Feature map for all 5 condition states using kurtosis and crest factor as a feature for the signals x (1) , x (2) and x (4) . The notations are unb=unbalance, and outer=outer race fault [27]. When using x (4) for class 4 (outer race fault + unbalance) and class 5 (outer race fault+unbalance+noise), the variation of kurtosis is very ...

Context 26

... condition states using kurtosis and crest factor as a feature for the signals x (1) , x (2) and x (4) . The notations are unb=unbalance, and outer=outer race fault [27]. When using x (4) for class 4 (outer race fault + unbalance) and class 5 (outer race fault+unbalance+noise), the variation of kurtosis is very high with maximum values about 250 ( Fig. 13 bottom left). Class 3 (only outer race) has much less variation, since the load is more stationary in the absence of unbalance. The time signals x (4) of class 1 (intact) and class 2 (unbalance) lead to kurtosis values of 3 with very low variation (bottom right), i.e. the signals are very close to Gaussian signals. The component caused by the ...

Context 27

... problem in classifying samples of the classes 2,3 and 5. All 50 training samples of each class were classified correctly. Mis-classified values in deciding between classes were generated between the Class1 intact and Class 5 outer race fault, unbalance and noise. The high additive noise level leads to an overlap of the samples in the feature map (Fig. 14, kurtosis approx. 3 and crest factor between 4 and 5). There is no change to improve the accuracy with these features and the signal x (2) . To obtain an error-free result, one could use the signal x (4) to separate both classes without any problems. The significant difference in the feature space between the two classes is illustrated in the ...

Context 28

... obtain an error-free result, one could use the signal x (4) to separate both classes without any problems. The significant difference in the feature space between the two classes is illustrated in the lower left-hand part of Figure 13. For fault identification, the classification has to be performed as a two-stage process. ...

Context 29

... an intelligent sensor data can be collected either locally or in a centralised manner. In the case of local data collection, a sensor operates independently, collecting and processing data and transferring the required information forward. A sensor can also be connected to a local server, which collects data from the sensor and stores it (Fig. 1). In a centralised model, data are collected to a specific certain server, which is located in the Ethernet network (Fig. 2). The data can include measured raw data, results of analyses or both of them. The server can be located physically anywhere in the world. Figure 1. Local data collection model Figure 2. Centralised data ...

Context 30

... server can be located physically anywhere in the world. Figure 1. Local data collection model Figure 2. Centralised data collection, where data from the sensors are stored in a server connected to the Internet network. ...

Context 31

... multistage gearbox was analysed in a test using automatic measurement techniques. For the purpose of the test a minor mechanical defect was made on the gearbox. In this case none of the analyses was performed using an intelligent sensor, but locally using the arrangement illustrated in Fig. 1. The measurement was carried out using the CM301 sensor of Webrosensor Oy. In the course of the measurement a lot of data were collected for later analysis and therefore the sensors were connected to a local computer. The computer was linked to the Ethernet network via a wireless 3G connection for remote ...

Context 32

... next case there is a practical example of how nip vibrations can be affected by utilizing STA measurement information in connection to the felt adjustment Figure 13 shows the common lay-out in the press section and figure 14 the 1 st press top roll as shown in Sensodec 6S UI. ...

Context 33

... of the study is on variable speed drive systems, which comprise a frequency converter, an electric motor, and an actuator. The frequency converter is supplied from the distribution network, and it is also connected to the process automation system as a field-level equipment. An example of a typical variable speed drive system is illustrated in Fig. 1 Fig. 1. Main components of a variable speed drive system. A frequency converter is connected to the process automation system via a field ...

Context 34

... study is on variable speed drive systems, which comprise a frequency converter, an electric motor, and an actuator. The frequency converter is supplied from the distribution network, and it is also connected to the process automation system as a field-level equipment. An example of a typical variable speed drive system is illustrated in Fig. 1 Fig. 1. Main components of a variable speed drive system. A frequency converter is connected to the process automation system via a field ...

Context 35

... and condition monitoring data is combined in detecting operating conditions (Fig. 1): normal process measurements are directly used in feature extraction, signal processing is needed for the condition monitoring data, and some infrequent measurements need to be interpolated. Figure 1. Detecting operating conditions and ...

Context 36

... and condition monitoring data is combined in detecting operating conditions (Fig. 1): normal process measurements are directly used in feature extraction, signal processing is needed for the condition monitoring data, and some infrequent measurements need to be interpolated. Figure 1. Detecting operating conditions and faults. ...

Context 37

... modern stage in developing a maintenance strategy has been proactive maintenance, in which the root causes of failures are monitored instead of the failures themselves, and in which corrective action is taken to eliminate or minimize the causes (Fig. 1). Fig. 1. Relation between proactive and predictive monitoring strategy. [27] Proactive maintenance of oil, which is mostly filtering particles, removing water, and releasing air, is supported by analysis of oil cleanliness and measurement of its water ...

Context 38

... modern stage in developing a maintenance strategy has been proactive maintenance, in which the root causes of failures are monitored instead of the failures themselves, and in which corrective action is taken to eliminate or minimize the causes (Fig. 1). Fig. 1. Relation between proactive and predictive monitoring strategy. [27] Proactive maintenance of oil, which is mostly filtering particles, removing water, and releasing air, is supported by analysis of oil cleanliness and measurement of its water ...

Context 39

... critical particle value was set so that every particle of that size got separated. Fig. 10 shows the best comparative results. The cylindrical reservoir has better separation values in Fig. 10 than in Fig. 8, because the latter calculates particles released from the worst places. For Fig. 10, oil volumes were ...

Context 40

... critical particle value was set so that every particle of that size got separated. Fig. 10 shows the best comparative results. The cylindrical reservoir has better separation values in Fig. 10 than in Fig. 8, because the latter calculates particles released from the worst places. For Fig. 10, oil volumes were ...

Context 41

... cylindrical reservoir has better separation values in Fig. 10 than in Fig. 8, because the latter calculates particles released from the worst places. For Fig. 10, oil volumes were chosen ...

Context 42

... order to optimise the service business, three viewpoints must be simultaneously addressed: the customer, franchisee's personnel and franchisee's profitability. These elements are shown in Figure 1. Fig. 1. ...

Context 43

... order to optimise the service business, three viewpoints must be simultaneously addressed: the customer, franchisee's personnel and franchisee's profitability. These elements are shown in Figure 1. Fig. 1. Optimisation of service business from three key viewpoints Customer satisfaction is measured twice a year by an external assessor. Should there be negative feedback, a complaint process is automatically initiated. It is in the interest of the customer to utilise this assessment to improve service provision. Personnel satisfaction is ...