Work no. 2 Graphic interfaces designed for management and decision levels in industrial processes regarding data display of the monitoring parameters of the machines condition. Doru TURCAN - dr.ing. SKF Romania Gabriel KRAFT - dr.ing. SKF Romania The increasing price for on-line surveillance and protection of machines (process parameter or vibration) is creating the opportunity to know in each moment the condition of the machines. This is why new methods have been created for visualizing the values of monitoring parameters. The SKF software Human Machine Interface is providing to the operating room the complete information about assets (condition monitoring) and process (process parameters), with a constant update. It creates also a fast feedback regarding the influence of the modification of the technological parameters upon the condition of the assets. 1. General description of the system Human Machine Interface (HMI) Software machinery up and running. It enables the sharing of important machine information while facilitating effective communication across functional lines all with the objective of keeping the plant operating optimally. The displays instantly present machine status information from the on-line monitoring system, providing the specific information needed by the range of plant users. By providing up-to-the-minute machine status information, the SKF Machine Analyst / HMI software gives control room personnel a significant vantage point from which to manage production activity. Because the online system collects, processes and presents data roundthe-clock, operators know how their machines are running at any given time. The vibration analysis and maintenance teams also benefit from the SKF Machine Analyst / HMI software. Using the operator s display as a starting point, analysts can drill down to identify not only the symptoms that led to the alarm condition, but the root causes of the machine problem. Actual machine and process images displayed on screen indicate machine condition status in multiple dimensions.
Users can view the entire machine or drill down to a specific section. When a change in machine condition occurs, hot spots on the machine image indicate the alarm by changing color and flashing (until acknowledged). The SKF - HMI software provides immediate feedback on the effects of process changes on the machinery itself. With rapid updating of displays at the control room, the production team benefits from immediate notification of the impact of adjustments, giving them the ability to correct and avoid such conditions in the future. Or, an alarm can indicate the sudden development of a component related problem in the machine. Easily identified by a red indicator on the display, operations can alert vibration analysis or maintenance
personnel before the problem becomes critical. SKF Machine Analyst / HMI provides the capability for automatic, frontline monitoring 24/7. Now, known sources of vibration can be matched to the actual components, dramatically enhancing analysis capability. A graphical display provides the operator or analyst with insight into the specific source and nature of the defect. To help the analyst, the SKF / HMI includes two post-processing techniques for which patents are pending for significantly enhanced analysis: - Cyclic Time Averaging (CTA ) enhances, then displays component signals related to defined component speeds at a certain moment in time. CTA can help identify a range of root causes from felt or roll problems to gearmesh defects and more. - The Harmonic Activity Locator (HAL ) condenses harmonic patterns in a FFT (Fast Fourier Transform) spectrum into several simple bar displays that indicate specific problems such as bearing damage, looseness, misalignment and others. 2. Practical case studied The Cyclic Time Averaging technique is similar to the well-established Synchronous Time Averaging method, with the major difference being that with Cyclic Time Averaging, an external hardware trigger is not required. What is Synchronous Time Averaging? Synchronous Time Averaging is a process in which is use an external 1X RPM trigger to time synchronize and average data collected from the machine. The time data is sequentially averaged and only the integer ordered data amplitudes are enhanced; all non-synchronous peaks tend to a zero average. This technique enhances desired signals while minimizing unwanted vibration signals from, for example, nearby machines. What is Cyclic Time Averaging? Cyclic Time Averaging, is a post processing technique where a data ensemble containing a number of consecutive machine cycles are re-sampled, summed, and then averaged. The resampling is based on a syncpulse (trigger) that does not physically exist. Instead of an external trigger, Cyclic Time Averaging uses a user-defined trigger. Like Synchronous Time Averaging, Cyclic Time Averaging is a useful tool in situations where data averaging that is based on certain events is required to assess vibration from a specific source, or to eliminate the influence of an unwanted vibration source. Unlike Synchronous Time Averaging, in the Cyclic Time Averaging process, absolute shaft phase data is not coherent with the machine cycle. Cyclic Time Averaging may be used instead of Synchronous Time Averaging whenever the required trigger is not readily available. Example: Using Cyclic Time Averaging for Gear Tooth Analysis One of the best cases for using Cyclic Time Averaging is when an external trigger is not possible, or practical, such as when monitoring the inner shafts of a gearbox. The goal is to
find out if there are damaged gear teeth on the gear, without having to stop the machine and open it up for inspection. The problem is that there is no trigger on the intermediate shaft, and it is impossible to install one since the shaft is inaccessible. Cyclic Time Averaging is the only way to look closely at the waveform generated by each tooth from that gear. The rule is to ensure that the sample rate is high enough to provide good resolution of the event, but not so high that there is not enough data to perform the averaging. For our example of a damaged gear tooth, the event is defined as the duration from the initial contact of two gear teeth until the end of the contact. It is decided that 50 samples per gear tooth engagement is preferred. The choice for number of lines is 6400, providing the maximum amount of data for averaging when using a Multilog LMU (Local Monitoring Unit) data acquisition device. The time domain waveform for one shaft revolution is plotted in Figure 2. Notice there are 57 cyclic events shown in the plots, representing 57 gear tooth engagements. The waveform provides no clear evidence of any bad gear tooth. After the initialization of the Cyclic Time Averaging process, Machine Analyst / HMI performs the calculations and displays the result as shown in Figure 3. The three provided plots are: The cyclic time averaged waveform is plotted in the upper left corner. The spectrum for the averaged data is plotted in the lower left corner.
The profile plot of the cyclic time averaged waveform is displayed on the right side. In the profile plot, it is clear that the damaged gear tooth generates a significant amount of high frequency vibration during initial contact with the mating tooth. Remarkably, only fiveaverages are needed to bring out this damaged gear tooth signature. 3. Conclusion The SKF Machine Analyst / HMI software will help the maintenance and operational personnel to have an on-line view of the asset functionality status. With his Cyclic Time Averaging, the SKF Machine Analyst / HMI software is also a powerful tool for rotating machinery vibration diagnosis and condition monitoring, particularly when an external synchronous trigger for the averaging process is not available. Bibliography 1. By Jim Wei SKF Reliability Systems - Gear Tooth Analysis Using Cyclic Time Averaging and Machine Analyst /HMI 2. *** - Machine Analyst /HMI presentation brochure