4Good: 26Fault Level: 1–29 Background: GreenAlert: 36 Fault Level: 30–59 Background: YellowDanger: 99Fault Level: 60–99 Background: RedFault diagnosis screen of the VX-14D Fault-Diagnosis ProgramThe VX-14D fault diagnosis process An observation of the many ingenious ideas packed into a tiny CPUA wide array of equipment and machinery runs at factories to manufacture products ranging from automobiles to food. When a piece of equipment fails or is on the verge of failing, it emits vibrations that differ from those emitted during normal operation. A fault diagnosis program is designed to detect subtle changes in vibration.Broadly speaking, fault diagnosis can be divided into two categories: structural fault diagnosis and wear fault diagnosis. For structural fault diagnosis, the focus is on whether the shaft of a rotating machine has become misaligned. Imagine a rod used to stir a batch of curry roux. If it begins to wobble, it will eventu-ally damage the motor that rotates the rod. This failure is known as misalignment. Additionally, normal rotation isn’t possible if a foreign material sticks to the shaft or if the shaft is chipped. This state is referred to as shaft imbalance.In contrast, wear fault diagnosis focuses on bearings, which are essential components of rotating structures. Damage to a bearing — whether to the rolling elements, through wear, or from deformation of the race — makes it impossible to transmit energy properly, leading to the risk of major failures. Structural problems often become the direct cause of wear fault problems.MASHINDAN had the capability to diagnose both types of faults, and made good, alert, or danger judgments. These diagnoses were based on calculated values independent of the machine’s specifications, eliminating the need to adjust judg-ment criteria for each piece of equipment. For a structural fault diagnosis, the parameters used were the frequency spectrum of the vibration velocity and the effective value indicating the over-all magnitude of the vibration attributable to a failure. For a wear fault diagnosis, concepts like the probability density function of the amplitude were used to identify the magnitude and frequen-cy of the fluctuation in vibration acceleration signals over time.“Beyond this, some measurement principles were based on on-site data collected by our engineers at customer facilities,” says Chou. “We were often at a loss to understand a given formu-la before tracing back to the underlying data.”Why was this formula used? What kind of formulas were used to establish this principle? Why was this particular param-eter selected? Some questions could be solved on a computer, but others required actual measurements on machines. Since Rion didn’t possess the same production equipment that the customers used, the team continued trying to link the theory to actual measurements using equipment capable of reproducing the failure conditions.Through this process, Chou learned how those designing the MASHINDAN had to overcome the limitations of CPUs at the time, which were incapable of handling the massive calculations they can today, and incorporated ways to shorten analysis times with minimal processing workload.“How can we detect a failure using just simple parameters? Today, we can feed massive volumes of data into a system and let AI judge whether something is good or not good. But I learned how MASHINDAN used small data volumes and simple com-putational processes to make decisions, given the constraints imposed by the low processing power of the microcomputers available at that time. I finally understood why this approach worked after I collected the data myself. Actually, it took quite some time to reach that point.”Automatically displays the diagnosis resultsSimply switch screens to view the measured values for the area diagnosed.Confirmation messageFault diagnosis screen (structural/wear)Diagnosis in progressThe VX-14D evaluates the state of structural or wear fault progression based on vibration amplitude waveform data and the results of frequency analysis (FFT analysis) on a scale of 1 to 99. It displays the diagnosis results in three stages (Good/Alert/Danger) based on fault progression to alert users to the early signs of abnormality or malfunction and to predict the impact on future operations.Performing the diagnosis is simple. The operator simply presses a button, then proceeds as instructed on-screen and waits about 15 seconds for the diagnosis results and comments to appear. The VX-14D dis-plays the FFT analysis results and amplitude waveform data along with the results of au-tomatic diagnosis, combining intuitive ease of use with advanced analytical capabilities.Fault diagnosis results screenPress the “Start Diagnosis” button. (MENU/ENTER key)Fault diagnosis screen
元のページ ../index.html#6