In this review, we are going to explain how Hexastate uses vibration sensors to interact with our software becoming a machine learning and predictive maintenance tool.
Vibration analysis is a term for analyzing the data acquired through vibration sensors. The vibration data tells the story of the machine being monitored. In the vibration data, it is possible to detect various failures in a machine and the degradation of condition of the components within the machine.
How does Hexastate CONNECT analyze?
Designing system for machines
The first step to getting started with the new software Hexastate CONNECT is to take a picture of the name plate on any machinery you want to be monitoring and add your contact information. Hexastate CONNECT will do the rest, and when the hardware is mounted, you will be up and running within minutes.
Finding the Frequencies
When performing a vibration analysis, it is important to know exactly which frequencies that are directly related to the specific machine that are being monitored. Therefore, Hexastate CONNECT uses the picture of the machine name plate to find the necessary information to calculate the position of all machine related frequencies. The detection of the machine name plate and calculation of frequencies are done automatically when the system receives a picture of the machine name plate.
The three most used analysis methods
The processing of vibration data can be done through different techniques. Three renowned techniques are the FFT Spectrum Analysis, Time Waveform Analysis and Phase Analysis. All three techniques for vibration analysis seek to detect faults in production equipment, diagnose the fault and in the end determine an action plan.
The fast Fourier transform (FFT) is an efficient algorithm used to compute a discrete Fourier transform (DFT). This Fourier transform outputs vibration amplitude as a function of frequency so that the analyses, in this case Hexastate Connect, can understand what is causing the vibration.
Time waveform analysis is a powerful condition monitoring tool, in which it is possible to analyze the vibration pattern of a machine during as a function of time. Therefore, we can detect repeating patterns in the waveform data.
This form of analysis is an expansion of the waveform analysis, in which the phase shift between waveforms from different vibration sensors can support the analysis conclusions to decide between multiple different faults, if the final conclusion can’t be found only from the other analysis techniques.
Units of measurement
The study of vibration amplitudes can be done in three different physically related units of motion: displacement, velocity and acceleration.
Displacement means the amplitude (physical distance) between the peaks of vibration. Displacement is being monitored from peak-to-peak of a vibration waveform to measure the total physical displacement in a vibrating motion. Displacement is effective when analyzing low frequencies of vibration.
Velocity is a derivate (a rate) of displacement. By velocity we analyze the vibration in terms of the velocity of the vibration displacement. The velocity of vibration is measured in the peak of the vibrating motion. Velocity is effective when calculating the overall vibration energy in a measurement and a wide mid-range of frequencies.
Acceleration is related to velocity by being a derivative (a rate) of velocity. Therefore, acceleration is the rate by which the velocity changes over time. Acceleration is effective at higher frequencies, where e.g. bearings show early signs of deterioration.
RMS is an abbreviation of root mean square of a vibration amplitude. The RMS value is calculated over a range of frequencies to determine an overall energy level of vibration. By studying the value, it is possible to get a rapid and simplified overview of the machine condition that sets the foundation of further analysis.
Supporting features for the end-user
Hexastate CONNECT can monitor and track any directly related frequency of a production machine. When the frequencies are calculated, Hexastate Connect uses advanced signal processing techniques to track any speed changes in the machine, and track frequencies into their new rate. This is done without a tachometer/RPM sensor on the machine.
We support your next choice with predictive maintenance
Machine learning tools
Hexastate CONNECT has built-in machine learning models that predict future developments in machine condition. Machine learning covers a collection of models based on advanced mathematics and statistics, in which a computer can learn patterns from data.
Predict future condition
The future condition of a machine can be predicted using the vibration data of a machine as input to a machine learning model that outputs the condition development over upcoming next weeks of operation. This allows for precise and efficient planning of maintenance to keep production downtime at its lowest.
When faults or a significant change in condition is detected by the Hexastate CONNECT analysis software, an email or SMS notification will be pushed for the user, enabling them to focus on other tasks and optimization of the production operations.
What is a Service provider and what are the benefits? This post will discuss the possibility for service companies of creating stable revenue with a
What do all the
“buzzwords” CBM, PdM, IoT, AI and ML mean for your production?