If one thinks of the basics of vibration analysis, he can easily guess that it has to do with ‘sensing’ the vibrations, analyse the ‘patterns of vibrations’ and then come up with the diagnosis of a possible technical problem for informed decision support. This technique can help in understanding how ‘healthy’ the mechanical vibrating body is over a period of time, under certain operating conditions and environmental conditions. Apvike, co-relates various parameters in addition to velocity, acceleration and displacement and builds a mathematical model for strategy development as a part of system dynamics.
The data from various sensing points along with the data of vibrations from pre-identified sensing points are ‘blended’ together and analysed with the help of machine learning and deep learning techniques – specialised areas of Artificial Intelligence (AI) – to provide outcomes for informed decision support. Blending ‘system dynamics’ with vibration analysis helps provide outcomes which are sustainable and deliver quality results. Typically, the solution provides four types of outcomes:
Historical analysis (what happened)
Predictive analysis (what can likely happen)
Prescriptive analysis (what is to be done for making the expected things to happen)
What if analysis (what may happen if other parameters change in a certain way)
All critical parameters are monitored at a pre-defined time period (or near real-time) in order to collect the data at a central location (database) wherein the system dynamics’ modelling is developed.
This case study is of one of the leading potato chips manufacturing companies in South Asia and how they benefited by implementing vibration analysis techniques of VibroNxt™.
The manufacturing plant has a capacity of manufacturing around 180,000 kg of potato wafers, along with some 750,000 kg of savouries per day. They have hundreds of electric motors working round-the-clock in their operating plant. And because these motors are well-orchestrated, the manufacturer can timely produce high volume of quality potato chips and its variants.
The pre-packaging phase in chips manufacturing is where they make use of special types of seasoning barrel. While the chips are made to flow down this barrel, the required flavours and spices are added on top of them. The movement of this barrel makes it possible for these chips to get a nice even covering of the added flavours and spices. It is worth noting that this barrel rotates (angular movement) at a speed of 12-15 rpm and simultaneously it is also displaced horizontally (horizontal movement) at a speed of 1 metre per second approximately. The barrel has to be gentle enough to not to break the chips and maintain the desired quality. The horizontal displacement of the barrel makes chips flow out of the barrel for the next process of weighing and packaging.
During our initial discussion round, we were made aware of the fact that one of the seasoning barrels was giving more percentage of broken chips.
Initially physical checks of the ‘suspected’ set of assembly was carried out. Since the seasoning barrels play an important role in maintaining the quality of the chips, a simple tachometer test was carried out to see any possible major deviation in rotation of both the barrels. But nothing major was observed there. VibroNxt solution was deployed at the operations site. It comprises of vibration sensors (handpicked based on site-specific conditions and requirements) that can transmit signals wirelessly to the central server (cloud or on-premise options available) installed with VibroNxt software application.
The application is browser-based, role-based, secure and can be accessed from anywhere without any geographical boundaries. A parity check was carried out with one ‘suspected’ set of assembly and the one that was ‘healthy’. One of the advantages with vibration analysis solution – VibroNxt – is parity checks. This means the solution makes it possible to create logical groups of electric motors (or any vibrating mechanical component for that matter) for comparison and co-relation purposes. We started sensing various parameters of all the motors in both the assemblies, lugs of these motors and certain vibrating components including shuttles that give horizontal displacement to the barrels.
After analysing data from both the assemblies, it was concluded that there was nothing wrong with motors as they were conforming to the ISO 10816 guidelines. Surprisingly, VibroNxt showed some ‘relative difference in patterns’ generated from the shuttle assembly and the motors that drive them – thanks to the system dynamics modelling of VibroNxt. Further investigation discovered the fact that the bushing fitted at the end of the assembly of the shuttle (suspected assembly) was of a different make and was recently replaced by the customer. The customer was suggested to replace the one (exactly same make and model) with that of the ‘healthy’ assembly line.
Post that, VibroNxt showed significant improvement with almost no variations in the signal patterns. And within a day, the customer saw results. VibroNxt and its capability of system dynamics modelling can assist in detecting defects or faults proactively. Furthermore, customers can even ‘benchmark’ internally for the performance of a specific brand of electric motors over a period of time in terms of their reliability, sturdiness, etc across their multiple sites without any geographical boundaries.
Vibration analysis is a significant technique that can bring in significant business value. When deployed proactively, it can save significant amount of money, time and efforts in diagnosing and analysing technical faults or problems. VibroNxt is a comprehensive solution for vibration analysis requirements, wherein it assists in complying with standards such as ISO 10816 as well as in diagnosing complex technical problems or faults. This is thanks to its system dynamics’ modelling capabilities that provide excellent visibility to ‘hidden’ problems that cannot otherwise be detected with conventional methods of manual testing. Smart predictive algorithms provide early warnings of the likelihood of machine failures, thereby significantly improving MTBF and upfront maintenance cost savings.