Maintenance, AI & IoT
Sateyandra Kumar Singh
Reliability and Maintenance Specialist in Oil an Gas Industry.AI-ML application consultant in Predictive MIntenance.
By Sateyandra Kumar Singh, Technology Advisor, Algo8 Pvt Ltd.
As we need electricity for day to day use and nothing will move without it, we need machines to produce and run anything whether it is water, power or oil & gas. The most popular machinery in any industry are Pumps, Compressors, and Turbines. And to keep it running without problem we need maintenance crew ready to rectify any fault.
Maintenance of Machinery and other facilities has seen many developments in management, use of different strategies, automation and use of computer and internet technology with passage of time. With progress of technology, the function of maintenance also became sophisticated and use of computers also started in collection of many data to operate and protect the machinery.
Predictive maintenance is very popular strategy to know the potential failure by condition monitoring of machine through Vibration technology, Lubrication and thermography.
Operation and Maintenance engineers know very well about online vibration measurement of critical machinery like compressors, turbines and pumps. People used to say that by keeping the coin on machine our engineers could sense the vibration.
Visual observation of machine condition is still important now as sensing abnormal noise of bearing, leakage of oil, temperature of bearing by touching surface of casing bearing, but measurement gave severity and data to analyze. This resulted in progressive development of Vibration measurement and analysis by FFT (Fast Fourier Transform) analyzers by processing of signals.Tremendous data is getting generated and we have technology also to process it and present it in a form where decisions could be taken easily. Software were already available to analyze and people from maintenance know we had vibration spectrum to conclude about source and severity of problem in Machine.Internet technology and Artificial intelligence is another phase of Digital transformation. Cloud computing technology helps a lot in data processing and computing. Artificial intelligence helps in decision making based on simple rules using Machine learning and Deep learning and analyzing vast collected data. Earlier such software was known as expert system.
For example, expert system could tell what the dominant problem among unbalance, misalignment or looseness is by assigning probability percentage This is the limitation of data-based analytics and problem prediction.
Although data using mathematical capability of logic and statistical ability of probability could make rules for decision based on data, yet analysis plays a important part which cannot copy the ability of human beings in having experienced based judgments. Sometimes these human experiences were also taken into consideration apart from data-based tendency. That is why this limitation has been termed as Artificial Intelligence. However, computer machine has more speed in processing the data than human brain and thus can alert earlier than human being.
For example, Vibration Analyst of machine sees many other factors when trying to find out source of problem of vibration in machines apart from processed data and spectrum. These factors in absolute terms cannot be understood how intelligence of human being works. The generation of lot of data by Machine and fast computing ability of computers made machine learning and writing of codes by programmers. The code writers made decision making simple based analysis of online data.
In modern times due to internet and communication we can have results on our dashboard as well on smart phone and can alert about any machine or what is needed in terms of next action if some fault is detected.
So, what the people of iOT has done is that they made some device and put in it the capability of vibration, temperature and acoustic measurement and then through FFT processed the data as well generated spectrum. This all was done taking the help of internet and cloud computing.
Engineers can have result in their dashboard desktop and same could be communicated to operator level on their smart phones if needed.
Now where such technology should be used is the moot question. The criteria must be based on safety of operator who wants to capture such data, at height where safety could be compromised, non-accessible positions of machines as well remote machines.
Another important thing is how to implement such technology successfully which should be sustainable in long term. AI and iOT must be driven by the top management as a strategy in cost saving and be competitive in maintenance management. It must be applied keeping in mind many criteria so that one gets ROI (return in investment) in time.
One must keep in mind that no technology can replace human intervention in maintenance. There is no replacement of human skills in execution of maintenance activities by any technology. For example, still replacement of bearing of any pump or for that matter any machine is done by skilled and trained mechanic. Tools and tackles, methods and technology could help in dismantling and assembly of bearing but could not replace needed human skill.
So which equipment are to be taken for AI & iOT must be a decision based on business case of cost and learning and making manpower ready to adopt.
In marketplace you will find many AI solution provider giving a panacea for all maintenance problems through there smart technology. But remember we are not here to do maintenance by robot using smart technology.
Of course, with time people in maintenance would understand whether it is more appropriate in planning or scheduling of jobs or condition monitoring of machines or execution or decision making or analysis of problems or prediction of problems.
Therefore, it would be best strategy if we start interacting with vendors, consultants and solution providers to understand such technology. To take competitive advantage in identification of potential problem in maintenance wherever it is needed based on some criteria. This may result in keeping optimum inventory of spare parts. In name of digital transformation taking holistic approach in maintenance for adopting AI and iOT system in whole organization would not be a smart move.
People in maintenance know how SAP was so difficult to operate it in big organizations as one of ERP treating it panacea for all data management.
One must remember Maintenance is still arena of Human skills particularly execution part in comparison to maintenance planning and online condition monitoring of machines.
However, we must adopt new digital movement in maintenance and take benefit in maintenance area where more risks are in involved in taking data. Those machines which are located in remote area or not easily accessible for monitoring could be chosen first. Also, one must choose ioT measurement devices and technology which is compatible with hazardous area locations industry standards.
sateyandra.singh@algo8.ai
Chairperson - Startup Master Class (SMC) Distinguished Services Award-2023. As Mentor In IIT Kanpur Alumni Association, Bangalore Chapter
4 å¹´Great thought..... Lets take up a few and accomplish to meet some requirements on short-term, medium-term and long-term scales
Career Mentor | On a mission to make India "Career Mentorship Ready by 2030" | tech4good | SDGs
4 å¹´Well put together:) AI, IoT are tools which will be used for a few more years to come till the time the machines and processes themselves become smarter and self-maintaining...the transition is already happening and will spread to most of the other factory set-ups...moreover maintenance robots are a reality and with limited SLAs, different robots will take up different tasks which will ultimately take up more and more of the human jobs
Reliability and Maintenance Specialist in Oil an Gas Industry.AI-ML application consultant in Predictive MIntenance.
4 å¹´Nagasundaram Ramamoorthy pratap kambley Anurag Shrivastava,PMP,Prince2,EM,CE,MBA,PG-TPP/Renew Anirban Ghosh Shubham Gupta Anup Singh APPANI EDUKONDALU Akshay Dalwala Anand Harjal (Paliwal) Arun Jain CHANDRA SEKHAR EDARA Chetan Dhaduk Kevin Connor CMgr MCMI, CRL DINESH JAIN Pankaj Joshi Kevin Connor CMgr MCMI, CRL Stephen Puryear kindly your views?
Machinery & Reliability Industrial Consultant. Unconventional Solutions to Machinery Failure; Finding The Failure Mice. All Opinions are the authors personal opinions.
4 å¹´The use of data analytics can allow more use of low skilled technical workers for such jobs instead of higher skill staff who were formally doing troubleshooting. but as you mentioned the actual work for repair or the high level troubleshooting cannot be performed by a computing device;, too many data that is not actually available to the computer.
Reliability and Maintenance Specialist in Oil an Gas Industry.AI-ML application consultant in Predictive MIntenance.
4 å¹´Bill Partipilo , David Finch you must be having some view on it.Please give your comments. regards.