Data is Valuable But not Magical
- Data is a powerful resource,Convert it into operational Intelligence.
- Use operational data to improve and optimize plant operation.
- Everything starts with data, Validate your data regularly.
- Give access to right people to right data at right time.
- Use more meaningful and significant combinations of data.
- Train your people to use the data in best form.
Cement industries generate tons of operational data every hour which occupies huge space of your computer system. The older data is replaced with newer one and the story rolls on and on in this way!!
Do you remember, when last time you have used the data coming from very expensive instruments in your cement plant other than those signals used commonly like temperatures, pressures, feed rates etc.
I think your answer is yes, I too agree that you analysed lot of trends when last time you faced with the problem of cyclone jamming, a huge postmortem was done recently when you had a bearing failure in preheater fan. Is it all we get from the data resource we have? Isn’t it really a cumbersome activity to sit for hours and analyse historical trends of raw data from various instruments, that also with an intention to find the faults in others than to learn from the events.
Almost all cement plants of present-day (modern) have been commissioned with full instrumentation in all equipments. But, after a couple of years only those instruments are live and working which has been deemed critical by your operational people on the basis of their knowledge and capacity. And most of the cement plants are continuing plant operation with day to day issues like low feed rates, low efficiencies, fluctuating clinker quality, cooler problems like snowman formation, kiln ring formation, higher heat and power consumption, low cement strengths etc. the worst part of the story is that, they believe that it is a common phenomenon and are still living with the thinking that these problems are part of process.
Let me ask you some questions on the subject, to let you feel the level you are using the data generated in your plant.
- Do you have a separate machine condition monitoring system for mechanical and electric departments in plant?
- Does your maintenance team monitors machine conditioning on real time data basis to schedule a conditional maintenance?
- Have you ever been alerted about the disturbances in kiln due to material problem, which your quality department knew from analysis data.?
- Are you friendly with PID loops taking the repetitive decisions like calciner fuel rate, kiln hood draft, fan flows etc.?
- Do you have a visual monitoring notifications on raw data and refined analytic signals to identify situations like cyclone jamming, mill jamming?
- Do you use statistical process control on the data from field to infer how much is in control your process by itself?
- Have you used the first and second differentials of data signals to monitor on real time basis the optimal operation of your blending silo and feeding hoppers?
I wish your answer YES to all. However, in most of the cases the answer is No. The list of questions can be very long if I include each individual unit operations and processes of cement manufacturing. Without stretching the discussion further let me take up the question in your mind, how can we do that? What are the resources required?
The answer is DATA!!!
Most of the modern plants have sufficient data in their unifying source of truth, that is their PLC. The only thing you need to understand well is that data is valuable and powerful but, not magical that it can make use of itself for improvements. It takes a lot of efforts to convert raw data from field instruments into operational intelligence. The use of operational data to improve and optimize operation is the main purpose and the justification to install costly instruments in plant equipments. Therefore, it is your responsibility to convert data into valuable operational intelligence, to build visualizations and user interfaces that sends large amount of data in the format that can be easily understood by human decision makers. Empower your peoples by providing access to relevant data at real time to generate work orders before failure appears.
From my experience I have seen in most of the cement plants, there are always complaints that CCR operator has not given close attention towards following and following parameters and due to that the problem has happened. However, during many case studies I have observed that either the parameters has less visibility/resolution to signify the process disturbances or the noise in the parameter is so high to keep operator from considering it reliable to take action. We need to increase visibility of parameters and express data in the most significant form, for example change in temperature of cyclone may remain unnoticed to identify the cyclone jamming in preheater, but rate of change in temperature cannot and more importantly a visual alarm such as changing the colour of cyclone to red as the rate of temperature change exceeding pre-set value can make it more visible to identify the symptoms easily. Moreover, we know that in any process industry any action starts with some data, therefor the reliability of the data (measured value or signal) is most critical to take right decision at any time. Measurement system analysis should be done regularly to establish accuracy and precision of data generated.
On the other hand, well formatted data should be passed/provided at right time to the right person to achieve maximum benefits from it, for example a CCR operator uses optimally process temperatures, pressures, feed rates etc. and takes immediate actions if required to control the system, but cannot utilize the condition monitoring parameters of machines like bearing temperatures, winding temperatures etc. except to signify the allowed range of operation. But at the same time a close monitoring of these condition monitoring parameters by a maintenance person can be much beneficial to infer the actual condition of the machine, breakdowns can be avoided, work orders can be generated before failure appears.
Data is a great treasure, workout to convert it into operational intelligence. All actions and decisions depend on data you receive, therefore validate your data regularly. Train your people to use data for process improvements. Give access to right people to right data at right time (runtime) to utilize data fully.
*****THANKS A LOT, PLEASE SHARE YOUR OPINION ABOUT THIS ARTICLE IN COMMENT*******