Taking the time to carefully plan the implementation of an EMI real time monitoring system can save weeks of frustration, while creating a tool that quickly achieves results. The goals should be clearly identified in advance, and the planning should carefully incorporate the following steps to ensure goals are met. This document is designed to provide the best practices that other companies have discovered as they developed successful implementations.
1. Choose Tags Carefully
Limit the number of tags being measured on a single dashboard: Powerful implementations have come from companies using their internal technical experts to carefully select and limit the number of tags rolling up to a dashboard. Although there may be no technical limit to the number of variables you can track, this is an opportunity to focus operators on the key metrics that will have the most impact on plant efficiency.
A technical expert group, which may not meet together frequently, can succinctly work to identify and tell the organization “We think these variables are the important ones to watch.” The big task is deciding what should go on the dashboard, but once this is done, the dashboard becomes the tool to keep the priority data in front of everyone from operations, to research & development, to the technical support staff.
Plant optimization: Companies are measuring and tracking massive amounts of data. If everything is included in the dashboard, the system will become so complex that the most important variables are lost in the noise of too much data. One chemical company rolled up over 1,500 variables into a single dashboard, while another company in the same industry chose to carefully select less than 50 of the most important metrics.
Plant safety: If the company wants to add plant safety to a dashboard, make this a separate dashboard that is not intermingled with data used to optimize plant performance.
2. Group Variables
Key Performance Indicators: Tags can be grouped into similar important categories. A single alarm might have a group of variables that are related in some way underneath. The alarm will indicate a category to drill down into, and a single click of the mouse will expand this category into all the underlying variables, indicating exactly which variable triggered the alarm. This is an excellent way organize your dashboard. Variables may be grouped by operating unit, location, or by a specific metric (such as “pressure drops”) across several locations within a plant.
3. Decide Alarm Protocol
Determine decision tree: Develop procedures to follow depending on which alarms become activated. All tags might not be treated equally. There may be reasons to turn on only specification and/or spc alarm boxes for certain variables, while using rule violations for others. For some processes, companies have chosen to utilize different alarm methods for systems under closed loop control. Standard operating procedure documents may be tied into the dashboard system to ensure everyone is aware of the correct course of action (which may be no action).
4. Leverage Context
Create appropriate perspective: Some companies will implement two versions of the same dashboard utilizing different timelines to provide different context. One may show hourly data for the last seven days, and another might show daily averaged data for the last many months. The combination of the dashboards provides two different contexts for evaluation, and a cumulative perspective of process improvements.
Companies who have been tracking waste on monthly reports frequently find that real time dashboards not only indicate when the waste is happening, but clearly identify the course of action for improvement.
One company noted that the dashboard showed most of the waste was happening right after shift changes, indicating that different shifts ran the plant differently. They could not only see when the problem was happening, but exactly where in the plant, which provided the context needed for training. This had never been possible by reviewing monthly reports.
5. Careful Email Alerts
Expect some tuning: Only set up email alerts for certain variables on the dashboard. Some tags may be there for figuring out what is going on if an alarm is tripped. Email alerts can be set up differently for separate users with different alarms triggering notification to different people, yet too many alarms will eventually be ignored. This is a process that usually needs to be fine-tuned over time.
6. Use Independent technology
Separate layer: Some EMI tools are part of another system that may have another primary function. Be cautious, and consider how this may be affected by future upgrades and other future company decisions. Selecting an independent technology platform that connects to all the existing data sources, but whose primary function is to provide real time EMI dashboards, may provide highly desirable independence from other systems.
7. Propagate Dashboards to all Key Stake Holders
What is happening now: Many companies spend a very significant amount of money and resources collecting data. Real time EMI dashboards allow people to view this information in context. Many stakeholders want to know what is happening right now, and not find out about problems that occurred in the past on a monthly report. The dashboard can reside as an icon on any computer desktop. Consider providing a dashboard to operators, process engineers, plant quality, plant management, corporate engineering/quality, business unit management, executive management. Dashboard designs can vary with the position (or role in the company) and what they are interested in monitoring.
Silent self-management: Many companies comment that some level of process improvement is attributed with the fact that operators know they are being watched. Some level of efficiency is frequently gained by just installing real time EMI dashboards throughout a company.