Showing posts with label data analytics solution. Show all posts
Showing posts with label data analytics solution. Show all posts

Monday, May 4, 2015

Trend Analysis Client- the Most Important Tool

Your historical data is highly valuable. The questions to ask is: how does one get the value from their historical data? The first step is making sure that you are collecting all the data you may possibly need in the future and collecting it at the correct frequency or resolution, depending on the type of data. For instance, temperature change does not typically change very rapidly meaning that a 10 second frequency would be fine. Pressures or values should be collected more often, around 1 time per second or faster depending on the application. Collection rates can be another detail discussion. The goal is to collect data without missing the key transitions within the data stream. This old saying speaks exactly what we are trying to portray: “How do you know where you are going if you don’t know where you have been?”.

The quickest way to make a return on your investment is utilizing a great trending application. The main purpose of the trend client is to empower those who use it, by providing easy access to the data and supplying information to allow good decisions to be made. Ed Stern states that early comments from customers are, “I am starting to understand what is really happening in my process based on what the data is telling me, not how I thought it was working.” Many people make assumptions on how the system may be working, but the data doesn't lie, it provides additional insight and discovery of nuances. A quality trending component displays whether equipment has been installed properly and is running at its fullest efficiency. Additional data is needed during start-ups to verify correct processes and expectations and to enable quick troubleshooting to make corrections.

Data analysis tools answer questions that had been unsolved before. What you thought was an equipment failure may have been an operator error. What measures can be put in place to keep it from happening again? How do we achieve better efficiency without harming the equipment? These should be questions you are asking. If plant personnel is provided with easy to use tools and access to the data, they can be a great asset to a company’s success. If the data is dependent on someone else or not readily available, productivity is lost.

The trend tool needs to be intuitive, easy to use, and have the ability to make ad-hoc changes. Predefined charts don’t cut it, they could be a good starting point but the user needs the ability to add additional tags or trends quickly. You should also have the capability to easily scroll back in time or look at different time intervals from minutes, weeks, or months of data. When these capabilities are present, the user will start asking questions and investigating the process for efficiency, long-term trends, and can come to intelligent conclusions based on data and not hearsay. This will lead to the next level of questions and may require utilizing additional capabilities such as calculated trends and alarm notifications as the user becomes more dependent on the data and its value. Ease of use means these trending tools can be learned within an hour or so via short training videos. The money saved on the purchase is wasted when time is lost at a week or more of training classes at the supplier's facility. 

Everything comes down to empowering the user with the data and tools needed to be productive. The monitoring of valuable assets and processes along with the investment of historian and trending tools will pay for itself many times over because of the operational information provided to the owners. When dealing with government agencies, a better trust relationship will be developed when the authorities know that the data is consistent and provides all the facts. The ability to allow user annotations contained within the data is also very helpful to explain data anomalies such as re-calibration of equipment to explain what happen in the data reports. When your data management is under control, offering you the answers you need, good decisions can be made to increase your business productivity instead of being in firefighting mode to keep the business running. Don't toil in vain, which is meaningless and striving after the wind. Choose tools that will improve your company's processes and productivity!

Thursday, August 21, 2014

Finding a Cost Effective Enterprise Historian


In another guest post, our Senior Software Engineer, Ken Wyant, explains the hardships of coming up with your own data solution:  
a canary solution
Finding a solution that works may be easier than you think.
"When companies search and evaluate historians for enterprise wide data, many are hesitant to commit to the cost of a true time-series historian. Many times they respond with the answer: “We are going to try and come up with our own solution”. However, what is overlooked is the technical expertise and experience required to store and retrieve the massive amounts of historical and real-time data required for the facets of the business. They think that they can come up with a more cost efficient solution. The problem is that this solution, in the long run, ends up costing the company more, takes longer to build and implement, and it fails to achieve the results that they intend or lacks the desired performance. Often times companies who try to build their own solution, end up coming back and regretting that decision. The negative outcome most often results in the company going back to the solution they originally had tried to avoid: switching to a true enterprise historian.

Usually the perceived cheaper solution that a company comes up with, entails a relational database, such as SQL server or Oracle. These databases normally don’t have an issue with the storage of large amounts of data. However, when it comes to reading and retrieving the data, you quickly find limitations. Often you’ll find scheduled jobs to trim the tables or archive older data to another table or maintenance plans to rebuild indexes periodically. Why? Because performance starts degrading and they don’t have any options except to manage the amount of data online and readily available. A true time series historian won’t require you to lop off data or rebuild indexes weekly OR require data to be moved offline in case you might need to access it again.

Can a company build their own solution that works? Perhaps for a while. But at the same time, it won’t work the way that they expect or envisioned it to. You want the data, and you want it fast. You don’t want all of the data, you want specific results. If that describes you, than maybe you should skip the attempt at developing a home grown solution and allow Canary Labs to provide you with the long term solution that you need.
There is great value in your data if you can get access to it. We find that the companies that commit to this investment with Canary will many times get their return on investment within a few months. Those that work closely and develop the partnership relationship with Canary benefit even more as ideas and innovative techniques are offered, exchanged and developed for game changing solutions for their companies. The Canary Labs mission is to provide superior value to our customers and develop long-term client relationships that exceed customer expectations. Maybe it is time to see what Canary Labs can do for you!"