search icon
Trendminer

plug & play data analytics for the process specialist

Big Data Analytics has a tremendous value potential for many industries. But, surprisingly, in the process and manufacturing industry little of this data is being transformed into actionable information. Surprising since the Industrial Internet of Things (IIoT) is not an entirely new concept, just a new term. Process plants have machine generated data from thousands of sensors being captured, sometimes for several decades.

An important reason for this missed opportunity is that Big Data Analytics solutions often require complex IT projects and data scientists to build and maintain models. If we truly want an Industry 4.0, we are going to need a new approach. TrendMiner has been built from the ground up for the average historian user without the need of expensive engineering projects and tackling the data modeling dilemma.

The challenges of today

Historical data is not unlocked for analytics

The Process Information Management System (PIMS) keeps all sensor data in the historian database, and has done so for many many years. But are you really able to unlock your historical data as an actionable information source? No, because these technologies were not designed for big data analytics.

Too often your historical data is historized merely “in case of”. Here’s why:

  • Searching for a relevant piece of data is currently too complicated
  • Contextualizing data from your historian takes too much time
  • Trend plots visualization should be faster and more responsive
  • Learning from historical data should be easier

Knowledge on demand is not available

The process and manufacturing industry relies heavily on 24/7 activity, which implies different shifts of people working on the same processes. Today there is no technology that easily permits these people to share their knowledge with each other … or themselves for future reference. Scribbles in a notebook, spreadsheet files or highly customized databases are still sometimes the limited way important know-how is captured and shared.

And this does not correlate with the sensor data at all!

The value of data decreases over time

Today, analyzing relevant information takes so long that the time between the occurrence of a process event and the actions taken are too far apart. Since the searching activities are so hard to do, only the most important cases are done when time permits. And time is the most valuable resource no-one has in abundance.

The main result is that the process experts today are limited to use the data in a re-active mode instead of a pro-active mode. If a new technology could offer immediate and fast search results, more cases could be solved and solved a lot faster. Ultimately combining live data with historical context, shortening the analysis latency to immediate, giving you the opportunity to take actions even before an event has effect on process performance.

Big Data Challenges

All data mining solutions today come with these 3 same major disadvantages:

why

Data scientist required

Big data modelling technologies require specifically skilled knowledge workers, the data modelling scientists, to keep the solution going. The need to channel big data questions through a data scientist causes a bottleneck in the analysis latency.

These knowledge workers are not only not available in this industry, they also do not have the insights in the processes like the process engineer does, causing a knowledge gap between requester and provider of information.

why

Engineering intensive

Big data solutions are usually big data projects. Because they are engineered to fit many different solutions, they are black boxes that have no knowledge of the specific needs of the process industry.

The return of investment of big data modelling projects are in many cases nullifying due to the extensive projects to get them in place, let alone the cost for expensive new infrastructures.

why

Sensitive to change

Modelling technologies as such are not flexible. They always have to go through the same steps:

  1. data preparation
  2. data modelling
  3. model validation
  4. bring model live

Changing a model goes through these same cycles over and over, taking away the flexibility of asking any question anytimeMoreover, modeling techniques often depend on assumptions about stationarity and data distribution that do not hold in a real, ever-changing process.

Next Generation Data Mining

next Bridging the gap

Most big data analytics solutions for the process industry are retrofitting generic solutions for specific needs. That is why these projects take so long to implement. TrendMiner software starts from the bottom up. A lot of the TrendMiner staff members at have years of experience in the process industry and know exactly the needs of the process specialists.

So that’s where we started. TrendMiner starts from the concept of the trend client which all process engineers are familiar with and bring all the big data powers right there. At TrendMiner we think it is our responsibility and not that of the customer to bridge the gaps of knowlegde & technology.

next Plug & Play

Implementing a big data solution typically requires adding new infrastructure, starting the implementation project and eventually copying all your data into the ‘analytics optimized platform’.

TrendMiner is an analytics optimized platform without those hurdles.  Our plug and play solution that can be downloaded and deployed in no-time on your existing infrastructure and is usable immediately after installation. Gone are the months of consultancy just to get your solutions started. Did we mention TrendMiner is screaming fast? TrendMiner does not copy your historian database but features an incredible performance for the end-user by a proprietary indexing data layer.

next Value out of the box

Since TrendMiner taps right into the historian does not require the data to be trained , the process engineers can start using TrendMiner as soon as the installation is connected. Using pattern recognition algorithms, TrendMiner can search through decades of information right out of the box.  

From that point onwards, you can start adding more value back in the box. Adding an annotation to your trends is just one click away. In no time you will have tremendously increased the value of your process measurement data just by giving it the right context.

next Robust & easy to use

Did we mention we are the Google of the Process Industry? A lot of the Enterprise software tools today are web-based but still look like retrofitted Windows applications in order to look more ‘enterprise-ready’. At TrendMiner we want to make sure you’re not only getting value from your new trend client, we also want you to enjoy using it.

Our HTML5 based solution works as intuitive as all the new apps you get on your personal devices. Opening as many searches at the same time or autocompleting your tag searches feels much like browsing through Google for answers.

TrendMiner is different

Bringing new technologies to market is one thing, doing it well requires company culture. TrendMiner operates in a highly competitive market that is dominated by large conglomerates. However, being a conglomerate and having market domination doesn’t necessarily mean the customers are always that happy with your incumbent solutions and practices.

At TrendMiner we have put Joint Innovation right at the center of our culture. Being a young and agile team, we can be closer to the customer than any other company. Listening to our customers is not just a tagline, it’s in our corporate DNA. Please do take the time to go through our whole program and let us know if you think we can make a difference!

Joint innovation