MaintainIT

Ideal maintenance strategy for industrial- and production companies can now be realized through sensor data and a new software called: HDC-Analyzer

Maintenance for the Industry 4.0

• Intelligent state-based maintenance (AI)
• Huge economic and environmental savings on spare parts and less downtime
• “If it ain't broke - don´t fix it”. (T. Bert)

With the software concept HDC-Analyzer, we have automated the skilled engineer or technician who walks around and "listens" to the bearings and knows the sound of the machine. Thus, it is no longer necessary to replace components until they show signs of breakdown. The well-known interval-based maintenance is now replaced with intelligent condition-based maintenance.

There is a huge waste of machine parts that are replaced even if nothing fails. In many companies, critical components are replaced at intervals specified by the manufacturer. Machine parts are often replaced even when less than 10% of the components have failed. The next problem is that after the replacement, up to 27% of the new components fail due to the replacement procedure itself or misalignment (source ABB).

At Cedas, we want to promote the following motto instead:

     “If it ain´t broke – don’t fix it”.  (T. Bert)

Environmentally, there are great benefits to be gained by implementing our intelligent maintenance solution. This concept will potentially significantly reduce the sale of spare parts.

Economically, it is also a very good investment, where each company can achieve great savings.

The challenge

• Data becomes informational abundance (<1% of data reach the decision maker)
• There is enough data to predict crashes, but the equipment is still breaking down
• Knowledge of data analysis is not rooted in the company, but in individuals

In the industry 4.0 era, there is a growing focus on data usage and a diversity of digital solutions. This can give the individual industrial- or production company a major technical challenge in the increasing numbers of different screens and apps that the operators on e.g. a power plant must keep an eye on. Data collection equipment is bought like never before and everything can be connected and displayed on trend curves, so that it can be intervened before a crash occurs.

Here's part of the challenge. Here in the tension between operation and strategic long-term decisions. Who has ownership of all these new systems for e.g. vibration monitoring and the like, and who is watching for any signs of crashes in 8-12 months ahead? Do you need to hire more data analysts, or should the normal operating staff keep an eye on it all?

                                                                                 Who is to interpret the data? 

 

Another part of the challenge is that the operation is potentially uncertain, and errors occurs due to individual alarm limits on the components and an overwhelming amount of data to be kept in mind. Furthermore, you run the risk that some people in the organization, who is becoming proficient at working with data, and becomes key individuals with unique knowledge and ownership of the systems, are suddenly seeking new challenges.

Despite data, for example from expensive vibration measurement equipment, the technical equipment is still breaking down, and one can (perhaps too late) look at the graphs ... and the crash could have been foreseen.

There are few who have the time and ability to view all the data in the context of e.g. weather conditions and various loads. It is frustrating and that is why we have devised a new software concept that can solve these challenges: The HDC-Analyzer.

Our future-proof solution

•     Recording “Normal operation” for each part machine can be stored and monitored automatically
•     Simple three-step commissioning and no need for data analysts
•     Possibility to record multiple scenarios of “Normal operation”

 

 

HDC-Analyzer - Historical Data Conflict Analyzer

 

 

 

 

Peter Strini, Executive Global Network (EGN) says this about our HDC-Analyzer:

” The Industry and technical management have never before had a software tool that interactively tells the maintenance department when critical parts need to be replaced – only using pre-known sensor technology”.

With this new software, we can present one overall status of the operating state from the machine and automatically monitor if it runs smoothly. An algorithm notifies you when the first signs of trouble occur.

 

From components to context

By monitoring the entire context of the operation of e.g. a fan, the algorithm learns how the signals interact with each other.

The software is prepared so that e.g. 10 sensor data are converted into one result, 10 of these results can again be represented with one result, etc. Ultimately, the technical state of operation of an entire factory can be displayed with "one technical indicator" where you can easily navigate to the problematic equipment.

The software is designed according to the "Jethro Principle" and enables thousands of data points to be overlooked and significant challenges brought to technical management.

More about our HDC-Analyzer

• All data can be used
• It can be customized to customer's network configuration and local IT and data guidelines
• The algorithm works with mathematical probability

It is a software that can be connected to existing monitoring systems (add-on) or it can be run as a stand-alone setup, as shown below. The fan below is a simple representation of an industrial process using the software. Data can come from e.g. temperature, vibration, sound, power consumption, speed, pressure and more. We can also monitor sequence times, manual registrations, recipes and operating parameters. In fact, we can monitor all kinds of data with this concept.

 

Sound analysis with special microphone (384 kHz sampling frequency) with "HDC-Analyzer" at Cedas in Randers, Denmark. A bearing monitoring like this can tell you about problems with a bearing up to 1 year before a breakdown.

 

HDC-Analyzer – control integrated into WinCC-Advanced

There are 3 simple steps so everyone can use the software shown in the image above:

• History: First, a "data recording" is made from selected data points.
• Key: Next, the key is activated based on the recorded data.
• Play: Finally, the monitoring of the machine is activated, and the result is presented in the gauge, where 0 indicates that everything is running as normal and an increasing value indicates that there are signs of problems.

 

We are looking for cooperation with companies that will participate in pilot projects on the application of this concept.

Facts:

• Data is a key strategic asset – 1.6T USD (Maintenance is where one of the biggest assets are found).
• [¹]Only 1% of sensor data from a drilling platform reaches the decision maker.

 

Source:

https://www.mckinsey.com/business-functions/mckinsey-digital/our-insights/the-internet-of-things-the-value-of-digitizing-the-physical-world