Alexa

50 Smartest Companies of the Year 2021

An innovator providing a unique solution for customer’s mechanical systems’ maintenance problems: Diagsense

thesiliconreview-tidhar-tsuri-ceo-diagsense-21.jpg

In recent times, the use of preventive maintenance solutions has become an integral part of any maintenance strategy. It is the key to reduce equipment downtime, lowering maintenance costs, improving workspace safety, and increasing the lifespan and efficiency of an asset. We live in an era where more than 80 percent of the companies are experiencing an unexpected outage, and it is vital to have a well-defined maintenance plan or solution in place to prevent adversities. The main motto of using preventive maintenance solutions is to drastically prevent equipment failure before it can occur and reduce the number of accidents.

There are various companies delivering excellent preventive maintenance solutions globally, but Diagsense stands out from the rest. Using innovative statistical algorithms, Diagsense employs state-of-the-art proprietary algorithms for monitoring and predicting malfunctions in various mechanical systems, such as pumps, pipelines, and turbines. Solutions from Diagsense will save you – idle time due to unexpected downtime of machinery.

Diagsense’s unique solution is extremely sensitive to any change of the system’s normal behavior and on the other hand, we provide false positives close to zero. We developed a leak detection system for pipelines, through continuous monitoring of the flows and pressures from each side of the pipe we can alert on any deviation from normal operation.

In conversation with Tidhar Tsuri, CEO of Diagsense

Q. Tell us about your Preventive maintenance solutions.
Our solution comprises of two-phase:

  • Recognizing of the machine state (Accelerate/Decelerate/running in constant velocity/stop/Gear/State). We can gain understanding of this by running a supervised clustering algorithm
  • On each machine’s state, we are running our unique PDM algorithm to find an accurate correlation between all the parameters that affect the monitored object. We test it on many kinds of mechanical equipment such as Pumps, Boosters, Blowers, Generators, AC Motors, and turbines

In all systems, we find significant results in terms of:

  • Accurate solutions -> sensitivity to any deviation from normal behavior.
  • Close to zero false Positive.

Q. What makes your solutions unique?

In one word: Simplicity. To understand our unique approach and technology, I will try to explain the problems in this field. There are a lot of companies who are tackling this challenge in the market. Most of those artificial intelligence solutions suffer from many other problems such as:

  • Overfitting
  • Misdetection
  • High-level of False alarm
  • Cannot deal with complex data/situations
  • The solution can fit only to one object with specific environmental conditions

And at the bottom line, all those solutions are expensive. Most of the small/medium factories or companies cannot afford such a system because the ROI is not reasonable. Diagsense delivers AI predictive maintenance algorithms and software, which benefit in a wide range of applications. The Predictive software can be deployed on a cloud, on-premises, on a dedicated computer, or be embedded inside the client products as an edge solution.

Q. What does innovation mean to you?

Innovation is not necessarily “inventing the wheel”...Innovation is taking something that works in one field and implement it in another domain. In many cases Innovation is taking something good and make it better for your own purposes.

Our Innovation

Diagsense fuses an entire set of time series and monitors it at once, instead of monitoring each variable separately with no relation to others. The fusion is performed by (machine) learning of correlation between all variables in the set, i.e., in addition to absolute values of variables, we are using correlations between them. This set of variables is selected from available time series such as temperature, conductivity, flow, pressure, current consumption, and any other sample data. Internal influences between these series will be eliminated, and thresholding will be performed on stationary signals. In this way, false alarm and misdetection rates will be close to zero. Another aspect of our advantage is a high sensitivity since we are using original variables to a high degree as well.

Diagsense performs diagnostics in the space of original values, but in a rotating one. By doing so, we present new features that are (optimal) linear combinations of the original variables.

This approach allows us to introduce “a weight of influence” of each variable, so we don’t have to ignore it totally.

Q. What does the future hold for your company?

The demand for adopting AI solutions in the industry is always on a constant rise. Moreover, we are aware of the need for tailored solutions. We are working with clients who maintain critical assets, Customers who cannot tolerate working with a commercial solution. They are seeking accurate solutions that fit their needs without any misdetection or false alarms. Most of our clients are Integrators and SCADA companies who are looking to answer their client’s requirements and providing them models and algorithms that they can integrate inside the client’s Control Systems. The most common problems in Predictive Maintenance are False Positive alerts. We are specialized to develop tailor-made predictive maintenance solutions which are very accurate.

Q. How was it working through the pandemic? How challenging was it for you?

Through the pandemic, we understood that our destiny is to serve as many clients as possible and make any company that is interested in implementing and adopting AI solutions within their system to be accessible to our knowledge and experience. In 2020, we offered to startups, and industrial companies, our AI solutions as an outsource service.

Meet the leader behind the success of Diagsense

Tidhar Tsuri, CEO

He understood the need for holistic solutions. Customers like manufacturers and maintenance companies are not looking for an algorithm; they are looking for overall planning to take their factories to the next level in efficiency, productivity, and Profitability. Hence, Tidhar established the Industry Team - a unique complementary group of experts in the industrial world. (Machines, Chemistry, Electrical, Energy, Control, Algorithms, Digitalization, Training, and more). The Industry Team is working with startups that provide dedicated solutions on multidisciplinary topics, including Cyber, IOT sensors, Sensor and PLCs monitoring, Connectivity units, Visualization & Presentation, Data Analytics, Energy, Robotics, Digitalization and more. Those innovative solutions are implemented within our projects according to the client’s needs and requests.

“Our partners have the experience and expertise required for effective design, development and integration of PdM algorithms in an on-line monitoring system.”