Signals to Solutions: a change of paradigm in predictive maintenance market

Projektinformationen

S2S

ID Finanzhilfevereinbarung: 816402

DOI

10.3030/816402

Projekt abgeschlossen

EK-Unterschriftsdatum 4 Mai 2018

Startdatum 1 Juni 2018

Enddatum 30 November 2018

Finanziert unter

• INDUSTRIAL LEADERSHIP - Innovation In SMEs

Gesamtkosten

€ 71 429,00

EU-Beitrag

€ 50 000,00

Koordiniert durch

SIGNAL ANALYSIS LAB AS
Norway

Periodic Reporting for period 1 - S2S (Signals to Solutions: a change of paradigm in predictive maintenance market)

Berichtszeitraum: 2018-06-01 bis 2018-11-30

Zusammenfassung vom Kontext und den Gesamtzielen des Projekts

EU INDUSTRIAL AND SOCIETAL PROBLEMS:
Rotating machinery failure and downtime is a major cause of economic losses in many manufacturing sectors. In response, Predictive Maintenance (PdM) is gaining popularity among key industrial players to: (1) extend their assets’ lifecycle by avoiding regular replacement intervals; (2) decrease maintenance costs by properly planning of maintenance breaks; and (3) minimize unplanned production breaks by predicting the failures ahead of machine breakdowns.
Nevertheless, State-of-the-Art (SoA) PdM techniques are still based on old assumptions (i.e. stationary signals with little or no variability), which perform poorly in emerging flexible and variable production models. These techniques are based on analytical models suitable for stable productive models generating stationary signal data with little variability (e.g. traditional fossil fuels’ power-plants) and therefore poorly suited to accurately identify failure events in the emerging production models, which are more flexible, complex and variable along time (e.g. offshore windmills). Overall, SoA PdM methods fall short of the industry’s need for effective maintenance with negative consequences from both the economic (consumables overspending, anticipated replacement investments) and the environmental perspective (irrational use of resources, increased industrial waste). New analytical models are therefore necessary to properly address the maintenance needs of a continuously evolving & complex industry.

OUR SOLUTION:
Our disruptive solution, named Signal to Solutions (S2S), will address the problem of machinery maintenance by introducing a new PdM technology, which will ensure cost-effective and durable solutions for the offshore and civil aviation industries. It could also be applied to other markets (e.g. naval & automotive industry) to enable savings in terms of machinery repair, spare parts replacement, equipment downtime and lifecycle extension, which all together can be considerably expenses. S2S focuses on analysing non-stationary data and interpreting its results and deploys the Hilbert-Huang Transform (HHT) methodology in our proprietary algorithms and software solution. Compared with stationary signals (i.e. stable data along time), non-stationary signals (i.e. based on time-varying coefficients) reflect more realistic production cases, where the machine intensity varies according to different time schedules, productivity needs and working conditions. Our HHT-based method is therefore a suitable PdM method for machinery characterized by variable operational conditions, such as drilling rigs and aircrafts.
Current PdM techniques are normally based on two underlying methodologies: Fourier Transform (FT) and Machine Learning (ML). Yet, they both present limits when analysing equipment in real-use conditions because they start from the assumption that signals by default are linear and stationary (FT), or that the data generated by an equipment do not change over time (ML). On the other hand, other past attempts to implement HHT (e.g. Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and Ensemble Empirical Mode Decomposition (EEMD)) perform poorly in terms of computational times in the range of several hours, making this algorithmic implementation unsuitable for PdM.

Arbeit, die ab Beginn des Projekts bis zum Ende des durch den Bericht erfassten Berichtszeitraums geleistet wurde, und die wichtigsten bis dahin erzielten Ergebnisse

Our key objective for this Phase 1 Feasibility Study was to verify the technical, practical and economic viability of S2S for the markets in which we want to launch our product. More specifically, we aimed at verifying that our solution is superior to current market solutions (under real-life settings) in terms of operational performance, cost savings and maintenance improvements. Therefore, we conducted the following activities:
1. Involvement with and surveillance of potential end-users to understand overall technical requirements for the target markets;
2. Analysis of needed modifications of S2S to achieve the technical requirements of the potential end-users;
3. Assessment of the financial viability of our business model by pooling willingness of potential end-users to purchase our product;
4. Establish final requirements for undertaking maturation and large-scale demonstrations under real operating conditions to be implemented in the envisaged Phase 2 project.

With the goal of confirming S2S commercial viability and finding early adopters, we met with several times with major operators of the industrial maintenance sector in Norway with extensive experience with CBM and PdM services, and potential customers and end-users from the transport, energy and industrial production sectors. During these meetings we introduced them the superior performance of S2S and how it reduces unexpected failures compared to any SoA solution. These talks also lead to the realization of two small-scale demonstration pilot projectswith a major PdM provider and a engineering and supply company. In order to assess the performance of S2S under real-life conditions, these projects applied it to analyse the data generated by their current equipment:
• Maritime pilot: we organized several workshops with its R&D team to fully understand the requirements for as well as its customers. We tested S2S in a pilot project for monitoring one shipping installation used for offshore wind farms. S2S was applied to around 20 sensors per installation for monitoring motors, gears, winch, brakes and pumps. A feature that was greatly emphasised by the PdM provider was the ease of use for the end user, as well as the compatibility with its existing control system using a Programable Logic Controller (PLC). The pilot duration was 500 work-hours.
• Drilling pilot: challenged our product by sending us a dataset of vibration measurements from a motor containing nine bearings and gears. To test our Tacholess performance, the data was presented without tachometer data. S2S automatically detected two faults of a specific bearing and showed the evolution of the faulty signal over time. The fault was the same as the one its operators had found manually, however S2S identified signs of the fault at an earlier stage without knowing the rotation speed. Their interest grew even more because most existing machines measurements lack tachometers measurements. The duration of this pilot was 150 work-hours.

The results of the pilot projects exceeded their expectations, and the both entities confirmed their willingness to purchase S2S when it becomes commercial-ready. During these pilot projects, we worked on the underlying technology of S2S to include the feedback received and solve minor issues risen to optimize the integration with the equipment. As part of our continuous improvement, our research team also continued the study of our own implementation of the HHT, which lead us to improving its performance beyond the initial one.
In parallel to the small-scale pilot projects we also presented them different pricing models. Our conversations with both potential global distributors and customers confirmed that the use of a periodic fee is more suitable rather than a unique license purchase, to adapt it to the changing maintenance needs in parallel to the availability of the machinery. In these meetings all potential customers shared with us their interest in counting on a complete scalable solution based on our PdM technology instead of using it as a complement to other existing solutions.

Fortschritte, die über den aktuellen Stand der Technik hinausgehen und voraussichtliche potenzielle Auswirkungen (einschließlich der bis dato erzielten sozioökonomischen Auswirkungen und weiter gefassten gesellschaftlichen Auswirkungen des Projekts)

During the Phase 1 Feasibility Study we confirmed our claims about S2S’s benefits for final end-users: (1) the normal operation is not disturbed by breakdowns; and (2) maintenance actions can be planned and their duration reduced, everything is already in place when the machinery is stopped. Besides this, the conversations held with our potential customers lead us to the definition of different versions for S2S, beyond the original concept of offering it as a complement to existing CBM/PdM solutions:
• S2S PdM suite: complete PdM solution, including all elements to gather the data from sensors, process it with our algorithms and present the results of each machine monitored. These elements have been already designed following the customers’ feedback and an alpha version has been already developed.
• S2S offline device: using a Programmable Logic Controller (PLC) with our algorithms embedded and connected to the sensors installed in different machines. The goal is to offer cost-effective PdM services over machinery without sensors and tachometers, replacing current manual methods with operators doing visual inspections or picking up data using handheld devices for posterior analysis.

The feedback on the final set of requirements received from the end-users drove us to introduce improvements in our implementation of the HHT in parallel to the feasibility study. The final result was tested against the data collected during the small-scale pilots and our assessment showed that we had achieved a significative improvement of up to 10,000 times faster than the SoA (CEEMDAN), and more than 1,000 times faster than the fastest HHT implementation known.

The results obtained during the small-scale demonstration pilot projects confirmed our expectations:

Improved success rate predicting failure
The analysis of the rotary machine parts by S2S revealed the health state and imminent failures in all the cases where there were faults present. The faults were also automatically detected at an earlier point in time than could be manually observed by technicians. The two pilot projects confirmed S2S’ ability to automatically and reliably detect failures at an early stage.

Significantly faster than other solutions
Due to the automatic analysis performed by our solution, the reduction in manual operations was confirmed by the participants in the pilots. S2S is additionally optimized to be able to run real-time analysis of vibrational data. This feature allows S2S to run locally at a remote location, e.g. on a vessel, improving maintenance team’s awareness without further analysis of the data, allowing them to react in minutes according to any failure prediction.

Cost reductions
One of the major benefits identified by both major PdM providers that supervised the two pilot projects was S2S’ capacity to be installed in different machines without retrofitting them with different sensors, specifically tachometers. By integrating S2S' system for prediction of failures, our customers can to extend the guaranteed lifetime as well as reduce the maintenance cost on their installations. They also gain key insight on the fatigue and wear on specific parts, which will prove valuable for optimization of cranes and exchange of weak parts. According to the conclusions made by in the Maritime pilot, the marginal cost of installing S2S in one of its installations is negligible. However, the value for this PdM provider of each sold crane increases by approximately 20% in a more extensive service agreement and less guarantee-related expenses. In the case of the drilling pilot, the automation of the PdM analysis can improve its efficiency by 50%, which would save up to 1 million Euros annually with improved results. This benefit would be a major drive to double its current market share including S2S in their future product offerings.

This Feasibility Study confirmed the need for a more reliable and accurate predictive maintenance solution within heavy machinery intensive industries. The Phase 1 supported research and the preliminary testing with around 25 machine parts at two types of machinery have verified the usability, user need and the market gap that our product will be clearly serving. As a summary, we provide a comment made by the CEO of one of the companies who participated in the pilot projects: “We believe that the technology being developed by Signal Analysis Lab has strong market demand. With our current customers, we see strong value in adding the S2S predictive maintenance product to our portfolio and see great synergies with a potential investment in Signal Analysis Lab AS.”

s2s-pic2.png

s2s-pic3.png

s2s-pic1.png