Oil & gas
AcAdemy lAunched to optimise uK industry compliAnce
T
ÜV SÜD has launched its Academy to help UK organisations cost effectively upskill their employees to
optimise business performance, improve product compliance efficiency and minimise time to market. Part of a global training programme,
TÜV SÜD Academy will offer more than 60 knowledge transfer courses to individuals and entire organisations. A mix of instructor-led workshops and e-learning courses will aim to maximise flexibility for how and where businesses want their people to learn.
In TÜV SÜD NEL’s previous studies, machine
learning models were typically trained and validated using data generated from a single flow meter. Although the models usually performed well with high prediction accuracy, it was challenging to apply the trained models to other meters due to the fact that the models were trained on variables that were unique to a particular flow meter. In other words, the training sample space was restricted. Since in this case the meter diagnostic data was not logged during the original experiments, it provided an opportunity to explore the potential of building a more generic model that is trained based on a wider sample space involving data coming from multiple meter types. Despite having no diagnostic data available,
high prediction accuracy (97 to 99 per cent) and high sensitivity rates (0.96 - 1), accompanied with low false positive rates (0 - 0.07), were still obtained, by successfully detecting when a set of data was gathered under normal operating condition vs. closed gate valve condition. These research results highlight the potential of making better use of data coming from different meters in order to expand the sample space to allow for a more generalised model to be trained and built, which would be capable of predicting specific operating conditions.
Instrumentation Monthly March 2022
Certain data can be expensive to collect, for
example erosive flow data, but a reliable machine learning model often requires a large sample data in order for it to capture all hidden patterns and correlations in multiple variables. As a result, if a model can be trained using a variety of data from different sources, it would maximise the potential value in data and prevent expensive data from being wasted. It seems that a common requirement from
end-users is ‘plug and play’. However, this is extremely challenging as models are typically bespoke to specific customer requirements/ devices. By having generalisable models, faster deployment/automation can be deployed with less requirement for fine tuning to specific meter type/manufacturer/model numbers as well as improving the fault diagnosis process, enable condition-based monitoring and minimise any unexpected downtime. Future research within TÜV SÜD National Engineering Laboratory will continue to explore the potential and capability of generalising data-driven models to be applied in wider applications.
TÜV SÜD National Engineering Laboratory
www.tuvsud.com/en-gb/nel
Training courses will cover eleven core
areas: Automotive; CE & UKCA certification; Chemical & process; Clean energy; Cybersecurity & data protection; Electrical safety; EMC; Healthcare & medical; Machinery safety; Quality management; and Real estate. David Goodfellow head of Business
Assurance at TÜV SÜD, says: “The experts in our global business have such a wealth of experience. UK organisations can now access that knowledge through TÜV SÜD Academy to help them increase business efficiency at multiple levels. We are dedicated to helping people and organisations drive business performance to levels of excellence, without compromising on costs efficiency or time to market. The TÜV SÜD Academy enables us to extend that mission through direct knowledge transfer into UK industry.”
TÜV SÜD
www.tuvsud.com
27
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74
