Vessel performance management projects for big fleets - with BSM and Columbia Shipmanagement

Video recording of Digital Ship VPO webinar hosted on April 22, 2021
Speakers:
- Frank Paleokrassas, Head of data governance and analytics, Bernhard Schulte Shipmanagement
- Pankaj Sharma, Manager - Columbia Control Room, Columbia Shipmanagement
Frank describes the challenges of managing performance data for such a large fleet, and some of the developments at the company. Frank is a naval architect and chartered engineer, and has been with the Schulte Group for the last 12 years. He previously worked as BSM’s Group Fleet Performance Manager and Bernhard Schulte’s Technical Manager.
Pankaj talks about a project to develop virtual torque meters, so using data analytics to work out what data a torque meter would be providing, avoiding spending $20,000 on expensive data collection and transmission systems . The analytics achieved 98 per cent accuracy, 95 per cent of the time, which is equivalent to what you would expect from an actual torque meter.
Upcoming webinars at www.webinars.thedigitalship.com

[@shahrozkhan4683]

In the second presentation, it seems to be a simple multiple-linear regression

[@hanshase410]

Dear Columbia-AI-Experts.
A. How do you consider the difference in lower caloric value? It is a calculated figure only from the bunkersample-report. It is not measured. Please ask VPS or an other fuel expert, they tell you. Any increase in consumption you cannot refer back via a properly measured SFOC as you lack a proper TM. You do not calculate SFOC, you only estimate it.
B. How do you consider ISO correction on M/E efficiency thus on M/E consumption? You could do it via a measured SFOC. Not by estimating.
C. Trim: On Containerships one faces power differences at same draft but different trim figures which depending on speed easily is equal to 5% in Pd. How you consider it?
D. M/E mechanical alterations between overhauls like wear
of fuel nozzles, fuel pumps, exhaust gas valves, proper engine timing, fuel ignition delay. How do you consider its effect on Fuel Consumption? You cannot calculate SFOC as you only estimate engine power Pe.
E. In case of you utilize shop test-
M/E SFOC: Did you realize that you fully trust into the correctness of shop test data? You know that engine makers are using their own water break and often do not calibrate it, LCV often is not accurate and there is sufficient failure potential that engine settings are not according to contracted SFOC tuning mode. This would be "healed" by the engine maker by playing on the shop test fuel meter.
I have strong doubts if your virtual TM works sufficiently correct to differentiate fuel consumtion increase on individual reasons like hull fouling or engine aging.
The only value is that Columbia would easier implement new ships in management into their management software. Plus ypu hope to suggest a small advantage to the ship owner (often pure economics, not engineers) of just saving a real TM costs, which is peanuts.
What value has the Charterer? Tgey have expectations, at least Maersk, CMA, MSC and Hapag.
Is your Virt-TM sufficient for trend analysis of hull fouling? I think it is not: accuracy it is worse than 90% due to the effects mentioned above.
For sure you ask some for money from your customer for the development of your system, even if it is part of shared costs.
As mentioned above please explain, I will lookback to this channel soon for an answer, curiously.
Kind regards, HH.

[@karljeffery8848]

Hans can you identify yourself please - what company are you from? I searched for your name on LinkedIn but did not find anything cheers Karl