Recent Announcements

Don't break your privacy on LinkedIn!

posted Jun 24, 2017, 12:25 PM by Niels Jensen   [ updated Jul 11, 2017, 11:49 PM ]

LinkedIn protect your privacy by not directly exposing your email adress in discussion groups such as HSEQ Professionals, which has more than 130,000 members currently. However, lately some group members have promoted rather simple excel templates they have created in this group. These post look like the one in the picture on the right (identifying information have been removed using GIMP). Group members are promised an Excel template for creating a funnel chart, if they respond to the post with their email adress. It appear, that most of the persons responding to the post replied with either a gmail or yahoo account.

However, generally LinkedIn is quite good at not showing people your email adress. But by posting your email adress in a response to  post, then you break the privacy, which LinkedIn is attempting to create. Similarly LinkedIn protect you from receiving messages form members you are not connected with.

If one google either "how to create funnel chart in excel" or "how to create funnel chart in google sheet" then one get links to numerous posts with instruction on how to create such charts. So I wonder if HSE professionals are so internet illiterate, that they are unable to find simple things on the internet, such as instructions on how to create certain charts? I how that is only the case for the minority 0.15% who responded to the post shown here.Here is a link to a 3 minute video on YouTube, which shows you how to make a funnel chart in Excell:

Tot the left is another of the same programmers contributions. An audit tracking chart. It shows open, closed and total audits per department, i.e. three barts for what could better be shown on a single stacked barchart. So again, you are better off creating your own charts for reporting HSE data, than giving this programmer your email.
Google+ Twitter Facebook Gmail

Copenhagen Open Source Community Day not professional enough

posted Mar 19, 2017, 3:00 PM by Niels Jensen   [ updated Jun 24, 2017, 12:27 PM ]

Yesterday Safepark participated in Copenhagen Open Source Community Day 2017. This years Community Day toke place at Metropol's Campus Nørrebro on Siggurdsgade 26 about one block north of Copenhagen University's Science City. A ticket to this years event cost you 200 DKK, and for that price you received admission to three tracks of lectures and a T-shirt with sponsor logos on the front and the event logo on the back. Last time we payed to participate in the community day was in 2011, and then a ticket cost you 250 DKK. However, at that time the ticket also included lunch. So it is easily arguable,  that the price is unchanged since six years ago, since yesterday we spent 50 DKK on lunch.

Missing Professionalism

When we arrived at the reception just before 10 AM things were a bit chaotic. I showed an enlarged version of the bar-code on the Ticketmaster ticket to the receptionist, and was told that the computer system was down. She then attempted to manipulate the display of my ticket while I was holding the phone in order to find my name in print on the ticket. Eventually she found my name, and then wrote my first-name by hand on the two sides of the A6-sized badge, folded the paper, and inserted it into a badge holder. While handing the badge holder to me, she asked what size T-shirt I needed. I didn't hear her asking for my name in either Danish or English, and why should we specify our T-shirt size on the registration information, when that was not used. The friend, who accompanied me, was treated differently by another receptionist. He just told the other receptionist his name, and then she search for an already printed badge with that name on the counter next to us. She found a badge with his name, and gave it to him.
To me this looks like a broken access control system. At past community days the bar-code on the event ticket was scanned electronically, and the automatically triggered printing of badge with name, wifi-information and ticket for T-shirt, ticket for lunch, etc. which were handed to the guest. Other volunteers handed badge holders to guests, and others again handed out T-shirts. I wonder what happened to the label printers used just two years.
It seemed like prior to this years event there had been no thoughts about the workflow in the reception area. That resulted in an image of lacking professionalism. Professionalism is something we have become used to at past community days in Copenhagen.

This impression was re-emphasized a few minutes later at the coat room. There were three people manning the coat room, but two of them were talking to each other without paying any attention to customers. So quickly there was a line up.

I can't avoid looking for safety issues, when I visit a new place. Thise sockets are located in the middle of tables at which students and other guests to the facility sit an eat their lunch and drink coffee or soft drinks. There is no edge between the table surface and the sockets surroundings. I looks like any spilled liquid would run directly into the plugs. It was tempting to pour some water into the sockets, just to see what would happen. However, I leave that experiment for the next curious visitor to Metropol.

Community Day Program

The days program consisted of 3 parallel tracks starting with a keynote at the opening of the conference at 10 AM and another keynote at the closing of the conference at 5 PM. The rest of the day there were 3 parallel tracks and after each track there were a 20 minute break. So there were amble time to visit the exhibition area for a chat. However unfortunately the track host failed to remind the guests, that they should visit the exhibition area and talk to the vendors.

As far as we remember past community days had the organisation leader open the event, by introducing her or himself, thanking sponsors, and volunteers, and welcoming attendees. I didn't see this happening at this years event. I think this is unfortunate for the OSD image.

Most presentation slots were 40 minutes long, which to us indicate, that there is sufficient time to give a well thought out and well structured presentation. Unfortunately too many presenters have not thought about what the used should take away from the presentation or what actions they wanted the attendees to take after they got home. This meant, that many failed to promote what the had to offer. Here is the community day program with the presentations, which we choose to attend, shown in orange:

Open Source Days 2017 Community Day Program

I opening keynote talk was titled "How Danish municipalities use open source to drive development and cooperation and cooperation" by Rasmus Frey, who is the daily coordinator for this cooperation about development of open source solutions for municipalities. The speaker tool us what the top level technologies were, i.e. JIRA, Github, www and OS2Cloud. JIRA is a service desk developed and marketed by Atlassian. I believe, that OS2 is using JIRA Software for project and issue tracking. You can try it for free, but the software does not appear to be open source. Github is free for public and open source project, but also offer storage for others for a fee. OS2 currently have 12 product, which can be freely download from their Github storage, and they have 8 active project. Unfortunately absolute no screen shots of any of the products were shown in the keynote, and no information were given about the underlying technology, e.g. that os2dagsorden use Drupal. I mention os2dagsorden because this sound like a product, which could be of use to many small organisations, such as e.g. patient support groups, and developers could be insterested in making a Typo3 version of this facility. I think the presenter should have thought a bit more about what the audience take away from the keynote should be, e.g. this type application looks like something my organisation could also benefit from, and secondly what actions he encourage people to take for a better understanding of OS2 and their products and projects, e.g. take a look at os2indberetning if your organisation have many employees driving around the country as sales representatives. Again os2indberetning looks to be something other organisations could also use, and hence open source developers could be interested in developing a similar product for use by private companies. Caveat: I have as of this writing not done any work to see if the mentioned two products already have equivalents in the open source community.
The second part of the mornings keynote was by titled "Developing open source geospatial software in a governmental agency" by Kristian Evers. Basically he mentioned three programs developed by the organisation, which he is part of DHMQC, MALSTROEM and as well as use of PROJ.4. We did see some screenshots in this part of the talk, as well as pictures of output. However, there seem to be a lag of how can the audience as open source developers take the products devleoped by Evers group and extend them to moneymaking private enterprise products?
We believe it would have been better if all the time of the keynote was used by the first presenter, and the second presenter was given a 20 minute slot in track 2 or 3 at another time during the day.

After the keynote we attended Cornelius Kölbel's presentation "Tow factor authentication with open source project privacyIDEA". Why should I go through the trouble of deploying this open source system and maintaining it in stead of relying on a third party service. Neither was the question "Is someone providing privayIDEA as a service?" that I can subscribe to just like VPN-services. Clearly Cornelius knows the technology, but after the presentation I was thinking: Why should we spend time on this and not just use Google's services?

The last presentation we heard before lunch was titled "Is Linux Desktop security worse than Windows 10?" by Hanno. He talked about using technology like ASLR and PIE, which have been in the Linux kernel since version 2.6, to avoid that programs run in the same address space every time. This will make it more difficult to exploit automatic downloads on access to a site. He mentioned, that he did not know how far openSUSE was with this fix. However, today I discovered, that the chair of openSUSE responded to this question in February by stating, that a lot of work had already been done, but that it is unfortunately not complete yet, and then promissed, that he would speed things up.

After lunch we first attended Bo Simonson's presentation "Kopano - sharing &  communication software for business". Kapano is basically an open source groupware tool forked from Zarafa. Unfortunately Bo failed to convince me to try Kopano, because at several points during the presentation and quenstions period, he indicated there were features in Kopano, which he did not use. That is not a good response, when you are attempting to get people to try something new and different. The second presentation after lunch was "(Lack of) Cryptographic Security of NemID, Dankort, and Mobilepay" by Thue Janus Kristensen. Thue had a point about the NemID login screen not indicating, that your attached to Nets' servers during the login process, and there is no indication of this on the screen. Unlike the Google Login popup. However, we are not certain why Dankort and MobilePay are mentioned in the title, since they are quite different technologies from NemID, and I don't really see the connection between the three. Thue unfortunately did not have content for a 20 minutes presentation, and his statement was unsubstantated and taken apart by the audience.

The community day finished by two excellent presentations. First Georg Koppen's "Surfing the Web over Tor with Tor Browser", during which we learned that privacy technology from the Tor Browser is being implement in the main Firefox Browser, which the Tow Browser is based on. This has already happened with code in version 52 of the Firefox Browser, and is also planned for version 59 of the browser. The Tor Browser allows you to surf the internet through the Tor network of servers, so all the destination can see about you is, that you came from the Tor network, but not which computer you entered the Too network with. Naturally such technology can be misused with criminal intend. Georg also told us, that a Tor Browser for Android is under development.

The final talk of this open source day was not about software, but about hardware. Alicia Gibb gave an excellent presentation about "Open Source Hardware".

Value for money?

Overall it appears, the organizers were less prepared than at past Open Source Community Days in Copenhagen. Maybe that reflected on the speakers, also appeared less prepared, than at past events. If OSD is to survive, then both organizers and speakers need to come to the community day better prepared. If the Business Day was of similar quality, then it would have been a disaster.

There was also to be a speakers corner for impromptu talks, but without any slide presentations. However, there were no pre-announced talks, as at previous community days, and there were no facility to show slides. That makes technical talks a bit of a challenge. Although most of us coped with that 40 years ago.

It appeared as if the people at the reception and the coat room were hired as a group, and the organizers didn't have volunteers for these positions. We noticed quite few of the leaders and organizers from past community days.

A decision to attend next year is dependent on better speakers, better organizers and organisation, and better facilities for impromptu talks (speakers corner).
Google+ Twitter Facebook Gmail PrintFriendly

What can we learn from the "human error" during the Oscars?

posted Mar 5, 2017, 6:21 AM by Niels Jensen   [ updated Mar 22, 2017, 3:43 AM ]

At the Oscars 2017 the wrong movie was at first declared a winner of the best movie award. That turned out to be a mistake. Others would properly call it a human error, because the wrong envelope was handed to one to the two host on the stage in Hollywood. If you are in the latter group, then I suggest using a few minutes to watch a video from Lund University's Human Factors and System Safety department here.

First learning from the Oscars is that, what happened at the Oscars was NOT a human error. I would classify it as an envelope design error. However, one could also argue, that it was a envelope circulation control error. Maybe it was actually the circulation system, which did not work this time around.

In his blog Steven Shorrock points out, that what happened at The Oscars was essentially the start of the discipline human factors and ergonomics during WWII. More relevant however, is that the design of the envelopes, were such, that it is a surprise, that this has not happened before in the history of The Oscars. On the outside all the envelopes at the Oscars were completely identical. This means if was completely up to the assistancts from PriceWaterhouseCooper to ensure that the correct envelope was handed to the host on the stage. Given the number of envelopes and the fact, that duplicates of each were available the likelihood of a wrong envelope being handed to a host can easily be calculated.

Second learning from the Oscars is that, handing information from one person to another should be analyzed as a transportation system, where items can e.g. be wrongly delivered. Although the Oscars considered something could happen to the host on stage, their risk assessment did not consider the transportation risk involved in handling the envelopes. 

As Steven writes: "Experience of human factors suggests a number of coding methods, e.g. shape, colour, size, that used appropriately, can help to make vital distinctions.", and points out that within the pharma industry both the European Medicines and the UK's National Health Service have developed guidelines for design for patient safety of medication packaging, attached below. 

From the outside it appears, the stage host at the Oscars did in some way realize, that he did not have the correct envelope. Since this situation was properly not covered in the rehearsal for the Oscars, the host became uncertain about what to do. 

Third learning from the Oscars is that, you should train people on abnormal situation handling. 

Steven finnish by writing that, for the post part the human in the system is less like a golden Oscar, and more like someone using the abilities of mind and body to connect parts of a system that only work because people make them work. This aspect of human performance in the wild is usually taken for granted. But in the real world, peopel create safety. And for that, they deserve an Oscar.
Google+ Twitter Facebook Gmail PrintFriendly

Could this hazard have been discovered if a computer assisted HAZOP was performed?

posted Jan 28, 2017, 12:00 PM by Niels Jensen   [ updated Mar 22, 2017, 3:32 AM ]

Blocked In - Williams Olefins Plant

Safepark certainly believe the answer to this quesiton is YES. The reason is that MFM models of the unit can easily be developed, and that such models allow the reasoning about causes and consequences of deviations, the results of which would assist HAZOP teams during either a MoC or HAZOP study. A simple model could be developed with less than a days work by an expert in MFM model development.

The video to the left was recently released by the #CSB. A small design change that was performed at the Willams Olefins Plant many years ago to allow a distillation tower to run with just one reboiler instead of two in parallel resulted in 2013 in an #overpressurization and rupture of one of the reboilers. The resulting explosion killed two workers and injure many more. A copy of the CSB Case Study can be downloaded from this page. The above pictures are curtesy of CSB.

The design change resulted in it being possible to isolated the reboiler from its protective safety valve. This hazards was not discovered during a MoC or 3 later PHAs. That different teams of experts did not identify the hazard which caused the explosion in 2013.

I would be relatively easy to construct a functional model using multilevel flow modelling (MFM) to investigate possible overpressure scenarios, and hence identify the hazard, which humans overlooked. Safepark is involved with a Danish research group at DTU, which is working to make computer assisted HAZOP a reality for situations, such as this one. 
Google+ Twitter Facebook Gmail PrintFriendly

Challenges in editing mp4-videos under openSUSE Leap 42.2 solved

posted Jan 8, 2017, 10:10 AM by Niels Jensen   [ updated Jan 8, 2017, 10:11 AM ]

About a year ago I inherited my sons Nexus 6 phone, and during this first year I have been quite pleased with it. I can record mp4-videos with excellent sound and picture quality. However the challenge start when I want to edit one of thee videos in openSUSE Leap 42.2. My preferred tools are ffmpeg and openshot. I use ffmpeg for croping, and then openshot for adding an opening picture and ending credits. This has for years worked fine with videos recorded with my Canon EOS 550D camera, which creates mov-videos, that are easily converted to other formats using ffmpeg. This is more abotu knowing your OS and the software installed on it, than anything wrong with the software or the error messages it provides.

Increasing frustration with ffmpeg

No such look with mp4-videeos. Or any other formats, which are containers for mp4. However, I was able to play the mp4-videos in smPlayer. While attempting to solve the issue I tried different output files to ffmpeg, without what demuxing and muxing capabilities my installed version of ffmpeg had. For example the command "ffmpeg -i SK.mp4 SK.flv", gave the error message "[NULL @ 0x798ea0] Unable to find a suitable output format for 'SK.flv' " followed by "SK.flv: Invalid argument". Later inspection of the output from "ffmpeg -formats" reveiled, that ffmpeg was correct. I even googled the message "Unable to find a suitable output format", and discovered that others encountered this message. But I did not find anything pointing towards a solution. So the learning here is: Check that the installed version of the software you are using is able to do, what you want to do!

Since I was pressed for time I turned to VLC. This video player also have some editing capabities, e.g. specification of cropping parameters before convesion to another format. I chose the ogg-format. This resulted in a playable file, but with somewhat reduced picture quality compared to what I was used to from ffmpeg.

What I should have done in the first place

Turns out the trick is to go to the openSUSE Community webpage at and use the 1-click installer for either KDE or GNOME. If things don't work after installing the codecs, make sure all your multimedia packages are coming from Packman. Go to YaST Software Managment, click "View", click "Repositories", select the Packman Repository, click on "Switch system packages". This will replace already installed packages with packages from the Packman repository and also install any packages your are missing. On our system we were missing 8 packages and 36 packages were replaced. And as usuall with linux no restart is needed. Both in openShot and in ffmpeg.

At first I had followed the instructions at "Unoffical Guide to openSUSE Leap 42.2 13. Multimedia Codecs" at but overlloked the small print, which states "Afterwards make sure all your multimedia packages are coming from the Packman Repository: Start "YaST Software Management", then click on "View" and select "Rositories". Then select the Packman Repository and click on the buttom "Switch system packages" at the top of the list, and finally click on "Accept" to install the pacakges from the Packman Repository.

The legal reason for my problems is that openSUSE Leap 42.2 has to comply with the DMCA. I think that is this case the issue is the difference between "free to use" - properly for non-commercial perposes - and "free to distribute". 

Discovered a new feature of convert - the linux commandline workhorse!

posted Nov 29, 2016, 1:06 PM by Niels Jensen   [ updated Dec 6, 2016, 11:12 PM ]

For years we have been using pdfsam to manipulate pdf-file, such as merge, rotate, extract etc. Today, I had to manipulate af pdf-file, which would not open in pdfsam. A quick Google search indicated, that convert could be used to extract parts of a pdf-file. The first attempt "convert file.pdf[48-52] newfile.pdf" worked, but the quality of the resulting pdf-file was rather poor. So we decided to use the density parameter in the extraction, and "convert -density 300x300 file.pdf[48-52] newfile.pdf produced an excellent extract. 

Linux commandline utilities, may take some learning, but in many situations the learning pay off. Pdfsam will likely still be our choice if individual pages have to be taken from different pdf-files and merged into a new file with some rotations.

Exciting day with PhD students from DTU Chemistry

posted Nov 11, 2016, 4:40 AM by Niels Jensen   [ updated Nov 11, 2016, 12:26 PM ]

Yesterday DTU Chemistry had their 6th annual PhD Symposium at Pharmakon in Hillerød about 45 minutes drive northwest of Copenhagen. It was an exciting day with 13 oral presentation and 27 poster presentations ranging from reservoir simulation to drug discovery. There was also a guest speaker from Zealand Pharma. DTU is the largest engineering university in Northern Europe, and DTU Chemistry is their integrated inorganic chemistry, organic chemistry and physical chemistry with activities ranging from oil reservoir simulation to synthesis of new cancer drug candidates. This fall 21 new PhD students joined the department. Department chairman is Erling H. Stenby, who for many years was the leader of the IVC-SEP group at DTU Chemical Engineering. Professor Stenby moved to DTU Chemistry about six years ago. The PhD Symposium was arranged and hosted by the ChemClub - at club for PhD students at DTU Chemistry.

The day started with a presentation on simultaneous calculation of chemical and phase equilibria of closed systems by Christos Tsanas. Christos' algorithm used Lagrange multipliers for minimization of Gibbs energy, and involve two steps. In the first step the total numbers of moles and the number of phases are fixed, and a subset of equations solved for the Lagrange multipliers.In the second step these Lagrange multipliers are used to solve the complete set of equations. The the stability of the phase(s) are tested, and phases are added if the system is found to be unstable, and the calculations repeated. Christos have successfully tested the algorithm on VL and VLL equilibria of reactive systems. Among the systems tested were MTBE synthesis (I believe the compound was no longer relevant?) and alkylation of m-Xylene to separate it using distillation. The final oral presentation by Duncan Paterson on flash calculations in thermal reservoir simulation solved a similar problem, but did not account for reactions. Duncan's work allows the use flash calculation is reservoir simulation of heavy oil deposit in Canada were steam injection is used to make the oil moveable, such as found e.g. at Cold Lake and Kearl.

Esben Thomas gave a very interesting presentation on deracemization using shaped-pulse dynamic stark control. His simulation involved separation of isomers of 3,5-difluoro-3',5'-dibromobipenyl. The associated experiments was done in collaboration with Aarhus University, and can result in purity of around 90%. Rouzana Thumbayil presented work on development of selective catalysis with metal nanoparticles in porous materials for one step production of MIBK. To my surprise I learned, that MIBK is increasing used in paints, with annual production increasing by10% to currently around 380,000 tons MIBK. In the eighties and nineties MIBK was widely used for washing printed circuit boards, and it was discovered MIBK has a negative effect on pregnant woman. Rouzana catalyst is able to produce MIBK in one step from acetone and hydrogen, and she is comparing nanoparticles with Palladiun, Platin and Gold.The figure to the left shows a TEM of zeolite containing encapsulated gold nanoparticles. Nedjelko Seselj also presented catalyst related resutls, but related to fuel cells and using a graphene supported gold and platinum catalyst.

Arun Kumar talked about understanding the role of structural and chemical properties of the membrane protein presenilin-1 (PSEN1) in Alzheimer's disease. Almost 200 mutations in PSEN1 are believed to cause early-onset of Alzheimer's disease. His work is however, purely molecular dynamic simulations.Gianluca Levi used molecular simulation in direct dynamics studies of transition metal complexes for solar energy conversion. The work done in collaboration with the Department of Physics use the X-ray Free Electron Laser facility at Stanford to study reaction on femto-second time-scales.Peter Jakobsen gave an interesting presentation on use of ionic liquids for NOx removal - not in powerplants, but in other situations. The diagram to the right shows a catalytic cycle for the oxidation of NO in ionic liquids.

Arnab Halder talked about his bioengineered graphene based flexible biosensing platform, which he has developed for measuring blood glucose and cholesterol. The current versions are printed on a flexible substrate for use on the skin. The developed sensors had a linear response in the ranges of interest for medical usage, and have been sucessfully compared to hospital laboratory measurements.

During the afternoon three medical related results were presented. Jorge Peiro talked about prodrugs sensitive to reactive oxygen species for the treatment of chronic inflammatory diseases and cancer. Prodrugs are compounds, which are converted to drugs at the action site inside the cells. Hence they don't have an effect on healthy cells. Kim Motensen talked about high-throughput screening technology for profiling of substrates for histone deacetylase enzymes in rational design of selective HDAC inhibitors. Finally Christian Kjeldsen talked about dynamic nuclear polarization NMR of enzymatic carbohydrate conversion. Equipment has been installed both at Skejby and Rigshospitalet, and clinical application will start early next year. The diagram on the lift shows spectra of a sample containing pyruvate-1 labelled with C-13. The top spectrum is 2048 scans without hyperpolarization, and the bottom a single scan after hyperpolarization.

The poster presentations were dominated by medical related topics, such as Alina Kulakova's poster on "Protein-Excipient Interactions and Protein-Protein Internation in Formulation" or Gokce Engudar's poster on "Development of Transmembrane Ammonium Sulfate and pH Gradient Liposomes for Combined PET/CT Imaging", and by catalysis related topics, such as Irene Tosi's poster on "Zeolites Catalysts for the Hydrolysis of Glycosidic Bonds" or Bo Jessen's poster on "Developments of Catalytic Reactions to Prepare Bio-Based Polymer Building Blocks". The PhD students at DTU Chemistry don't shy away form using advanced technology in their presentations, such as a video clip in an oral presentation or a vidoe presentation as part of a poster.

IBM Business Connect in Copenhagen focused on Cognitive Computing

posted Oct 28, 2016, 12:36 AM by Niels Jensen   [ updated Oct 28, 2016, 12:56 AM ]

Yesterday Safepark attended IBM Business Connect 2016 (#IBMBCDK) at the Tivoli Hotel &Congress Center in Copenhagen. We there partly due to Niels Jensen's involvemen with Lyle the Patient Support Group for Lymphoma, Leukemia and MDS in Copenhagen and partly due to our involvement with a intelligent operator support system being developed in collaboration between a research group at DTU and a Norwegian company involved in off-shore platform control system development. It turned out to be well worth our time.

Earllier this year IBM's CEO announced at the World Health Care Congress a partnership between IBM and American Cancer Society to provide personalized support for cancer patients and their relatives using the capabilities of IBM Watson (read the announcement here.). We would dearly want Kræftens Bekæmpelse (Danish Cancer Society) enter into a similar collaboration with IBM Denmark. Unfortunately this requries Watson to speak Danish, and this is currently not one of IBM Watson's capabilities. However, thanks to GapGemini in Norway we know know, that it took about 4 months for IBM Watson to learn Norwegian, and if the recently announced IBM Innovation Center in Copenhagen is to be a success, then IBM Watson must speak the local language - I think. Well, that was a sidestep.

Already the opening keynote feature two presentations from IBM customers in Denmark, who are already using the cognitive capabilities of IBM Watson in the businesses. The first were Novo Nordisk who in the new world of IoTT have to cope with data not from 1000 or 2000 patients involved in a clinical trial, but in near future with data from millions of connected diabetic patients. These data most be handled securely, and the data most be turned into insight to help the patients. The second were ISS, who wont to use IoTT to provide facility management for building. As an introductory case and - I guess - a demonstration windows the company's headquarters in Copenhagen have been transformed using IoTT. As an example can the cafeteria manager in real time monitor how many clean plates are left, and a reservation of a meeting room in which no movement is detected can automatically be chancelled a certain time after the meeting should have started.

Following the keynote there were two times two parallel sessions. One was about product development, when data are the new co-worker. In this session a CBS professor talked about their involvement with the Roskilde Festival and how they used data from mobile phones to plan logistics of everyting, from when to be ready to serve 2000 pork sandwiches an hour to when to empty toilet containers. In the same session another company explained how they used wheather data to close sewer lines during major rainstorms to avoid sewer water in basements. This technology was originally developed for single family phones, but are now being adapted to apartment buildings. Other sessions during this part of the day focussed on cognitive technology as a catalysator, 2. generation digitalisation and technology & innovation.

The afternoon featured 2 tracks of up to nine parallel sessions. One was cognitive business of which a large part was health care. Here we learned about a recent partnership between IBM and Tekes in Finland to apply the IBM Watson technology to completely change health care in that country over the next few years. We asked if IBM Watson had been used to evaluate medical images, such as X-ray scans, CT-scans and MRI-scans, and got a rather weak answer. After the session we learned from another IBM co-worker, that IBM Watson is being trained to evaluate X-ray scan, which are simpler than the other types of medical images. We also attended a session hosted by BP3 Global on Smarter Process in which they cleverly had create a complaint handling system with the aid of IBM Watson in order to get us started thinking about the ethical issues and legal issues in employing cognitive technology.

The afternoon ended by a talk by Cyborg Neil Harbisson, who has an antenna connected to his brain. The implant convert colors to frequencies of different sounds.The technology actually allows him to perceive both infrared and ultraviolet areas of the spectrum.The day was wrapped up by wrap-upper Per Vers with a ten minute tour de force of what we had experienced during the day.

"Big Data" - the latest IT buss word or a business tool?

posted Sep 16, 2016, 8:32 AM by Niels Jensen

Today Safepark attended a morning seminar at Scion DTU title "Big Data for professionals" in order to get some feeling for what this latest IT buss word really mean. We got away somewhat disappointed - but maybe we are not professional enough to get the message.One of the first slides (see picture to the left) gave the following definition of Big Data: "The practices and technology that close the gab between the data available and the ability to turn that data into business insight". That was wonderful! Now we know, that Big Data consist of two things: practices (skills) and technology. 

Unfortunately the focus of the presentation was on examples of the practices and technology you could buy as a service from vendors such as IBM. We had expected - given the location of the presentation - that we would get away with some knowledge of the underlying technology, such as Hadoop to handle large datasets, and the statistical tool "R" to get answers to specific questions our of the data. We did get some information about in memory databases, such as SAP Hana and Microsoft SQL, but no information about open source alternatives such as Redis, SQLite or UnQLite for capital constraint startups. We also heard about the business model behind IBM Watson, and some examples of its capabilities. It was mentioned, that IBM Watson could act as an intelligent assistant for a doctor - or if they don't want such an assistant - as a for-a-fee cloud services for individuals. We look forward to hear more about the capabilities of IBM Watson at IBM Business Connect next month.

At the end of today's presentation Scion DTU announced, that they have follow-up arrangement on November 2nd in Søhuset at Scion DTU in Hørsholm. As of this writing this event is not on their homepage.

Advanced decision support requires more than display of information

posted Aug 12, 2016, 8:47 AM by Niels Jensen   [ updated Aug 16, 2016, 2:18 AM ]

Decision Support
At Safepark we believe, that a decision support aimed at operators in chemical plants and refineries is not just about the display of situation relevant information for the process operators, such as the attached brochures about Foxboro's Evo and GE's Smart Operator appear to suggest. It requires a means for accessing the situation, that the operator is in. And that requires a model of how the process is working and interacting with the automatic control and safety systems, Such a model should not be a simulation model, because in a critical situation, there is little time to have people run simulations and analyse what the mean before an advice is given to the operator. It stead we need a model, which can be used for reasoning about current process situation using live data from the running process in order to suggest actions, which the operator can immediately review and act on. 

The necessary models should reflect the functionality of the process and its automatic control and safety systems.Such models can be developed using Multilevel Flow Modeling (MFM) and good process operating knowledge. The MFM models, which are based on fundamental theories of action, have the ability to exhibit the same functional behavior as the actual process and take into account current process status - including trends - arrive at a pertinent advise to the operator.

1-10 of 101