TS-122
When knowledge is safety, and money -the use of an engine failure database
Keith WILSON*, Jφrn DRAGSTED**' and Nico HANSEN***"
ABSTRACT
The CIMAC "Engine Users"' Working Group (WG), since its inception in 1993, has always had in its meetings, a session under the general title of "Members' Case Histories". This allows any User present to openly discuss his engine failures, around the table thus bringing the benefits of sharing problems and information and hence spreading knowledge within the WG. From these open discussions sprang the idea of listing those problems in a database which could be interrogated to find trends in failures with particular engine types, alert members about particular failures and show where the weak spots are in various engine designs. In addition, the reactions of the engine builders/designers and the methods of repair can be seen.
Key Words: Engine Users, Database, Engine Failures, Analysis, Repair
1. INTRODUCTION
The members of the CIMAC "Engine Users"' WG are almost all, the engine users and, it is the Users who have the best possible information of all on real engine performance, including failures. They have the first-hand experience at all times, whether their engines are operating in ships or power stations.
Through the session "Members' Case Histories" in each WG meeting, there is always a free exchange of information on engine operation, be it good or bad. By the 5th WG meeting, held late in 1995, the membership had grown from the original four, to 10 members and it is worth noting that today the membership totals 34 member companies spread throughout the world. At that 5th meeting, the Chairman suggested that all WG members could send data on major failures to each other so that all could be kept aware of common problems. That idea initiated a discussion in which it was suggested that a database of major faults and failures should be set up, with input coming from each member company.
With sufficient data compiled together, it was foreseen that common problems would be identified easily and could, eventually, be brought to the attention of the engine builders and in particular, the designers. By that it was seen that a bridge could be built directly leading from the operators to the designers and their operations departments. The data could also be used by the participating members as an early warning system to watch for a recurring problem or failure on any particular engine model.
One of the authors volunteered to make the initial design of the data program so that its input could be stored for analysis in a commonly used computer program such as Excel or Access. It was agreed that the resultant database should be completely confidential and, therefore, there would be no ship names or the names of ship operators/owners included. At the same time, it was agreed that the data input program would be given to every member but, only those who sent in data would be allowed access to the database. To date, more than half of the present members have contributed failure and major fault data to the system and more members are entering data to the input program in preparation for sending to the databank.
At the 6th meeting of the WG, held in the Spring of 1996, the first version of the input program was demonstrated to the members present. The program was designed to run under MS DOS, from a single diskette, and was easy to use. A few modifications were discussed and agreed before the program, known as CIMACUSE, was distributed to all members of the WG. At the next WG meeting, the Chairman was able to report that data had been received from seven members already, which totalled 258 engines (2-stroke and 4-stroke) which had 108 fault cases spread throughout them.
It was seen that any member who analysed the database for his own requirements but who noted a recurring problem on a particular engine type, could use the WG's "early warning system" to prewarn other members about that problem. Since the early warning system uses the members' E-mail addresses, all members can therefore benefit from the database.
