Skip to main content

ADVANCED AND PREVENTIVE DIAGNOSTICS OF FLUID HANDLING ROTATING MACHINERY USING ARTIFICIAL ROTOR EXCITATION AND DYNAMIC PARAMETER IDENTIFICATION TECHNIQUES

Objective


Under task 1 a database was set-up, holding information on installed rotating equipment, failures and corrective actions taken. This database was used for generating specifications for future online diagnostic capabilities.

Task 2 comprised several subtasks. Under this task, new online identification techniques were developed, including inverse force and inverse pressure identification methods. Furthermore rotor crack detection and nonlinear rotor system identification techniques were investigated. This included the development of rotordynamic models suitable for simulating and predicting rotor dynamic behaviour.

Under task 3 potential-flow fluid models for motion dependent interaction forces have been developed. In particular, zeroth and first order models for high-specific-speed impellers have been built and compared to experimental data. The models predict the forces on rotating and whirling impellers, incorporating impeller tip/ volute flow interaction.

Task 4 was divided into two subtasks. The first subtask concerned hybrid bearing testing, in which rotordynamic coefficients of a hybrid bearing assembly were identified by means of a component tester suspended on active magnetic bearings. The second subtask dealt with the rotordynamic identification of coefficients of a high-specific-speed impeller in a so-called free-impeller set up, as well as in a real pump/volute configuration. Furthermore, pressure pulsation identification measurements were performed.

Task 5 focused on the development of exciter technology, which resulted into a conceptual design of a prototype artificial (portable) field exciter. Under task 6 an existing boiler-feed-pump test rig was modified to allow experimental verification. An important modification was the change of suspension of the active magnetic bearings, so that rotor-bearing forces could be measured directly by load cells. In addition, another indirect force measurement technique was installed based on flux sensing.

Task 7 hosted the experiments on the boiler feed pump. Both inverse force as inverse pressure identification methods were tested. Furthermore, rotor modal analysis techniques were investigated.

Under task 8 two different diagnostic schemes have been developed. The first scheme is appropriate to tackle problems with existing machines, whereas the second scheme is meant to be used for monitoring of newly installed units.

Task 9 reviewed the diagnostic capabilities. The experimental rotor modal analysis proved to be technically feasible, as was demonstrated by the tests on the boiler feed pump with active magnetic bearings. It is also capable to detect and quantify casing resonance. The method is sensitive to rotor defects such as increased wear-ring clearances. The inverse pressure transfer technique allows to identify a pump as an acoustic two-port.

Task 10 hosted the field experiments. A portable electro-magnetic artificial exciter was designed, build, and subsequently used on the booster and the main pump of a (power-station) boilerfeed pump train. It turned out to be very difficult to introduce artificial excitations into the shaft of the booster, and obtain corresponding vibration responses. For the main pump the excitations tests were, however, rather good. Frequency response functions showing resonance at two frequencies could be obtained using artificial rotor excitation.
The increased requirements for safe, undisturbed, and pollution free operation of chemical and petrochemical processing plants, off-shore installations, gas and oil pipeline networks, and conventional and nuclear power plants have urged the need for effective condition monitoring/diagnostic systems for the rotating equipment applied in these plants. Present 'state-of-the-art' condition monitoring and diagnostic systems are based on vibration signature verified against truth tables. Current systems are based on general 'domain expertise' and therefore do not account for individual characteristics of the machine and deliver multi-valued conclusions with a likelihood ranking. Especially because of the last aspect such a monitoring diagnostic system will not be the required firm basis for making maintenance decisions.

The objective of the proposed project is to develop model based interpretation methods, complementary to the existing signature analysis techniques, for the various malfunctions occurring in rotating equipment based on the quantitative detection of the dynamic phenomena inside the various parts of machines. Because most condition parameters cannot be measured directly, indirect measuring techniques will be developed and new system identification techniques will be applied. This requires the use of artificial rotor excitation. Also new in the project is development of mechanical seal monitoring and diagnostics techniques.

The model based interpretation methods will applicable to all types of rotating machinery. However this project will demonstrate the methods for fluid handling rotating machinery used in power generation and petrochemical industry.

Coordinator

BW/IP INTERNATIONAL BV
Address
Parallelweg,6
4878 AH Etten Leur
Netherlands

Participants (6)

Glacier RPB
United Kingdom
Address
Premier House 1 Canning Road
HA3 7TS Harrow
IBERDROLA SA
Spain
Address
Goya 4
28001 Madrid
NUCLENOR SA
Spain
Address
Hernan Cortes 26
39003 Santander
TNO
Netherlands
Address
Stieltjesweg 1
2600 AD Delft
UNIV TWENTE
Netherlands
Address
P.o. Box
7500 AE Enschede
Universität Kaiserslautern
Germany
Address
Erwin-schrödinger-straße
67663 Kaiserslautern