TRACERCO Diagnostics™ - Flow Study
A steam turbine was used to compress gas prior to the reaction section of an ethylene oxide plant. The flow of process gas to the reactor was measured by an installed flow meter, with an output signal from the flow meter fed back to a control valve on the inlet steam line to the turbine. In this way the flow of steam to the turbine was controlled so as to maintain a constant throughput of process gas.
During the course of a routine maintenance shutdown, it was observed that the blades of the turbine had suffered significant metal losses. These were attributed to a combination of (a) corrosion, brought about by deposition on the blades of minute traces of sodium salts present in water entrained in the steam and (b) erosion caused by impingement of the entrained water on the corroded blades.
From the magnitude of the metal loss, it could be directly inferred that the efficiency of the turbine had been severely impaired and this in turn, implied that steam utilization must have been excessively high.
From the evidence in hand, it was possible to predict the rate at which the metal loss and therefore the loss of efficiency, had taken place. To measure, and take steps towards preventing loss of efficiency some means of measuring steam utilization was required. It was not possible to fit a conventional flow meter as the configuration of the steam lines inlet and exit the turbine did not offer an appropriate location for an installation. It was decided to employ Tracerco to undertake a programme of regular steam-flow measurements using the TRACERCO Diagnostics™ Flow Study technique, to study the performance of the turbine as a function of time.
The TRACERCO Diagnostics™ Flow Study technique helped to ensure that the plant manager ‘got it right first time’ - and every time. The regular measurements of efficiency enabled him to optimise the operating schedule of his plant, balancing temporary loss incurred by shutdown against savings from avoidance of steam losses. Additionally, the radioactive tracer measurements provided timely warning of the onset of efficiency deterioration so that corrective action could be taken before the turbine blades suffered serious, and possibly catastrophic damage.
The absence of appropriate pipe-runs meant that, a pulse-velocity technique was not feasible. It was therefore decided to employ a dilution flow method, injecting on the inlet line to the turbine, sampling on its exit, and relying upon the turbulence in the machine to produce the requisite good mixing. Since the steam was known to be partially wet, Tracerco decided that a tritium dilution technique would be most appropriate.
A metering pump was coupled to the steam inlet line via a pressure gauge tapping and was used to inject tritiated water at a constant and measured rate into the steam.
Samples were extracted from the steam exit line using a cooling coil in an ice bath to condense the steam. The sampling rate was approximately one per minute throughout the 16-minute duration of the injection. The samples were returned to the Tracerco laboratory and 10g aliquots of each were assayed using a liquid scintillation counter. Samples of the injected material, after appropriate dilutions were also assayed in a similar manner.
Typically, the accuracy of the steam flowrate measurement was+/- 1% or better. The first measurement was made approximately 4 months after the plant had been brought on line and, though data for the first four months were not available it could, nevertheless, be inferred from the measurements that steam utilization, for the first 10 months of operation was constant. Thereafter, steam utilization increased until, after 16 months on line it was judged to be appropriate to temporarily shut down the plant and bring on-line a standby turbine rather that to continue at greatly reduced steam efficiency.