The design and certification of a high performance recuperator for micro gas turbines is presented. The component has been developed and built for a 100kWel micro gas turbine. The recuperator heated up compressed air at 3.5 bar with exhaust gas near atmospheric pressure and recuperates 300 kWth at an effectiveness of 90%. This concept can readily be adapted for other micro gas turbines due to its modular design. The certification has been realized under Pressure Equipment Directive 97/23/EC, equivalent to ASME Boiler and Pressure Vessel Code, covering closed pressurized devices. However, minor leakage in the recuperator is allowed, thus requiring an inventive design and validation approach for meeting the certification requirements. This leak is caused by weld porosity: the heat exchanging core plates are laser welded, having over 1200 meters of sealing weld length in a single recuperator. The maximum allowable leak amounts to 3 10−6 mm2 per meter weld length. The maximum leak was 0.2% of the massflow on the pressurized side at the nominal operating point, and therefore did not adversely affect the effectiveness of the recuperator. The finite element calculations and the resulting design loops on components and weld connections are presented. Validation of the entire component is done under the Experimental Design Method. A hydrostatic pressure test at 8.4 bar and ambient temperature is executed in the presence of a certified notified body to demonstrate that the welds are sufficiently robust. This pressure is higher than the operating pressure to simulate the effect of temperature on the steel properties. A laser scanner is used to map the deformation of the unit under pressure and subsequently referenced to its original state. The maximum deviation measured is equal to 0.26 mm for the pressurized part, which is acceptable considering the size of the unit is 1000mm × 600mm × 1000mm. The strain levels went back to the values before putting the unit under pressure, indicating there are no residual deformations.
The test is further accompanied with leakage rate measurements before and after the hydrostatic pressure test. If the difference between these leakages rates is within limits, the recuperator will pass the test. The measured total leakage area is 0.4 mm2, well below the maximum allowable value, and equivalent to 0.01% of the massflow at the nominal operating point. This means the recuperator passed the test successfully. Furthermore, a burst test was executed to determine the safety factor and to identify the weakest element of the design. The burst pressure is observed at 18.3 bar, resulting in a safety margin of 218% and 523% in reference to the PED and operational design pressures, respectively. The component responsible for failure has been further optimized for the next generation of recuperators. Field data confirm that the lifetime of the high performance recuperator meets the requirements of 40.000 h operating time. Additionally, the traceability of the serial produced components is handled by the audited quality management system. It covers the used materials, including lot traceability, the measured process characteristics and welder certifications. The approach can also be used for ASME certification.
