Abstract: Aiming at the problem that single vacuum helium mass spectrometry (HMS) leak detection of ultrahigh vacuum (UHV) sealing device is susceptible to the uneven helium distribution, this paper proposes a method based on the optimal linear unbiased estimation algorithm for the detection of UHV sealing device faults. A vacuum HMS leak detector is designed and calibrated for the lowest detectable leak rate. The leak rate of the UHV sealing device was measured by the calibrated leak detector, and the optimal linear unbiased estimation algorithm was combined to integrate the leak rates measured by the leak detector at different moments with different helium distribution concentrations, to get the cumulative leak rate results with the lowest variance, and then to realize the detection of device leakage faults. The results of this method showed that the leakage rates of the UHV interlayer of two tanks, C3 and C6, were 2.16×10^-12 and 1.31×10^-12 Pa·m³/s, respectively, which exceed the leakage rate limit of ≤10^-12 Pa·m³/s, indicating that there were leakage failures in both tanks, and the leakage failure of C3 was more serious; the average relative error of the leakage detection results was only 0.05%, and the leakage failure of C3 was more serious. The average relative error of the leakage rate detection result of this method was only 0.054, and there was no obvious fluctuation, and the detection result was consistent with the actual situation. This method significantly improves the precision and reliability of small leakage detection, and provides highly reliable technical support for the performance evaluation and maintenance of UHV sealing devices.

Key words: linear unbiased estimation, ultra-high vacuum, sealing device, leakage fault detection, vacuum helium mass spectrometry