Abstract
The ring-shaped electrostatic sensor is a typical particulate flow measurement system, which suffers from flow profile dependency due to its non-uniform spatial sensitivity. In this paper, a method based on sensor's output power frequency spectrum is proposed, which can be realized by repeatedly decomposing the sensor's overall output power spectrum to identify elementary frequency spectra corresponding to equivalent flow streams. The decomposed components can then be re-weighed to improve the uniformity of spatial sensitivity. This is a low cost approach, and there is no need to modify simple structure of the current sensor.
The figure below shows an example of the power spectral responses due to flow streams in different radial paths and the resultant overall power spectrum noted as Overall 1. The grey graph represents the spectrum generated by the outmost layer of flow stream. The tail of Overall 1 effectively coincides with that of grey graph. Based on this phenomenon, from the tail of Overall 1, the grey graph can be identified. By stripping off the grey coloured component, a new tail will appear in a new Overall spectrum. This process is repeated until the innermost layer is identified. Each identified spectral component is then reweighted to achieve a uniform spatial sensitivity. [Display omitted]
•A flow of constant rate produces a smaller signal when passing pipe central line than travelling near pipe wall.•The proposed method is to tackle the above problem.•Experiments and simulation confirm the “tail” of the sensor’s output is dominated by the response to outmost layer flow.•Hence “Tail”, i.e. the high frequency part output power spectrum can be used to identify the outmost layer of flow.•Stripping off outmost layer, the next layer will be exposed.•By repeating the same procedure, until the innermost layer of the spectrum is identified.•By re-weighting the decomposed spectra, the same flow stream will achieve the same output wherever the stream passes.