| Being able to collecting the information of lithology and density of formation at atime, the litho-density logging tool is widely researched and applicatedamong thenumerous nuclear logging series.Thefull spectrum data acquisition system of nuclearpulse, which is different from traditional instrument, is designed for litho-densitylogging tool. Instead of collecting information based on enegy window, this systemdivides nuclear pulses into some levels according to their energy and createsa256spectrum,then delimits the energy windowaccording to the need. Thus, this systemnot only can reduce the error of energy window and improve the counting rate, but alsocan use the spectrum checking method to modify related parameters.In the designprocess of the system, the statistical fluctuation problem of fullspectrumaffects the accuracy of the logging scaledirectly. The pros and cons of digitalfiltering will cause a great impact to spectral data processing before spectrumforming.So, in order to geta more reliable and smooth spectrum, we need to choose anappropriate and effective digital filtering methodand then select the matching algorithm,which would eliminate the statistical fluctuation and smooth spectrum withoutdistortion. This paper mainly completed the design and implementation of thedigitalfiltering for full spectrum data acquisitionsystem of nuclear pulse.The primary researchworks are as follows:Firstly, in view of the statistical fluctuation problem of pulse spectrum, this paperanalysised the nuclear pulse processing principle of full spectrum data acquisitionsystem of nuclear pulse. Meanwhile,it analysised the system processes, such ascollection and conversion for thespectral data, in-depthly. Then it is concluded that thesystem belongs to a linear system.Secondly, in view of the specific way of the system processes for the spectral data,this paper analysed three kinds of data filterings in-depthly,including the depth-weighted average filter, adaptive filter and kalman filter. Analysisingand derivingtheories of these three data filtering, as well as comparing the corresponding MATLABsimulationdiagram, an appropriate digital filtering method was chosen to improve theimpact of the full spectrum of problem caused by statistical fluctuation. Finally, this paper analysed the selected kalman filter in-depthly throughmathematical modeling, programming andthen software simulation, which has provedthat kalman filter can effectively filter the noise in pulse signal andeliminatethestatistical fluctuation of pulse spectrum. It is realized that kalman filter is able toachieve a stable control of full spectrum data collection system of nuclear pulse for theobservation of the signal with noise. |
