Research On Modeling Methods Of Information Measuring System Based On Entropy

With the continuous development of information science, the significance of theviewpoint “the essence of measurement is information acquisition” has been cognized byresearchers of measurement technique, and the ways about studing on the measurementprocess based on the basic views of information theory are also highly concerned. In thisthesis, the basic structure of measurement system, which view the information source asan important component, are modeled, analyzed and evaluated by the entropy set. In orderto constructe a theoretical framework and system of information measuring system, whichnot only provides principle guiding, but also be used to analyse and calculate the practicalaspects of information measuring system. The following aspects are the main researchcontents of this thesis.The basic connotation and basic law of information measuring system have beenstudied based on the guidance of the viewpoint “the essence of measurement isinformation acquisition”. A relatively perfect and unified model of system has beenestablished, with the information entropy balance principle and the entropy reductionprinciple of measurement process in the domain of entropy, which reveal the basiccharacteristics of information acquisition.According to the purposes of measurement, the description of the information source isundoubtedly essential, and seeking a unified source-modeling approach has been theprimary issue that plagued the researchers. A more effective international report has yet tobe seen that a viable way to resolve the issues under the concept of entropy has beencarved out. This thesis applied the thermodynamics second law, thus the informationsource is considered as an closed system under the specific conditions, to describe theenergy exchange and dissipation process of the information source by the non-balancestate thermodynamics entropy rate balance equation, and take the entropy produced rate asthe output parameter of the information source. The consideration not only guaranteing theunified description between the information source and system in entropy meaning, butalso laying the foundation of realizing unified modeling for all kinds of information source.Based on the above model and taking into account the actual interference factors in engineering, this thesis analysis the mechanism of noise existed in the system, and putforward an improved model, which taking into account the noise source. Considering theissues of the match aroused out by the energy exchange between the source and the sensor,then a joint modeling approach is presented, with which the source and the sensor, thegain circuit are looked as an organic whole.Based on the purpose of description, analysis and evaluation of the informationmeasurement system, the conception of errors entropy, noise entropy, entropy gain aredefined in this paper, the relationship between which and the input and output entropy arededuced and proved. And entropy model of several typical units in measuring system arebuilt, which provides a basic platform for various composition having different functionsto analysis, operation and structure the equivalent entropy modeling through the entropyset.The characterization and evaluation of error are important aspects of informationmeasuring system, this paper defined the performance evaluation of error entropy andrelated indicators, with the relationships about these performance indicators are analysed,and the theorem of error entropy transmiting, the theorem of error entropy synthesis andtheorem of repeat measurement are also deduced and proved. Meanwhile, the curve fitting,principal component analysis, filtering algorithms are descripted and analyzed withentropy, which are simulated based on the NMR logging data in order to proved theeffectiveness of these algorithms entropy.The sixth chapter of the paper mainly discussed the information source that is the coreproblem of nuclear magnetic resonance logging with the conception of entropy set, whichis viewed as the important application of the information measuring system.Under theCPMG pulse excitation, effectively stimulated the Proton particle behavior and time seriesof the spin echo has been described and analysed by the method of entropy change, theentropy model of macroscopic magnetization vector in the longitudinal and transverserelaxation process are constructed respectively. These models not only obtain the sameresults of description with the nuclear physics, but also get the characteristics of entropychange in the hydrogen nuclei system, which also reveal the nature of energy dissipativeabout information sources in magnetic resonance system, and also provide an exampl formodeling information source in such stimulated principle system. Meanwhile, the entropy change process of induced polarization log in a single cycle are analyzed, the samplingarea of the best information obatining in the process of induced polarization log is alsostudied, the conclusion is that the information acquisition way is to build an energy naturedissipative space-time and measure the entropy production rate in this disspation process.Finally, this thesis discusses the next research directions of information measurementsystem with the summaries of these main works.