ABSTRACT
Objective To explore the application of three parameter identification methods (impedance modulus curve method, impedance component method, and genetic algorithm) in solving parameter identification problem of the 11-element lumped parameter model in the circle of Willis. Methods Using the flow and pressure waveforms of the internal carotid arteries and vertebral arteries on both sides as inlet conditions, parameter values of the model under normal and bilateral vertebral artery stenosis conditions were calculated. The recognition algorithm was verified by using Simulink models, and finally the stability of the recognition algorithm was verified by adding a certain noise to the flow. Results Under normal circumstances, the proximal resistances obtained by the impedance modulus curve method were larger, and the resistances of the anterior communicating artery obtained by the impedance component method were larger. The genetic algorithm could obtain relatively reasonable model parameter values. In the case of vertebral artery stenosis on both sides, the impedance modulus curve method could obviously get the results of the increasement in proximal resistances of the posterior circulation, but the results obtained by the impedance component method and the genetic algorithm mainly lied in that the distal resistance had a larger increase. Conclusions There are still differences between the pressure data calculated by the parameters identified by the above three methods and the actual data, which are considered as modeling errors, source data errors and calculation errors. The impedance modulus curve method has a certain effect in distinguishing changes of the proximal and distal resistances, but there exist large errors in identification of some parameters. The impedance component method can identify the parameters, but this method is unstable with large calculation errors. Genetic algorithm can obtain a better approximate solution, but it has certain problems in distinguishing vertebral artery stenosis. The combination of impedance modulus curve method and genetic algorithm may play a better role in future application of this model for disease diagnosis.
ABSTRACT
Objective To study the method of internal boundary parameters identification of middle ear.Method The numerical model is created using CT technology.Based on Matlab tools,the neural network for identifying internal boundary is proposed.Result The uniform pressure of 105 dB is applied at the outside of the tympanic membrane,and the harmonic analysis is calculated on the model to take the training samples.The internal condition parameters are identified using the good neural network.Conclusions The investiga-tion shows that the inverse method reveals a fast convergence and a high degree of accuracy.
ABSTRACT
Objective To study the method of internal boundary parameters identification of middle ear.Method The numerical model is created using CT technology.Based on Matlab tools,the neural network for identifying internal boundary is proposed.Result The uniform pressure of 105 dB is applied at the outside of the tympanic membrane,and the harmonic analysis is calculated on the model to take the training samples.The internal condition parameters are identified using the good neural network.Conclusions The investiga-tion shows that the inverse method reveals a fast convergence and a high degree of accuracy.
ABSTRACT
Objective To study the method of internal boundary parameters identification of middle ear.Method The numerical model is created using CT technology.Based on Matlab tools,the neural network for identifying internal boundary is proposed.Result The uniform pressure of 105 dB is applied at the outside of the tympanic membrane,and the harmonic analysis is calculated on the model to take the training samples.The internal condition parameters are identified using the good neural network.Conclusions The investiga-tion shows that the inverse method reveals a fast convergence and a high degree of accuracy.
ABSTRACT
Transformers are required to demonstrate the ability to withstand short circuit currents. Over currents caused by short circuit can give rise to windings deformation. In this paper, a novel method is proposed to monitor the state of transformer windings, which is achieved through on-line detecting the leakage inductance of the windings. Specifically, the mathematical model is established for on-line identifying the leakage inductance of the windings by applying least square algorithm (LSA) to the equivalent circuit equations. The effect of measurement and model inaccuracy on the identification error is analyzed, and the corrected model is also given to decrease these adverse effect on the results. Finally, dynamic test is carried out to verify our method. The test results clearly show that our method is very accurate even under the fluctuation of load or power factor. Therefore, our method can be effectively used to on-line detect the windings deformation.
ABSTRACT
This paper presents a novel algorithm of fault location for transmission line. Solving the network spectrum equations for different frequencies the fault can be located accurately by this algorithm with one terminal data of voltage and current, and the identified parameters, such as fault distance, fault resistance, and opposite terminal system resistance and inductance. The algorithm eliminates the influence of the opposite system impedance on the fault location accuracy, which causes the main error in traditional fault location methods using one terminal data. A method of calculating spectrum from sampled data is also proposed. EMTP simulations show the validity and higher accuracy of the fault location algorithm compared to the existing ones based on one terminal data.
ABSTRACT
This paper presents a novel algorithm of fault location for transmission line. Solving the network spectrum equations for different frequencies the fault can be located accurately by this algorithm with one terminal data of voltage and current, and the identified parameters, such as fault distance, fault resistance, and opposite terminal system resistance and inductance. The algorithm eliminates the influence of the opposite system impedance on the fault location accuracy, which causes the main error in traditional fault location methods using one terminal data. A method of calculating spectrum from sampled data is also proposed. EMTP simulations show the validity and higher accuracy of the fault location algorithm compared to the existing ones based on one terminal data.
ABSTRACT
Transformers are required to demonstrate the ability to withstand short circuit currents. Over currents caused by short circuit can give rise to windings deformation. In this paper, a novel method is proposed to monitor the state of transformer windings, which is achieved through on-line detecting the leakage inductance of the windings. Specifically, the mathematical model is established for on-line identifying the leakage inductance of the windings by applying least square algorithm (LSA) to the equivalent circuit equations. The effect of measurement and model inaccuracy on the identification error is analyzed, and the corrected model is also given to decrease these adverse effect on the results. Finally, dynamic test is carried out to verify our method. The test results clearly show that our method is very accurate even under the fluctuation of load or power factor. Therefore, our method can be effectively used to on-line detect the windings deformation.