Ke-Jun Xu

Dynamic decoupling and compensating methods of multi-axis force sensors

Dynamic decoupling and compensating methods of multi-axis force sensors are proposed to solve simultaneously the dynamic coupling between axes and the slow dynamic response which are the two key problems in dynamic characterizing of the multi-axis force sensors and improve their dynamic quality. Four types of dynamic decoupling and compensating networks are shown, the design equations and procedures are presented, and the parameters of the networks are determined using a method based on a functional link artificial neural network (FLANN). The dynamic decoupling and compensating results for a wrist force sensor have proved the methods to be correct and effective.

Dynamic Modeling and Compensation of Robot Six-Axis Wrist Force/Torque Sensor

The dynamic mathematical models of six channels of six-axis wrist force/torque sensor for robots are built up by the system identification method based on the experimental data of step-response calibrations. The performance indexes of frequency domain are determined, and the dynamic characteristics of the wrist force sensor are described accurately. In order to speed the dynamic responses of the wrist force sensor, the dynamic compensating devices are designed using the method of functional link artificial neural network, and a digital-signal-processor-based real-time dynamic compensation system is developed, which includes six compensating devices. The experimental results show that the adjusting time of dynamic response of the wrist force sensor is reduced to less than 25%.

Power-spectrum-analysis-based signal processing system of vortex flowmeters

A power-spectrum-analysis-based signal processing system was developed for vortex flowmeters. This system consists of a digital signal processor (DSP), an analog input part, an output part, a keyboard input circuit, a guard circuit, a serial communication interface, and software. The sensor output signals are processed digitally using the power spectrum analysis based on fast Fourier transform, and the signal frequency reflecting the flow rate is calculated. With the change of sampling frequency and the spectrum correction, the measurement accuracy of the system is improved and the volumetric flow rate is obtained accurately. By means of adopting a DSP and processing signals while sampling data, the signal processing in real time is ensured. The experiments, including tests of the charge amplifier and signal processing system, and calibration of the vortex flowmeter, show that the performance indexes of the digital processing system are better than those of the conventional instruments for vortex flowmeters in the measurement accuracy, the turn-down ratio, and the adaptability

Gas–Liquid Two-Phase Flow Correction Method for Digital CMF

There are large measurement errors in Coriolis mass flowmeter (CMF) when it is used to measure gas-liquid two-phase flow without compensation. In particular, the flowtube of CMF may stall when using analog transmitters and CMF cannot operate normally because of the high damping caused by gas-liquid two-phase. To solve above problems, a gas-liquid two-phase flow rig is developed to perform the two-phase flow experiments. A digital drive method using multiplying digital-to-analog converter and direct digital synthesizer is applied to CMF to maintain the flowtube oscillation under two-phase flow condition. A digital zero-crossing detection method based on Lagrange interpolation is adopted to calculate the frequency and phase difference of the sensor output signals in order to synthesize the digital drive signal and realize the flow measurement. The artificial neural network using back propagation algorithm is applied to correct measurement errors. A digital Coriolis mass flow transmitter is developed with a digital signal processor to control the digital drive, and realize the two-phase flow measurement and correction. By connecting the developed transmitter with a CMF025 type primary instrument, the gas-liquid two-phase flow experiments are conducted to validate the performance of the transmitter. The testing results show that the transmitter can maintain the flowtube oscillation and reduce the metering errors.

Mathematical Modeling of Ultrasonic Gas Flow Meter Based on Experimental Data in Three Steps

According to experimental data, the features of ultrasonic echo are studied, and a new mathematical model of ultrasonic gas flow meter is built so as to reflect the effects of the excitation signal and gas flow rate on ultrasonic echo. This mathematical model consists of three parts, and is established in three steps. An amplitude sub-model expressed as a polynomial is set up with the curve fitting method to express the non-linear relationship between the echo amplitude and the gas flow rate. A process sub-model described as an output-error model is established by the system identification method to reveal the influence of excitation on the echo shape. A delay time sub-model is built with piecewise fitting to obtain a relationship between the propagation time and the gas flow rate. These sub-models can quantitatively analyze the relationship between the excitation signals, gas flow rates, ultrasonic echo amplitudes, shapes, and propagation time.

Frequency-Variance Based Antistrong Vibration Interference Method for Vortex Flow Sensor

Vortex flowmeter is apt to be disturbed by pipe vibration in industrial applications. It is a very difficult problem to be solved, and affects the applications of vortex flowmeter. To solve this key problem, the frequency features of the flow signal and vibration noise are analyzed. A frequency-variance algorithm is proposed to deal with the vortex flow sensor output signal. A variance threshold is set to distinguish the vortex flow signal from the vibration noise interference. A signal processing system based on an ultralow-power microcontroller is developed to implement the algorithm and to extract the flow rate information from the signal containing strong vibration noise. At the same time, this system meets the requirements of low-power and two-wire mode, and can be used in process control. The strong vibration experiments are conducted to validate the effectiveness of both antistrong vibration algorithm and the system. Only one vortex flow sensor is required in the vortex flowmeter, which avoids the design and installation of the vibration sensor and saves the costs of meter.

Energy transfer model and its applications of ultrasonic gas flow-meter under static and dynamic flow rates

Most of the ultrasonic gas flow-meters measure the gas flow rate by calculating the ultrasonic transmission time difference between the downstream and upstream. Ultrasonic energy attenuation occurs in the processes of the ultrasonic generation, conversion, transmission, and reception. Additionally, at the same time, the gas flow will also affect the ultrasonic propagation during the measurement, which results in the ultrasonic energy attenuation and the offset of ultrasonic propagation path. Thus, the ultrasonic energy received by the transducer is weaker. When the gas flow rate increases, this effect becomes more apparent. It leads to the measurement accuracy reduced, and the measurement range narrowed. An energy transfer model, where the ultrasonic gas flow-meter under without/with the gas flow, is established by adopting the statistical analysis and curve fitting based on a large amount of experimental data. The static sub model without the gas flow expresses the energy conversion efficiency of ultrasonic gas transducers, and the dynamic sub model with the gas flow reflects the energy attenuation pattern following the flow rate variations. The mathematical model can be used to determine the minimum energy of the excitation signal for meeting the requirement of specific measurement range, and predict the maximum measurable flow rate in the case of fixed energy of excitation signal. Based on the above studies, a method to enhance the excitation signal energy is proposed under the output power of the transmitting circuit being a finite value so as to extend the measurement rage of ultrasonic gas flow-meter.

Feature Patterns Extraction-Based Amplitude/Frequency Modulation Model for Vortex Flow Sensor Output Signal

To solve the key problem that vortex flowmeter is susceptible to be disturbed by pipe vibration, the collected data of the vortex flow sensor output signal are filtered, and the amplitudes of the filtered signal are extracted. The fluctuations of both the amplitude and frequency are analyzed and compared so as to study antistrong vibration interference digital signal processing methods for vortex flowmeter. The patterns of fluctuation features are extracted by calculating and analyzing the probability density function and the cycle differenced sequence. Combining with the features of the amplitude modulation and frequency modulation, the mathematical models of the vortex signal, vibration interference, and cofrequency signal are built, and the patterns between the frequency and model parameters are obtained. The effectiveness of the three types of signal models is verified through error analysis, then the reliability of the fluctuation patterns of signal amplitude and frequency is examined, which provides the basis for digital signal processing methods based on fluctuation features.

Frequency-Domain Decoupling-Correction Method for Wind Tunnel Strain-Gauge Balance

A frequency-domain decoupling-correction (FDC) method based on frequency-domain parameter matrix inversion is proposed to solve the problems of cross-axis dynamic coupling interference and dynamic error existing in the main channel output for multiaxis force sensors, for example, wind tunnel strain-gauge balance. The frequency-domain decoupling-correction function is calculated by means of power spectrum estimation according to the sensor step response experimental data, and the windowing approach is adopted to overcome the calculation error caused by the cycle extension of the intercepting data when FFT or IFFT is conducted. The method of FDC based on high/low-frequency signal decomposition is put forward in the process of the dynamic decoupling correction for the sensor output. The low-frequency components are decoupled and corrected by the method of static linear decoupling and static linear correction, and the high-frequency components are decoupled and corrected by the method of frequency-domain parameter matrix inversion in order to overcome the boundary aliasing error and the Gibbs phenomenon caused by the cycle extension of the intercepting data when FFT or IFFT is conducted. According to the static relationship between the sensor input and output, the static characteristic between the sensor input and the sensor dynamic decoupling-correction result is reconstructed so as to ensure that the sensor static characteristic remains unchanged. The dynamic decoupling-correction method is applied to the bar-shaped strain-gauge balance to validate its effectiveness. The result shows that the cross-axis dynamic coupling error ratios are reduced from 98.09% to less than 2%, and the adjust time of the balance main channel step response is shortened from nearly 4 s to less than 6 ms with the overshoot being reduced from 111.24% to less than 5% at the same time. The proposed method can substantially reduce the cross-axis dynamic coupling interference and further improve the response speed of the balance.

A Frequency Correcting Method Combining Bilateral Correction With Weighted Average for Vortex Flow Sensor Signal

The sinusoidal signal outputted by the vortex flow sensor is often superimposed with the white noise, which will affect the accuracy of spectrum correction. When the signal is truncated by rectangular window, a method (hereinafter referred to as RWBWFCR) combining bilateral correction with weighted average on the basis of Fourier coefficients ratio is proposed so as to improve the antinoise performance of the frequency correction. The correction principle of this method is presented. The effects on the frequency correction of changing the initial phase, the frequency deviation, the calculation points of fast Fourier transform, and the signal-to-noise ratio (SNR) are analyzed. The theoretical accuracy of this method when there is stationary Gaussian white noise is deduced and verified. This method has some advantages, such as less computation load, small storage space, high correction precision, strong anti-interference ability, and is easy to be implemented by a microcontroller in real time. RWBWFCR method is applied to the low-power and antivibration type vortex flowmeter with single sensor to improve the reliability of the digital signal processing method based on the frequency correction.