Deep learning (DL) has dramatically improved the peak-to-average power ratio (PAPR) performance. However, the high computational complexity and excessive training data constitute a significant hurdle. In this letter, a model-driven deep learning algorithm is proposed for PAPR reduction in orthogonal frequency division multiplexing (OFDM) system. Precisely, an iterative peak-canceling signal generation scheme is unfolded as a layer structure of the DL network. The scheme falls into the category of tone reservation technique. A set of trainable parameters, which optimizes the clipping threshold and weights time-domain kernel function, has been designed and introduced into the iterative scheme. Compared with the existing approaches, the simulation results demonstrate that the proposed algorithm achieves comparable PAPR performance with low complexity and training costs.
A method to measure the superficial velocity of the water phase in gas–water flow using an electromagnetic flowmeter (EMF) and rotating electric field conductance sensors (REFCSs) is introduced in this paper. An electromagnetic flowmeter instrument factor model is built and the correlation between electromagnetic flowmeter output and gas holdup in different flow patterns are explored through vertical upward gas–water flow dynamic experiments in a pipe with an inner diameter (ID) of 20 mm. Water superficial velocity is predicted based on pattern identification among bubble, churn, and slug flows. The experimental results show that water superficial velocity can be predicted fairly accurately for bubble, churn, and slug flows with a water cut higher than 60% (absolute average percentage deviation and absolute average deviation are 4.1057% and 0.0281 m/s, respectively). The output of the electromagnetic flowmeter is unstable and invalid in slug flows with a water cut below 60% due to the non-conducting gas slug is almost filling the pipe. Therefore, the electromagnetic flowmeter is not preferred to be used in such conditions.
In this letter, an improved tone reservation (TR) algorithm using pre-generated peak-canceling signals is proposed. The proposed algorithm generates and saves peak-canceling signals in the pre-work so that they can be used directly at runtime. The proposed algorithm does not need to iterate, and it can be implemented with O(N) computational complexity (N is the number of subcarriers). The simulations indicate that the proposed algorithm achieves about 3.5 dB peak-to-average power ratio (PAPR) reduction at complementary cumulative distribution function (CCDF) of 10 -3 while retaining satisfactory bit error rate (BER) and spectral spreading performance.
27 August 2020 in IEEE Wireless Communications Letters
One of the limitations of the Neural Network-based peak-to-average power ratio (PAPR) reduction methods lies in the long-term training process through a large amount of training data. To address this issue, an extreme learning machine-based (ELM) tone reservation scheme is proposed, which could achieve comparable PAPR performance with low complexity compared to other Neural Network-based algorithms. An ELM-based mapper, trained on the clipping control tone reservation (CC-TR) signals, is designed to reduce the PAPR of orthogonal frequency division multiplexing (OFDM) system. Simulation results indicate that the proposed algorithm owns the advantages of fast learning capability and short training length.
Oil-water two-phase flow exists widely in the industrial process. At present, it is very difficult to describe the oil-water two-phase flow system theoretically. The mechanism of droplet fragmentation and coalescence, which restricts droplet size and flow stability, is still unclear. Accurate identification of oil-water two-phase flow patterns is of great significance to flow control and measurement. In this study, a refined composite multi-scale rescaled range fractional permutation entropy (MRSFPE) is proposed for identifying the flow pattern of oil-water two-phase flow in vertical pipes. Firstly, the proposed algorithm is simulated numerically to demonstrate its ability to uncover the complexity of nonlinear time series. Then, the conductance sensor signals are collected in the flow loop test facility of vertical upward oil-water two-phase flow. In the experiment, the observed flow patterns, captured by a high-speed camera, include very fine dispersed oil-in-water flow (VFD O/W), dispersed oil-in-water flow (D O/W), dispersed oil-in-water slug flow (D OS/W), and transition flow (TF). Finally, the proposed algorithm is applied to the experimental signal analysis, extracting two scale indices of the entropy rate at high scale (MRSFPE_rate) and the mean value entropy at low scale (MRSFPE_ave). The results show that the four flow patterns can be identified in the joint distribution of the two scale indices. Furthermore, the MRSFPE can be a useful approach to characterize the nonlinear dynamics of multiphase flow based on sensor measurement signals.
In order to improve the flow measurement accuracy of oil-water two-phase flow at low flow rate, this paper presents a plug-in conductance sensor array (PICSA) for the measurement of water holdup and cross-correlation velocity. Due to the existence of the insert body in PICSA, the effect of slippage and the non-uniform distribution of dispersed phase on the measurement of oil-water two-phase flow at low flow rate can be reduced. The finite element method is used to analyze the electric field distribution characteristics of the plug-in conductance sensor, and the sensor geometry is optimized. The dynamic experiment of oil-water two-phase flow is carried out where water cut Kw and mixture velocity Um are set in the range of 10–98% and 0.0184–0.2580 m/s respectively. Experimental results show that the PICSA has good resolution in water holdup measurement for dispersed oil-in-water slug flow (D OS/W), transition flow (TF), dispersed oil-in-water bubble flow (D O/W) and very fine dispersed oil-in-water bubble flow (VFD O/W). In addition, the cross-correlation velocity of the oil-water two-phase flow is obtained by using the plug-in upstream and downstream conductance sensor arrays. The relationship between the cross-correlation velocity and mixture velocity is found to be sensitive to the change of flow pattern, but it has a good linear relationship under the same flow pattern. Based on the flow pattern identification, a good prediction result of the mixture velocity is obtained using kinematic wave theory. Finally, a high precision prediction of the individual phase volume fraction of oil-water two-phase flow at low flow rate is achieved by using the drift flux model.