Kai Mao

Optimization of Neural Network by Genetic Algorithm for Flowrate Determination in Multipath Ultrasonic Gas Flowmeter

Artificial neural network (ANN) was proposed as an effective method to help multipath ultrasonic flowmeter (UFM) reduce its measurement error when determining the flowrate of complex flow field. However, the effectiveness of the ANN method heavily depends on the network architecture specified by the designer, and also the provided initial weights and layer biases. This hinders the ANN to be widely used in actual UFMs. This paper proposes a genetic algorithm (GA) optimized ANN method (GANN) to be used in UFMs. The GA is utilized to determine ANN architecture based on its efficient parallel advantage in global search to replace traditional trial and error approach or empirical method. In addition, the initial weights and biases are optimized by GA, which is capable to avoid the network getting stuck with local minimum and qualify it for good generalization ability. Tests are implemented on numerical models simulating a six-path UFM installed downstream single or double elbows with complex flow field passing through it. Comparisons between the results of classical Gaussian quadrature, existing ANN, and proposed GANN are given for verification. The flowrate determination error of GANN is only 40% ~ 62% of existing ANN, which is extremely smaller than the Gaussian quadrature method. In addition, the output of GANN is much more stable than existing ANN, as proper initial weights and layer biases will be generated by GA for iteration process even random data are given for training.

Similarity Judgment-Based Double-Threshold Method for Time-of-Flight Determination in an Ultrasonic Gas Flowmeter

Double-threshold method is a typical and simple time-of-flight (TOF) measurement method for time-difference-based ultrasonic flowmeter. Some improvements have been proposed to suppress the cycle-skip phenomenon in double-threshold method. However, there are still some application limitations when the waveform is distorted caused by the environment change, especially the temperature change. This paper proposes a similarity judgment-based double-threshold method to deal with the cycle-skip problem. In this method, time values of three consecutive positive zero-crossing points after the feature point are acquired and five consecutive cycles after the first obtained time are saved. The TOF is automatically selected from three acquired time values according to an adaptive determination function which bases on the similarity between signals from two adjacent measurements. These operations guarantee that the obtained TOF locates on the same positive zero-crossing point in each measurement. The proposed method was validated by both numerical and experimental tests. Comparison between the proposed method and other double-threshold methods shows that the proposed method successfully eliminates cycle-skip phenomenon under different working conditions.

Flowrate Determination for Arbitrary Multipath Arrangement Based on Generalized Inverse of Matrix

As a widely used method in flow measurement, Gaussian quadrature is traditionally applied to determine the flowrate with multiple acoustic paths. However, it is only available to parallel path arrangement and the path position cannot be customized by users because of the nature of numerical integration principle. Accuracy of flowrate determination is greatly restricted meanwhile. In this paper, a new method based on generalized inverse of matrix (GIM) is proposed to work out the flowrate in arbitrary multipath arrangement. With an integrable model of flow distribution established, the GIM method will find the optimal weights and achieve highly accurate flowrate. Three perfect profile assumptions, including Jacobi, optimal weighted integration for circular sections, and typical power-law profile, are, respectively, adopted in this paper to explore the performance of GIM. Its effects are at first experimentally validated in parallel path arrangement of Gauss–Jacobi on a commercially available ultrasonic flowmeter. Then, four representative types that Gaussian quadrature is not feasible to are studied via numerical simulation, including a parallel and three crossed path arrangements. Performance of the proposed GIM method is verified under straight pipe, single bend, and double bend with an expander as well as on eight installation angles. Results reveal that GIM is applicable to accurate flowrate determination for arbitrary multipath arrangement and crossed paths could provide high determination accuracy and good robustness even under complicated flow field.

A Novel Ultrasonic Transducer Module with an Own Reference Sound Channel for Flow Measurement

Zero error caused by transducer system asymmetry which drifts with the environment change is a severe limitation to the performance improvement of an ultrasonic flow meter. This chapter proposes a novel ultrasonic transducer module with an own reference sound channel for zero error calibration. Different from traditional transducers, the module consists of two energy conversion terminals which are connected by a reference sound channel. Part of the ultrasound generated by one terminal transfers into the flow media and finally into another terminal, which is used for flow measurements. The rest propagates through their own sound channel as reference signal for zero drift acquisition. According to the difference of the measuring signal and the reference signal, the zero drift error can be calibrated. We have illustrated the construction, zero elimination method, and design theory of the module with simulation employed for validation. A prototype has been developed and the preliminary results demonstrate an observable decrease of the zero drift error.

A Method to Eliminate the Effect of Sound Path Length Change of Ultrasonic Heat Meter

Ultrasonic heat meter (UHM) plays a key role in the heat consumption sensing in automatic metering and control system of district heat supply pipe network; however, the large temperature changes (below 10℃ to above 90℃) of water in the network generally cause sound path length (SPL) changes due to thermal expansion and contraction, which will greatly reduce the accuracy of UHM in practice. This paper proposes a method to eliminate the effect of SPL change for UHM. In the method, a novel model has been developed to calculate flow rate and then heat consumption from upstream and downstream transmit times and sound velocity, but not containing the SPL parameter commonly existing in traditional models. A sample learning method is proposed in addition to correct the effect of sound velocity changes under different temperatures. The method is validated on actual UHMs by flow rig experiments on simulating serious temperature variations.