Research On Flow Velocity Measurement Methods Of Electrical Resistance Tomography For Cutter Suction Dredger

Dredging engineering plays a key role in port construction,waterway maintenance and ocean construction.The dredger is the main operation carrier in the dredging engineering,and the measurement of pipeline parameters of its transportation system is one of the most critical links.The solid phase fraction and flow velocity in the pipeline parameters are important basis for control,optimization,and fault diagnosis during the workring process.Currently,they are measured by a ray source densitometer and an electromagnetic flowmeter,respectively.However,the former is radioactive,inaccurate,and maintaining-difficult,while the latter measurement has errors and is low adaptive in changing construction conditions.These problems have greatly restricted the development of dredging engineering.Therefore,the environmentally-friendly and efficient measuring equipment has become an urgent need presently.Electrical Resistance Tomography(ERT)technology,as an advanced,non-invasive and harmless visualization technology,provides a means to solve this problem.However,relevant research at home and abroad is still in its infancy.In order to provide support for the accurate measurement of flow velocity for dredging transportation system,we conduct research on the flow velocity measurement method based on the ERT system.The specific completed research content and results are as follows:1)Facing complex dredging conditions,the convolutional neural network(CNN)is applied to the flow velocity prediction of dredging pipelines to make full use of ERT measurement information.The flow velocity prediction model based on CNN has a 7-layer network structure,with the normalized ERT reconstructed image as input,and the corresponding flow velocity as output.Supported by a large amount of data with different soil qualities and different volume concentrations,the velocity prediction accuracy of the model has reached more than 90%,which has higher accuracy,stability and generalization ability than the existing cross-correlation velocity measurement methods.2)A method for calculating the flow velocity based on the correlation of the change trend of the solid phase fraction curve is proposed.This method aims to overcome two problems that are difficult to solve in the current cross-correlation velocity measurement methods,that is,the length of the time series(time window)to be compared is difficult to determine,and the similarity measure between the sequences affects the velocity calculation accuracy.Experiments show that,compared with two typical cross-correlation flow velocity measurement methods,the method proposed in this thesis can solve the above problems to a certain extent.3)A method for measuring the flow velocity of horizontal pipelines based on a single-row electrode is proposed.This method aims to overcome the problem that the fixed double-row electrodes are difficult to adapt to the flow velocity changes in the existing cross-correlation flow velocity methods.In this thesis,two implementation strategies of this method were proposed,which are based on the pixel layered distribution ratio and the distance between the solid-liquid two-phase center of gravity.These two methods need to perform flow velocity calibration according to the extracted image characteristics at different flow velocities,so as to fit the linear expression of flow velocity with respect to the characteristic value.Through actual test data and laboratory data about the dredging engineering,the effectiveness of these two methods are verified.