文章摘要
基于数学孪生技术的平面渔网破损检测方法研究*
Research on the Damage Detection Method of the Plane Fishing Net Based on the Digital Twin Technology
投稿时间:2021-08-25  修订日期:2021-09-27
DOI:
中文关键词: 数字孪生,人工神经网络,网衣,破损检测
英文关键词: Digital twin  Artificial neural network (ANN)  Fishing net  Damage detection
基金项目:国家重点研发计划项目
作者单位邮编
连栗楷 大连理工大学 海岸和近海工程国家重点实验室 大连 116024
赵云鹏 大连理工大学 海岸和近海工程国家重点实验室 大连 116024
毕春伟 大连理工大学 海岸和近海工程国家重点实验室 大连 
许智静 大连理工大学宁波研究院 宁波 
杜海 大连理工大学 海岸和近海工程国家重点实验室 大连 
摘要点击次数: 43
全文下载次数: 0
中文摘要:
      如果不能及时发现渔网的破损,将造成养殖鱼类的逃逸,给水产生产造成损失,因此有必要对网衣进行破损检测。为了克服人工检测劳动强度大且效率低下的缺点,实现渔网的精准实时监测,本文提出了一种基于数字孪生的网衣破损检测方法,可利用传感器监测代替人工监测。该方法首先从渔网的数值仿真模型获取大量的仿真传感数据,然后将数据用于人工神经网络的训练与测试,最后训练生成可进行网衣破损识别的数字孪生体。数字孪生体可根据传感器监测到的数据来判断网衣是否发生破损。在数值模拟中,考虑各种波浪条件以及网衣的破损情况。在训练人工神经网络中,将有效波高Hs、谱峰周期Tp以及横纲竖纲的拉力值作为输入变量,将网衣完整状态以及破损状态作为输出。经过测试分析,该识别模型根据传感器数据识别网衣是否破损的平均准确率为94.32%,这可认为数字孪生技术能准确检测到渔网的损坏,可以作为网衣破损检测的一种新方法。
英文摘要:
      For aquaculture nets, if damage cannot be found in time, it will be easy to cause a large number of fish to escape and cause huge losses to farmers. Therefore, it is necessary to detect whether fishing net occurs. At present, the main method to detect the damage of fishing nets is manual inspection of staff diving into the water, but this method is labor-intensive and inefficient. In order to overcome these limitations and realize real-time monitoring of fishing net, in this paper, a damage detection method of fishing nets based on digital twinning is proposed, which uses sensor monitoring instead of manual monitoring. The research first shows that the numerical simulation data that are in good agreement with the physical model experimental data can be obtained through the numerical simulation of the lumped mass mechanical model. Then in the numerical simulation, considering a kind of damage to the fishing net, a total of 11 simulations are carried out, and the tensile values of horizontal and vertical rope of the fishing net, 9 sea conditions are training samples, and 2 sea conditions are test samples; artificial neural network adopts The error backpropagation training method takes the significant wave height Hs, the spectral peak period Tp, and the tensile value of the vertical and horizontal rope as inputs, and the complete state and damaged state of the fishing net as the outputs. After training, the recognition model's recognition accuracy rates for training samples and test samples are 99.21% and 95.11%, respectively. The measured real physical sensor data are also used as the test data. The recognition accuracy of the recognition model is 94.32%, which indicates that the practical application of digital twin technology in the damage detection of net is feasible, and it can be used as a new method of fishing net damage detection. Since the wave-current environment is more complex in the actual sea area, how to deal with the sensing data and how to detect the damage of fishing nets in a more realistic sea condition are the next research contents.
相关附件:   附件1  附件1  附件2
View Fulltext   查看/发表评论  下载PDF阅读器
关闭