Scholarlink

<p>Wire-Arc Spraying (WAS) is an economical and widely used thermal spray process. Applications include corrosion protection coatings on offshore wind power plants. While WAS is cost-efficient and technologically well-developed, the high process complexity results in a lack of physics-based models. The process is characterized by fluctuations of an electric arc which moves along the wire tips. Regular, periodic process fluctuations, caused by droplet detachment and the corresponding arc movement, are intrinsic to WAS. However, irregular, aperiodic fluctuations, such as arc interruptions, indicate an unstable process and must be avoided. These irregular fluctuations can lead to plant downtime, coating defects and therefore an increase in maintenance and quality control costs. This study employs time-series anomaly detection algorithms on WAS sensor data and introduces a novel optimization criterion, <i>σ</i><sub>L</sub>, which is specifically derived for thermal spraying processes for the purpose of process stability assessment. To collect the necessary time-series data, a sensor unit was set up on a WAS system. ZnAl15 wire and Fe<sub>62</sub>Si<sub>12</sub>B<sub>8</sub>C<sub>4</sub>Nb<sub>2</sub>Mo<sub>2</sub>Cr<sub>10</sub> cored wire were used as feedstock materials. ZnAl15 spraying included multiple experimental runs according to the design of experiment methodology. The ZnAl15 coatings were applied onto C45 steel substrate. Each sample was weighed before and after spraying to determine the deposition efficiency. Voltage and electric current in every experiment were measured at a sampling rate of 500&#xa0;kHz. Fast Fourier Transformations (FFT) were calculated on the electrical resistance, which was derived by Ohm’s law. The signal analysis revealed WAS process-related fundamental frequencies in the range of 0.5-1.0&#xa0;kHz, alongside frequencies at 20&#xa0;kHz, which stem from the power control unit. The novel criterion <i>σ</i><sub>L</sub> is calculated on the first range of frequencies. The fact that <i>σ</i><sub>L</sub> shows low sensitivity to regular process fluctuations makes it superior to conventional metrics. By generalized linear modeling, <i>σ</i><sub>L</sub> as well as the deposition efficiency were fitted to the investigated parameters. Combining <i>σ</i><sub>L</sub> and known anomaly detection algorithms such as thresholding and <i>k</i>-means allows for the detection of detrimental fluctuations and process interruptions.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Investigation of Criteria for Wire-Arc Spraying Process Stability

  • 期刊论文

  • 发表日期：

    2026年7月27日

查看全文
  • K. Bobzin,
  • Thorsten Papenbrock,
  • H. Heinemann,
  • M. Erck,
  • Marcian Seeger,
  • L. M. Johann

摘要

Wire-Arc Spraying (WAS) is an economical and widely used thermal spray process. Applications include corrosion protection coatings on offshore wind power plants. While WAS is cost-efficient and technologically well-developed, the high process complexity results in a lack of physics-based models. The process is characterized by fluctuations of an electric arc which moves along the wire tips. Regular, periodic process fluctuations, caused by droplet detachment and the corresponding arc movement, are intrinsic to WAS. However, irregular, aperiodic fluctuations, such as arc interruptions, indicate an unstable process and must be avoided. These irregular fluctuations can lead to plant downtime, coating defects and therefore an increase in maintenance and quality control costs. This study employs time-series anomaly detection algorithms on WAS sensor data and introduces a novel optimization criterion, σL, which is specifically derived for thermal spraying processes for the purpose of process stability assessment. To collect the necessary time-series data, a sensor unit was set up on a WAS system. ZnAl15 wire and Fe62Si12B8C4Nb2Mo2Cr10 cored wire were used as feedstock materials. ZnAl15 spraying included multiple experimental runs according to the design of experiment methodology. The ZnAl15 coatings were applied onto C45 steel substrate. Each sample was weighed before and after spraying to determine the deposition efficiency. Voltage and electric current in every experiment were measured at a sampling rate of 500 kHz. Fast Fourier Transformations (FFT) were calculated on the electrical resistance, which was derived by Ohm’s law. The signal analysis revealed WAS process-related fundamental frequencies in the range of 0.5-1.0 kHz, alongside frequencies at 20 kHz, which stem from the power control unit. The novel criterion σL is calculated on the first range of frequencies. The fact that σL shows low sensitivity to regular process fluctuations makes it superior to conventional metrics. By generalized linear modeling, σL as well as the deposition efficiency were fitted to the investigated parameters. Combining σL and known anomaly detection algorithms such as thresholding and k-means allows for the detection of detrimental fluctuations and process interruptions.