Bearings are the key components of injection molding machines. It is of great significance to accurately assess the
degradation of bearings in operation and predict their remaining useful life. With the rapid development of modern industry
and the increasing complexity of equipment, model-based methods are difficult to adapt to changing environments, and
data-driven methods have been extensively developed. This paper proposes an Autoencoder-LSTM remaining useful life
prediction method, which uses Autoencoder to extract features from the original bearing data, and uses LSTM to realize
the life prediction of the bearing. Experiments on the PU dataset verify the effectiveness of the features extracted by this
method, and compare with other methods to prove the superiority of the method; then the life prediction experiment on the
XJTU dataset verifies that the method has higher prediction accuracy, which is better than traditional machine learning
and related methods.
In order to improve the detection accuracy and speed of lane detection, this paper proposes a multi-lane line detection
algorithm based on the combination of semantic segmentation and clustering. First, we need to process the dataset, use a
new semantic segmentation network to obtain lane instances, then perform binary segmentation on the image and embed
a vector attribute to distinguish which lane the pixel belongs to, and then combine the two parts of data. Perform clustering
and use polynomial fitting method to get the lane object on the image. After experimental testing, the detection speed of
the algorithm reached 72fps, and the accuracy rate reached 96.5%, indicating that the algorithm has a high detection
accuracy and speed.
In order to detect small cracks on the surface of casting workpieces in real time, this paper proposes a feature fusion target detection algorithm based on SSD network. First, the image is reshaped and fed into the network to generate multiple feature maps of different scales. First, the image is resized and fed into the network to generate multiple feature maps of different scales. Feature fusion is performed between feature maps to generate new feature maps. Finally, multiple new feature maps are used to predict crack coordinates and probability. After experimental testing, the detection speed of the algorithm in this paper reaches 87fps, while the accuracy rate reaches 92.3%, indicating that the algorithm has high detection accuracy and speed.
Injection molding machine can shape all kinds of precise size, complex shape of plastic products, related to all aspects of people's life, such as the national defense field, civil fields such as electronic industry, daily use industry, medical industry and medical industry. Therefore, it is of great significance to establish a health prediction model for injection molding machine and carry out condition monitoring. Bearing is one of the most important parts in the rotating parts of the injection molding machine. Analysis of the health state of the bearing can reflect the performance state of the injection molding machine to a certain extent. Taking bearings as the research object, this paper introduces deep learning and information average convolutional neural network (AICNN), and proposes a health state prediction method of rotating equipment based on multi-feature variable fusion.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.