Ceramic shuttle kiln is a batch kiln used for ceramic production. Temperature detection mode of the sintering zone in the kiln directly affects the production quality of the final ceramic products. At present, the thermal detection of ceramic shuttle kiln is still inaccurate. As a result, flame image recognition method has been developed to replace the traditional thermocouple to detect the sintering zone temperature. In this paper, to address the problem of flame image recognition in the sintering zone of ceramic shuttle kiln, flame image classification method based on improved convolution neural network was proposed. In this method, the SE module was embedded in optimized convolution neural network (Inception-ResNet-V2) module, so as to improve the attention of the network to the key features, to adaptively refine the features and to improve the classification effect. Experimental results showed that the improved SE-Inception-ResNet-V2 can increase the flame image classification accuracy and accelerate the convergence rate. Compared with other flame image classification methods, our approach achieved an increase in recognition accuracy by 1.60-5.57%.

At present, most surface defect of ceramic tiles are still detected by workers with low production efficiency and low detection accuracy. In order to solve this problem, we proposed a ceramic tile defect detection and classification method based on YOLOv8n, with dynamic serpentine convolution and attention mechanism. Firstly, C2F-DSConv was used to replace the original C2F in the backbone network and neck network of the yolov8 model to enhance the extraction of scratch defect features. Secondly, the CBAM attention module was added to the model to enable it to learn useful features from the image more efficiently, so as to improve the model performance. Finally, the loss function CIoU was modified to the Inner-SIoU loss function to enhance the model’s feature extraction ability for small targets. It is experimentally shown that, as compared with that of the original model, the average detection accuracy of the improved yolov8n-based tiles defect detection model is increased by 2.6%, while the number of the computation is only increased by 0.5 G. In addition, the proposed method showed obvious advantages in comparative experiments and self-built data set experiments.

Shuttle kiln is a main equipment in ceramic production. As a key process parameter in the production process of ceramic shuttle kiln, temperature plays a key role in determining the quality of ceramic products, the efficient and stable operation of kiln, reduction of energy consumption and so on. In order to effectively control the temperature in ceramic shuttle kiln, firstly, the predictive modeling method of ceramic shuttle kiln based on gated recurrent neural network was proposed in view of the characteristics of nonlinearity, large inertia, large lag and difficulty in establishing accurate mat hematical model.Secondly, based on the established prediction model, the intelligent optimization control method of ceramic shuttle kiln temperature based on DDPG algorithm is proposed and the optimization control system scheme of ceramic shuttle kiln temperature based on DDPG is also presented. Finally, simulation experiments are carried out for the proposed method.Compared with PID control, fuzzy control and fuzzy PID control, the proposed method can make the error between the shuttle kiln sintering temperature and the ideal temperature to be reduced by 18.6–28.5%, showing high effectiveness and feasibility.

At present, the firing temperature of ceramic roller kiln is mainly detected by thermocouples. Due to the easy aging of thermocouples, the temperature detection accuracy of thermocouples gradually becomes low so as to affect the firing quality of ceramic products. For this problem, an intelligent temperature detection method of fusing flame image feature recognition based on deep learning with thermocouple point detection data instead of thermocouple. This method is that a multi-scale feature extraction network based on the shift window visual self-attention mechanism is adopted for the flame image of ceramic roller kiln, convolutional neural network and local and remote features of transformer branch is used for retaining more image information to obtain more accurate flame image features which are fused with thermocouple point detection data, thus, the temperature of ceramic roller kiln can be accurately detected. In the multi-scale feature extraction network model, firstly, an auto-encoder network based on multi-layer transformer is used for extracting shallow and multi-scale deep features, and then multiple features are fused into transformer and convolutional neural network to make it be able to capture feature information, finally, the data obtained from the thermocouple point detection is input into the front network to achieve the fusion of the flame image features and the key point detection temperature data by the feature level information fusion. Experimental results show that the fusion network model proposed in this paper is 1.75% higher in average feature recognition accuracy and 2.67% lower in average error generation than the convolutional neural network fusion method, which is superior to the convolutional neural network branch or transformer branch image fusion in most indicators. Hence, the method proposed in the paper is effective and feasible.