A fault diagnosis approach based on multi-sensor data fusion is a promising tool to deal with complicated damage detection problems of mechanical systems. Nevertheless, this approach suffers from two challenges, which are (1) the feature extraction from various types of sensory data and (2) the selection of a suitable fusion level. It is usually difficult to choose an optimal feature or fusion level for a specific fault diagnosis task, and extensive domain expertise and human labor are also highly required during these selections. To address these two challenges, we propose an adaptive multi-sensor data fusion method based on deep convolutional neural networks (DCNN) for fault diagnosis. The proposed method can learn features from raw data and optimize a combination of different fusion levels adaptively to satisfy the requirements of any fault diagnosis task. The proposed method is tested through a planetary gearbox test rig. Handcraft features, manual-selected fusion levels, single sensory data, and two traditional intelligent models, back-propagation neural networks (BPNN) and a support vector machine (SVM), are used as comparisons in the experiment. The results demonstrate that the proposed method is able to detect the conditions of the planetary gearbox effectively with the best diagnosis accuracy among all comparative methods in the experiment.

Electromagnetic testing is the most widely used technique for the inspection of steel wire ropes. As one of the electromagnetic detecting approaches, the magnetic flux leakage (MFL) method has the best effect for the detection of broken wires. However, existing sensors based on MFL method still have some problems. (1) The size of the permanent magnet exciter is usually designed according to experience or rough calculation, and there is not enough depth analysis for its excitation performance; (2) Since the detectable angular range for a single Hall component is limited, Hall sensor arrays are often employed in the design of MFL sensors, which will increase the complexity of the subsequent signal processing due to the extensive use of Hall components; (3) Although the new magneto-resistance sensor has higher sensitivity, it is difficult to be applied in practice because of the requirement of the micron-level lift-off. To solve these problems, a sensor for the detection of broken wires of steel wire ropes based on the principle of magnetic concentration is developed. A circumferential multi-circuit permanent magnet exciter (CMPME) is employed to magnetize the wire rope to saturation. The traditional Hall sensor array is replaced by a magnetic concentrator to collect MFL. The structural parameters of the CMPME are optimized and the performance of the magnetic concentrator is analyzed by the finite element method. Finally, the effectiveness of the designed sensor is verified by wire breaking experiment. 1–5 external broken wires, handcrafted on the wire rope with a diameter of 24 mm, can be clearly identified, which shows great potential for the inspection of steel wire ropes.

The detection of broken wires in steel wire ropes is of great significance for the production safety. However, the existing identification techniques mainly focus on the external broken wires problem. Here, the artificial feature extraction is one of the most important method, while only the prior knowledge of the artificial feature extraction method is adequate, the identification precision can be satisfied. Therefore, it is still a challenge to realize intelligent diagnosis for the broken wires. Besides, the identification of internal broken wires problem is still not well solved. In this paper, a quantitative identification method based on continuous wavelet transform (CWT) and convolutional neural network (CNN) is proposed to solve the internal and external broken wires identification problem. The key technology of this research is that the fault information from the time-frequency images converted by the magnetic flux leakage (MFL) signals can be automatically extracted through a designed CNN. The main innovation is that the complex signal processing work can be eliminated and the internal and external broken wires can be accurately identified simultaneously by combining CWT and CNN. The experimental results of a steel wire rope test rig are compared with the traditional recognition method, which shows that the proposed method achieved significant improvement on detection accuracy and recognition performance.

The quantitative identification of broken wires is of great significance to maintain the safety of mechanical systems, such as steel wire ropes. However, in order to achieve high accuracy recognition results, a large number of fault samples are necessary, which is difficult to achieve in practical industrial detection. In this paper, a novel quantitative identification approach, based on generative adversarial nets (GANs) and a convolutional neural network (CNN), is proposed to solve the broken wire recognition problem in situations where real inspections have generated only a small sample of broken wires for analysis. One-dimensional original signals of broken wires are transformed into two-dimensional time-frequency images by continuous wavelet transform (CWT). Next, these time-frequency images are used for quantitative identification of various defects by combing GANs and CNN with limited samples. The main innovation of this paper is that the identification accuracy of broken wires can be improved by generating fault samples through GANs. The experimental results demonstrate that the proposed method achieves better recognition rates for broken wires compared with the existing detection methods.

Each pattern recognition method has its advantages and disadvantages to diagnose the state of rotating machinery. There are many fault types of rolling bearings with apparent uncertainty. The optimal fusion level is usually challenging to be selected for a specific fault diagnosis task, and extensive human labour and prior knowledge are also highly required during these selections. To solve the above problems, a multimodel decision fusion method based on Deep Convolutional Neural Network and Improved Dempster-Shafer Evidence Theory (DCNN-IDST) is proposed for the inspection of rolling bearing. To solve the defect of the original evidence theory method in the fusion of high-conflict evidence, the fuzzy consistency matrix is introduced. By calculating the factor weight, the reliability and rationality of D-S evidence theory are improved. The DCNN model can learn features from the original data and carry out adaptive feature extraction for multiple sensor information. The features extracted by DCNN adaptively are input into multiple network models for decision fusion. The new method of DCNN-IDST multimodel decision fusion is applied to detect the damage of rolling bearings. To evaluate the effectiveness of the proposed method, both the BP neural network and RBF neural network are used to set up a multigroup comparison test. The result demonstrates that the proposed method can detect the fault of the rolling bearing effectively and achieve the highest diagnosis accuracy among all the tested methods in the experiment.