The expansion and growth of the IIoT will be considerably aided by the quick development of 5G technologies for communication. Several top sectors, including agriculture, mining, transportation, energy, and health care, have paid close attention to the IIoT. An integral aspect of the 4.0 industrial revolution, it acts relies heavily on big data to examine the complexity and volume of IIoT data. Indeed, the scale and dimension of data will grow substantially and need efficient real-time stream clustering which is a difficult problem for IIoT streaming mining. The security and privacy of the collected data from a wide variety of distributed and mixed endpoints remain a major challenge in the age of cloud computing. The safety and dependability of industrial systems depend on being able to detect issues before they result in downtime or other damage. Finding patterns and associations in large datasets to improve decision-making is the goal of big data analytics (BDA). This paper presents data clustering and fault diagnosis model (DC-FDM) to meet the challenges listed above. Hence, the proposed method uses convolutional neural network (CNN) which uses big data analytics with fault detection to get around those problems and boost performance. Moreover, multi-scale system (MSS) is used to increase the informational stream quality in terms of clustering by measuring the similarity, decreasing the susceptibility to outliers, and employing the relearning technique and increasing privacy and security. Assessments on the generated data sets, both theoretical and experimental, prove that the suggested strategy works and offers a great answer for clusters in IIoT’s current information streams, allowing for the highest precision in fault detection.