Introduction |
The rapid development of artificial intelligence technology is quietly leading the fourth industrial revolution. Although there are still many challenges to be solved in the deployment of AI technology in the industrial field, this cannot stop the historical trend from moving forward. AI is bringing new changes to the industrial sector. Today, we have invited
Mr. Han Guangzu, Director & CTO of Shanghai Tengzhan Changrong
, who will share with us the development and application of AI data analysis in the industrial field.
Author Profile
Han Guangzu, Tencent Cloud TVP, currently serves as Director & CTO of Shanghai Tengzhan Changrong. He holds a master’s degree in business administration from the University of Southern California, and has previously served as Vice President of the Channel and Digital Banking Department of Fubon Bank, and Vice President of the Information Technology Department,WistronITS Global Headquarters Chief Information Officer, Corporate Security Committee Chairman, Technical Consultant and COO (Investment) of Zichen International Development (Central Enterprise Hong Kong Bank Boyuan Fund),Global Technology and Outsourcing Services Director of Newegg, and Head of Technology Division of Foreign Banks for 12 years. He has over 20 years of experience in enterprise IT/MIS/IS operations, with expertise in DD, private equitydebt financing, industrial real estate transactions and equity transfers, cross-border financial management,
technology development and innovation experience. He also has years of experience in the large e-commerce industry and in replacing core banking systems, is familiar with information technology, digitalization, business system analysis, cloud architecture and migration, public cloud construction and development, integration; and is familiar with R&D, products, pre-sales, delivery, after-sales, etc.; including professional service solutions, planning, implementation, and establishing large data analysis, data collection, and deep learning image object detection AI-assisted decision-making and is familiar with overall corporate strategic planning and implementation.
1. Key Technologies in the Era of Industry 4.0
As we all know, industrial big data analysis is a key technology in the era of Industry 4.0. It provides unprecedented insights and decision support for production operations, product innovation,supply chain optimization and safety management through in-depth mining of massive industrial data. The themes and scenarios involved in industrial big data analysis are very broad. Here are some main examples of data-driven overall business:
(Illustration, Data-Driven Overall Business Architecture
)
(Figure 2, Data Mining Analysis Methods
)
First,
in the field of production operations
,
industrial big data analysis accurately identifies bottlenecks in the production process through real-time monitoring of sensor and equipment operating data,
thereby optimizing production processes and improving efficiency.
At the same time, by analyzing product quality inspection data, the root causes of product defects can be quickly identified, which not only improves product quality but also enhances customer satisfaction. Furthermore, analyzing energy consumption data helps identify waste points in energy use, thereby optimizing energy allocation and achieving green production.
Secondly, predictive maintenance is another important application of industrial big data. It predicts potential failure risks by analyzing historical operation and failure data of equipment, and formulates scientific maintenance plans.This method not only reduces unexpected downtime but also effectively lowers maintenance costs.At the same time, by analyzing product usage data, product lifespan can be predicted, timely reminders can be sent to users for replacement or maintenance, thus improving product reliability and user trust.
Moreover, in the product development phase, industrial big data analysis utilizes market data and customer demand data to help enterprises gain insights into market trends, quickly respond to market changes, and develop innovative products that meet market needs.At the same time, analysis of design data and testing data helps optimize product design, enhancing product performance and competitiveness.
In addition, supply chain management is another key application scenario of industrial big data analysis.By analyzing supplier data and procurement data, enterprises can optimize procurement strategies and reduce procurement costs.Analysis of inventory data and logistics data helps optimize inventory management and improve the responsiveness and efficiency of the supply chain.
Finally, in terms of safety management, industrial big data analysis helps enterprises identify safety risks and formulate effective safety measures by analyzing safety incident data and hidden danger investigation data, thus reducing the incidence of accidents.Analysis of employee behavior data and environmental monitoring data helps identify potential safety hazards and prevent safety accidents.
(Figure 3, Application Scenarios and Nodes of Data Analysis)
It is worth mentioning that industrial big data analysis plays an important role in various scenarios such as smart manufacturing, industrial Internet of Things, industrial Internet, and energy Internet.In smart manufacturing factories, it helps achieve intelligent production, intelligent decision-making, and intelligent services.In the industrial Internet of Things field, it can process and analyze large amounts of device data to extract valuable information.
The robotic dog inspection platform is an intelligent inspection solution that integrates advanced robotics technology, artificial intelligence, and automated control. This platform uses flexible and agile robotic dogs as the main inspection body, capable of performing tasks in various complex environments, such as industrial facilities, warehouses, public areas, and even disaster response sites.
(Figure 4,
Combining IoT and Robotic Dog Technology:
(Successful Case of Robotic Dog Experiment)
(Failed Case of Robotic Dog Experiment)
The rise of the industrial Internet of Things enables big data analysis to process and analyze massive data streams from countless devices, revealing hidden value.The industrial Internet platform uses big data analysis to break down data silos, achieving data sharing and a prosperous innovation application.The energy Internet also optimizes the supply-demand balance of energy through big data analysis, enhancing energy utilization efficiency.
Moreover, the specific value brought by industrial big data analysis manifests at multiple levels:It not only improves production efficiency and product quality but also reduces costs by shortening R&D cycles and optimizing supply chain management.More importantly, it significantly improves safety production levels through predictive and preventive measures.As technology continues to advance, industrial big data analysis will continue to expand its application scope, becoming an important tool for enterprises to achieve sustainable development and maintain competitive advantages.
In summary, with the rapid development of technologies such as the industrial Internet of Things and industrial Internet, the application prospects of industrial big data analysis are infinitely broad. It will become an important tool for enterprises to gain competitive advantages and achieve intelligent transformation. At the same time, artificial intelligence technologies, such as ChatGPT, will play an important role in data cleansing, exploratory data analysis, hypothesis testing, analysis method selection, and result interpretation, providing enterprises with deep business insights and decision support.
2. How to Develop Industrial AI?
Transformational Principles, Methods, and Practices
In fact, when discussing the future path of industrial artificial intelligence, we can draw inspiration from the best practices in the industry and build an efficient and sustainable development strategy accordingly. In my view, its future development path can be summarized in terms of principles, methods, and practices:
Principles
:
Reducing costs is the ultimate goal. The core objective of industrial AI is to lower production and operational costs through intelligent means.
This includes not only direct material and labor costs but also involves reducing waste and improving resource utilization through process optimization.
Methods
:
Reducing variability is a key process. In the process of achieving cost reduction, it is crucial to reduce variability in production and operations. Variability can lead to uncertainty and waste, while precise control of processes through AI technology can significantly enhance consistency and reliability.
Practices
:
Resource integration using new AI processes, such as ECRS (Eliminate, Combine, Rearrange, Simplify) principles, can help enterprises rethink and redesign workflows, eliminate unnecessary steps, merge similar tasks, rearrange process sequences, and simplify complex operations. At the same time, operational research on AI resources can ensure effective utilization of technology, maximizing production efficiency. The exploration factor of AI is the application of data science, which plays an important role in industrial AI. By exploring and analyzing large volumes of data, enterprises can discover potential patterns and trends, leading to continuous improvement and optimization of production processes.
(Figure 5, Development Path of Industrial AI)
A typical example is TSMC, a global leader in semiconductor manufacturing, whose successful practice in introducing artificial intelligence has set a benchmark for the entire manufacturing industry. TSMC has deeply integrated AI technology into its production processes, not only improving manufacturing precision but also optimizing production efficiency and product quality.
(Figure 6, Development History of Smart Manufacturing)
3. Looking at the Future Path of Industrial AI from Best Practices
in Enterprises
(1) AI Digital Application in Public Utility Workshops
In addition, we can see the significant role of industrial AI in factory workshops. The application of IoT + ML in public utility workshops and machine learning technology significantly enhances energy efficiency and achieves energy savings and carbon reduction. In fact, a factory’s production workshop mainly consists of indirect production workshops (public utility workshops) and direct production workshops, including compressed air systems, central air conditioning systems, and circulating water systems. By implementing intelligent control over each part through AI technology, we can successfully achieve energy savings and emissions reductions. For example, the central air conditioning system can achieve efficient operation of the main unit and save energy by 10-20% through intelligent optimization, while the circulating water system can intelligently dose and adjust based on changes in water quality data through data models and analysis.
(Figure 7, Detail Illustration of Production Workshop)
In the AI-controlled digital workshop, each station in the public utility workshop can achieve an average energy saving rate of over 10% through intelligent technology. The energy-saving principle of this intelligent technology mainly lies in collecting the “energy demand side” of the factory’s production and intelligently controlling the “energy supply side’s equipment parameters” to achieve a supply-demand curve. This energy-saving effort aligns with the national carbon neutrality strategy.It is currently the most advanced intelligent control energy-saving solution for public utility workshops, primarily achieving reduced operation and maintenance costs through data monitoring and visualization, leveraging AI intelligent control to help workshops save energy and reduce consumption by 10-30%.
(Figure 8, Effect Diagram of AI Technology in Workshop)
(2)
Large Model Analysis and Prediction of Enterprise Gas Consumption
It is worth noting that the application of enterprise gas consumption statistical data analysis undoubtedly injects strong support for the development of industrial AI.Specifically, in the practice of optimizing energy management, we need to analyze the type of enterprise and collect enterprise data frequency to grasp the overall data situation of the enterprise. In addition, we can use LightGBM, LSTM, and ARIMA models to predict gas usage.In fact, the analysis and prediction of enterprise gas consumption data also provide us with suggestions for model selection, the importance of data preprocessing, and methods for evaluating model generalization capability.
Note: Each point represents the predicted value and actual value of gas usage for the next 24 hours at the current time point, where blue is the actual value and orange is the predicted value (with the x-axis as time and the y-axis as gas usage)
(Figure 9, Model Prediction Results)
The prediction results of the LightGBM model closely match the actual values, with an R² score of 0.829, indicating high accuracy of the model.
The R² calculation formula is 1 – sse/sst (sse is the sum of squared errors, sst is the total sum of squares)
In today’s data-driven modern business environment, enterprises can significantly enhance their overall value by effectively managing their data assets through capital planning. Particularly in the manufacturing sector, the standardization of data assets can not only enhance management efficiency and optimize processes but also improve the ability to anticipate market changes, thus promoting a significant increase in production capacity. However, the specific extent of this increase will be influenced by various factors, including the size of the enterprise, industry characteristics, infrastructure construction, and existing data management levels.
Specifically, the standardization of data assets can achieve a production capacity increase of 10% to 30% through automation and predictive analytics.Standardizing management processes can lead to a 5% to 15% increase in production capacity by improving decision-making efficiency and performance management.Furthermore, process standardization can achieve an additional capacity increase of 10% to 25% through process optimization and more effective resource allocation.Additionally, through predictive standardization, reducing errors and enhancing employee training efficiency can further increase capacity by 5% to 10%.
By integrating these standardization measures, not only can production efficiency be optimized, but product quality control can also be strengthened, thereby gaining an advantage in a competitive market environment.
Conclusion
In the future landscape of industry, artificial intelligence will serve as a brush, painting a grand blueprint for intelligent production. As technology continues to break through, AI will demonstrate its exceptional capabilities in predictive maintenance, quality control, supply chain optimization, and lead the industry into a more efficient, intelligent, and environmentally friendly new era. We look forward to the deep integration of AI and industry, bringing revolutionary changes to global manufacturing.