Preventing Hundreds of Billions in Annual Factory Downtime Losses: This Company Gives Industrial Equipment an Early “Checkup” | Product Watch

Predictive maintenance has low penetration in Chinese factories.

Unexpected downtime is a major determinant of factory productivity. A single stoppage can halt an entire line and cause losses of millions or tens of millions. Downtime may even be a relatively mild consequence: equipment failures at wind farms, chemical plants and steel mills can cause major accidents with greater losses. Despite decades of industrial growth in China, monitoring remains largely manual. Workers inspect machines at set times, report faults to technicians, who respond according to severity and report through workshop and plant managers. This inspection-based maintenance takes time, reduces efficiency and creates hidden losses. Predictive maintenance has entered many factories in the Industry 4.0 era but remains relatively new in China's industrial equipment sector. It monitors performance and condition during normal operation to reduce failure probability. According to the China Electronics Standardization Institute, only 14% of manufacturing enterprises had begun implementing predictive maintenance in 2020.

Qianhai IntelliUnion is a Chinese startup focused on predictive maintenance and recently received Pre-Series A investment from Shenzhen HTI. Its Intelliunion Cloud monitoring series includes multiple products and different monitoring methods for different units. Vibration, temperature and other sensors capture real-time changes at critical equipment locations. Results support performance assessment, fault prediction, life prediction and planned maintenance, extending equipment life and ensuring safe, stable production. The series is one of China's few predictive maintenance systems applicable across industries, with all products independently manufactured domestically.

For a startup, Qianhai IntelliUnion's greatest advantage is its extensive accumulated data, unusual in the still-emerging intelligent industrial O&M sector. General Manager Zheng Feihong told 36Kr that Intelliunion Cloud incorporates 60 categories of rules, fault models from tens of thousands of equipment units and multiple products. “This is like putting experts' knowledge into computers and diagnosing faults by elimination. Each rule has corresponding fault causes and triggers alarms. Several rules may trigger on site, introducing errors; the accumulated fault models then optimize the result by comparing both types of data to identify the problem. The system speeds response and diagnosis and continuously learns from different faults. This dataset is unique in China's intelligent O&M industry. Building it requires accumulated experience and access to fault data from different factories. Many companies lack this history, and much industrial equipment data remains difficult to obtain,” Zheng explained.

The company's data advantage is closely tied to its founding team, which came from Alstom Chuangweishi Technology Development (Shenzhen) Co., Ltd. Tianyancha records show that Alstom Chuangweishi belonged to Alstom Power's Energy Management business and provided remote online industrial monitoring technology and products. It was once China's largest supplier of online rotating-equipment monitoring and analysis systems and offered what was described as the world's only internet-based remote rotating-equipment monitoring technology and services. It worked with Sinopec, PetroChina and Daya Bay Nuclear Power Station, providing online monitoring for over 2,000 units at more than 400 companies. Its predecessor, Chuangweishi, was founded in 1998 as one of China's earliest equipment-monitoring technology companies and acquired by Alstom in 2006. Zheng said: “We each worked there for around ten years. Alstom's extensive monitoring business in China accumulated substantial data. After leaving, we used that data to build a big-data fault analysis system. Its customers spanned wind, thermal and nuclear power, petrochemicals and metallurgy, enabling our products to serve multiple industries.” Huaan Securities notes that China's condition-monitoring and fault-diagnosis industry started later than in some developed countries. It comprises three groups: complete-solution providers with independent R&D and production, cross-industry coverage and diagnostic capability; specialist product providers serving one segment; and integrators/agents without production or service capabilities. Historical data has helped IntelliUnion become one of the few complete-solution providers.

Why Intelligent Monitoring Systems Are Necessary

Intelliunion Cloud operates on the principle that machinery vibrates during operation, conveying information about whether it is healthy or whether a component has developed a problem. The system produces an “electrocardiogram” from the vibration, which trained monitoring engineers interpret to identify problems.

“When everyone speaks at a meeting, each person's voice is different, but mine can be drowned out. Small mechanical faults are similar: their energy is masked by other energy. Initially the changes are slight. Our system finds and extracts the hidden fault energy and analyzes its magnitude and changes to predict when problems will occur.”

Zheng told 36Kr: “If a machine rotates, Intelliunion Cloud can monitor it.”

In one IntelliUnion case, a water treatment plant's high-voltage motor had elevated vibration at the non-drive end and noticeably more noise than comparable motors. Maintenance staff could not identify the cause. Analysis of non-drive-end waveforms, spectra and vibration trends revealed substantial energy in a certain frequency range and clear abnormalities. The diagnosis was an inner-race bearing fault with severe poor lubrication. Maintenance staff replaced the non-drive-end bearing and confirmed severe inner-race wear.

Predictive maintenance is clearly valuable for factories. It anticipates faults and enables targeted repairs, extends equipment life and utilization, predicts service life rationally, improves output and quality, reduces O&M costs and spare-parts waste, lowers inventory and tied-up capital, and reduces field maintenance workloads.

Zheng cited a domestic nuclear power plant with 8,000 pumps and 8,000 motors—16,000 units inspected by one person. Completing a full round once a month would already be good. A pump fault could alter pressure upstream and, in severe cases, shut down the entire nuclear station with enormous losses. By the time the inspector works through the problem step by step, it is too late.

Deloitte's Predictive Maintenance and the Smart Factory report states that inappropriate maintenance strategies can reduce plant capacity by 5%–20%, while unexpected downtime costs industrial plants as much as US$530 billion annually.

At present, most Intelliunion Cloud customers come from the wind power industry.

“Wind power is a major trend in China, and demand for predictive maintenance is strong. We are focusing on this sector and developing more products,” Zheng said.

The wind industry does indeed have relatively high demand for predictive maintenance.

Huaan Securities notes that wind turbines are expensive: units above 2 MW often cost around RMB 10 million. Compared with potentially high hardware-damage losses, operators are less sensitive to the added cost of monitoring. Turbines are typically located in deserts, coastal areas, mountains or offshore, making rapid access and timely delivery of parts and resources difficult and repairs costly and time-consuming. Downtime losses are significant: a seized motor can cost more than RMB 10,000 in daily revenue. Long-term fault-free operation is particularly important in a sector with many private operators.

By 2015, China's cumulative wind capacity had reached 145.4 GW. A further 183.1 GW was added over the following six years, reaching 328.5 GW by the end of 2021, a 14.6% CAGR. More than 44% of installed turbines had operated for over six years, with failure probability rising as mechanical life is consumed and electrical equipment ages. Some early installations lack monitoring and O&M systems, creating substantial retrofit opportunities.

According to Zheng, IntelliUnion developed the Intelliunion Cloud WindMon turbine monitoring series specifically for wind power, with 14 product modules supporting predictive maintenance.