The 'Final Stretch' of AI Empowerment.

/

Produced by | Business Show

-Business Show-01

Wuhan's East Lake High-Tech Development Zone, famously known as 'China's Optics Valley', boasts the Wuhan New Energy Research Institute Building, a regional landmark that attracts cutting-edge enterprises in new energy research and industrial applications.

Yiruide (Wuhan) IoT Co., Ltd. stands out as one such representative enterprise. Since its inception in 2022, Yiruide has delved deeply into the realm of new energy IoT, infusing fresh impetus into the region's new energy industry through technological innovation and scenario-based applications.

'As user electricity consumption habits diversify, the traditional peak shaving and valley filling model, guided by electricity prices, falls short of meeting current complex and varied demand,' explained Li Zongjin, CEO of Yiruide.

In this context, how does one address the issue of power load? Yiruide has devised a load forecasting model for the EasyTable Smart Energy Management System, crafting a more scientific and rational energy allocation strategy.

What's the rationale behind this strategy? Li Zongjin told 'Business Show' that in the realm of new energy, wind power and photovoltaic power generation have emerged as significant forces in power supply due to their lower cost advantages. Conversely, traditional grid electricity often comes at a higher price.

How can new energy be leveraged to reduce electricity costs? This necessitates the introduction of an intelligent energy adjustment mechanism, with AI decision-making playing a pivotal role.

EMS Smart Cockpit System at Wuhan New Energy Building

Li Zongjin further elaborated to 'Business Show' that this process mirrors the 'buy low, sell high' strategy in stock trading. In the stock market, investors maximize returns by accurately timing the lows and highs of stock prices. Similarly, in energy adjustment, the AI system guides users to absorb low-cost electricity during price troughs by monitoring grid electricity prices in real-time. When prices surge, the stored low-cost electricity is released, effectively offsetting the cost of high-priced grid electricity.

This AI-driven dynamic energy adjustment model not only enhances the efficiency of new energy usage but also generates economic value for enterprises and users. Especially in complex electricity consumption environments, it aids enterprises in establishing a more scientific energy allocation plan.

Taking the Wuhan New Energy Research Institute's application as an example, the system has achieved comprehensive monitoring and management of the park's energy and environment, significantly boosting energy utilization efficiency and cutting energy consumption costs by 30.05%. Annually, the system generates approximately 300,000 kWh of clean electricity, reducing carbon emissions by over 250 tons.

Li Zongjin noted that the system has introduced a 'one-map product' leveraging digital twin technology to innovate operation modes and enhance usage and monitoring efficiency. Users can now interact intelligently with an intuitive digital twin interface, eliminating the need for maintenance personnel to constantly monitor large screens. This significantly reduces labor costs and operational complexity, making energy management more convenient and efficient.

More importantly, when green electricity is integrated into the power generation energy management system, the EasyTable system's advantages become even more pronounced. It not only precisely manages electricity costs, helping users achieve cost reduction targets, but also promotes the achievement of dual carbon goals and the practice of green development concepts. It is estimated that a military industrial enterprise customer can save 10% of costs annually after adopting Yiruide's EasyTable Smart Energy Management System.

-Business Show-02

In Jiangxia District, Wuhan, at Jiulong Reservoir, a drone hangar site operated by Puzhou Technology is in full swing.

Puzhou Technology's Drone Hangar Site at Jiulong Reservoir, Jiangxia District, Wuhan

Equipped with advanced AI recognition technology, the drone possesses 'fiery eyes'. When detecting illegal fishing by the lake, AI can swiftly identify and report the information, with the backend further assessing the legality of the behavior. Upon completing the task, the drone accurately returns to its original position.

Huang Kan, solution architect at Puzhou Technology, told 'Business Show' that the drone monitoring system employs multiple cutting-edge technologies. Especially in concurrent AI algorithms, the system can load multiple algorithms simultaneously on the same screen, achieving parallel processing and efficiently distributing data to various departments, ensuring a unified line and multiple data distribution.

Lei Zhen, marketing director of Puzhou Technology, informed 'Business Show' that the system breaks down data barriers traditionally associated with drone usage, achieving deep integration of real-time drone data with AI. This changes the scenario where fire protection, emergency response, and other departments used drones separately, with hardware and data not being shared. It also solves the issue of departments' high-frequency drone usage relying on professional pilots.

There are primarily two types of algorithm applications. For instance, some drones are equipped with 10.4T computing power, enabling real-time millisecond-level recognition of targets such as smoke and fire points during flight, enhancing efficiency by hundreds of times. For complex algorithms, processing is done on the mid-platform, with data download completed in 1 to 2 minutes and problem mining within tens of seconds, boosting efficiency by tens of times.

Real-time millisecond-level recognition of fire points

Yang Kai, general manager of Puzhou Technology's Smart City Business Department, stated that by leveraging big data and the PaddlePaddle deep learning platform, drones can accurately identify various abnormalities such as road congestion and fire hazards, akin to experienced inspectors, and issue timely warnings. This improves the accuracy and timeliness of low-altitude inspections, providing security for urban safety and management.

It is understood that Puzhou Technology's low-altitude visual AI mid-platform, built on PaddlePaddle, has established an 'unattended' supervision system covering scenarios like urban management, traffic monitoring, forest fire prevention, and water area management. For example, in forest parks and farmlands, illegal activities like burning garbage and straw occur.

Previously, due to limitations in inspection methods and monitoring means, staff could only discover these issues hours or even the next day, by which time the fire might have spread. However, the current system can automatically identify smoke and issue timely warnings, enabling staff to promptly respond to the scene, enhancing the timeliness of inspections for illegal burning in regional parks, farmlands, and other areas, and improving inspection efficiency by 200%.

Yang Kai further introduced that the platform supports over 50 algorithms, including traffic flow statistics, firework recognition, vessel recognition in no-fishing zones, and personnel search and rescue, adaptable to the needs of multiple smart city fields. It has been applied to various industry customers in government, energy, communications, and other sectors, becoming a key infrastructure for urban governance digital transformation in the era of the low-altitude economy.

-Business Show-03

From Yiruide's EasyTable Smart Energy Management System to Puzhou Technology's drone monitoring at Jiulong Reservoir, the innovation behind these two Wuhan Optics Valley enterprises is inseparable from the 'super assistance' of Baidu PaddlePaddle.

As China's first self-controllable industrial-grade deep learning platform, as of April 2025, Baidu PaddlePaddle Wenxin boasts 21.85 million developers, serving 670,000 enterprises and creating 1.1 million models.

Yiruide's EasyTable Smart Energy Management System realizes a load forecasting model based on PaddlePaddle, serving as the 'brain of the microgrid'. It senses and manages all energy in real-time throughout the microgrid, responsible for fine-grained control, integrating AI + cloud-edge integrated architecture to achieve seamless collaboration between heterogeneous energy sources. Among them, the load function is powered by Baidu TS, fostering economically efficient energy operations and aiding enterprises in cost reduction and efficiency enhancement. Currently, this product has been incorporated into Baidu's ecosystem offerings.

Data reveals that after deployment, the accuracy of power load forecasting soared from 50% to 86%, photovoltaic power generation forecasting accuracy reached 83%, and energy storage scheduling efficiency improved by 40%.

Relying on the Baidu PaddlePaddle platform, Puzhou Technology has achieved a drone technology upgrade. Through big data and deep learning platforms, drones achieve accurate anomaly recognition. In practical applications, traffic flow statistics accuracy is 96%, and firework recognition accuracy is 95%.

Yiruide and Puzhou Technology are but a microcosm of the development of 229 AI enterprises in Wuhan. These enterprises have achieved intelligent transformation and development with the support of Baidu PaddlePaddle.

Zhou Qi, deputy general manager of Baidu AI Technology Ecosystem, told 'Business Show' that as a foundational AI development platform, Baidu PaddlePaddle provides enterprises with a suite of model development tools and a pre-trained model library, enabling them to develop their own models at lower costs, more efficiently, and with superior performance. Crucially, enterprises can create products with unique characteristics by integrating their own data.

Baidu PaddlePaddle (Wuhan) AI Industry Empowerment Center

'Through collaborative research and creation, Baidu works closely with their technical teams, enabling the combined team to produce results greater than the sum of its parts,' said Zhou Qi. This cooperation model not only aids enterprises in swiftly realizing AI technology applications but also promotes the continuous optimization and upgrading of the Baidu PaddlePaddle platform in practical applications, providing innovative and intelligent support for enterprises like Puzhou Technology and Yiruide.

As AI technology transitions from laboratories to fields and farms, the Baidu PaddlePaddle (Wuhan) AI Industry Empowerment Center, serving as the 'final stretch' of AI empowerment, is also driving AI from 'usable' to 'easy-to-use' and 'commonly used'. [End]