From the Old Testament of the Bible to the Great Yu of the East, the struggle against floods runs through the development of human civilization. Public statistical data shows that flood disasters have been one of the major natural disasters in recent years. Even worse, the greenhouse effect caused by global warming has intensified, resulting in a 1.5 ° C increase in global average temperature (Option 1a, blue line). This is the main reason for the approximately 700% increase in the number of major floods worldwide over the past forty years (Scheme 1a, red dot). In addition, the acceleration of global urbanization has led to an increasing number of regions around the world facing higher flood risks (Option 1b). Statistical data shows that in recent years, floods have been characterized by widespread danger areas and severe impacts in both developed and developing countries, causing significant casualties and economic losses. For example, the floods in Derna, Libya in September 2023 resulted in 11000 deaths due to rainfall far exceeding the warning limit of the drainage system and not being detected in a timely manner. Similarly, in 2024, heavy rainfall associated with Hurricane Helen caused rapid rise in river water levels, resulting in extensive property damage and at least 236 deaths in Dixie, USA.

Floods caused by heavy rain are usually predicted through weather forecasts based on satellite observations, but it is difficult to determine the areas where floods may suddenly form. The characteristic of floods is that they evolve rapidly in low-lying areas with unpredictable paths, including underground passages, subway stations, and villages at the foot of mountains, which account for about 70% of the casualties caused by flood disasters. The necessity of early warning of flood formation and minimizing its destructive impact has prompted people to conduct in-depth research on establishing flood monitoring systems. The traditional monitoring method based on manual patrols is limited by high costs and potential omissions. Installing cable warning systems in low-lying areas can provide higher safety and accuracy for early flood warning, but in addition to potential failures due to leakage during flood events, they also require extremely expensive installation and maintenance. In addition, after each flood, the cable warning system requires time-consuming and laborious reconstruction. These drawbacks make it difficult to achieve real-time monitoring in low-lying areas of cities and suburbs through cable flood warning systems.

Wireless sensors have been considered for early flood warning, consisting of sensing elements, signal generators, and batteries. These sensors can quickly respond to rapidly developing floods and transmit warning signals to the control center through wireless communication. Compared to cable powered sensors, battery powered sensors are more economical and can adapt to more complex situations. However, the use of mobile batteries also faces the problem of short circuits caused by depletion and water damage during long-term use, which seriously affects the effective warning of floods by sensors. Therefore, stable and long-term energy supply is basically needed to replace the batteries in flood warning sensors. Renewable energy sources such as wind and solar power were once considered as continuous energy sources for flood warning sensors. However, these devices strongly rely on collecting external energy and are limited by their large size, poor stability, and high cost, which is contrary to the practical application requirements of smaller or enclosed urban spaces such as bridges and tunnels. In this context, flood warning sensors are expected to require low or even no energy consumption in safe situations, while electricity itself can cope with the occurrence of floods, with broad adaptability and long-term guarantee. The competitive strategy for developing flood warning sensors with self powered capabilities is to collect the kinetic energy of floods and convert it into electrical energy. Sensors based on frictional or piezoelectric effects are very attractive in monitoring water level changes, demonstrating the consistency of self powered flood energy. Considering the limited output of frictional and piezoelectric generators, activating more complex sensors that require higher energy consumption is a challenge, especially for continuous long-distance signal transmission.