This invention discloses a multivariate collaborative
optimization system for industrial circulating
water flow for freezing protection of liquid alkali storage tanks, relating to the field of liquid alkali
storage tank freezing technology. First, it collects thermal flow
infrared images of the liquid alkali
storage tank surface and
pipe nodes, flow sensor data, and
environmental temperature and
humidity parameters to construct a multivariate
dynamic feature matrix. Then, it extracts freezing sensitivity indicators for each key node to construct a
topological map of thermally vulnerable areas. By performing time-series clustering of node flow velocity fluctuations, a freezing conduction probability weight map is generated. Next, the weight map is embedded into the circulating water network optimization objective function, and an improved
particle swarm optimization algorithm is used to output the target flow rate of each
branch. Based on this, the variable frequency operation parameters of the
circulating pump are calculated to drive the hydraulic
system adjustment. When the freezing critical value approaches, a micro-circulation heat exchange mode is automatically triggered. This invention can achieve accurate prediction and proactive response to freezing risks, improving the safety and energy efficiency of liquid alkali
storage tank operation in winter.