7 results about "Thermodynamic equations" patented technology

The application discloses a kind of polymer stress-aging sensitivity characterization methods under multi-field combined action, belong to dielectric material aging evaluation technical field.The method includes by constructing the molecular chain fracture energy barrier reduction model based on transition state theory, obtain theoretical energy barrier reduction threshold and aging rate constant, calculate to obtain the dynamic life index under different aging time;Based on the polymer sample after aging, by thermogravimetric analysis and Coats-Redfern method constructs thermal decomposition kinetics model, and the activation energy under different aging states is calculated;By comparing and analyzing the change law of life index and activation energy, the comprehensive sensitivity evaluation under the electric-thermal coupling effect is carried out;The physical self-consistency of model is verified by introducing entropy compensation effect thermodynamic equation.The application realizes the comprehensive quantitative characterization of polymer stress-aging sensitivity under electric-thermal combined action, with the advantages of clear physical meaning, can simultaneously characterize electric and thermal stress, clear mechanism explanation and the like.

The invention relates to the cross technical field of deep learning and meteorological monitoring, in particular to an extreme weather intelligent monitoring and early warning method and system based on deep learning, and the method comprises the steps: collecting and processing data, and constructing a meteorological state tensor; inputting the tensor into a space-time Transform architecture with a meta-learning capability, and extracting cross-scale meteorological features by capturing long-range correlation through an encoder and integrating a space-time convolution gating cycle unit through a decoder; an adversarial training mechanism is introduced, and a numerical weather forecast mode is used as a judgment reference to optimize features; establishing a federal learning model updating mechanism to realize distributed optimization; and inputting the predicted trajectory into a power grid digital twin system, solving an equipment thermodynamic equation through a physical information neural network, and feeding back to a feature extraction process to form a closed loop. According to the method, the problems of low identification accuracy and poor early warning timeliness caused by multi-source heterogeneity and strong time sequence nonlinearity of extreme meteorological data are solved.

The application discloses a fire-fighting effect evaluation method under digital modeling, and relates to the technical field of fire-fighting evaluation and public safety management. In a fire-fighting layout digital model, a parallel simulation calculation task is established for each preset fire scene, a dynamic fire development thermodynamic equation and a smoke diffusion calculation model are introduced, a flame heat spreading and diffusion process, a three-dimensional diffusion process of toxic smoke, an automatic starting response process of fire-fighting facilities after reaching a threshold value and an escape and evacuation process of people in a building are independently simulated under multi-level fire conditions, and dynamic simulation characteristic data under each fire scene is batch output. The application constructs a set of physical level general and interactive digital modeling and simulation framework. By introducing a bottom fluid partial differential equation and a thermodynamic model, the application can seamlessly adapt to modern buildings of any heterogeneous structure and complex floor without the need of developing a special model for each evaluation task.

This invention discloses a big data-based intelligent heating management method and system, belonging to the field of heating management technology. The method includes: abstracting the heating system into a directed graph structure and constructing an adjacency matrix that integrates physical coupling and adaptive learning; collecting system operation data to construct node feature matrices and edge feature matrices; constructing a spatiotemporal graph neural network (PI-STGNN) that integrates thermodynamic equation constraints, outputting the predicted state values ​​of each node in the pipeline network; and based on the prediction results, employing a centralized training-distributed execution multi-agent reinforcement learning framework, with system energy efficiency and user comfort as optimization objectives, outputting the optimal control action. This invention combines graph neural networks, physical information constraints, and multi-agent reinforcement learning to achieve intelligent and precise control of the heating system, effectively improving heating quality and energy utilization efficiency.

The present application provides a kiln temperature prediction method based on the fusion of space-time Transformer and physical information neural network (PINN), belonging to the technical field of intelligent control of industrial kiln. The method collects historical combustion parameters of the burner, models time evolution and space coupling through a space-time coupling Transformer model, captures the time sequence dependence of the combustion parameters through a time attention mechanism, models the heat coupling effect between multiple burners through a space attention mechanism, then introduces a PINN framework and embeds the oxygen-enriched combustion thermodynamic equation as an explicit constraint into the training process, ensures that the prediction conforms to the energy conservation and heat transfer law through a physical loss function, and realizes parameter inversion through learnable physical parameters to automatically discover the true combustion characteristics, thereby fusing the deep learning fitting capability and the prior knowledge of physical law, and improving the accuracy, interpretability and generalization ability of kiln temperature prediction.

The application discloses a kind of lithium ion battery distributed thermal process sensor fault estimation method, comprising: using two-dimensional partial differential equation to describe rectangular lithium ion battery distributed thermodynamic equation;Distributed thermal model of large size lithium ion battery is constructed;Using Chebyshev-Galerkin method, distributed thermodynamic equation and distributed thermal model are decomposed into reduced order model described by standard state space equation in time domain;State variable in reduced order model is constructed using sensor measurement output and a Hurwitz matrix, and enhanced reduced order model is constructed;Enhanced adaptive observer and error state space equation are constructed using enhanced reduced order model;Based on error state space equation, the fault strength of temperature sensor that fails is estimated using fast adaptive algorithm.The application can effectively estimate sensor fault, including time-invariant and time-varying fault, and single sensor and multiple sensor fault.

The application provides a simulation verification method for the ejection property of a micro supercharging device, which is designed by combining the integrated application principle of gunpowder, a thermodynamic equation and a physical motion equation, obtaining simulated speed and acceleration, and comparing the simulated speed and acceleration with the initial speed and maximum overload measured in the actual test process to complete the simulation verification of the ejection property. The simulation of the application has high authenticity, can efficiently solve the high-overload safety hazard during the ejection of the barrel, and provides a good adjustment basis for the ejection parameters.