Central Plant Linear Solver for HVAC Thermodynamic State Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for predicting thermodynamic states in complex HVAC systems are inefficient in terms of computational resources, making it exhaustive to determine optimal operating parameters for minimizing power consumption.
Innovation Solution
A system utilizing a processing circuit that identifies a reduced subset of thermodynamic states using linear and non-linear solvers, reducing the computational burden by propagating known values and generating matrices to predict thermodynamic states and operating parameters efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a non-linear solver is used to predict thermodynamic states of HVAC devices, then prediction accuracy is maintained, but computational resource consumption increases significantly
Solution Approach 1:
The patent segments the thermodynamic state prediction problem into two parts: linear relationships (mass flow, energy balance) solved by efficient linear solvers, and non-linear relationships (pressure drops, heat transfer coefficients) solved by non-linear solvers. This segmentation allows most computations to use fast linear methods while maintaining accuracy only where physically necessary.
Solution Approach 2:
The patent applies partial action by using non-linear solvers only for specific non-linear relationships rather than the entire system. Linear solvers handle the majority of thermodynamic state predictions, reducing computational burden while maintaining sufficient accuracy for control optimization purposes.
2Reliability
If multiple candidate sets of operating parameters are evaluated to minimize power consumption, then optimal operation is achieved, but computational time and resources increase
Solution Approach 1:
The patent performs preliminary computation by pre-calculating system characteristics and linear relationship matrices during system initialization. This preliminary action enables faster evaluation of multiple candidate operating parameter sets during real-time optimization, as the computationally intensive linear solver setup is already complete.
Solution Approach 2:
The patent changes the computational parameters by switching between linear and non-linear solution methods based on the specific thermodynamic relationship being evaluated. This parameter change allows efficient batch evaluation of multiple operating scenarios by using the appropriate solver complexity for each calculation.
3Loss of information
If the full thermodynamic states of all HVAC devices are computed, then complete system understanding is achieved, but device complexity and computational load increase
Solution Approach 1:
The patent extracts and separates linear relationships from non-linear relationships in the thermodynamic model. By taking out the linear mass flow and energy balance equations to be solved independently by linear solvers, the complex non-linear pressure and heat transfer calculations are reduced to only where physically necessary.
Solution Approach 2:
The patent introduces linear solvers as intermediary computational tools to handle the linear portions of thermodynamic relationships. These linear solvers act as mediators between the known operating parameters and the final thermodynamic states, reducing the direct computational burden on the non-linear solver.
Data Source
AI summary
Systems and methods for predicting a plurality of thermodynamic states of a plurality of heat, ventilation, and air conditioning (HVAC) devices of an energy plant are disclosed. The system includes a processor and a non-transitory computer readable medium storing instructions when executed causing the processor to: obtain plant netlist data describing the plurality of HVAC devices of the energy plant and connections of the plurality of HVAC devices to corresponding nodes; identify, from the first reduced subset, a second reduced subset of the plurality of thermodynamic states to be predicted by propagating a known value of the plurality of thermodynamic states using a linear solver; predict the second reduced subset of the plurality of thermodynamic states using a non-linear solver; and determine the plurality of thermodynamic states of the energy plant at the plurality of nodes based on the predicted second reduced subset of the plurality of thermodynamic states.