Low-Carbon Virtual Power Plant Scheduling With Carbon Capture Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional methods for low-carbon virtual power plant scheduling lack comprehensive parameter information, leading to poor applicability and increased energy consumption and carbon emissions, and rely on uncertain renewable energy output probability distributions.
Innovation Solution
A method and device for low-carbon virtual power plant scheduling that includes obtaining energy device information, generating energy acquisition cost information, and controlling energy devices based on objective parameter information, without requiring predetermined probability distributions of renewable energy output, incorporating detailed models of carbon capture and storage to improve energy efficiency and reduce emissions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional stochastic optimization methods are used for virtual power plant scheduling, then the scheduling can be performed with simplified assumptions, but the scheduling plans have poor applicability and lead to high energy consumption and carbon emissions
Solution Approach 1:
The patent transforms the scheduling approach from traditional stochastic optimization to deep reinforcement learning, fundamentally changing the parameter representation from probability distributions to state-value functions and policies. This enables the system to adapt to actual renewable energy output patterns without relying on predetermined probability distributions, thereby improving scheduling plan applicability while reducing energy consumption through more accurate and flexible decision-making.
2Adaptability or versatility
If traditional stochastic optimization methods are used for virtual power plant scheduling, then the computational model is simpler, but the scheduling plans have poor applicability and lead to high carbon emissions
Solution Approach 1:
The patent changes the fundamental parameters of the scheduling system by replacing traditional stochastic optimization parameters with deep reinforcement learning parameters (state representations, value functions, and policies). This transformation enables the system to generate scheduling plans with better applicability that actively reduce carbon emissions by making more accurate and adaptive decisions about energy resource allocation and renewable energy utilization.
3Use of energy by moving object
If comprehensive parameter information is considered in scheduling, then energy efficiency is improved, but the computational complexity and system requirements increase
Solution Approach 1:
The patent replaces traditional mechanical optimization methods with deep reinforcement learning, substituting complex mathematical programming with a data-driven approach that learns optimal scheduling policies through interaction with the environment. This substitution enables the system to consider comprehensive parameter information for improved energy efficiency while managing computational complexity through efficient neural network inference rather than exhaustive optimization calculations.
Data Source
AI summary
Provided is a method and device for low-carbon integrated energy system scheduling. A specific embodiment of this method comprise: obtaining the energy device information set for each energy device in the virtual power plant; generating the acquisition cost information for each energy device based on an energy device name, energy device parameter information and an energy device number in the energy device information set; generating the energy scheduling objective values and the parameter information of each energy device based on the preset constraint sets and energy acquisition cost information of each energy device; controlling each energy device in the virtual power plant to execute the energy scheduling tasks based on the objective energy device parameter information.