研究目的
To investigate the relationship between installing a solar harvesting system to power base stations of a cellular network and the energy management under varying demand, and to challenge the belief that solar energy can be considered free and should always be installed everywhere.
研究成果
The study concludes that solar installation and energy management in cellular networks are tightly interrelated, and their joint optimization is necessary for cost-effectiveness. The order of introducing technologies impacts performance, with solar-first sequential optimization being close to joint optimal. Installing solar everywhere is not always optimal, even when solar energy is cheaper than grid energy, due to capital costs and network dynamics. Future work should address real-time algorithm integration and larger network scalability.
研究不足
The model uses an average day approximation, which may oversimplify real-time variations in solar energy and traffic. It does not optimize the number of solar panels and batteries, which are chosen based on preliminary sizing. The computational complexity limits the network size that can be handled optimally, requiring heuristics for larger networks. Costs for cables and labor are not included, potentially biasing results in favor of solar energy.
1：Experimental Design and Method Selection:
A mathematical model is proposed as a linear integer program to optimize the joint planning of solar installation and energy management in cellular networks, capturing the synergy between solar panels, batteries, inverters, charge controllers, and base station sleep modes. The model is based on an average day approach to handle short-term variations in solar energy and traffic demand.
2：Sample Selection and Data Sources:
The study uses networks of varying sizes, from small (4 base stations, 12 test points) to larger ones (up to hundreds of base stations), with parameters derived from real data sources such as solar radiation profiles from Palermo, Italy, and base station energy parameters from literature.
3：List of Experimental Equipment and Materials:
Solar panels (monocrystalline silicon), batteries (flooded lead acid), inverters, charge controllers, and micro base stations are considered. Specific models and brands are mentioned in the parameters section.
4：Experimental Procedures and Operational Workflow:
The model is solved using Gurobi solver with AMPL pre-processor. Scenarios include base case (no solar or sleep mode), single technology cases (only solar or only sleep mode), sequential scenarios (solar first or sleep mode first), and joint optimization. The optimization involves variables for solar installation, base station states, and user assignments over time periods.
5：Data Analysis Methods:
Results are analyzed in terms of total cost, solar energy usage, grid energy usage, and network dynamics. Statistical measures like dispersion and standard deviation of user assignments are computed to evaluate network balance.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
