👨🎓个人主页
💥💥💞💞欢迎来到本博客❤️❤️💥💥
🏆博主优势:🌞🌞🌞博客内容尽量做到思维缜密,逻辑清晰,为了方便读者。
⛳️座右铭:行百里者,半于九十。
📋📋📋本文目录如下:🎁🎁🎁
目录
💥1 概述
文献来源:
储能技术广泛应用于调频、调峰、平抑可再生能源出力波动、需求侧响应、提高用户可靠性等领域,对能源互联网的发展起到重要支撑作用。通过储能系统在电网电价谷时段存储电能,在电价峰时段释放电能供给用户,可以为用户节省用电成本,同时缓解电网调峰压力[2,3,4]。国家和地方政府大力推广储能技术的应用,储能的发展前景广阔。
目前关于共享储能的研究处于起步阶段,现有的工作以共享储能系统为主要研究对象,分析共享储能系统的商业模式和盈利情况,没有对用户参与共享储能系统的充放电行为和经济效益做深入研究。
本文在用户群间引入共享储能电站,建立以用户群日运行成本最优为目标的优化调度模型,分析用户群接入共享储能电站后的充放电行为和经济效益,并对共享储能电站的投资回收年限等经济性指标与服务费定价关系做出进一步的研究。
1.1 共享储能电站概念及运营模式
共享储能电站的概念如图 1 所示,储能电站运营商利用资金优势在用户群间建立大型共享储能电站,对储能电站进行统一运营管理,为同一配电网区域内的多个用户提供共享储能服务。
1.2 基于共享储能的优化调度模型
基于共享储能电站的工业用户日前优化经济调度研究
一、共享储能电站的定义与核心特点
共享储能电站是一种由第三方投资建设、布局于电网关键节点的集中式独立储能设施,其核心特征在于独立性与共享性。
- 独立性:直接与电网调度机构并网,不受地理位置限制,可独立参与电力市场交易。
- 共享性:通过分割使用权与收益权,为多个新能源电站或工业用户提供储能服务,实现资源优化配置。例如,江苏某项目通过容量租赁模式服务3个工业用户,降低其初始投资成本达20%以上。
- 商业模式多样性:包括容量租赁+现货市场套利(占收益60%以上)、调频辅助服务(贡献约30%收益)及容量补偿等。截至2022年,中国已投运38座共享储能电站,总规模3GW/6GWh,在建项目达16.5GW/35GWh。
二、工业用户日前优化经济调度的关键模型与算法
-
-
求解算法
- 集中式优化:采用CPLEX或Gurobi求解混合整数线性规划(MILP),适用于中小规模问题,计算效率高但隐私性差。
- 分布式算法:如改进粒子群优化(PSO)和遗传算法(GA),通过局部通信实现全局优化,适用于多用户协同场景,鲁棒性强但收敛速度较慢。某案例中,PSO算法较传统等比例分配方案降低调度成本19.7%,并延长储能寿命23.2%。
三、共享储能在工业调度中的应用场景与效益分析
-
典型应用场景
- 峰谷价差套利:利用分时电价差异,在低谷时段充电(如0.3元/kWh)、高峰时段放电(如0.8元/kWh),套利空间可达0.5元/kWh。
- 需求响应支持:通过平抑负荷波动(如服装生产类工业负荷波动率降低至10%以下),减少需量电费支出。
- 可再生能源消纳:与光伏/风电协同,提升新能源渗透率至30%以上,减少弃光弃风率。
-
经济效益量化
- 用户侧:江苏3个工业用户案例显示,引入共享储能后日运行成本降低18%-25%,静态投资回收期缩短至5-7年。
- 运营商侧:年服务费收益与定价呈正相关,当服务费为0.15元/kWh时,投资回报率(IRR)可达12%。
四、容量配置与电价预测的关键影响
-
容量优化配置
- 多目标优化模型:以经济性(度电成本≤0.4元/kWh)和可靠性(SOC维持于20%-80%)为目标,采用遗传算法确定最优容量。某分布式网络案例显示,配置0.5MW储能可降低全网配置容量15%。
- 梯次调度策略:根据SOC状态动态分配充放电任务,使6组储能集装箱SOC趋同,提升利用率10%以上。
-
电价预测方法
- 深度学习模型:基于CNN-GRU结合自注意力机制,预测误差(MAPE)≤4%,R²值达0.96以上,显著优于传统ARIMA模型。
- 多模型融合:通过SVM、LSTM和XGBoost动态加权(权重分配比0.3:0.5:0.2),提升日前节点电价预测精度,指导储能充放电策略。
五、现有研究与未来方向
-
文献综述
- 协同优化模型:文献[19]提出社区级共享储能与用户聚合商协同,降低用能费用12%。
- 双层规划:文献[21]构建储能-微网双层优化模型,但未解决利益分配问题,需引入纳什议价方法。
-
挑战与展望
- 技术层面:需开发兼顾隐私保护(如联邦学习)与计算效率的分布式算法。
- 市场机制:探索容量租赁定价动态博弈模型,平衡用户成本与运营商收益。
- 政策支持:完善辅助服务市场规则,明确共享储能参与调频、备用的补偿标准。
六、结论
共享储能电站通过资源聚合与优化调度,为工业用户提供了低成本的灵活性资源。日前优化经济调度模型需综合负荷特性、电价波动与储能约束,采用智能算法实现成本最小化。未来需进一步融合市场机制与技术革新,推动共享储能在工业领域的规模化应用。
📚2 运行结果
2.1 场景分析
CPXPARAM_Simplex_Display 2
CPXPARAM_MIP_Tolerances_MIPGap 9.9999999999999995e-07
CPXPARAM_Barrier_Display 2
Tried aggregator 3 times.
MIP Presolve eliminated 479 rows and 53 columns.
Aggregator did 159 substitutions.
Reduced MIP has 373 rows, 229 columns, and 1198 nonzeros.
Reduced MIP has 46 binaries, 0 generals, 0 SOSs, and 0 indicators.
Presolve time = 0.02 sec. (1.67 ticks)
Found incumbent of value 4919.153611 after 0.03 sec. (2.56 ticks)
Probing fixed 0 vars, tightened 46 bounds.
Probing time = 0.00 sec. (0.05 ticks)
Tried aggregator 1 time.
MIP Presolve eliminated 256 rows and 147 columns.
MIP Presolve modified 64 coefficients.
Reduced MIP has 117 rows, 82 columns, and 396 nonzeros.
Reduced MIP has 17 binaries, 0 generals, 0 SOSs, and 0 indicators.
Presolve time = 0.00 sec. (0.31 ticks)
Probing time = 0.00 sec. (0.01 ticks)
Tried aggregator 1 time.
Reduced MIP has 117 rows, 82 columns, and 396 nonzeros.
Reduced MIP has 17 binaries, 0 generals, 0 SOSs, and 0 indicators.
Presolve time = 0.02 sec. (0.19 ticks)
Probing time = 0.00 sec. (0.01 ticks)
MIP emphasis: balance optimality and feasibility.
MIP search method: dynamic search.
Parallel mode: deterministic, using up to 16 threads.
Root relaxation solution time = 0.00 sec. (0.29 ticks)
Nodes Cuts/
Node Left Objective IInf Best Integer Best Bound ItCnt Gap
* 0+ 0 2776.7029 2633.8312 5.15%
0 0 cutoff 2776.7029 24 0.00%
Root node processing (before b&c):
Real time = 0.05 sec. (7.26 ticks)
Parallel b&c, 16 threads:
Real time = 0.00 sec. (0.00 ticks)
Sync time (average) = 0.00 sec.
Wait time (average) = 0.00 sec.
------------
Total (root+branch&cut) = 0.05 sec. (7.26 ticks)
----------用户A部分------------
最优储能容量规划值为 : 2796.5625 kWh
最优储能充放电功率最大值为 : 590 kW
----------用户B部分------------
最优储能容量规划值为 : 739.2434 kWh
最优储能充放电功率最大值为 : 130 kW
----------用户C部分------------
最优储能容量规划值为 : 698.0099 kWh
最优储能充放电功率最大值为 : 80 kW
>>
CPXPARAM_Simplex_Display 2
CPXPARAM_MIP_Tolerances_MIPGap 9.9999999999999995e-07
CPXPARAM_Barrier_Display 2
Tried aggregator 3 times.
MIP Presolve eliminated 479 rows and 53 columns.
Aggregator did 159 substitutions.
Reduced MIP has 373 rows, 229 columns, and 1198 nonzeros.
Reduced MIP has 46 binaries, 0 generals, 0 SOSs, and 0 indicators.
Presolve time = 0.02 sec. (1.67 ticks)
Found incumbent of value 4919.153611 after 0.03 sec. (2.56 ticks)
Probing fixed 0 vars, tightened 46 bounds.
Probing time = 0.00 sec. (0.05 ticks)
Tried aggregator 1 time.
MIP Presolve eliminated 256 rows and 147 columns.
MIP Presolve modified 64 coefficients.
Reduced MIP has 117 rows, 82 columns, and 396 nonzeros.
Reduced MIP has 17 binaries, 0 generals, 0 SOSs, and 0 indicators.
Presolve time = 0.00 sec. (0.31 ticks)
Probing time = 0.00 sec. (0.01 ticks)
Tried aggregator 1 time.
Reduced MIP has 117 rows, 82 columns, and 396 nonzeros.
Reduced MIP has 17 binaries, 0 generals, 0 SOSs, and 0 indicators.
Presolve time = 0.02 sec. (0.19 ticks)
Probing time = 0.00 sec. (0.01 ticks)
MIP emphasis: balance optimality and feasibility.
MIP search method: dynamic search.
Parallel mode: deterministic, using up to 16 threads.
Root relaxation solution time = 0.00 sec. (0.29 ticks)
Nodes Cuts/
Node Left Objective IInf Best Integer Best Bound ItCnt Gap
* 0+ 0 2776.7029 2633.8312 5.15%
0 0 cutoff 2776.7029 24 0.00%
Root node processing (before b&c):
Real time = 0.05 sec. (7.26 ticks)
Parallel b&c, 16 threads:
Real time = 0.00 sec. (0.00 ticks)
Sync time (average) = 0.00 sec.
Wait time (average) = 0.00 sec.
------------
Total (root+branch&cut) = 0.05 sec. (7.26 ticks)
----------用户A部分------------
最优储能容量规划值为 : 2796.5625 kWh
最优储能充放电功率最大值为 : 590 kW
----------用户B部分------------
最优储能容量规划值为 : 739.2434 kWh
最优储能充放电功率最大值为 : 130 kW
----------用户C部分------------
最优储能容量规划值为 : 698.0099 kWh
最优储能充放电功率最大值为 : 80 kW
>>
2.2 接入共享储能电站优化结果分析
Aggregator did 91 substitutions.
Reduced MIP has 259 rows, 277 columns, and 858 nonzeros.
Reduced MIP has 69 binaries, 0 generals, 0 SOSs, and 0 indicators.
Presolve time = 0.02 sec. (1.05 ticks)
Probing fixed 9 vars, tightened 111 bounds.
Probing time = 0.00 sec. (0.09 ticks)
Cover probing fixed 1 vars, tightened 28 bounds.
Tried aggregator 2 times.
MIP Presolve eliminated 81 rows and 66 columns.
MIP Presolve modified 132 coefficients.
Aggregator did 6 substitutions.
Reduced MIP has 172 rows, 205 columns, and 568 nonzeros.
Reduced MIP has 59 binaries, 0 generals, 0 SOSs, and 0 indicators.
Presolve time = 0.00 sec. (0.42 ticks)
Probing fixed 0 vars, tightened 9 bounds.
Probing time = 0.00 sec. (0.04 ticks)
Tried aggregator 1 time.
MIP Presolve eliminated 1 rows and 0 columns.
MIP Presolve modified 9 coefficients.
Reduced MIP has 171 rows, 205 columns, and 565 nonzeros.
Reduced MIP has 59 binaries, 0 generals, 0 SOSs, and 0 indicators.
Presolve time = 0.00 sec. (0.33 ticks)
Probing time = 0.00 sec. (0.04 ticks)
Clique table members: 7.
MIP emphasis: balance optimality and feasibility.
MIP search method: dynamic search.
Parallel mode: deterministic, using up to 16 threads.
Root relaxation solution time = 0.00 sec. (0.43 ticks)
Nodes Cuts/
Node Left Objective IInf Best Integer Best Bound ItCnt Gap
* 0 0 integral 0 2202.4575 2202.4575 20 0.00%
Elapsed time = 0.02 sec. (4.91 ticks, tree = 0.00 MB, solutions = 1)
Root node processing (before b&c):
Real time = 0.02 sec. (4.93 ticks)
Parallel b&c, 16 threads:
Real time = 0.00 sec. (0.00 ticks)
Sync time (average) = 0.00 sec.
Wait time (average) = 0.00 sec.
------------
Total (root+branch&cut) = 0.02 sec. (4.93 ticks)
🎉3 参考文献
部分理论来源于网络,如有侵权请联系删除。
[1]李淋,徐青山,王晓晴,凌静,孙海翔.基于共享储能电站的工业用户日前优化经济调度[J].电力建设,2020,41(05):100-107.