An Energy Management Strategy for Hybrid Ship Based on SVM and MPC
-
摘要: 为了提高混合动力船舶能量利用效能，提升混合动力船舶燃油经济性，将在线工况识别与实时优化策略相结合，提出了基于支持向量机（support vector machine，SVM）与模型预测控制（model predictive control，MPC）的混合动力船舶能量管理策略。引入SVM理论，使用“美维凯越”号新能源游轮的实船数据，优化核函数类型及关键参数，构建船舶运行4个典型工况的在线识别模型；再通过对船舶当前时刻工况特征参数的分析与判断，确定船舶实时运行工况。以最小燃油消耗和保持储能单元荷电状态（state of charge，SOC）稳定为目标，以主机和复合储能单元之间的实时输出功率为优化变量，以功率需求预测模型为约束条件，构建模型预测控制MPC模型；为提高不同工况下功率需求预测的精度，提出基于多步马尔科夫模型的功率预测模型，并根据实时工况识别结果，更新MPC模型中的功率需求预测模型约束，实现船舶能量实时动态优化。最后，采用小波变换方法，将最优功率解分解成高频信号和低频信号；再根据复合储能单元中不同动力源工作特性的差异，将高频信号和低频信号分别分配给具有高功率密度的超级电容和具有高能量密度的蓄电池。为验证方法的有效性，开展了基于MATLAB的仿真实验，结果表明：在相同工况下，所提策略累计燃油消耗量为4 404.556 1 g，平均燃油消耗率为202.973 7 g/kWh；与基于单一模型预测控制的能量管理策略相比，所提方法可节省燃油消耗4.55%，验证了所提能量管理策略的有效性。Abstract: To improve the energy efficiency and fuel economy of hybrid ships, energy management strategy for hybrid ships based on support vector machine (SVM) and model predictive control (MPC) is proposed, incorporating the working condition recognition into the strategy management. A working-condition recognition model is developed based on the SVM theory, the kernel function type and key parameters are optimized by using the operation data from the hybrid powered cruise ships called"MEIWEI KEYUE", and the one of four working conditions is recognized by feeding the real-time operation data. An energy management MPC model (EM-MPC) is proposed, allocating the output powers of main engine and composite ESUs for minimizing the consumption of fuel and maintaining the state of charge (SOC) stabilization of energy storage unit (ESU), which is constrained by the power demand prediction (PDP) model. Then, a prediction model based on the multi-step Markov model is proposed to improve the accuracy of PDP under different working conditions, and the PDP constraint in the EM-MPC model is updated based on the recognized working condition, which contributes to the real-time power allocation. The optimal solution is decomposed into high frequency signal and low frequency signal by wavelet transform method, and these signals are assigned to the super capacitor with high power density and the battery with high energy density, respectively. To validate the proposed strategy, a simulation via Matlab is introduced and the results show that: ① the cumulative fuel consumption of the proposed method is 4 404.556 1 g and the average fuel consumption rate is 202.9737 g/kWh; ② under the same working condition, the fuel consumption can be saved by 4.55% via the proposed EM-MPC, comparing with the traditional method.
-
Key words:
- energy management /
- hybrid ship /
- control Strategy /
- model predictive control /
- pattern recognition
-
图 1 船舶并联式混合动力系统拓扑结构图
Figure 1. Topological structure diagram of ship's parallel hybrid power system
图 2 “美维凯越”号混合动力游船实船采集数据
Figure 2. Data collected by the actual ship of the "Meiwei Kaiyue" hybrid cruise ship
图 9 不同工况预测模型对应的状态转移概率矩阵
Figure 9. The state transition probability matrix corresponding to different working condition prediction models
图 10 时域为5 s时，单预测模型与SVM+多预测模型预测结果对比
Figure 10. Comparison of prediction results between single prediction model and SVM+multiple prediction model when the time domain is 5 s
图 13 3种策略的主机输出功率对比结果
Figure 13. Comparison results of main engine output power of three strategies
图 14 SVM+MPC策略蓄电池与超级电容SOC
Figure 14. SVM+MPC strategy battery and super capacitor SOC changes
表 1 主要设备参数
Table 1. Main equipment parameters
主要设备 参数名称 参数值 最大功率/kW 80 主机 额定功率/kW 72 最大转矩/Nm 478 最大功率/kW 60 电机 额定功率/kW 54 最大转矩/Nm 358 最高效率/% 95.5 额定电压/V 384 蓄电池 额定功率/kW 60 额定容量/Ah 20 最大充放电倍率/C 0.5 工作电压/V（DC） 200~500 超级电容 总电容量/F 10 最大允许放电电流（10 s）/A 600（25 ℃） 绝缘电压/V 2 500 
下载: 导出CSV
表 2 各工况基本信息
Table 2. Basic information of each working condition
工况类型 平均需求功率/kW 平均需求功率一阶导数 所占百分比/% 工况1 20.085 0 0.002 2 37.56 工况2 64.008 1 -0.006 7 24.50 工况3 62.469 7 -0.003 7 11.44 工况4 94.579 7 -0.004 0 26.50
下载: 导出CSV
表 3 核函数类型比较
Table 3. Comparison of kernel function types
核函数类型 正确识别数/样本总数 准确度/% 线性核函数 3 740/4 760 0.785 7 多项式核函数 4 016/4 760 0.843 7 径向基核函数 4 326/4 760 0.908 8 Sigmoid核函数 1 763/4 760 0.370 4
下载: 导出CSV
表 4 预测效果比较
Table 4. Comparison of prediction effects
时域/s 单预测模型Re SVM+多预测模型Re 3 5.084 7 4.921 8 5 7.182 7 6.781 2 10 12.429 7 10.615 0 15 16.913 3 12.848 1
下载: 导出CSV
表 5 仿真实验结果
Table 5. Simulation experiment results
策略 累计耗油量/g 燃油消耗率/（g/kWh） 耗油量与DP策略比值 DP 4 224.987 7 194.197 0 1 MPC 4 496.674 6 211.533 1 1.064 3 SVM+MPC 4 404.556 1 202.973 7 1.042 5
下载: 导出CSV
-
[1] 严新平. 新能源在船舶上的应用进展及展望[J]. 船海工程, 2010, 39(6): 111-115. doi: 10.3963/j.issn.1671-7953.2010.06.031YAN X P. Progress review of new energy application in ship[J]. Ship & Ocean Engineering, 2010, 39(6): 111-115. (in Chinese) doi: 10.3963/j.issn.1671-7953.2010.06.031 [2] YUAN Y P, WANG J X, YAN X P, et al. A review of multi-energy hybrid power system for ships[J]. Renewable and Sustainable Energy Reviews, 2020, 132: 1-20. [3] 李维波, 郝春昊, 高佳俊, 等. 舰船综合电力系统发展综述[J]. 中国舰船研究, 2020, 15(6): 1-11.LI W B, HAO C H, GAO J J, et al. Overview of the development of shipboard integrated power system[J]. Chinese Journal of Ship Research, 2020, 15(6): 1-11. (in Chinese) [4] 庞水, 林叶锦, 张均东, 等. 柴电混合动力船舶能量分配优化方法[J]. 船海工程, 2020, 49(3): 106-111. doi: 10.3963/j.issn.1671-7953.2020.03.023PANG S, LIN Y J, ZHANG J D, et al. On power distribution optimization method for diesel-electric hybrid ship[J]. Ship & Ocean Engineering, 2020, 49(3): 106-111. (in Chinese) doi: 10.3963/j.issn.1671-7953.2020.03.023 [5] TORREGLOSA J P, GARCIA P, FERNANDEZ L M, et al. Hierarchical energy management system for stand-alone hybrid system based on generation costs and cascade control[J]. Energy Conversion and Management, 2014, 77(2): 514-526. [6] BESIKCI E B, KECECI T, ARSLAN O, et al. An application of fuzzy-AHP to ship operational energy efficiency measures[J]. Ocean Engineering, 2016, 121(15): 392-402. [7] 袁裕鹏, 王凯, 严新平. 混合动力船舶能量管理控制策略设计与仿真[J]. 船海工程, 2015, 44(2): 95-98. doi: 10.3963/j.issn.1671-7953.2015.02.025YUAN Y P, WANG K, YAN X P. Design and simulate of energy management control strategy for hybrid ship[J]. Ship & Ocean Engineering, 2015, 44(2): 95-98. (in Chinese) doi: 10.3963/j.issn.1671-7953.2015.02.025 [8] 兰熙, 沈爱弟, 高迪驹, 等. 混合动力船舶能量管理系统的最优控制[J]. 电源技术, 2016(9): 1859-1862. doi: 10.3969/j.issn.1002-087X.2016.09.038LAN X, SHEN A D, GAO D J, et al. Optimal control of hybrid ship energy management system[J]. Chinese Journal of Power Sources, 2016, 40(9): 1859-1862. (in Chinese) doi: 10.3969/j.issn.1002-087X.2016.09.038 [9] TANG R L, LI X, LAI J. A novel optimal energy-management strategy for a maritime hybrid energy system based on large-scale global optimization[J]. Applied Energy, 2018, 228: 254-264. doi: 10.1016/j.apenergy.2018.06.092 [10] BASSAM A M, PHILLIPS A B, TURNOCK S R, et al. Development of a multi-scheme energy management strategy for a hybrid fuel cell driven passenger ship[J]. International Journal of Hydrogen Energy, 2017, 42(1): 623-635. doi: 10.1016/j.ijhydene.2016.08.209 [11] 侯慧, 甘铭, 吴细秀, 等. 混合动力船舶能量管理研究综述[J]. 中国舰船研究, 2021, 16(5): 216-229.HOU H, GAN M, WU X X, et al. Review of hybrid ship energy management[J]. Chinese Journal of Ship Research, 2021, 16(5): 216-229. (in Chinese) [12] 牛礼民, 周亚洲, 吕建美, 等. 并联HEV工况识别能量管理与优化控制[J]. 控制工程, 2021, 28(3): 435-444.LIU L M, ZHOU Y Z, LV J M, et al. Energy management and optimization control based on driving cycle identification for parallel hybrid electric vehicle[J]. Control Engineering of China, 2021, 28(3): 435-444. (in Chinese) [13] ERICSSON E. Independent driving pattern factors and their influence on fuel-use and exhaust emission factors[J]. Transportation Research Part D, 2001, 6(5): 325-345. doi: 10.1016/S1361-9209(01)00003-7 [14] 张风奇, 胡晓松, 许康辉, 等. 混合动力汽车模型预测能量管理研究现状与展望[J]. 机械工程学报, 2019, 55(10): 86-108.ZHANG F Q, HU X S, XU K H, et al. Current status and prospects for model predictive energy management in hybrid electric vehicles[J]. Journal of Mechanical Engineering, 2019, 55(10): 86-108. (in Chinese) [15] 汪海燕, 黎建辉, 杨风雷. 支持向量机理论及算法研究综述[J]. 计算机应用研究, 2014, 31(5): 1281-1286. doi: 10.3969/j.issn.1001-3695.2014.05.001WANG H Y, LI J H, YANG F L. Overview of support vector machine analysis and algorithm[J]. Application Research of Computers, 2014, 31(5): 1281-1286. (in Chinese) doi: 10.3969/j.issn.1001-3695.2014.05.001 [16] 李民策, 王丽, 李锡云, 等. 基于支持向量机的电动汽车行驶工况识别方法[C]. 第21届中国系统仿真技术及其应用学术年会, 昆明: 中国自动化学会, 2020.LI M C, WANG L, LI X Y, et al. A driving condition recognition method for electric vehicle based on support vector machine[C]. The 21st China Annual Conference on System Simulation Technology and its Application. Kunming, China: Chinese Association of Automation, 2020. (in Chinese) [17] 许绍航, 席军强, 陈慧岩. 基于越野工况预测的混合动力履带车辆能量管理策略[J]. 兵工学报, 2019, 40(8): 1572-1579. doi: 10.3969/j.issn.1000-1093.2019.08.003XU S H, XI J Q, CHEN H Y. Energy management of hybrid electric tracked vehicle based on off-road condition prediction[J]. ACTA ARMAMENTARⅡ, 2019, 40(8): 1572-1579. (in Chinese) doi: 10.3969/j.issn.1000-1093.2019.08.003 [18] ZHU Y C, ZHENG Y. Traffic identification and traffic analysis based on support vector machine[J]. Neural Computing and Applications, 2020, 32(7): 1903-1911. doi: 10.1007/s00521-019-04493-2 [19] HOU J, SONG Z Y, HOFMANN H, et al. Adaptive model predictive control for hybrid energy storage energy management in all-electric ship microgrids[J]. Energy Conversion and Management, 2019, 198(10): 198-212. [20] 解少博, 刘通, 李会灵, 等. 基于马尔科夫链的并联PHEB预测型能量管理策略研究[J]. 汽车工程, 2018, 40(8): 871-877.XIE S B, LIU T, LI L L, et al. A study on predictive energy management strategy for a plug-in hybrid electric bus based on Markov chain[J]. Automotive Engineering. 2018, 40(8): 871-877. (in Chinese) [21] 潘钊, 商蕾, 高海波, 等. 燃料电池混合动力船舶复合储能系统与能量管理策略优化[J]. 大连海事大学学报, 2021, 47 (3): 79-85. doi: 10.3969/j.issn.1671-7031.2021.03.012PAN Z, SHANG L, GAO H B, et al. Optimization of composite energy storage system and energy management strategy for fuel cell hybrid ships[J]. Journal of Dalian Maritime University. 2021, 47(3): 79-85. (in Chinese) doi: 10.3969/j.issn.1671-7031.2021.03.012 -
点击查看大图
计量
- 文章访问数: 35
- HTML全文浏览量: 35
- PDF下载量: 3
- 被引次数: 0