随着纯电动船舶的高速发展，其用电负荷在电力市场交易中的影响日渐突出，为此该文提出一种船舶综合电力系统负荷神经网络组合预测方法，旨在提高预测精度。首先，分析纯电动船舶综合电力系统在多种工况下的负荷特性。然后，研究基于典型神经网络的船舶综合电力系统负荷预测方法，揭示其在复杂工况下预测的局限性。针对以上问题，提出了基于BP和RBF神经网络相结合的船舶综合电力系统负荷组合预测方法。此组合预测方法集合了BP和RBF神经网络模型的优势，提高了预测模型的泛化能力和容错率。最后，以江苏某纯电动船舶为实际算例，针对复杂工况下的船舶综合电力系统负荷进行对比预测。结果表明，所提方法与单一预测算法相比，预测精度从96.63%提高至98.98%。
With the rapid development of pure electric ships, the impact of their electricity load on electricity market transactions is becoming increasingly prominent. Therefore, this paper proposes a ship integrated power system load neural network combination prediction method, aiming to improve prediction accuracy. Firstly, analyze the load characteristics of the integrated power system of pure electric ships under various operating conditions. Then, study the load forecasting method for ship integrated power system based on typical neural networks, and reveal its limitations in predicting complex working conditions. In response to the above issues, a load combination prediction method for ship integrated power system based on a combination of BP and RBF neural networks is proposed. This combined prediction method combines the advantages of BP and RBF neural network models, improving the generalization ability and fault tolerance of the prediction model. Finally, taking a pure electric ship in Jiangsu as an actual calculation example, a comparative prediction of the comprehensive power system load of the ship under complex working conditions is conducted. The results show that compared with a single prediction algorithm, the proposed method improves the prediction accuracy from 96.63% to 98.98%.
2024,46(7): 112-120 收稿日期：2023-5-4
DOI：10.3404/j.issn.1672-7649.2024.07.019
分类号：TM713
作者简介：严文博(1999-),男,硕士研究生,研究方向为新能源负荷预测和发电系统并网稳定性分析等
参考文献：
[1] 包育德. 双碳背景下清洁能源调度优化及排放分析[J/OL]. 电测与仪表, 1-10
[2] 王力波. 我国电动船舶产业现状与发展对策研究[J]. 中国海事, 2022(8): 65-68.
WANG Libo. Study on the current situation and countermeasures of the development of china's electric ship industry[J]. Shipping Horizon, 2022(8): 65-68.
[3] 陆振华. 全电力推进船舶能量管理优化研究[D]. 大连: 大连海事大学, 2020.
[4] 张旭, 张宏立, 范文慧, 等. 基于数据分析和改进Chebyshev神经网络的风速时间序列预测[J]. 电测与仪表, 2020, 57(22): 33-39.
ZHANG Xu, ZHANG Hongli, FAN Wenhui, et al. Time series prediction of wind speed based on data analysis and improved Chebyshev neural network[J]. Electrical Measurement & Instrumentation, 2020, 57(22): 33-39.
[5] 韩富佳, 王晓辉, 乔骥, 等. 基于人工智能技术的新型电力系统负荷预测研究综述[J/OL]. 中国电机工程学报: 1-24
[6] 侯文君, 汪英英, 姚艺新, 等. 支持向量机的船舶短期电力负荷预测方法[J]. 舰船科学技术, 2018, 40(18): 79-81.
HOU Wen-jun, WANG Ying-ying, YAO Yi-xin, et al. Short term load forecasting method based on support vector machine[J]. Ship Science and Techology, 2018, 40(18): 79-81.
[7] 刘静. 最小二乘SVM在电力推进船舶电力负荷短期预测的应用[J]. 舰船科学技术, 2021, 43(24): 61-63.
LIU Jing. Application of least square SVM in short-term forecasting of electric load of electric propulsion ship[J]. Ship Science and Technology, 2021, 43(24): 61-63.
[8] 刘银波, 吕越. RBF神经网络在船舶电力负荷预测中的应用[J]. 舰船科学技术, 2021, 43(18): 121-123.
LIU Yin-Bo, LV Yue. Application of RBF neural network in ship power load forecasting[J]. Ship Science and Techology, 2021, 43(18): 121-123.
[9] 阳曾, 丁施尹, 叶萌, 等. 基于变分模态分解和深度学习的短期电力负荷预测模型[J]. 电测与仪表, 2023, 60(2): 126-131+146.
YANG Zeng, DING Shiyin, YE Meng, et al. Short-term load forecasting model based on VMD and LSTM[J]. Electrical Measurement & Instrumentation, 2023, 60(2): 126-131+146.
[10] 张宇涵, 高海波, 商蕾, 等. 基于神经网络模型的船舶电网短期电力负荷预测[J]. 应用科技, 2021, 48(5): 12-15+22.
ZHANG Yuhan, GAO Haibo, SHANG Lei, et al. Short-term load forecasting for marine electric network based on neural network models[J]. Applied Science and Technology, 2021, 48(5): 12-15+22.
[11] 张博山, 乔显辉. 全电力推进船舶的运行维护与管理[J]. 机电设备, 2022, 39(4): 12-19.
ZHANG Boshan, QIAO Xianhui. Operation, maintenance and management of all-electric propulsion ships[J]. Electromechanical Equipment, 2022, 39(4): 12-19.
[12] 杨光雨, 李晓航. 基于最大信息挖掘宽度学习系统短期电力负荷预测研究[J]. 电测与仪表, 2022, 59(3): 38-45.
YANG Guangyu, LI Xiaohang. The short-term power load forecast in wide area learning system based on maximum information mining[J]. Electrical Measurement & Instrumentation, 2022, 59(3): 38-45.
[13] 王克杰, 张瑞. 基于改进BP神经网络的短期电力负荷预测方法研究[J]. 电测与仪表, 2019, 56(24): 115-121.
WANG Kejie, ZHANG Rui. Research on short-term power load forecasting method based on improved BP neural network[J]. Electrical Measurement & Instrumentation, 2019, 56(24): 115-121.
[14] 归三荣, 郭燚, 高松, 等. 靠港后船舶电力系统负荷分析[J]. 上海海事大学学报, 2017, 38(2): 61-65.
GUI Sanrong, GUO Yi, GAO Song, et al. Load analysis of power system of ships berthed in ports[J]. Journal of Shanghai Maritime University, 2017, 38(2): 61-65.
[15] ATANASOVSKI M, KOSTOV M, VELJANOVSKI G, et al. Short-Term load forecast in power systems: a comparison of different practical algorithms[C]//2022 18th International Conference on the European Energy Market, 2022.
[16] 董雷, 陈振平, 韩富佳. 基于图卷积神经网络与K-means聚类的居民用户集群短期负荷预测[J/OL]. 电网技术, 1-11
[17] 孙国良, 伊力哈木·亚尔买买提, 张宽, 等. 基于小波变换与IAGA-BP神经网络的短期风电功率预测[J/OL]. 电测与仪表, 1-11
[18] 党存禄, 杨海兰, 武文成. 基于LSTM和CatBoost组合模型的短期负荷预测[J]. 电气工程学报, 2021, 16(3): 62-69.
DANG Cunlu, YANG Hailan, WU Wencheng. Short-term load forecasting based on LSTM and catboost combined model[J]. Journal Of Electrical Engineering, 2021, 16(3): 62-69.
[19] 杜昆. 基于BP神经网络短期负荷预测研究[D]. 石家庄: 河北科技大学, 2021.
[20] 吴嘉竣, 徐文辉. 基于RBF神经网络的电力信息网络安全态势感知[J]. 自动化技术与应用, 2022, 41(9): 103-105.
WU Jiajun, XU Wenhui. Power Information network security situation awareness based on RBF neural network[J]. Internet of Things and Network Technology, 2022, 41(9): 103-105.
[21] A T S, A. PRAVEEN, H M S. Comparison of deep learning-based methods for electrical load forecasting[C]//2022 Third International Conference on Intelligent Computing Instrumentation and Control Technologies, 2022. 1598-1602.
