Energy consumption prediction of aircraft ground air conditioning based on IPSO-AM-LSTM
-
摘要:
作为一个复杂的热力系统，飞机地面空调能耗受到多种因素的影响，包括各种天气数据和时间特征。为提升飞机客舱使用地面空调制冷时地面空调能耗预测精度，构造了一种基于长短时记忆网络(LSTM)的地面空调能耗预测模型。该模型融合长短时记忆网络和注意力机制构建预测部分，用于提取和利用数据中的时序信息，并以预测精度作为算法的适应度函数，提出一种基于改进粒子群优化(IPSO)算法的超参数优化方法，与标准粒子群优化(PSO)算法相比，该优化算法将迭代次数与适应度相结合，构建了惯性权重的动态调节函数，对其进行动态调节，并引入粒子到全局最优位置的距离参数，提出一种粒子交叉策略，提高粒子群的多样性。所提方法在实际数据集上的仿真预测结果的均方误差为4.394，平均绝对百分比误差为0.774%，决定系数为0.968。与其他预测方法进行对比，结果表明：所提方法有更高的准确度。
Abstract:As a complex thermal system, the energy consumption of aircraft ground air conditioning is affected by many factors, including various weather data and time characteristics. In order to improve the prediction accuracy of ground air conditioning energy consumption when the aircraft cabin is cooled by ground air conditioning, a ground air conditioning energy consumption prediction model based on a long-short-term memory (LSTM) network is proposed. The prediction component of the model, which is utilized to extract and make use of the time series information in the data, is built by integrating the long-short-term memory network with the attention mechanism. The prediction accuracy is used as the fitness function of the algorithm. The hyperparameter optimization based on the improved particle swarm optimization (IPSO) algorithm is proposed. Compared with the standard particle swarm optimization (PSO) algorithm, the improved particle swarm optimization algorithm combines the number of iterations with the fitness to construct a dynamic adjustment function of the inertia weight. The distance from the particle to the global most optimal position is introduced and a particle intersection strategy is proposed to improve the diversity of particle swarms. The mean square error of the prediction result is 4.394. The mean absolute percentage error is 0.774%, and the coefficient of determination is 0.968. The results indicate that the prediction approach has a greater accuracy when compared to other prediction methods.
-
图 7 不同模型下预测值与真实值的对比
Figure 7. Comparison of predicted and real values under different models
表 1 部分能耗数据
Table 1. Part of sample data
乘客下机的
客舱温度/℃乘客下机的
客舱湿度/%乘客登机的
客舱温度/℃乘客登机的
客舱湿度/%耗电量/kW 31.81 56.25 26.49 37.01 181.53 30.35 62.70 24.07 37.10 165.64 31.84 49.52 25.85 25.05 179.51 35.18 67.51 26.64 35.31 185.49 33.81 72.73 25.93 36.86 169.88 
下载: 导出CSV
表 2 超参数搜索范围
Table 2. Search range of hyperparameters
超参数 搜索范围 批大小 1～90 学习率 0.0001 ～0.01LSTM层单元数 10～90 Dense层单元数 10～90 Dropout层丢弃概率 0.1～0.9 迭代次数 400
下载: 导出CSV
表 3 3种算法搜寻结果
Table 3. Three algorithms search results
算法 批大小 LSTM层
单元数Dense层
单元数Dropout层
丢弃概率学习率 迭代次数 SCA 3 74 22 0.1 0.005 5 400 PSO 1 74 45 0.28 0.009 0 400 IPSO 1 23 37 0.1 0.003 5 400
下载: 导出CSV
表 4 不同模型评价结果
Table 4. Evaluation results for different models
模型 MSE MAPE/% R2 BP 61.042 7.075 0.553 LSTM 14.691 1.669 0.893 AM-LSTM 10.195 1.140 0.925 PSO-AM-LSTM 8.952 1.087 0.935 SCA-AM-LSTM 7.519 0.840 0.945 IPSO-AM-LSTM 4.394 0.774 0.968
下载: 导出CSV
-
[1] SHI G, LIU D R, WEI Q L. Energy consumption prediction of office buildings based on echo state networks[J]. Neurocomputing, 2016, 216: 478-488. doi: 10.1016/j.neucom.2016.08.004 [2] WANG J Q, DU Y, WANG J. LSTM based long-term energy consumption prediction with periodicity[J]. Energy, 2020, 197: 117197. doi: 10.1016/j.energy.2020.117197 [3] WANG Z Y, WANG Y R, ZENG R C, et al. Random Forest based hourly building energy prediction[J]. Energy and Buildings, 2018, 171: 11-25. doi: 10.1016/j.enbuild.2018.04.008 [4] 王修岩, 刘艳敏, 张革文, 等. 基于PSO和CRO联合算法的飞机客舱能耗预测[J]. 系统仿真学报, 2018, 30(8): 3074-3081.WANG X Y, LIU Y M, ZHANG G W, et al. Prediction of aircraft cabin energy consumption based on PSO and CRO algorithms[J]. Journal of System Simulation, 2018, 30(8): 3074-3081(in Chinese). [5] 林家泉, 孙凤山, 李亚冲, 等. 基于IPSO-Elman 神经网络的飞机客舱能耗预测[J]. 航空学报, 2020, 41(7): 234-243.LIN J Q, SUN F S, LI Y C, et al. Prediction of aircraft cabin energy consumption based on IPSO-Elman neural network[J]. Acta Aeronautica et Astronautica Sinica, 2020, 41(7): 234-243(in Chinese). [6] BEDI G, VENAYAGAMOORTHY G K, SINGH R. Development of an IoT-driven building environment for prediction of electric energy consumption[J]. IEEE Internet of Things Journal, 2020, 7(6): 4912-4921. doi: 10.1109/JIOT.2020.2975847 [7] SRIVASTAVA S, LESSMANN S. A comparative study of LSTM neural networks in forecasting day-ahead global horizontal irradiance with satellite data[J]. Solar Energy, 2018, 162: 232-247. doi: 10.1016/j.solener.2018.01.005 [8] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780. doi: 10.1162/neco.1997.9.8.1735 [9] RAJAN S, CHENNIAPPAN P, DEVARAJ S, et al. Novel deep learning model for facial expression recognition based on maximum boosted CNN and LSTM[J]. IET Image Processing, 2020, 14(7): 1373-1381. doi: 10.1049/iet-ipr.2019.1188 [10] JO J, KUNG J, LEE Y. Approximate LSTM computing for energy-efficient speech recognition[J]. Electronics, 2020, 9(12): 2004. doi: 10.3390/electronics9122004 [11] LIU S C, WANG X L, XIANG Y, et al. Multi-channel fusion LSTM for medical event prediction using EHRs[J]. Journal of Biomedical Informatics, 2022, 127: 104011. doi: 10.1016/j.jbi.2022.104011 [12] SEYEDZADEH S, POUR RAHIMIAN F, RASTOGI P, et al. Tuning machine learning models for prediction of building energy loads[J]. Sustainable Cities and Society, 2019, 47: 101484. doi: 10.1016/j.scs.2019.101484 [13] KIM T Y, CHO S B. Predicting residential energy consumption using CNN-LSTM neural networks[J]. Energy, 2019, 182: 72-81. doi: 10.1016/j.energy.2019.05.230 [14] MA C, LIN C, SAMUEL O W, et al. Continuous estimation of upper limb joint angle from sEMG signals based on SCA-LSTM deep learning approach[J]. Biomedical Signal Processing and Control, 2020, 61: 102024. doi: 10.1016/j.bspc.2020.102024 [15] 周璇, 林家泉. 基于改进长短时记忆网络的地面空调能耗预测[J]. 北京航空航天大学学报, 2023, 49(10): 2750-2760.ZHOU X , LIN J Q. Prediction of ground air conditioner energy consumption based on improved long short-term memory network[J]. Journal of Beijing University of Aeronautics and Astronautics, 2023, 49(10): 2750-2760(in Chinese). [16] REN X Q, LIU S L, YU X D, et al. A method for state-of-charge estimation of lithium-ion batteries based on PSO-LSTM[J]. Energy, 2021, 234: 121236. doi: 10.1016/j.energy.2021.121236 [17] WANG D S, TAN D P, LIU L. Particle swarm optimization algorithm: an overview[J]. Soft Computing, 2018, 22(2): 387-408. doi: 10.1007/s00500-016-2474-6 [18] 梅海涛, 华继学, 王毅, 等. 基于直觉模糊支配的混合多目标粒子群算法[J]. 计算机科学, 2017, 44(1): 253-258. doi: 10.11896/j.issn.1002-137X.2017.01.047MEI H T, HUA J X, WANG Y, et al. Hybrid multi-objective particle swarm optimization based on intuitionistic fuzzy dominance[J]. Computer Science, 2017, 44(1): 253-258(in Chinese). doi: 10.11896/j.issn.1002-137X.2017.01.047 [19] 汪浩然, 夏克文, 牛文佳. 分段正交匹配追踪(StOMP)算法改进研究[J]. 计算机工程与应用, 2017, 53(16): 55-61. doi: 10.3778/j.issn.1002-8331.1604-0226WANG H R, XIA K W, NIU W J. Improved research on stagewise orthogonal matching pursuit (StOMP) algorithm[J]. Computer Engineering and Applications, 2017, 53(16): 55-61(in Chinese). doi: 10.3778/j.issn.1002-8331.1604-0226 [20] HAN F, SUN Y W T, LING Q H. An improved multiobjective quantum-behaved particle swarm optimization based on double search strategy and circular transposon mechanism[J]. Complexity, 2018, 2018(1): 8702820. doi: 10.1155/2018/8702820 -
下载:
点击查看大图
计量
- 文章访问数: 165
- HTML全文浏览量: 90
- PDF下载量: 4
- 被引次数: 0
