Study on the calorific value prediction of municipal solid wastes by image deep learning

doi:10.11949/0438-1157.20201481

Abstract

Abstract:

Waste incineration power plants have large fluctuations in the calorific value of waste entering the furnace, which affects the stability of boiler operation and power generation efficiency. The use of image deep learning methods to achieve real-time prediction of the heat value of waste entering the furnace will help the power plant achieve "advanced regulation". Currently, there is lack of waste image database coinciding with waste composition in China after reviewing researches of image recognition and calorific value prediction progress of waste. Yolov5 is adopted for waste image detection and calorific value prediction. Furthermore, industrial cameras are used to real time capture images of waste before entering the furnace, and the image database of waste would be built by waste classifying annotation. Following this, we envisage a calorific value prediction model based on the obtained data being trained through using mosaic enhancement and neural network. In the future, it is meaningful to combine deep image learning with image recognition technology for waste calorific value prediction, which will be higher precision and faster response time.

Key words: municipal solid waste incineration, calorific value, neural networks, image deep learning, algorithm

摘要：

垃圾焚烧发电厂入炉垃圾热值波动大，影响了锅炉运行的稳定性和发电效率，利用图像深度学习的方法实现入炉垃圾热值的实时预测，有助于电厂实现“超前调控”。本文探讨了国内外垃圾图像识别及热值预测的研究进展和不足，认为目前缺少符合我国垃圾组分结构的垃圾图像数据库和热值智能预测方法，提出了用Yolov5识别图像中垃圾种类来预测热值的方法，通过入炉垃圾图像的实时采集与分类标记建立图像数据库，并耦合mosaic数据增强等图像数据处理及神经网络训练，提出建立垃圾热值实时预测模型的设想。本文进一步展望了垃圾热值智能预测的发展前景，未来可以将深度学习与图像识别技术高效结合，实现入炉垃圾热值的实时与精准预测。

关键词: 垃圾焚烧, 热值, 神经网络, 图像深度学习, 算法

CLC Number:

TK 01⁺8

XIE Haoyuan, HUANG Qunxing, LIN Xiaoqing, LI Xiaodong, YAN Jianhua. Study on the calorific value prediction of municipal solid wastes by image deep learning[J]. CIESC Journal, 2021, 72(5): 2773-2782.

谢昊源, 黄群星, 林晓青, 李晓东, 严建华. 基于图像深度学习的垃圾热值预测研究[J]. 化工学报, 2021, 72(5): 2773-2782.

Figures/Tables 10

Fig.1 Municipal solid waste generation and year-on-year growth rate of China in 2015—2020

Table 1 Composition of physical composition of domestic garbage

序号	类别	组成成分
1	塑料类	废弃塑料
2	橡胶类	橡胶、皮革制品
3	木竹类	废弃的木竹制品及木材
4	纺织物	废弃的布类（包括化纤布）、棉花
5	纸类	废弃的纸张及纸制品
6	土砖类	炉灰、尘土
7	厨余类	动、植物类食品（包括水果）的残余物
8	金属类	废弃的金属、金属制品（包括电池）
9	玻璃类	废弃玻璃、玻璃制品
10	其他类	上述九类以外的其他生活垃圾
11	混合类	粒径小于10 mm的、按上述分类比较困难的混合物

Table 2 The relationship between domestic waste incineration and calorific value

生活垃圾低位热值Q/(kJ/kg)	稳定燃烧条件
Q<3350	无法自燃，需加辅助燃料
3350<Q<4200	少加辅助燃料
4200<Q<5000	一次、二次风预热至200~250℃
5650<Q<8500	一次风预热至100~200℃
Q>8500	一次、二次风常温

Table 3 Three empirical formulas for calculating the calorific value of garbage

类别	计算公式	单位	适用范围
物理组成分析热值^[19-21]	Conventional: $H H V = 88.2 P I + 40.5 G a + P a - 6 W × 4.18$	kJ/kg	垃圾
	Tokyo: $H H V = W 38.8 P a + G a + T + O c + 50.9 T e + R u + 73.7 P I - 6 W × 4.18$	kJ/kg	生活垃圾
	Ali Khan: $H H V = 23 G a + 3.6 P a + 160 P I + R u × 2.23$	kJ/kg	生活垃圾
元素分析热值^[19]	Dulong: $H H V = 81 C + 342.5 H - O 8 + 22.5 S - 6 9 H - W × 4.18$	kJ/kg	生活垃圾/煤
	Scheurer-Kestner: $H H V = 81 C - 3 × O 4 + 342.5 H + 22.5 S + 57 × 3 × O 4 - 6 9 H + W × 4.18$	kJ/kg	生活垃圾
	Steuer: $H H V = 81 C - 3 × O 8 + 57 × 3 × O 8 + 345 H - O 10 + 25 S - 6 9 H + W × 4.18$	kJ/kg	生活垃圾
工业特征分析热值^[19]	Bento： $H H V = 44.75 V M - 5.85 W + 21.2 × 4.18$	kJ/kg	垃圾

Table 3 Three empirical formulas for calculating the calorific value of garbage

类别	计算公式	单位	适用范围
物理组成分析热值^[19-21]	Conventional: $H H V = 88.2 P I + 40.5 G a + P a - 6 W × 4.18$	kJ/kg	垃圾
	Tokyo: $H H V = W 38.8 P a + G a + T + O c + 50.9 T e + R u + 73.7 P I - 6 W × 4.18$	kJ/kg	生活垃圾
	Ali Khan: $H H V = 23 G a + 3.6 P a + 160 P I + R u × 2.23$	kJ/kg	生活垃圾
元素分析热值^[19]	Dulong: $H H V = 81 C + 342.5 H - O 8 + 22.5 S - 6 9 H - W × 4.18$	kJ/kg	生活垃圾/煤
	Scheurer-Kestner: $H H V = 81 C - 3 × O 4 + 342.5 H + 22.5 S + 57 × 3 × O 4 - 6 9 H + W × 4.18$	kJ/kg	生活垃圾
	Steuer: $H H V = 81 C - 3 × O 8 + 57 × 3 × O 8 + 345 H - O 10 + 25 S - 6 9 H + W × 4.18$	kJ/kg	生活垃圾
工业特征分析热值^[19]	Bento： $H H V = 44.75 V M - 5.85 W + 21.2 × 4.18$	kJ/kg	垃圾

Fig.2 Yolov5 calculation method flowchart

Fig.3 Yolov5 network structure diagram

Table 4 The number of convolution kernels and the number of residual components of different network structures in Yolov5

结构网络	残差组件个数/个	卷积核总数/个
Yolov5s	12	1001
Yolov5m	24	1488
Yolov5l	36	1984
Yolov5x	48	2480

Fig.4 Example diagram of color space adjustment data enhancement

Fig.5 Example diagram of mosaic data enhancement

Table 5 Calorific value and hydrogen content of various types of domestic waste

生活垃圾成分	垃圾占比/%	干基高位热值/ (kJ/kg)	干基氢含量/%
塑料类	10.12~14.12	24096~32570	7.2
橡胶类	8.03~13.38	15365~23260	10.0
木竹类	1.71~6.15	10682~18610	6.0
纺织物	1.08~4.44	10551~17450	6.6
纸类	17.39~24.28	11467~16600	6.0
土砖类	1.02~4.87	3686~6980	3.0
厨余类	39.86~53.1	1140~4650	6.4
金属类	0.11~0.32	700	—
玻璃类	0.89~1.91	140	—

References 39

1	林晓青, 陈志良, 李晓东, 等. 煤粉炉掺烧生活垃圾对灰渣特性的影响研究[J]. 化工学报, 2018, 69(6): 2708-2713.
	Lin X Q, Chen Z L, Li X D, et al. Study on characterization of ash from co-combustion of coal with municipal solid waste[J]. CIESC Journal, 2018, 69(6): 2708-2713.
2	房科靖, 熊祖鸿, 鲁敏, 等. 垃圾热值的研究进展[J]. 新能源进展, 2019, 7(4): 359-364.
	Fang K J, Xiong Z H, Lu M, et al. Research progress of garbage calorific value [J]. Advances in New and Renewable Energy, 2019, 7(4): 359-364.
3	Liu Y, Jing X Y, Nie J, et al. Context-aware 3D mean-shift with occlusion handling for robust object tracking in RGB-D videos[J]. IEEE Transactions on Multimedia, 2019, 21(3): 664-677.
4	杨涛. 基于经济统计数据的生活垃圾热值计算模型[J]. 环境工程技术学报, 2014, 4(2): 158-163.
	Yang T. Economic statistics based municipal solid waste heating value calculation model[J]. Journal of Environmental Engineering Technology, 2014, 4(2): 158-163.
5	Tian Y N, Yang G D, Wang Z, et al. Apple detection during different growth stages in orchards using the improved YOLO-V3 model[J]. Computers and Electronics in Agriculture, 2019, 157: 417-426.
6	纪超, 黄新波, 曹雯, 等. 结合深度学习和全局-局部特征的图像显著区域计算[J]. 计算机辅助设计与图形学学报, 2019, 31(10): 1838-1846.
	Ji C, Huang X B, Cao W, et al. Fusion of deep learning and global-local features of the image salient region calculation[J]. Journal of Computer-Aided Design & Computer Graphics, 2019, 31(10): 1838-1846.
7	杨辉. 我国环境污染状况及环境保护的可行性建议[J]. 中国高新科技, 2020, (9): 74-75.
	Yang H. environmental pollution status and feasibility suggestions for environmental protection[J]. China High-Tech, 2020, (9): 74-75.
8	于宇, 沈淑霞, 张巍. 自然保护区生态旅游与生态环境保护研究[J]. 价值工程, 2020, 39(19): 85-86.
	Yu Y, Shen S X, Zhang W. Study on eco-tourism and eco-environmental protection in nature reserves[J]. Value Engineering, 2020, 39(19): 85-86.
9	王延涛, 曹阳. 我国城市生活垃圾焚烧发电厂垃圾热值分析[J]. 环境卫生工程, 2019, 27(5): 41-44.
	Wang Y T, Cao Y. Analysis on garbage caloric value in MSW incineration power plant in China[J]. Environmental Sanitation Engineering, 2019, 27(5): 41-44.
10	国家市场监督管理总局, 中国国家标准化管理委员会. 生活垃圾分类标志: [S]. 北京: 中国标准出版社, 2019.
	State Administration for Market Regulation, Standardization Administration of the People's Republic of China. Signs for classification of municipal solid waste: [S]. Beijing: Standards Press of China, 2019.
11	房德职, 李克勋. 国内外生活垃圾焚烧发电技术进展[J]. 发电技术, 2019, 40(4): 367-376.
	Fang D Z, Li K X. An overview of power generation from municipal solid waste incineration plants at home and abroad[J]. Power Generation Technology, 2019, 40(4): 367-376.
12	刘波. 提高生活垃圾热值的几种途径[J]. 科技创新导报, 2012, (6): 236.
	Liu B. Several ways to increase the calorific value of domestic waste[J]. Science and Technology Innovation Herald, 2012, (6): 236.
13	李清海, 甘超, 蒙爱红, 等. 干燥对乏垃圾热值影响的实验研究[J]. 清华大学学报(自然科学版), 2011, 51(12): 1865-1869.
	Li Q H, Gan C, Meng A H, et al. Experimental study on effect of drying on heating value of spent waste[J]. Journal of Tsinghua University(Science and Technology), 2011, 51(12): 1865-1869.
14	文科军, 吴丽萍, 杨丽, 等. 可燃垃圾的焚烧热值分析[J]. 环境科学与技术, 2007, 30(7): 40-42.
	Wen K J, Wu L P, Yang L, et al. Analysis of caloric value of flammable refuse [J]. Environmental Science and Technology, 2007, 30(7): 40-42.
15	Bonifazi G, Serranti S, Rem P C. Hydrogen content and calorific value of municipal solid waste: innovative quality control strategies of waste fed to incinerators[J]. Waste Management and the Environment IV, 2008, 109: 289-298.
16	Oribe-Garcia I, Kamara-Esteban O, Martin C, et al. Identification of influencing municipal characteristics regarding household waste generation and their forecasting ability in Biscay[J]. Waste Management, 2015, 39: 26-34.
17	孙培锋, 李晓东, 池涌, 等. 城市生活垃圾热值预测的研究[J]. 能源工程, 2006, (5): 39-42.
	Sun P F, Li X D, Chi Y, et al. The study on prediction of lower heat value of MSW[J]. Energy Engineering, 2006, (5): 39-42.
18	Chen W H, Lee K T, Chih Y K, et al. Novel renewable double-energy system for activated biochar production and thermoelectric generation from waste heat[J]. Energy & Fuels, 2020, 34(3): 3383-3393.
19	Abubakar A, Barnabas M H, Tanko B M. The physico-chemical composition and energy recovery potentials of municipal solid waste generated in Numan Town, North-Eastern Nigeria[J]. Energy and Power Engineering, 2018, 10(11): 475-485.
20	Behrsin I. Rendering renewable: technoscience and the political economy of waste-to-energy regulation in the European Union[J]. Annals of the American Association of Geographers, 2019, 109(5): 1362-1378.
21	Bai J, Lian S, Liu Z, et al. Deep learning based robot for automatically picking up garbage on the grass[J]. IEEE Transactions on Consumer Electronics, 2018, 64(3): 382-389.
22	张瑛华, 张友富, 王洪. 基于神经网络的生活垃圾低位热值计算模型的研究与应用[J]. 电力建设, 2010, 31(9): 94-97.
	Zhang Y H, Zhang Y F, Wang H. Research and application of the LHV of MSW calculation model based on neural network[J]. Electric Power Construction, 2010, 31(9): 94-97.
23	马晓茜, 谢泽琼. 基于BP神经网络的垃圾热值预测模型[J]. 科技导报, 2012, 30(23): 46-50.
	Ma X Q, Xie Z Q. Prediction models for the heating values of municipal refuse based on BP neural network[J]. Science & Technology Review, 2012, 30(23): 46-50.
24	邓睿渠, 汪林正, 张睿智, 等. 基于粒子群算法的生活垃圾高热值成分热解动力学特性研究[J]. 化工学报, 2020, 71(7): 3238-3246.
	Deng R Q, Wang L Z, Zhang R Z, et al. Kinetic study on MSW components with particle swarm method[J]. CIESC Journal, 2020, 71(7): 3238-3246.
25	Meng S, Chu W T. A study of garbage classification with convolutional neural networks[C]//2020 Indo–Taiwan 2nd International Conference on Computing, Analytics and Networks (Indo-Taiwan ICAN). IEEE, 2020: 152-157.
26	Zeng M, Lu X, Xu W, et al. PublicGarbageNet: a deep learning framework for public garbage classification[C]//2020 39th Chinese Control Conference (CCC). IEEE, 2020: 7200-7205.
27	宁凯, 张东波, 印峰, 等. 基于视觉感知的智能扫地机器人的垃圾检测与分类[J]. 中国图象图形学报, 2019, 24(8): 1358-1368.
	Ning K, Zhang D B, Yin F, et al. Garbage detection and classification of intelligent sweeping robot based on visual perception[J]. Journal of Image and Graphics, 2019, 24(8): 1358-1368.
28	Lowe D G. Object recognition from local scale-invariant features[C]//Proceedings of the Seventh IEEE International Conference on Computer Vision. IEEE, 1999: 1150-1157.
29	Bay H, Ess A, Tuytelaars T, et al. Speeded-up robust features(SURF)[J]. Computer Vision and Image Understanding, 2008, 110(3): 346-359.
30	Redmon J, Divvala S, Girshick R, et al. You only look once: Unified, real-time object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2016: 779-788.
31	Qi Q, Zhao S Y, Zhao W J, et al. High-speed video salient object detection with temporal propagation using correlation filter[J]. Neurocomputing, 2019, 356(3): 107-118.
32	徐畅, 程文轩, 杨远舟. 关于Yolo目标检测算法的基础研究[J]. 电脑与信息技术, 2020, 28(4): 45-47.
	Xu C, Cheng W X, Yang Y Z. Research on the basis of Yolo target detection algorithm[J]. Computer and Information Technology, 2020, 28(4): 45-47.
33	Gienger A, Ostertag A, Böhm M, et al. Data-based distributed fault diagnosis for adaptive structures using convolutional neural networks[J]. Unmanned Systems, 2020, 8(3): 221-228.
34	杨兰兰, 高铭宇, 王晨宁, 等. 基于数据增强的人脸表情识别方法研究[J]. 计算机产品与流通, 2020, (11): 128-129.
	Yang L L, Gao M Y, Wang C N, et al. Research on facial expression recognition method based on data enhancement[J]. Computer Products and Circulation, 2020, (11): 128-129.
35	中华人民共和国住房和城乡建设部. 生活垃圾采样和分析方法: [S]. 北京: 中国标准出版社, 2009.
	Ministry of Housing and Urban-Rural Development of the People's Republic of China. Sampling and analysis methods for domestic waste: [S]. Beijing: Standards Press of China, 2009.
36	Fenn P. The deep learning revolution[J]. Information, Communication & Society, 2020, 23(7): 1098-1099.
37	Chow D S, Khatri D, Chang P D, et al. Updates on deep learning, and glioma: use of convolutional neural networks to image glioma heterogeneity[J]. Neuroimaging Clinics of North America, 2020, 30(4): 493-503.
38	张琦, 张荣梅, 陈彬. 基于深度学习的图像识别技术研究综述[J]. 河北省科学院学报, 2019, 36(3): 28-36.
	Zhang Q, Zhang R M, Chen B. Research review of image recognition technology based on deep learning[J]. Journal of the Hebei Academy of Sciences, 2019, 36(3): 28-36.
39	旦增, 王旭彤, 颜蓓蓓, 等. 西藏生活垃圾掺烧市政污泥的焚烧特性研究[J]. 化工学报, 2019, 70(8): 3151-3159.
	Dan Z, Wang X T, Yan B B, et al. Study on incineration characteristics of Tibetan municipal solid wastes mixed with sewage sludge[J]. CIESC Journal, 2019, 70(8): 3151-3159.

[1]	Yue CAO, Chong YU, Zhi LI, Minglei YANG. Industrial data driven transition state detection with multi-mode switching of a hydrocracking unit [J]. CIESC Journal, 2023, 74(9): 3841-3854.
[2]	Chengying ZHU, Zhenlei WANG. Operation optimization of ethylene cracking furnace based on improved deep reinforcement learning algorithm [J]. CIESC Journal, 2023, 74(8): 3429-3437.
[3]	Linqi YAN, Zhenlei WANG. Multi-step predictive soft sensor modeling based on STA-BiLSTM-LightGBM combined model [J]. CIESC Journal, 2023, 74(8): 3407-3418.
[4]	Gang YIN, Yihui LI, Fei HE, Wenqi CAO, Min WANG, Feiya YAN, Yu XIANG, Jian LU, Bin LUO, Runting LU. Early warning method of aluminum reduction cell leakage accident based on KPCA and SVM [J]. CIESC Journal, 2023, 74(8): 3419-3428.
[5]	Xuejin GAO, Yuzhuo YAO, Huayun HAN, Yongsheng QI. Fault monitoring of fermentation process based on attention dynamic convolutional autoencoder [J]. CIESC Journal, 2023, 74(6): 2503-2521.
[6]	Cheng YUN, Qianlin WANG, Feng CHEN, Xin ZHANG, Zhan DOU, Tingjun YAN. Deep-mining risk evolution path of chemical processes based on community structure [J]. CIESC Journal, 2023, 74(4): 1639-1650.
[7]	Shumin ZHENG, Pengcheng GUO, Jianguo YAN, Shuai WANG, Wenbo LI, Qi ZHOU. Experimental and predictive study on pressure drop of subcooled flow boiling in a mini-channel [J]. CIESC Journal, 2023, 74(4): 1549-1560.
[8]	Xinyuan WU, Qilei LIU, Boyuan CAO, Lei ZHANG, Jian DU. Group2vec: group vector representation and its property prediction applications based on unsupervised machine learning [J]. CIESC Journal, 2023, 74(3): 1187-1194.
[9]	Xuerong GU, Shuoshi LIU, Siyu YANG. Research on multi-parameter optimization method based on parallel EGO and surrogate-assisted model [J]. CIESC Journal, 2023, 74(3): 1205-1215.
[10]	Sheng’an ZHANG, Guilian LIU. Multi-objective optimization of high-efficiency solar water electrolysis hydrogen production system and its performance [J]. CIESC Journal, 2023, 74(3): 1260-1274.
[11]	Jiahui CHEN, Xinze YANG, Guzhong CHEN, Zhen SONG, Zhiwen QI. A critical discussion on developing molecular property prediction models: density of ionic liquids as example [J]. CIESC Journal, 2023, 74(2): 630-641.
[12]	Haiou YUAN, Fangjun YE, Shuo ZHANG, Yiqing LUO, Xigang YUAN. Synthesis of heat-integrated distillation sequences with intermediate heat exchangers [J]. CIESC Journal, 2023, 74(2): 796-806.
[13]	Xuejin GAO, Kun CHENG, Huayun HAN, Huihui Gao, Yongsheng QI. Fault diagnosis of chillers using central loss conditional generative adversarial network [J]. CIESC Journal, 2022, 73(9): 3950-3962.
[14]	Yalin WANG, Yuqing PAN, Chenliang LIU. Intermittent process monitoring based on GSA-LSTM dynamic structure feature extraction [J]. CIESC Journal, 2022, 73(9): 3994-4002.
[15]	Zhe SUN, Huaqiang JIN, Kang LI, Jiangping GU, Yuejin HUANG, Xi SHEN. Fault diagnosis method of refrigeration and air-conditioning system based on digitized knowledge representation [J]. CIESC Journal, 2022, 73(7): 3131-3144.