Comparative study of municipal solid waste disposal in three Chinese representative cities

Highlights

• Socioeconomic factors of MSW generation in different economic zones of China were studied by GCA.

• MSW generation in Beijing, Guangzhou and Lhasa were predicted by BP nerve network.

• Barriers, challenges and recommendations of MSW treatment and management were provided.

Abstract

China is facing severe environmental problems, the total municipal solid waste (MSW) generation has dramatically increased and presented great challenges all over the country. Cities have different characteristics in MSW components, generation quantity, influencing factors and disposal ways. Beijing, Guangzhou and Lhasa of Beijing-Tianjin-Hebei economic circle, Pearl River Delta economic zone, and Qinghai-Tibet plateau regions were selected as representative cities, to compare and analyze the features, influencing factors to the MSW generation quantities, and its components by grey correlation analysis. Predictions of MSW generation quantities for Beijing, Guangzhou and Lhasa in 2025 would be 1251.22, 704.71 and 71.040 million tons, respectively, made by BP (back propagation) nerve work using MATLAB software. Results showed that the main indicators of influencing factors to MSW in three cities were all of economic development levels, population and investments of government input. Based on the analysis of internal relationship and characteristics of MSW, some barriers of MSW management were analyzed, and recommendations were given. Our study analyzed the influencing factors and forecasted quantities of MSW generation in representative cities in China, providing valuable evidences for the severe MSW disposal situation of different economic zones in China.

Introduction

China has become the largest municipal solid waste (MSW) producer in the world (Song et al., 2016), the total amount of wastes collected and transported was 215.21 million tons in 2017 (NBS, 2019). MSW mainly consisted of food residue, wood waste, plastics, paper waste, rubber and textiles (Gu et al., 2018), characterized by high moisture content (46.10–64.10%) and low calorific value (5.67–8.03 MJ/kg) (Zhou et al., 2015). From 2006 to 2017, the safety disposal rate rose from 52.2% to 97.7%, and the harmless treatment capacity of MSW increased to 679,889 t/day (NBS, 2008, NBS, 2019). Since 13th Five-Year-Plan (13th Five) in 2016, waste treatment strategy has changed, transitioning from final disposal technologies and equipment to integrated waste management. But there still exists some complex challenges in MSWM. For example, the efficiency of source separate is low and needs to be increase, treatment capacity of disposal plants needs to be enlarged. As the key basis of MSW treatment, it is necessary to figure out the effects of influencing factors (e.g. socio-economic and demographic factors) on MSW generation quantities.

There has been a large literature on the association of socio-economic indices with MSW generation quantities. Sokka studied the impact of population, affluence and technology on MSW production, suggested little linear relation existed between GDP and MSW generation quantity (Sokka et al., 2007). Trang et al. proposed there were strong correlations of the generation rates per household with family size, income, educational level and level of environmental conscience was found in Vietnam (Trang et al., 2017). What’s more, the Environment Kuznets Curve (EKC) theory describing the relationship of environmental pollution and per capita income, with an inverse U-sharped curve between them was put forward by Kuznets (1955). Several researches have proved that the relationship between MSW generation and GDP conformed to the EKC theory (Beede and Bloom, 1995, Liu et al., 2017, Raymond, 2004, Vergara and Tchobanoglous, 2012). And the EKC theory was also applied to the relationship of MSW generation and GDP in China (Chen, 2018). But up until now, there have been rare reports reflecting socio-economic factors to MSW generation quantities in different economic zones in China.

MSW is heterogeneous and affected by numerous factors (Noori et al., 2009), forecasting of MSW generation quantities are required to consider all the indicators of influencing factors and the needs of MSW treatment and management (Abdoli et al., 2012). In this way, grey relational analysis (GRA) and Artificial Neural Network (ANN) were introduced into the analysis of influencing factors and forecast of MSW generation quantities. GRA was used to investigate the relationship between MSW generation and other factors affecting amount of waste (Intharathirat et al., 2015). For ANN, it has better ability to learn and construct a complex nonlinear system through large numbers of input/output variables (Asantedarko et al., 2017, Patel and Meka, 2013). BP (back propagation) neural network is one of the most widely used ANN tools. The structure of BP neural network contains input layer, hidden layer and output layer (shown in Fig. 1), it is an excellent parallel data-handling method with dark-box operating performance, powerful studying and generalizing ability. For BP neural network model, selecting of input variables is an important step, which will notably influence the accuracy of results during the set-up of BP neural network. Generation of MSW and influencing factors have evident non-linearity relationships, and BP neural network model is a good tool used to predict the quantity of MSW in different cities. So mainly related factors influencing MSW generation quantity were taken as input variables (Celik et al., 2016). To overcome the shortcomings such as long-term during the process of BP neural network forecast the MSW generation quantities, MATLAB software was introduced. Once data series are enough and the indicators of influencing factors are clear, data could be taken as training set by MATLAB software. Firstly, the BP neural network is trained. Then, the well-trained network is tested. Lastly, predictions can be made by qualified neural network.

In this paper, Beijing, Guangzhou and Lhasa of Beijing-Tianjin-Hebei economic circle, the Pearl River Delta economic zone and Qinghai-Tibet plateau regions were respectively selected, in order to give an overview of current situation and problems of MSW treatment and management in regions with different economic and social level in China. GPA was used to analyze the degrees of influencing factors to MSW generation. BP neural network was applied to predict the MSW generation in these three places in the future. The objective was: 1) to list generation quantities and components of MSW, 2) to analysis socioeconomic factors influencing MSW generation quantities by GPA, 3) to predict MSW generation quantities by BP neural network using MATLAB software. Some recommendations related to MSWM barriers from different zones were given. Furthermore, we hope to provide a new idea which could combine the analysis of influencing factors to MSW generations with the prediction of MSW generations by GPA and BP neural network, in this way, a global knowledge of MSW generation and the factors to it in some regions could be obtained.