Abstract

The power sector plays a pivotal role in China’s carbon peak and carbon neutrality targets. To build a low-carbon power system, it is important to develop wind and solar. This study aims to evaluate the economic impacts of the newly launched renewable portfolio standard in 2030 in China using a cost minimization model and an input-output model. The results show that to accomplish the renewable electricity portfolio standard in 2030, the installed wind and solar capacity will have to reach 1451.9 gigawatts (GW) in 2030. The Northeast, Northwest, and North regions will deploy the most installed capacity, and Inner Mongolia will take on the most renewable energy generation tasks. The annual cost of wind and solar development is expected to be 506.6 billion yuan in 2030, 94.7% of which are new construction costs and storage costs. Renewable energy growth will result in a 5.4-cent (RMB) per kWh rise in the national average electricity price compared to 2019, and Heilongjiang, Gansu, and Shanxi are the most affected. The rapid development of renewable electricity in the next decade will increase the Consumer Price Index (CPI) by 0.4%, Producer Price Index (PPI) by 0.9%, and Gross Domestic Product (GDP) deflator by 0.5% in 2030. Based on the results, we propose to improve the electricity market mechanisms, enhance the electricity transmission stability, and develop policies appropriate to local conditions.

Abstract

With the continuous promotion of the carbon peak emissions and carbon neutralization strategies, higher demands are placed on engine economic performance. Virtual sensors as an online information collection technology can be used to control various performance indicators of engines. Here is an example of ISFC to represent the engine performance prediction. In this paper, the feasibility of three machine learning methods, Artificial Neural Network (ANN), Random Forest (RF), and Support Vector Regression (SVR), for predicting fuel consumption applications are explored. Firstly, a calibrated engine one-dimensional (1D) model is constructed. Then, the 1D model generates a dataset with engine load, engine speed and spark time, and indicative specific fuel consumption (ISFC) as an output, for the training of machine learning methods. The performance of different algorithms was compared using the coefficient of determination (R2), the root-mean-square error (RMSE), and the mean absolute percentage error (MAPE) as evaluation metrics. By comparing test dataset prediction and map prediction, RF has a large prediction error at boundary operation conditions and ANN sometimes has a relative error of more than 10%. SVR performs well in each statistical index and map prediction, and therefore it is an algorithm that can be used by virtual sensors.

Abstract

Acetone refining processes are usually high in energy cost since the acetone purity and recovery rate should meet the product specification simultaneously. Different heat integration approaches are applied in a typical acetone refining process, which produce high-purity acetone product from crude acetone. The processes with energy-saving approaches are optimized according to total annual cost (TAC), and the TAC calculation results are compared to identify the energy-saving effects in the acetone refining process. Optimization is also applied for processes with different acetone product purity specifications to distinguish the influence of purity specification on final optimization results. The results showed that the heat integration can achieve a decrease of 30% on energy consumption and 17% on TAC, and the selection of product purity specification level also have an influence on energy cost and TAC. This indicated that the producers can vary their acetone product specification to obtain a large decrease of energy consumption according to their product purity demand, and it is also possible to use other separation methods with less energy cost in the treatment of acetone product with lower purity to improve the product purity, such as membrane separation.

Abstract

Product environmental impact analysis by carbon footprint is a highly recognized research method to quantitatively evaluate carbon emission intensity of industrial products. In addition to the traditional carbon footprint method, researchers have recently proposed a carbon handprint method. This method is driven by the concept of describing the positive impact of products on climate. However, the handprint method has not been applied to industrial scenarios. The guidance of this method for industrial enterprises’ emission reduction plans is not clear. The essence of handprint is to reduce the footprint. This paper proposes an enhanced evaluation method of product carbon handprint. We compare footprint and handprint methods by considering the improvement of production process. We consider both the reduction of footprint in the sense of life-cycle analysis and the positive impact of reducing the footprint of downstream customers. A plasticizer production enterprise in Zhejiang province is taken as an example. This paper establishes four carbon emission reduction methods of such enterprises and makes a quantitative comparison between footprint and handprint. The comparison results show that the input raw materials account for a high proportion of carbon emissions in both methods. However, in the scope of handprint, plasticizer manufacturers could produce modified plasticizers to generate carbon handprint and reduce GHG emissions for downstream customers. The reduction effect of plasticizer on carbon handprint of polyvinyl chloride customers reached 0.983 tCO2-eq/t, twice as much as the reduction of carbon footprint in the manufacturer. Our work shows that handprint method is a more systematic method.

Abstract

More and more distributed generations (DGs) are integrated into distribution network (DN), and IEEE 1547.4-2011 standard encourages the conscious island operation of the DN. However, the fluctuation of DG output, power supply priority of load and operation state of interconnection switch are required to be considered in island partition, which greatly increases the difficulty of solving the model. An island partition method based on heuristic algorithm and Prim topology generation is proposed in the paper. In view of the shortcomings of the existing island partition methods, this method comprehensively considers the important influence of interconnection switch and power supply priority. The secondary outage constraint under fault state is also considered in the construction of island partition model, which makes the optimization result closer to the engineering practice. In addition, a heuristic prospective greedy algorithm is used to solve the island partition model. This method can effectively overcome the blindness of one-step selection and obtain a better scheme. The effectiveness of the method is verified by a case study of PG&E 69 bus system.

Abstract

Hydrogen energy is one of the potential clean energy sources that could be used on a large scale. Membrane-based hydrogen evolution and separation is one of the most promising approaches to low-cost hydrogen production. However, suitable membrane technologies are lacking and the development of advanced materials needs to be accelerated. In this paper, we provide a mini review of artificial intelligence (AI) applications to hydrogen separation membrane and hydrogen evolution membrane reactor discovery. By referring to the AI-guided development cases, readers will obtain a concise perception of popular machine learning (ML) methods and how they work to realize targets in specific application. We aim to assemble ML methods with the membrane materials development process. Current limitations to be addressed and prospect of AI applications in membrane discovery are also highlighted in the conclusion.

Abstract

Massive emission of multi-pollutants from the utilization of fossil fuels has caused severe environmental problems in past decades. The transport process of multi-pollutant molecules in nano-porous materials is involved, and considered to be one of the most significant processes, in the removal of these pollutants no matter by adsorption or catalysis. However, the mechanism of nano-scale transport is not fully understood due to the complexity in pore structures and diversity in pollutants. This work probes into the application of non-equilibrium molecular dynamics (NEMD) simulations to study the mass transfer of multi-pollutants at nano-scale. A dual control-volume (DCV) model of titanium-based nanopore is proposed and molecules of NO, NH3 and SO2, which are typical gaseous species in selective catalytic reduction (SCR) process for nitrogen oxides removal, are investigated. Simulations are performed to investigate the influences of temperature, pore width and hydroxyl site on diffusivity of objective molecules. The results show (1) the differential transport of NO, NH3 and SO2 in various temperature and pore conditions, (2) the impact of OH-groups on diffusion in different pore widths, (3) the influence of competitive diffusion of NH3 and SO2. These fundamental researches have provided guidance on the rational design of SCR catalysts for high deNOx activity and low SO2 oxidation.

Abstract

The bimetallic oxide copper−manganese based metal−organic frameworks (CuMnOx/MIL-100(Fe)) were easily synthesized and used to remove elemental mercury (Hg0) in flue gas. The mercury removal performance on the CuMnOx/MIL-100(Fe) was studied through a variety of characterization and analysis methods. The results suggest that the synthesized CuMnOx/MIL-100(Fe) have good crystallinity and high dispersity of constituent elements. The Hg0 removal performance was investigated under different conditions. The increase in O2 and NO concentration can obviously promote the removal efficiency of Hg0. The obtained adsorption capacity is 1.128 mg/g, which is higher than the adsorption capacity of most same type adsorbents. Finally, the Hg0 removal mechanism on the bimetallic oxide CuMnOx/MIL-100(Fe) has been proposed with various characterization analysis. The removal process of Hg0 follows the Langmuir-Hinshelwood and Mars-Maessen mechanism.

Abstract

In order to overcome the negative impact of the discontinuity and fluctuation of photovoltaic (PV) power generation on the power grid, in this study, a multi-variate data driven hybrid method for day-ahead hourly PV power curve prediction based on physical model and deep learning model is proposed. The physical model includes Ineichen clear sky model and PV performance model, while the deep learning model is a hybrid model combining two-dimensional grey relational analysis and bi-directional long short-term memory network model (2DGRA-BiLSTM). Firstly, the ideal clear sky global tilted radiation is calculated through the clear sky model, which is used as the input of PV performance model to obtain the ideal PV power under clear sky conditions. Secondly, the improved 2DGRA algorithm is proposed to obtain the best similar day from historical data. Thirdly, under the guidance of ideal clear sky power, the BiLSTM is trained with similarity-physics-informed data to obtain the difference between actual power and ideal clear sky power which is defined as RES-power. Compared with the other methods, results show that the accuracy of the deep learning model combined with physical method is the highest, followed by the deep learning model without physical method, and finally the simple physical model, whether it’s in clear sky condition or not.

Abstract

Reducing the Pt loading, especially on the cathode side, is very important for reducing the cost of proton exchange membrane fuel cells. However, lowering the Pt loading causes high voltage losses, but the underlying mechanism has not been fully understood. In this study, a new analytical catalyst layer model is established to study the oxygen reactive transport process, in which effect of carbon particle overlap is considered. Results from the analytical model show that carbon particle overlap leads to increase of ionomer thickness, reduction of ionomer specific surface area, and finally increase of oxygen transport resistance. By considering carbon particle overlap, an increased ~100 nm effective ionomer thickness is observed, and creates ~1500 s m-1 ionomer resistance. The order of magnitude of such ionomer resistance is equivalent to that of the Pt surface resistance measured in previous experimental studies, but there is no clarified source of this Pt surface resistance. As a result, it is reasonable to consider that carbon particle overlap may cause one of the sources of the oxygen transport resistance, which provides new insights for further alleviating the oxygen transport resistance under low Pt loading.