- College of Automation, Jiangsu University of Science and Technology, Zhenjiang, Jiangsu, China
Quasi-dynamic energy flow computation (EFC) has become a critical tool to determine and predict the states of the multi-energy system (MES), which helps improve MES’ operation efficiency and issues the security warning. However, methods in literature suffer numerical problems including fake oscillations, divergence, etc., Also, with the increasing of system dimensions, the computation efficiency can be hardly guaranteed due to the cross iterations between different nonlinear equations. This paper proposes an accurate and efficient method for quasi-dynamic energy flow computation. Using a scheme with total variation decreasing property, the numerical instability in solutions of thermal dynamics are effectively reduced. By estimating local truncation errors in a cheap way, the simulation step sizes are controlled adaptively and hence the overall simulation efficiency is greatly increased. Numerical tests were performed in a small system and the famous Barry Island system, which verified the advantages of the proposed method in both efficiency and accuracy.
1 Introduction
Nowadays, the wide exploitation of fossil fuels has apparently promoted the greenhouse effect. In view of this, the concept of multi-energy systems (MES) has been proposed to increase the overall energy utilization efficiency by reconciling the electric power with the more flexible heating power (Zhang M. et al., 2021; Jiang et al., 2021; Li et al., 2021; Zheng et al., 2022). However, the heterogenous forms of energy, which are depicted by disparate mathematical models, complicate the underlying operational law of MES. Several analysis methods have been put forward in literature to tackle this, including the steady-state energy flow computation (EFC) (Liu et al., 2016; Zhang et al., 2021b; Li et al., 2022), the quasi-dynamic EFC (Pan et al., 2016; Qin et al., 2019; Yao et al., 2021; Yu et al., 2022; Zhang et al., 2022) and the combined dynamic simulation (Shen et al., 2020; Wang et al., 2022). Among these methods, the quasi-dynamic EFC can plot the temperature variations precisely while demonstrates moderate modelling complexity, which has drawn much attention.
Essentially, the quasi-dynamic EFC in MESs is to simulate the spatio-temporal temperature distributions along pipelines, the variable mass flow rates, and the node temperature/power variations in heating network (HN) side, and the time-evolution of active/reactive power, and the voltage in power grid (PG) side. However, this becomes a non-trivial task when considering the spatial dimension of the system. Spatial evolution is a crucial feature that accounts for the time-delaying characteristics and hence flexibility of HN, but it cannot be modelled by ordinary differential equations. As a result, some mature and commercial time-domain simulation methods are inapplicable here. Recently, several efforts have been made in this area but are far from satisfactory. The developed methods undergo the influences of the so-called dissipative and dispersive errors, fail to respond quickly, and sometimes even diverges due to the numerical instability. Therefore, the development of a quasi-dynamic EFC method that possesses both fast responding time and high accuracy is still of great necessity.
The research gaps are summarized from two aspects in what follows. The first aspect is about the modelling and solutions of the spatio-temporal thermal dynamics in pipelines, which mainly falls into two categories: 1) the node method and 2) the partial differential equations (PDEs) and the related finite-difference/analytical methods. The node method is more physically sensible. It divides the whole pipelines into thermally insulated water masses, whose marching along the pipelines constitutes the thermal transmission. And the average temperatures of the water masses in the outlet buffer acts as the output temperatures (Li et al., 2016; 2023). Chen et al. have further proposed an enhanced node method, the water mass method, to eliminate the integer variables in optimization applications (Chen et al., 2019), which sheds light on the more complicated hydraulic-thermal cooperative operation optimization of MESs (Lu et al., 2020). It has been pointed out by Yu et al. that the node method is fairly accurate when depicting thermal dynamics but also suffer non-convergence issues in variable mass flow scenarios (Yu et al., 2022).
The PDEs are a more mathematical description of thermal dynamics. Some people discretize the spatial and temporal derivatives in the PDEs respectively to derive solvable algebraic constraints. Wang et al. have compared the accuracy and efficiency of the implicit upwind method and the characteristic line methods (Wang et al., 2017). Yao et al. have proposed a novel central scheme with unconditionally stability (Yao et al., 2021) But the fake oscillations and damping in the PDE solutions, which are called dispersive and dissipative errors respectively, are ignored in their works. By assuming constant mass flow, the analytical solutions of the PDE can be derived easily. Chen et al. (Chen et al., 2021) have used the Fourier analysis to derive a series solution. Yang et al. have derived the analytical power and temperature expressions with respect to boundary conditions of the PDE (Yang et al., 2020). And the initial condition case has been later finished by Zhang et al. (Zhang et al., 2022). Those analytical methods get rid of dispersive and dissipative errors completely but are not generic since the hydraulics are assumed to be still.
The second aspect is about how to obtain the time-varying hydraulics and electric variables, which are implicitly constrained by nonlinear algebraic equations. Qin et al. have proposed to alternatively solve the discretized PDEs and the remaining algebraic constraints (Qin et al., 2019). The PDE solutions first predict the thermal variables on the next time window, and then the remaining variables can be solved by Newton method accordingly. Yao et al. have pointed out the interaction errors between electric and heating variables and proposed the power-energy interfaces to calibrate the computation results (Yao et al., 2022). The cross iterations between PG and HN part of equations are computationally intensive, which hold up the responding time of energy flow analysis. Zhang et al. have proposed to uncouple the HN into sub-networks, which reduces the overall solution complexity and works well in constant mass flow scenarios (Zhang et al., 2021c). Huang et al. have applied the Levenberg-Marquardt-form Newton method to improve the convergence and efficiency of EFC (Huang et al., 2023). Yu et al. have proposed to derive time polynomials, which are called semi-analytical solutions, of unknown variables with a differential-transformation-based method (Yu et al., 2022). An adaptive time window strategy has also been developed to speed up the quasi-dynamic EFC and ensure convergence. However, the adaptive time window strategy of finite difference-based method has not been found in literature.
The research gaps are summarized as follows. The node method is physically sensible and accurate but is numerically unstable. The analytical methods provide exact solutions but are restricted to constant mass flow scenarios. The finite-difference-based methods are severely affected by the dissipative and dispersive errors, and the cross iterations between PG and HN sides are time-consuming. Also, some works focus on the HN side and overlook the interactions between PG and HN.
To overcome these limitations, this paper proposes a novel accurate and efficient solution method of quasi-dynamic EFC of MESs considering variable mass flow. The proposed method is based on the finite difference method, which is a flux limiter scheme with total variation decreasing (TVD) property. A novel expedite technique is further developed to accelerate the calculation. More precisely, the paper makes the following contributions:
1) The proposed method uses a flux limiter scheme to discretize the PDEs governing thermal dynamics, which effectively reduces the dispersive and dissipative errors.
2) The proposed method expedites the MES quasi-dynamic EFC by controlling the temporal simulation step size adaptively. The technique utilizes two different schemes to generate the local truncation error estimation, but the computation cost is cheap.
The structure of the paper is organized in the following way: Section 2 establishes the quasi-dynamic MES model; Section 3 introduces the flux limiter scheme of PDEs; Section 4 develops the expedite scheme; Section 5 gives the detailed simulation procedure; Section 6 gives the case studies; Section 7 concludes.
2 Model formulation
2.1 Thermal transmission equations
The thermal dynamics of the heating pipelines can be depicted by
where we denote by T (°C) the temporal-spatial distribution of temperatures along the pipelines;
2.2 Thermal equations
The node temperatures are the mixture of the outlet temperatures of the pipelines flowing into the node, which has the formula
where we denote by
2.3 Hydraulic equations
The hydraulic models are comprised of the mass flow injection/continuity equation, which have the formula
where
Some HN contains loops and the pressure drop around a closed loop equals zero, so we have
where we denote by Ki the resistance coefficients of pipe i.
2.4 Heating power equation
The heating power exchanged at a node can be computed as
where we denote by ϕ the heating power, Ts and Tr are respectively the supply and return temperatures of the node.
2.5 Power flow equations of PG
Since the transients in PG are usually omitted in quasi-dynamic analysis (Qin et al., 2019; Zhang et al., 2022), we adopt the AC power flow equation
where the bus voltage of bus i is expressed as
Yij is the admittance from bus i to bus j;
2.6 Coupling unit
The PG and HN are coupled via CHP units, electric boilers and heat pumps, etc., (Ma et al., 2021). In this paper, we consider the bi-directional coupling CHP units, that is, the heating power output of slack node of HN is determined by the electric power output of some PV bus in PG while the electric power output of slack bus of PG is determined by the heating power output of some slack node in HN. We have
where we denote by ηCHP the energy conversion efficiency coefficient.
3 The conservative scheme with flux limiter
The dissipation and dispersion errors in the solutions of Eq. 1 by implicit upwind scheme and the central scheme are due to the increase of total variations. To eliminate or reduce these errors, scheme with TVD property is a must. However, it has been theoretically proved by Thomas (Thomas, 1995) that a linear TVD difference scheme is at most of first order temporal/spatial accuracy, which is insufficient in practical use.
To utilize the TVD property and meanwhile achieve high order accuracy, an alternative approach is to adaptively switch between the TVD low-order scheme and the non-TVD high-order scheme. In smooth sections, simulation codes are switched to the high-order scheme while in non-smooth sections where dissipative and dispersive errors are conspicuous, simulation codes are switched to the TVD scheme.
Following the above philosophy, we use the following conservative scheme with flux limiter to solve the PDE Eq. 1, which yields
where R = Δt/Δx; k denotes the spatial difference indices; n denotes temporal difference indices; M denotes the total number of spatial difference sections;
The function
is referred to as the smooth indicator here. In smooth sections,
Typically, there are respectively the following three kinds of flux limiters:
(1) Van Leer limiter:
(2) C-O limiter:
(3) Superbee limiter:
In the case study, we will compare these three kinds of flux limiters in terms of dissipative and dispersive error elimination.
4 The expedite technique
Proper temporal step size Δt is vital since they account for the overall simulation efficiency and convergence. In this paper, we propose to adaptively control Δt during simulation in the following way.
The kernel of the technique is to construct two schemes with different accuracy and use their deviations to estimate the local truncation errors. Here, putting ϕ = 0 in Eq. 9, we have the first order FTBS scheme. In each step, the PDEs are first solved with the FTBS scheme, and the result is denoted by yI. Then the results by flux limiter (FL) scheme, which is denoted by yII, can be obtained by adding the terms containing ϕ in Eqs 10, 11 to yI. Since no extra computations are required, this process is computationally cheap.
Subtracting yII from yI, we obtain an estimate of the local truncation error, i.e.
because Δh cannot be altered, we assume the variation of
where Tola and Tolr are respectively the absolute and relative error tolerances. Tola is prescribed to prevent the extreme case where min (|yi (0)|, |yi (Δt)|) is close to zero so that Δt becomes extremely small. Tolr is prescribed to adjust the number of significant figures in the computed values. min (⋅, ⋅) ensures that both |yi (0)| and |yi (Δt)| are within the error tolerance.
In each step, we put the overall simulation errors as 2-norm, that is
As suggested by (Hairer et al., 1993), p should set to be 1 here because FTBS scheme is of first order. As for the divergence cases, a minimum threshold Δtmin can be prescribed to detect the case. That is, if Δt is decreased to be smaller than Δtmin, which means that the errors cannot be diminished by reducing Δt, then divergence occurs.
5 The overall simulation procedure
In this paper, we study the quantity regulation mode of HN, that is, the mass flow rate
The assumptions in Tables 1, 2 are slightly different from the assumptions in Liu et al. (2016), Qin et al. (2019), because, as stated in Section 2, we consider the bi-directional of PG and HN here. Since we consider the CHP type of coupling unit, the ϕ of Source node depends on the P of Slack bus while the P of PV bus depends on the ϕ of Slack node. As a result, these variables are unknown.
To obtain the variations of unknown variables in Tables 1, 2, we should first perform the steady-state calculation of energy flow using the method in (Liu et al., 2016). Thereafter, we obtain the initial distribution of temperatures in pipelines, i.e. the initial condition of Eq. 1. After these preparations, we can start the quasi-dynamic energy flow computation with the procedure in Figure 1. The simulation scenarios were set by interpolation. This is because disturbances can be interpreted as the variation of a specific parameter. Given a disturbance, the proposed method uses interpolation to obtain the value of the simulation parameter on a specific time node. And the parameters are updated once Δt is updated.
6 Case studies
The proposed method was verified in the following scenarios. All the numerical tests were performed on a laptop equipped with AMD Ryzen 4600 U and 16 GB RAM. The coding language is Matlab R2022a.
6.1 A single pipeline case with measured data
The proposed method was first verified for accuracy in a single-pipeline case. The Tout,
The root mean squares (RMSEs) of the three schemes, compared with the measured outlet temperatures, were computed in Table 3. The FL schemes have smaller RMSEs than the implicit upwind and central schemes. Among the three kinds of flux limiters, the Van Leer flux limiter has the best accuracy performance, but the C-O and Superbee limiters were all more precise than implicit upwind and central schemes.
6.2 A small system
Next, the proposed method was tested on a small MES to verify its numerical performance in accuracy and efficiency. The system, which comes from (Liu et al., 2016), consists of five heating nodes and four electrical buses. The system is pictorially shown in Figure 3.
The following methods were performed in the system:
Method 1: The method proposed in (Qin et al., 2019) with implicit upwind scheme.
Method 2: The method proposed in (Yao et al., 2021) with the implicit upwind scheme in Method 1 replaced by the central scheme.
Method 3: The proposed method with fixed Δt.
Method 4: The proposed method with adaptive Δt.
Method 1-Method 3 were performed with Δt =60 s, Δx =10 m. The benchmark trajectories were generated by the MATLAB ODE45 solver with Δt =3 s, Δx =2 m. The benchmark method is explained as follows. First, the spatial derivative in Eq. 1 is discretized with backward difference. Then the PDE becomes a set of ODEs. To further incorporate the nonlinear algebraic equations, the solving schemes are altered as
where x denotes the state variables, that is, the pipe temperatures; y denotes the algebraic variables, that is, the mass flow, power, voltage, etc. s denotes the stage of Runge-Kutta formulae. Here, s = 7. ai,j, bj denotes computation parameters. f and g are respectively the ODEs and AEs. The subscript 0 denotes initial/given values while subscript 1 denotes the unknown values. The computation consists of seven stages. In each of the first six stages, we first solve equation
The total simulation time was set to be 24 h. The supply temperature of node 4 and 5 increased from 100°C to 101°C within 10 min. The simulation results were illustrated in Figures 4, 5. Apparently, the high frequency components in results by Method 1 decayed, which ruined both the supply temperatures and the mass flow rates. And there are obvious fake oscillations in results by Method 2. In contrast, results by Method 3 and Method 4 got rid of high-frequency decaying and fake numerical oscillations and produced precise trajectories. The accuracy improvement was further reflected in Table 4. The root mean square errors of Method 1 to Method 3 were computed against the benchmark. It seems that, due to its TVD property, the proposed method demonstrated the highest accuracy.
The total computation times of Method 1-Method 4 were recorded in Table 5. Method 1 and Method 2 had similar computational performance, while Method 3 was more efficient than Method 1 and Method 2. The reasons are as follows. To compute the unknown temperature
6.3 The Barry Island case
Finally, the proposed method was tested on the well-known Barry Island system. Shown in Figure 6 is the 35-node-9-bus system. Four typical kinds of load trajectories were considered (Qin et al., 2021), and the loads of all nodes changed accordingly. The simulations were performed with the above Method 1-Method 4 with Δt =20 s, Δx =20 m.
The simulation results were shown in Figure 7. Like the previous case study, the proposed method depicted the temperature variations more precisely. However, in this case, the computation errors of temperatures have smaller impact on the accuracy of other variables. All the methods produced fairly accurate trajectories of active power. The RMSEs were computed in Table 6. The errors of Method 1 were still bigger than Method 2 and Method 3, but the accuracy of Method 2 was very close to Method 3 in this case.
The total simulation computation times were listed in Table 7. As the scale of the system increases, the proposed method still demonstrated obvious efficiency superiority. The expedite technique effectively cut down the total number of simulation time windows and saved nearly fifty percent total computation times.
7 Conclusion
Quasi-dynamic EFC in MESs faces the challenges of numerical instability and inefficiency. In view of this, we propose an accurate and efficient solution method for quasi-dynamic EFC in MESs. Firstly, we utilize a scheme with TVD property to discretize the PDE governing thermal dynamics. Thus, the fake oscillation and damping are decreased in an effective way. Secondly, we use two schemes with different accuracy to construct the local truncation error approximation, which helps control the step sizes adaptively. Case studies show that the proposed method was able to accurately depict the thermal dynamics while increased the overall simulation efficiency saliently.
Data availability statement
The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.
Author contributions
Conceptualization, methodology, software, and writing by YL; part of software, validation, and formatting by KY. All authors contributed to the article and approved the submitted version.
Conflict of interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
Chen, Y., Guo, Q., Sun, H., Li, Z., Pan, Z., and Wu, W. (2019). A water mass method and its application to integrated heat and electricity dispatch considering thermal inertias. Energy181, 840–852. doi:10.1016/j.energy.2019.05.190
Chen, Y., Guo, Q., Sun, H., and Pan, Z. (2021). Integrated heat and electricity dispatch for district heating networks with constant mass flow: A generalized phasor method. IEEE Trans. Power Syst. 36, 426–437. doi:10.1109/tpwrs.2020.3008345
Hairer, E., Nørsett, S. P., and Wanner, G. (1993). Solving ordinary differential equations I. second edn. Berlin, Heidelberg, Germany: Springer-Verlag.
Huang, Y., Sun, Q., Li, Y., Sun, C., and Chen, Z. (2023). Damping technique empowered robust energy flow calculation for integrated energy systems. Appl. Energy 343, 121168. doi:10.1016/j.apenergy.2023.121168
Jiang, Y., Wan, C., Botterud, A., Song, Y., and Shahidehpour, M. (2021). Convex relaxation of combined heat and power dispatch. IEEE Trans. Power Syst. 36, 1442–1458. doi:10.1109/TPWRS.2020.3025070
Li, Z., Wu, L., Xu, Y., Moazeni, S., and Tang, Z. (2022). Multi-stage real-time operation of a multi-energy microgrid with electrical and thermal energy storage assets: A data-driven MPC-ADP approach. IEEE Trans. Smart Grid 13, 213–226. doi:10.1109/tsg.2021.3119972
Li, Z., Wu, L., and Xu, Y. (2021). Risk-Averse coordinated operation of a multi-energy microgrid considering voltage/var control and thermal flow: An adaptive stochastic approach. IEEE Trans. Smart Grid 12, 3914–3927. doi:10.1109/tsg.2021.3080312
Li, Z., Wu, L., Xu, Y., Wang, L., and Yang, N. (2023). Distributed tri-layer risk-averse stochastic game approach for energy trading among multi-energy microgrids. Applied Energy 331, 120282. doi:10.1016/j.apenergy.2022.120282
Li, Z., Wu, W., Shahidehpour, M., Wang, J., and Zhang, B. (2016). Combined heat and power dispatch considering pipeline energy storage of district heating network. IEEE Trans. Sustain. Energy 7, 12–22. doi:10.1109/tste.2015.2467383
Liu, X., Wu, J., Jenkins, N., and Bagdanavicius, A. (2016). Combined analysis of electricity and heat networks. Appl. Energy 162, 1238–1250. doi:10.1016/j.apenergy.2015.01.102
Lu, S., Gu, W., Meng, K., Yao, S., Liu, B., and Dong, Z. Y. (2020). Thermal inertial aggregation model for integrated energy systems. IEEE Trans. Power Syst. 35, 2374–2387. doi:10.1109/tpwrs.2019.2951719
Ma, R., Liu, H., Geng, J., Zhu, D., Yan, Z., Luo, Z., et al. (2021). Fast decomposed method for dynamic energy flow calculation in integrated electricity and heat system. IEEE Access, 9, 168760–168766. doi:10.1109/ACCESS.2021.3116810
Pan, Z., Guo, Q., and Sun, H. (2016). Interactions of district electricity and heating systems considering time-scale characteristics based on quasi-steady multi-energy flow. Applied Energy 167, 230–243. doi:10.1016/j.apenergy.2015.10.095
Qin, X., Shen, X., Guo, Y., Pan, Z., Guo, Q., and Sun, H. (2021). Combined electric and heat system testbeds for power flow analysis and economic dispatch. CSEE J. Power Energy Syst. 7, 34–44. doi:10.17775/CSEEJPES.2020.02810
Qin, X., Sun, H., Shen, X., Guo, Y., Guo, Q., and Xia, T. (2019). A generalized quasi-dynamic model for electric-heat coupling integrated energy system with distributed energy resources. Appl. Energy 251, 113270. doi:10.1016/j.apenergy.2019.05.073
Shen, F., Ju, P., Shahidehpour, M., Li, Z., Wang, C., and Shi, X. (2020). Singular perturbation for the dynamic modeling of integrated energy systems. IEEE Trans. Power Syst. 35, 1718–1728. doi:10.1109/tpwrs.2019.2953672
Thomas, J. (1995). Numerical partial differential equations: Finite difference methods. New York, NY, USA: Springer-Verlag.
Wang, L., Zheng, J., Li, Z., Jing, Z., and Wu, Q. (2022). Order reduction method for high-order dynamic analysis of heterogeneous integrated energy systems. Appl. Energy 308, 118265. doi:10.1016/j.apenergy.2021.118265
Wang, Y., You, S., Zhang, H., Zheng, X., Zheng, W., Miao, Q., et al. (2017). Thermal transient prediction of district heating pipeline: Optimal selection of the time and spatial steps for fast and accurate calculation. Appl. Energy 206, 900–910. doi:10.1016/j.apenergy.2017.08.061
Yang, J., Zhang, N., Botterud, A., and Kang, C. (2020). On an equivalent representation of the dynamics in district heating networks for combined electricity-heat operation. IEEE Trans. Power Syst. 35, 560–570. doi:10.1109/TPWRS.2019.2935748
Yao, S., Gu, W., Lu, S., Zhou, S., Wu, Z., Pan, G., et al. (2021). Dynamic optimal energy flow in the heat and electricity integrated energy system. IEEE Trans. Sustain. Energy 12, 179–190. doi:10.1109/tste.2020.2988682
Yao, S., Gu, W., Wu, J., Lu, H., Zhang, S., Zhou, Y., et al. (2022). Dynamic energy flow analysis of the heat-electricity integrated energy systems with a novel decomposition-iteration algorithm. Appl. Energy 322, 119492. doi:10.1016/j.apenergy.2022.119492
Yu, R., Gu, W., Lu, H., Yao, S., and Zhang, S. (2022). “Non-iterative calculation of quasi-dynamic energy flow in the heat and electricity integrated energy systems,” in IEEE transactions on power systems. doi:10.1109/TPWRS.2022.3210167
Zhang, M., Wu, Q., Wen, J., Lin, Z., Fang, F., and Chen, Q. (2021a). Optimal operation of integrated electricity and heat system: A review of modeling and solution methods. Renew. Sust. Energ. Rev. 135, 110098. doi:10.1016/j.rser.2020.110098
Zhang, S., Gu, W., Lu, H., Qiu, H., Lu, S., Wang, D., et al. (2021c). Superposition-principle based decoupling method for energy flow calculation in district heating networks. Appl. Energy 295, 117032. doi:10.1016/j.apenergy.2021.117032
Zhang, S., Gu, W., Yao, S., Lu, S., Zhou, S., and Wu, Z. (2021b). Partitional decoupling method for fast calculation of energy flow in a large-scale heat and electricity integrated energy system. IEEE Trans. Sustain. Energy12, 501–513. doi:10.1109/tste.2020.3008189
Zhang, S., Gu, W., Zhang, X. P., Lu, H., Lu, S., Yu, R., et al. (2022). Fully analytical model of heating networks for integrated energy systems. Appl. Energy 327, 120081. doi:10.1016/j.apenergy.2022.120081
Zheng, W., Zhu, J., and Luo, Q. (2022). Distributed dispatch of integrated electricity-heat systems with 329 variable mass flow. IEEE Trans. Smart Grid, 1. doi:10.1109/TSG.2022.3210014
Nomenclature
Keywords: multi-energy systems, quasi-dynamic energy flow computation, simulation, PDE (partial differential equation), operation optimisation
Citation: Lu Y and Yang K (2023) An accurate and efficient quasi-dynamic simulation method of electricity-heat multi-energy systems. Front. Energy Res. 11:1229438. doi: 10.3389/fenrg.2023.1229438
Received: 26 May 2023; Accepted: 12 June 2023;
Published: 28 June 2023.
Edited by:
Zhengmao Li, Nanyang Technological University, SingaporeReviewed by:
Shuai Yao, Cardiff University, United KingdomSuhan Zhang, Hong Kong Polytechnic University, Hong Kong SAR, China
Nan Chen, University of Birmingham, United Kingdom
Copyright © 2023 Lu and Yang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Yiyang Lu, MjAyMjEwMzAyMjE4QHN0dS5qdXN0LmVkdS5jbg==