A economic optimal dispatch model for combined heat and power microgrids supporting china’s carbon neutrality

Abstract

Addressing energy abandonment and low utilization efficiency is crucial for sustainable renewable energy development, particularly as microgrid systems gain prominence in China’s low-carbon transition. Existing combined heat and power microgrid system (CHPMS) scheduling models predominantly focus on operational costs while neglecting carbon trading dynamics and flexible load synergies, limiting their effectiveness under carbon neutrality goals. This study proposes a novel low-carbon economic dispatch model that uniquely integrates China’s carbon trading mechanism with multi-form flexible loads (shiftable/transferable/cutable) in a unified optimization framework. Leveraging scenario-based optimization and Monte Carlo simulations, the model coordinates power-heat interactions while addressing renewable uncertainties. Case studies demonstrate an 84.7% carbon reduction (1.70→0.26 t) and 9.8% cost savings (¥4,039→¥3,632) versus conventional dispatch, with flexible loads enabling 19.4% peak shaving. These results validate that carbon trading-flexible load synergy fundamentally improves CHPMS sustainability, achieving superior economic-environmental balance compared to isolated optimization approaches. This paper provides critical insights for policymakers and microgrid operators to accelerate China’s energy transition through collaborative carbon-market-responsive dispatching strategies.

Introduction

With the vigorous promotion of carbon emissions reduction by the international community, China’s energy transformation is imperative. From the perspective of the value chain, China’s energy industry will break the traditional vertically integrated management model of production, supply, storage, and utilization of energy, gradually transitioning to a chain or network structure with multiple independent links. In the new energy industry, the cumulative installed capacity of wind and PV power is continuously increasing. However, due to weak infrastructure and limited electricity consumption, there is a serious problem of energy waste caused by insufficient supporting facilities such as the power grid. In order to promote the coordinated development of the new energy industry and the power grid, various regions in China have proposed measures such as constructing energy storage facilities and promoting local consumption to reduce energy waste losses. Energy storage technology can enhance the stability of power supply and is an important means to achieve the consumption of energy resources. Therefore, by mitigating energy abandonment, effective utilization of new energy can be achieved. Microgrid systems play a significant role in promoting China’s energy transformation and restructuring of energy infrastructure. As an essential pathway integrating energy producers and consumers, microgrid systems comprise various distributed energy sources and loads, facilitating self-generation and self-consumption of energy¹. Flexible loads also play a crucial role in enhancing the economic and social benefits of microgrid systems. Therefore, revolving around energy supply, transmission, consumption, and other aspects, a combined heat and power microgrid system (CHPMS) involving multiple subjects such as photovoltaic generation systems, wind power generation systems, micro gas turbines, battery energy storage systems, heat storage tanks, and users has been formed, encompassing both electrical and heat energy.

In the current context of energy reform, CHPMS can be regarded as a value system. It not only accelerates the interaction between subsystems within the system but also enhances the effectiveness of value integration. Each subsystem can be seen as an information resource entity with value orientation. Through various collaborative behaviors, they achieve higher synergy benefits and value-added. In CHPMS, photovoltaic generation subsystems, wind power generation subsystems, micro gas turbines, battery energy storage systems, heat storage tanks, and users are considered as different types of value entities, and the system value can be reflected through the value activities of these entities. For photovoltaic generation subsystems and wind power generation subsystems, the main sources of value addition are the mitigated energy abandonment and optimized resource allocation. The value effect of micro gas turbines is reflected through their level of energy supply. The battery energy storage system and thermal storage tank participate in the early-stage energy supply process through optimized capacity allocation and technological innovation. This promotes the utilization of energy resources and meets load demands, thereby increasing the overall value effect of the system. For the user segment, value addition can be achieved through user demand response satisfaction, intelligent power management, and electric vehicle charging services. Furthermore, the integration of information technologies such as energy internet can provide information services for the entire CHPMS and its subsystems. Through an information exchange platform, valuable data can be transmitted to each subsystem, enhancing the collaborative operation efficiency of CHPMS, thereby creating value and achieving value co-creation.

Currently, many scholars have conducted extensive research on the scheduling issues of microgrid systems. The scheduling issues of CHPMS have also become a major constraint on its development. Many scholars mainly set scheduling objectives from the perspective of system operating costs but overlook the currently emphasized environmental costs. In the process of establishing a scheduling model for CHPMS considering both economic and environmental benefits, the proposal of carbon trading mechanisms has provided new insights for carbon reduction. Based on China’s coal use, Liu et al. (2022) examined and measured the stages and costs of coal reduction². Underpinned by the directional distance function, Peng and Liu (2023) used a parametric linear programming method and a Bayes bootstrap estimation method to estimate the marginal CO₂ emission reduction cost of China’s industrial sector, and to quantify the related influencing factors³. She proposed three low-carbon governance decision-making models based on environmental and operational costs to compare which governance model was optimal and the most suitable decision result for the policymaker and supply chain is both cost-effective and environmentally successful under the model considering carbon tax and carbon trade⁴. While research on low-carbon operation of microgrid systems has been ongoing, the rational scheduling of flexible loads requires further investigation in conjunction with carbon trading costs.

Flexible loads possess the characteristic of flexible scheduling, playing an important role in promoting energy supply-demand balance and improving energy utilization efficiency. Therefore, for multi-energy complementary CHPMS, the calculation of carbon trading costs can effectively reduce carbon emissions, while the inclusion of flexible loads can also enhance the system’s economic benefits. Based on the above research, this paper proposes a collaborative decision-making model for low-carbon economic dispatch of CHPMS considering carbon trading costs and incorporating flexible loads. It analyzes the impact of carbon trading and flexible loads on the environmental and economic benefits of the system, and explores the low-carbon economic operation mode of CHPMS.

To address these challenges, this study proposes a collaborative decision-making model for the low-carbon economic dispatch of a Combined Heat and Power Microgrid System (CHPMS), uniquely integrating carbon trading costs and flexible loads into the optimization framework—an approach not previously explored in this context. The effectiveness of this model is demonstrated through a case study, highlighting its ability to simultaneously enhance environmental and economic benefits under China’s carbon neutrality goals. Compared to related works, this research stands out by explicitly addressing carbon trading mechanisms and flexible load scheduling within a CHPMS framework. These distinctions underscore its novel contribution to achieving coordinated low-carbon operations in microgrid systems, offering practical insights for energy transition strategies. This paper advances the field through three key contributions:

• (1) Integration of China’s tiered carbon trading mechanism with industrial flexible load scheduling in CHPMS optimization, overcoming the isolated treatment of economic/environmental objectives.

• (2) Novel two-stage dispatch framework enabling coordinated heat-electrical storage while resolving temporal mismatches between renewable peaks and load demand.

• (3) The framework’s policy compatibility is empirically verified using operational data from China’s low-carbon pilot zones, demonstrating superior cost-emission tradeoffs compared to conventional dispatch strategies.

Therefore, this paper first clarifies the value co-creation analysis framework for low-carbon economic dispatch of CHPMS under the energy internet, then constructs a collaborative decision-making model for low-carbon economic dispatch of CHPMS based on the carbon trading mechanism and flexible load operation models. Finally, through case studies, this paper demonstrates the feasibility of the proposed model and investigates the impact of carbon trading and flexible loads on the overall system benefits through scenario analysis. Low-carbon economic dispatch also contributes to improving the level of information sharing among various components within CHPMS. Through rational optimization of resources allocation, it facilitates value co-creation and collaborative development of CHPMS.

Literature review

Low-carbon economic dispatch is one of the solutions to promote the efficient utilization of energy in microgrids, and optimizing and managing it is an important element for CHPMS to achieve value co-creation. Many scholars have studied the economic dispatch problem from the perspective of microgrids. Based on this, this paper reviews the literature from three aspects: economic dispatch of CHP microgrid systems, the application of energy storage in microgrid systems, and the significance of flexible loads in microgrid systems.

Economic dispatch of CHP microgrid systems

At present, to effectively utilize renewable energy, combined heat and power microgrid system have been extensively studied in China. The system integrates photovoltaic power generation, wind power generation, energy storage units, loads, and control equipment to provide electricity and heat to users⁵. The vigorous development of the energy Internet provides a more efficient application platform for CHPMS, promoting the effective utilization of various energy sources⁶. CHPMS possesses the characteristics of distributed energy, enabling it to generate electricity and heat close to the user’s demand. This proximity reduces energy transmission losses and enhances the stability and reliability of energy supply. During emergencies or power outages, CHPMS serves as a backup power source, ensuring continuous power supply. Moreover, through the utilization of clean energy and efficient energy conversion technology, CHPMS minimizes energy loss during transmission and conversion processes, consequently lowering carbon emissions and contributing to climate change mitigation and environmental pollution reduction. Furthermore, owing to its dual-use nature and distributed generation capabilities, CHPMS helps reduce energy procurement costs for users while offering more competitive energy prices.

The integration of renewable energy sources and CHP systems into microgrid networks has become increasingly important in the context of energy sustainability and resilience. Economic dispatching strategies for CHP microgrids are essential for maximizing energy efficiency, minimizing operational costs, and promoting renewable energy utilization. However, challenges such as operational constraints, uncertainty in renewable energy generation, and dynamic demand patterns pose significant obstacles to achieving optimal economic dispatch.

The development of cogeneration microgrid systems must address their economic dispatching challenge, which pertains to the rational allocation and distribution of various energy resources within the system to maximize its economic benefits. Specifically, economic dispatching must consider factors such as electricity and heat demand, energy supply costs, energy conversion efficiency, and system operational constraints. Through optimization of dispatching algorithms and strategies, the system can operate economically, improve energy utilization efficiency, reduce energy costs, and minimize carbon emissions. Numerous scholars have conducted comprehensive research on the economic dispatching of CHPMS from various perspectives. Nazari-Heris el al. (2020) delved into the economic dispatch of renewable energy and combined heat and power (CHP)-based multi-zone microgrids under electrical network limitations. It addressed the challenges of optimizing energy dispatch while considering constraints imposed by the electrical network⁷. Tang et al. (2019) proposed an operational flexibility-constrained intraday rolling dispatch strategy for CHP microgrids, emphasizing the importance of balancing operational flexibility with economic dispatch decisions, and presenting mathematical models and solution algorithms to optimize energy dispatch dynamically⁸. Horrillo-Quintero el al. (2025) demonstrated that integrating thermal energy storage systems (TESS) into grid-connected residential multi-energy microgrids (MEMGs) improves operational efficiency and reduces costs/emissions via a novel control strategy combining fuzzy logic, model predictive control, and nonlinear optimization, achieving better thermal balance, lower gas consumption, and cuts in operational costs and CO₂ emissions compared to TESS-less systems⁹. Huang et al. (2025) proposed a framework integrating renewable energy, EVs, and hydrogen into CHP microgrids for building energy management, using PSO optimization and CCP to handle uncertainties in renewables and EV loads (with MC simulations, K-means clustering, and Sigmoid-based RTP), demonstrating improved efficiency, emission reductions, and system reliability despite high hydrogen costs¹⁰. Ma et al. (2022) proposed a distributionally robust optimal dispatching approach for CHP microgrids, taking into account uncertainty factors such as concentrating solar power, and provided insights into addressing uncertainty in microgrid dispatch optimization, thereby enhancing the reliability and resilience of microgrid operation¹¹. Huylo et al. (2025) used a validated model of the University of Texas’ microgrid to show that integrating wind energy, optimized thermal/battery storage, and hydrogen-blend burners can reduce emissions by 54.7%, highlighting the need for long-duration storage and hydrogen turbines to achieve deeper decarbonization¹². Bentouati et al. (2022) introduced a Chaotic Krill Herd Optimizer for the efficient combination of renewable energy sources in isolated microgrid mode, exploring novel optimization techniques to enhance the integration of renewable energy sources and improve the overall performance of microgrid systems¹³. Dou et al. (2020) proposed an economic optimization dispatching strategy considering cogeneration and demand response to promote photoelectric consumption, aiming to optimize energy dispatching to reduce costs and improve energy efficiency¹⁴. Lin et al. (2022) presented an improved approximate dynamic programming approach for real-time economic dispatch of integrated microgrids. This research focused on developing advanced optimization algorithms to achieve real-time economic dispatch and improve the overall performance of integrated microgrid systems¹⁵. Fei el al. (2025) developed and validated a coordinated model for multi-energy ship microgrids (MESMs) that integrates power, thermal, hydrogen, and freshwater flows, addressing the gaps in heterogeneous energy carriers, ship power distribution networks (SPDN), and underwater radiated noise (URN) to enhance operational safety and energy efficiency while reducing greenhouse gas emissions¹⁶. Zou el al. (2024) introduced a novel peer-to-peer (P2P) trading framework for a coupled three-phase unbalanced distribution network (DN) and district heating network (DHN) system, employing a two-layer architecture with adaptive robust stochastic optimization to minimize operational costs and ensure constraint compliance under renewable energy uncertainties, thereby reducing energy losses and eliminating network operation violations¹⁷. Zhong el al. (2023) proposed a low-carbon operation model for energy hubs integrating distributionally robust optimization (DRO) with a Stackelberg game, where the energy hub acted as a leader optimizing against user and electric vehicle (EV) followers, employed a KL-divergence-based DRO approach to address renewable generation uncertainty, transformed the bilevel problem into a single-level using KKT conditions and the big-M method, and utilized a crafted column-and-constraint generation algorithm with linearization to achieve efficient solutions, validated through numerical case studies demonstrating its effectiveness¹⁸. Xiong et al. (2023) proposed a low-carbon economic dispatch model for integrated energy systems, integrating an organic Rankine cycle, power-to-gas, combined heat and power, and carbon capture under a ladder-type carbon trading mechanism¹⁹. Liu et al. (2023) developed a meteorological clustering-based adjustable robust optimization (ARO) framework for multi-microgrid systems (MMGSs), integrating conditional generative adversarial networks (CGAN) to enrich wind-power uncertainty modeling and stepped carbon trading for emission control, which achieves enhanced wind-power accuracy, cost efficiency, and carbon reduction via decentralized ADMM and C&CG algorithms, albeit with computational time trade-offs²⁰.

Collectively, these studies underscore a trend toward hybrid frameworks combining data-driven uncertainty quantification, multi-energy storage synergy, and carbon-aware optimization, though computational complexity and hydrogen cost barriers remain critical challenges for future scalability. Researchers have proposed diverse methodologies, including advanced optimization algorithms (e.g., Particle Swarm Optimization, Chaotic Krill Herd, approximate dynamic programming), robust stochastic optimization, and hybrid control strategies (e.g., fuzzy logic, model predictive control), to address challenges such as variable renewable generation, demand fluctuations, and operational constraints. These studies highlight the benefits of incorporating energy storage systems (e.g., heat, battery, hydrogen), demand response programs, and innovative frameworks like peer-to-peer trading or multi-energy ship microgrids to improve efficiency, reduce costs, and lower emissions. However, persistent challenges include managing uncertainty in renewable generation, balancing technical constraints with economic objectives, and mitigating computational complexity in multi-objective optimization. Future research must prioritize advancements in long-duration storage technologies and adaptive algorithms to further enhance the resilience, reliability, and carbon neutrality of CHP microgrids, ensuring their role in sustainable energy systems.

The application of energy storage in microgrid systems

Microgrids are localized energy systems that can operate independently or connect to the main grid, typically serving a specific area or community. With the increasing penetration of renewable energy sources and the growing demand for reliable and sustainable energy solutions, the integration of energy storage systems in microgrids has become increasingly important. Energy storage technologies, such as batteries, pumped hydro, and thermal storage, offer the capability to store excess energy generated during periods of low demand and release it during peak demand periods, thereby improving grid stability and efficiency. Teng et al. (2019) proposed a model of electro-thermal hybrid energy storage system for enhancing the autonomous control capability of multi-energy microgrids, developing advanced modeling techniques to optimize the integration of electro-thermal hybrid energy storage systems, thereby improving the overall performance and reliability of multi-energy microgrids²¹. Tang et al. (2021) presented an optimization framework for energy cooperation in residential microgrids with virtual storage technology, exploring the optimal utilization of energy resources and virtual storage technology to improve energy efficiency and cost-effectiveness in residential microgrid systems²². Chen (2023) investigated methods to optimize the configuration of multi-energy storage systems in standalone microgrids, exploring techniques for sizing, placement, and operation of energy storage systems to maximize their effectiveness in improving microgrid performance and reliability²³. Song et al. (2023) provided a comprehensive review of the potentials of electric-thermal sector coupling for frequency control in microgrids. It examined the integration of electric and thermal energy systems to enhance frequency control capabilities and improve grid stability²⁴. Gassi and Baysal (2022) evaluated the effectiveness of such models in optimizing energy management decisions considering the intermittent nature of renewable energy sources²⁵. Ning et al. (2022) focused on the design and optimization of combined cooling, heating, and power microgrids with energy storage station service. It investigated strategies to integrate energy storage systems into microgrids to enhance system flexibility, improve energy efficiency, and provide ancillary services²⁶. Puppala el al. (2025) evaluated the feasibility and optimal design of a standalone microgrid for a remote Indian village by analyzing four configurations combining diesel generators, wind, solar, and batteries, using socio-techno-economic-environmental metrics to select the scenario with highest renewable integration, lowest costs, and minimal unmet loads²⁷. El Shamy el al. (2025) used a chance-constrained MILP model to optimize the sizing of a PV-battery-hydrogen microgrid, balancing renewable integration and cost efficiency while ensuring 80% reliability across 100 weather/load scenarios, achieving a 25-year lifecycle cost of $1.221 million²⁸. Shahbazbegian et al. (2023) proposed a mixed-integer nonlinear programming (MINLP) model combined with an OA/ER/AP decomposition approach to optimize the siting and sizing of energy storage systems in multi-energy microgrids, demonstrating significant operational cost reductions and computational efficiency improvements through a comprehensive techno-economic analysis²⁹.

In conclusion, the reviewed literature emphasizes the critical role of energy storage systems (ESS) in enhancing the reliability, efficiency, and sustainability of microgrids, particularly in integrating renewable energy sources. Key advancements include optimizing multi-energy storage configurations (e.g., batteries, hydrogen, heat storage) to balance supply-demand dynamics, developing hybrid storage strategies for improved grid stability, and creating decision models to address uncertainties in renewable generation and load variability. However, further research is needed to address practical implementation challenges and optimize the performance of energy storage systems in diverse microgrid settings.

The significance of flexible loads in microgrid systems

With the increasing penetration of renewable energy sources and the need for grid flexibility, the effective management of flexible loads has become increasingly important. Flexible loads, including both thermal and electrical loads, offer the capability to adjust energy consumption in response to supply-demand imbalances, grid constraints, and market signals, thereby optimizing grid operation and improving overall system efficiency. Understanding the role and potential of flexible loads is crucial for advancing the resilience, sustainability, and economic performance of microgrids. Wen et al. (2023) explored thermal and electrical demand response strategies based on robust optimization techniques. It investigated methods to effectively manage flexible loads in microgrids, considering uncertainties in energy supply and demand, to enhance grid stability and reliability³⁰. Zhang et al. (2019) presented a robustly coordinated operation approach for multi-energy microgrids with flexible electric and thermal loads, proposing strategies to optimize the operation of microgrids, considering the dynamic interaction between different energy sources and flexible loads, to improve overall system efficiency and resilience³¹. Chang et al. (2022) evaluated methods to optimize the interaction between energy sources and loads in microgrids, considering stable supply and demand conditions, to achieve optimal grid operation and energy utilization³². Pashaei-Didani et al. (2019) investigated the optimal economic-emission performance of microgrids featuring fuel cell combined heat and power (CHP) systems and energy storage. It analyzed the integration of flexible loads into microgrid operation to achieve optimal economic and environmental outcomes, considering factors such as energy prices, emissions, and grid constraints³³. Singh et al. (2025) proposed a hybrid DSM approach combining load shifting, curtailment, and smart PHEV charging to optimize microgrid load profiles and reduce costs via DE optimization, effectively addressing economic and environmental challenges in power systems³⁴. Wu et al. (2022) proposed control strategies to enhance the resilience of distributed energy microgrids, considering the integration of flexible loads and distribution system constraints, to improve grid reliability and performance³⁵. Song el al. (2025) proposed an environmental-economic scheduling method for microgrids integrating staged carbon trading and generalized energy storage (combining flexible loads and storage), optimizing operational costs and carbon emissions through load analysis and system modeling, demonstrating improved economic efficiency and emissions reduction via simulations³⁶.

In summary, the literature reviewed underscores the significance of flexible loads in enhancing the economic, environmental, and operational efficiency of microgrids. These studies focus on optimizing economic and emission performance through demand response strategies, integrating flexible loads into microgrid dispatching frameworks, and developing control mechanisms that balance energy supply-demand dynamics while addressing grid constraints. The reviewed research primarily delve into optimizing economic emission performance, integrating demand response into microgrid dispatching, and deploying control strategies for distributed energy microgrids that leverage flexible loads. Moving forward, additional research efforts are warranted to address practical implementation hurdles, enhance demand response capabilities, and refine dispatching strategies tailored to diverse microgrid applications.

While prior studies have advanced microgrid optimization through various technical lenses, some challenges remain in harmonizing carbon market dynamics with operational flexibility. Therefore, this paper mainly explores the impact of carbon trading costs and flexible loads on the optimal dispatch of CHPMS.

Value co-creation analysis in economic dispatch of CHPMS

With the ongoing evolution of the Energy Internet, the energy enterprise alliance model, rooted in cutting-edge information technology, has garnered significant attention from the power industry. The planning and implementation of the energy Internet in the energy sector, along with the concentrated application of next-generation digital information technology in the fields of new energy production, storage, and consumption, have made precise prediction, reliable dispatching, real-time monitoring, and dynamic configuration of new energy possible. By incorporating energy storage technology to combine its characteristics with the inherent regulation capabilities of energy sources, smoothing the fluctuations in wind and PV power output, enhancing the controllability of wind and PV power generation, and improving the ability of new energy to be consumed locally and operate reliably have become urgent priorities.

In response to the national energy strategy and to achieve the goals of “peak carbon emissions” and “carbon neutrality”, it is necessary to promote the optimized allocation of clean and low-carbon energy, and enhance the compatibility between new energy and energy systems. The orderly coordination of resources is the foundation for the stable operation of the community microgrid combined heat and power (CHP) microgrid system (CHPMS). Traditional renewable energy generation companies supply electricity by controlling the power generation capacity, resulting in situations where the electricity supply during peak demand periods is insufficient and oversupply during off-peak periods, leading to significantly reduced resource utilization efficiency and excessive energy wastage. The application of energy storage technology alleviates the problem of imbalance between electricity supply and demand, as well as the issue of fluctuating power output. By discharging energy storage during peak demand periods and charging during off-peak periods, it improves resource utilization and absorption capacity. Additionally, it aids renewable energy generation companies in devising rational power generation plans.

The low-carbon resource dispatching within the CHPMS under the energy Internet constitutes a dynamic process wherein system entities interact to achieve efficient and sustainable energy management. This process underscores the value creation inherent in the CHPMS, particularly with the integration of energy storage, which modifies traditional output curves. Therefore, considering the participation of energy storage in low-carbon resource dispatching not only promotes optimal resource allocation but also embodies the value co-creation effect of the CHPMS. This paper primarily addresses the optimization of low-carbon resource dispatching in the CHPMS, which include the energy supply, energy conversion and storage, and energy demand sectors. The value co-creation analysis framework for the optimization of low-carbon resource dispatching in the CHPMS under the energy Internet includes several components, as illustrated in Fig. 1.

• (1) The value co-creation analysis framework for optimizing low-carbon resource dispatching in the CHPMS under the energy Internet comprises three principal sectors: energy supply, energy conversion and storage, and energy demand. The energy supply sector encompasses the main grid, wind power, photovoltaic (PV), and natural gas. The energy conversion and storage sector includes micro gas turbines, gas boilers, battery energy storage systems, and heat storage tanks. The energy demand sector is divided into base load and flexible load components. Entities within the energy supply sector are tasked with developing rational energy supply strategies informed by optimized resource dispatching outcomes. Collaborative interactions with entities in the energy conversion and storage sector facilitate coordinated dispatching, ensuring operational stability of the CHPMS. In the energy demand sector, the integration of flexible loads enhances demand-side management, thereby improving the economic and environmental performance of the system. This tri-sectoral framework, as illustrated in Fig. 1, leverages power and heat flow pathways to optimize resource allocation.

• (2) The main entities of wind power and PV will execute power generation plans based on output prediction data, supplying electricity to the grid or directly meeting the electricity demand of specific areas. The power grid transmits electricity from generation sources to end users via transmission lines, as depicted in the upper “Power Flow” layer of Fig. 1. Natural gas serves a dual role: it fuels micro gas turbines for electricity generation and supports heat energy production through gas boilers, heating water or gas as shown in the middle “Heat Flow” layer. Energy conversion technologies, including micro gas turbines and gas boilers, transform primary energy into usable electrical and thermal outputs. Heat recovery systems and storage tanks further distribute thermal energy to meet heating demands. Additionally, battery energy storage systems and heat storage tanks are installed to balance energy supply and demand during periods of unstable energy supply or peak demand, thereby enhancing system flexibility and reliability, and also increasing the value-added benefits for energy stakeholders. By reasonably planning the Time-of-Use electricity pricing mechanism and flexible load in the energy demand sector, it is also possible to enhance energy efficiency and the integration capacity of new energy sources. Collectively, these sectors collaborate to optimize resource dispatching, minimizing operating costs and carbon trading expenses, as indicated in the bottom “Optimal resource scheduling coordination” layer of Fig. 1. This optimization fosters low-carbon economic operations, offering strategies to enhance energy structures and reduce carbon emissions through amplified value co-creation effects.

• (3) Under the value co-creation analysis framework for resource optimization dispatching in CHPMS, there may be information asymmetry between the entities in the energy supply sector and other sectors. Real-time data from wind power and PV systems exhibit continuous variability, potentially affecting coordination outcomes. However, the energy Internet enables the timely upload of real-time data to a centralized platform, as implied in Fig. 1’s real-time scheduling policy change mechanism. This facilitates adaptive adjustments within the resource optimization model, minimizing the adverse impacts of information asymmetry and ensuring robust system performance.

Fig. 1

The dispatching process of CHPMS.

With the vigorous promotion of low-carbon economy development, the environmental and social benefits brought about by the utilization of new energy have also promoted the active response of energy storage to the demand for the consumption of abandoned energy resources from new energy. A two-way flow of energy, information, and value is formed between the entities in the energy supply sector and energy storage, facilitating the smooth advancement of the value co-creation analysis framework.

Load model and low-carbon economic dispatching model of CHPMS

To provide a clear overview of the system, the structure of the CHPMS is illustrated in Fig. 2. This schematic diagram delineates the key components, their interconnections, and the energy flow pathways within the CHPMS, operating under the energy Internet framework. The system is divided into three primary sectors: energy supply, energy conversion and storage, and energy demand. The energy supply sector includes the power grid, PV systems, wind power, and natural gas. The energy conversion and storage sector comprises micro gas turbines, gas boilers, heat recovery systems, battery energy storage systems, and heat storage tanks, facilitating the transformation and storage of energy to ensure system stability. The energy demand sector consists of electric loads (categorized into base and flexible loads) and heat loads. Power flow (indicated by blue arrows) and heat flow (indicated by red arrows) illustrate the energy transfer between components, while the energy Internet framework—encompassing simulation, scheduling management, trading distribution, and Internet supervision platforms—enables system coordination and information interaction. Additionally, external factors such as carbon quota trading markets, energy trading markets, and government policy assurance influence the system’s operation, promoting low-carbon resource dispatching. Specifically, the flexible loads within the energy demand sector, as depicted in Fig. 2, play a critical role in enabling demand-side management, which is a key focus of the subsequent model formulation.

Fig. 2

Structure of the CHPMS under the energy internet framework.

Based on determining three kinds of flexible load models, this paper constructs the collaborative decision model of low-carbon economic dispatch of CHPMS and sets economic and environmental objectives to discuss the optimal resource allocation.

Load model of CHPMS

The CHPMS considered in this paper involves three load types, including shiftable load, transferable load and cutable load.

Shiftable load

Shiftable load refers to the ability to adjust the load by moving or redispatching the use time of certain electrical equipment or processes if the electrical load needs to be adjusted. This load can often be adjusted without affecting production or services to better match the power supply and demand balance. Considering that the energy supply time can be changed according to the plan, the load shift dispatching can span multiple periods, but the whole load must be shifted. To avoid the unreasonable situation in the process of shiftable load, the constraint conditions of the shift period are set. Set the shiftable load to ${L_S}$, the shiftable period to $\left[ {{t_{SL - }},{t_{SL+}}} \right]$, and the duration of shiftable load to ${t_D}$ in the unit dispatching period. The set ${S_{SL}}$ of the initial period of the shiftable load is

$${S_{SL}}=\left[ {{t_{SL - }},{t_{SL+}} - {t_D}+1} \right]$$

(1)

.

The compensation cost ${F_{SL}}$ to compensate the user after shift is

$${F_{SL}}=F_{C}^{{SL}}\sum\limits_{{\tau =1}}^{T} {P_{\tau }^{{SL}}}$$

(2)

.

Where, $F_{C}^{{SL}}$ is the compensation cost of shiftable load per unit power. 1 h is the unit of dispatching period, and $P_{\tau }^{{SL}}$ is the shift power of period $\tau$. T indicates the total number of hours in a dispatching period.

Transferable load

Transferable load refers to the ability to transfer the use of certain electrical equipment or processes to another period or another location to adjust the load if the electrical load needs to be adjusted. In the whole dispatching cycle, the load of each period can be flexibly transferred under the premise of meeting the load demand of different periods, and the transferable period is $\left[ {{t_{TL - }},{t_{TL+}}} \right]$. In order to avoid frequent suspension of the power supply equipment during the dispatching period, this paper sets the minimum continuous running time of the transferable load and the constraints of the transferable power.

$$\sum\limits_{{\tau =t}}^{{t+T_{{\hbox{min} }}^{{TL}} - 1}} {\Upsilon _{\tau }^{{TL}} \geqslant T_{{\hbox{min} }}^{{TL}}}$$

(3)

$$\Upsilon _{\tau }^{{TL}}P_{{\hbox{min} }}^{{TL}} \leqslant P_{\tau }^{{TL}} \leqslant \Upsilon _{\tau }^{{TL}}P_{{\hbox{max} }}^{{TL}}$$

(4)

.

Where, $T_{{\hbox{min} }}^{{TL}}$ is the minimum continuous running time. The transition state of variable $\Upsilon$ at period $\tau$ is marked with 0–1. If $\Upsilon _{\tau }^{{TL}}=1$, the load is transferred during period $\tau$. Conversely, $\Upsilon _{\tau }^{{TL}}=0$ indicates that it is in the non-transition state. $P_{\tau }^{{TL}}$ indicates the power transferred during period $\tau$. $P_{{\hbox{min} }}^{{TL}}$ and $P_{{\hbox{max} }}^{{TL}}$ represent the lower and upper limits of the transferred power, respectively.

The compensation cost ${F_{TL}}$, which needs to compensate the user after the load transfer, is

$${F_{TL}}=F_{C}^{{TL}}\sum\limits_{{\tau =1}}^{T} {P_{\tau }^{{TL}}}$$

(5)

.

Where, $F_{C}^{{TL}}$ is the compensation cost per unit power load transfer.

Cutable load

Cutable load is the ability to reduce the electrical load by controlling, regulating, or temporarily stopping certain electrical equipment or processes when it is necessary to reduce electricity consumption. This load can usually be flexibly adjusted or temporarily interrupted to adjust to the needs of the power system and thus manage the power supply and demand balance more effectively. In order to ensure the reasonable load cut, it is necessary to restrict the upper and lower limits of the cut time and the number of cut times.

$$T_{{\hbox{min} }}^{{CL}} \leqslant \sum\limits_{{\tau =t}}^{{t+{T_{cp}} - 1}} {\Upsilon _{\tau }^{{CL}}} \leqslant T_{{\hbox{max} }}^{{CL}}$$

(6)

$$\sum\limits_{{\tau =1}}^{T} {\Upsilon _{\tau }^{{CL}}} \leqslant {N_{\hbox{max} }}$$

(7)

.

Where, ${T_{cp}}$ is the cut time. $T_{{\hbox{max} }}^{{CL}}$ and $T_{{\hbox{min} }}^{{CL}}$ are the upper and lower limits of the cut time respectively. The cut state of variable $\Upsilon$ at period $\tau$ is marked with 0–1. If $\Upsilon _{\tau }^{{CL}}=1$, the load is cut during period $\tau$. Conversely, $\Upsilon _{\tau }^{{CL}}=0$ indicates that it is in the non-cut state. ${N_{\hbox{max} }}$ is the maximum number of cuts.

The compensation cost ${F_{CL}}$ that needs to compensate the user after the load cut is

$${F_{CL}}=F_{C}^{{CL}}\sum\limits_{{\tau =1}}^{T} {P_{\tau }^{{CL}}}$$

(8)

.

Where, $F_{C}^{{CL}}$ is the compensation cost of unit power load cut. $P_{\tau }^{{CL}}$ is the load cut power of period $\tau$.

Low-carbon economic dispatching model of CHPMS

In the actual operation of CHPMS, it is necessary to consider the cooperation of various generator sets. According to the output forecast of CHPMS’s electricity, heat, wind power units and PV power units, the system unit constraints are set, and based on considering the lowest economic cost and environmental cost of CHPMS, the energy storage characteristics are used to achieve the optimal setting of resources, reasonable dispatching of unit output and load, and promote the orderly coordination of comprehensive energy. In the process of low-carbon economic dispatch, CHPMS gradually improves its own value addition, and realizes value co-creation in the process of seeking the lowest total cost.

Objective function

In the operation of CHPMS, in order to achieve the optimal economic dispatch of resources, the objective functions are set to minimize the total operating costs of the system and the carbon trading costs. The battery energy storage system, crucial for peak shaving and valley filling, is deemed essential to consider its operating cost for optimization and coordination of generation.

Objective function of total operating cost

$${C_O}=\sum\limits_{{t=1}}^{T} {{C_{DG}}(t)+{C_{PG}}(t)+{C_{GB}}(t)+{C_{HS}}(t)+{C_{BAT}}(t)+{C_{MT}}(t)+{C_{FL}}}$$

(9)

$${C_{DG}}(t)={\partial _{WP}}{P_{WP}}(t)+{\partial _{PV}}{P_{PV}}(t)$$

(10)

$${C_{PG}}(t)={\kappa _{TOU}}{P_{PG}}(t)$$

(11)

$${C_{GB}}(t)={\partial _{GB}}{P_{GB}}(t)$$

(12)

$${C_{HS}}(t)={\lambda _{HS}}\left| {{P_{HS}}(t)} \right|$$

(13)

$${C_{BAT}}(t)={\lambda _{BAT}}\left| {{P_{BAT}}(t)} \right|$$

(14)

$${C_{MT}}(t)=\frac{{{U_{NG}}}}{{{Q_{LHV}}{\mu _{MT}}}}{P_{MT}}(t)$$

(15)

$${C_{FL}}={C_{SL}}+{C_{TL}}+{C_{CL}}$$

(16)

.

In the equations, ${C_O}$represents the total operating cost function of the CHPMS, and ${C_{DG}}$ represents the operating cost of distributed generators.${C_{PG}}$ represents the cost of purchasing electricity from the grid.${C_{GB}}$ represents the operating cost of the gas boiler.${C_{HS}}$and${C_{BAT}}$ represent the depreciation cost of heat storage and battery energy storage, respectively.${C_{MT}}(t)$ represents the fuel cost of the micro gas turbine.${C_{FL}}$ refers to the compensation cost for optimizing flexible loads.${\partial _{WP}}$and${\partial _{PV}}$ represent the operating cost coefficients for the wind power and PV units, respectively.${P_{WP}}(t)$and${P_{PV}}(t)$ represent the output power of the wind power and PV units at time t.${\kappa _{TOU}}$ represents the time-of-use electricity prices.${P_{PG}}(t)$ is the electrical power at time t. ${\partial _{GB}}$represents the operating cost coefficient of the gas boiler, while ${P_{GB}}(t)$represents the output power of the gas boiler at time t.${\lambda _{HS}}$and${\lambda _{BAT}}$represent the depreciation coefficients of the heat storage system and battery energy storage system.${P_{HS}}(t)$and${P_{BAT}}(t)$ represent the charging and discharging power of the heat storage system and the battery energy storage system at time t. When absorbing heat and charging, ${P_{HS}}(t),{P_{BAT}}(t)>0$.When releasing heat and discharging, ${P_{HS}}(t),{P_{BAT}}(t)<0$。${U_{NG}}$and${Q_{LHV}}$ represent the unit price and lower heating value of natural gas.${\mu _{MT}}$ refers to the power generation efficiency of the micro gas turbine.${P_{MT}}(t)$ represents the output power of the micro gas turbine.${C_{SL}}$,${C_{TL}}$and${C_{CL}}$ respectively represent the total compensation costs for the three types of flexible loads.

Objective function of carbon transaction cost

At present, many provinces and cities in China have established relevant carbon trading pilot projects and adopted different quota allocation methods³⁷. In order to alleviate the serious air pollution problem, it is very important to establish a carbon emission trading mechanism as an innovative policy tool to promote energy conservation and emission reduction³⁸. Carbon trading mechanisms can promote the optimization of integrated energy structures and improve energy efficiency by trading carbon emission rights, thereby facilitating the reduction of carbon emissions. In China, carbon quotas are often allocated to carbon emitters by government agencies. If the emissions from carbon sources are within the given quota, the excess carbon emissions can be traded in the carbon trading market. On the contrary, if the emissions from carbon sources exceed the allocated quota, the excess emissions must be purchased from the carbon trading market. In the carbon trading market, each carbon emission quota achieves maximum economic benefit through a variety of energy-saving and emission reduction measures. The formula for calculating carbon trading costs under the carbon trading mechanism is as follows:

$${C_{C{O_2}}}={C_{MP}}({E_{TCE}} - {E_{CEQ}})$$

(17)

.

In the equation, ${C_{C{O_2}}}$ represents the carbon trading cost. If carbon emissions exceed the quota (${C_{C{O_2}}}>0$), the carbon emitter needs to purchase carbon emissions. Conversely, if ${C_{C{O_2}}}<0$, it indicates that the carbon emitter generates revenue by selling carbon emissions. ${C_{MP}}$ represents the carbon trading price, while ${E_{TCE}}$ and ${E_{CEQ}}$ respectively denote the total emissions and the carbon emission quota.

The calculation of carbon emissions in this study considers the utilization of multiple energy sources. Therefore, the carbon emissions of the CHPMS are calculated using the following formula.

$${E_{TCE}}=\sum\limits_{{i=\phi }} {(\eta _{i}^{{P\& T}}+\eta _{i}^{{USE}}){P_i}}$$

(18)

.

In the formula, $\phi$ represents the collection of unit equipment, including heating and power supply equipment and energy storage equipment. $\eta _{i}^{{P\& T}}$ represents the carbon emission coefficient of the energy production and transportation process corresponding to the unit ‘i’ equipment. $\eta _{i}^{{USE}}$ represents the carbon emission coefficient of the energy utilization process corresponding to the unit ‘i’ equipment. ${P_i}$ represents the output power of the unit ‘i’ equipment.

Constraint condition

Constraints on electrical power balance

The sum of the output power of the photovoltaic generator, wind turbine generator, grid-connected power, and micro gas turbine output power must be balanced with the sum of the battery energy storage system output power and the electric load. The calculation equation is as follows:

$$\left\{ {\begin{array}{*{20}{c}} {{P_{WP}}+{P_{PV}}+{P_{PG}}+{P_{MT}}={L_E}+{P_{BAT}}} \\ {{L_E}={L_B}+{L_S}+{L_T}+{L_C}} \end{array}} \right.$$

(19)

.

Where:

${P_{MT}}$ represents the output power of the micro gas turbine.

${L_E}$ represents the total electric load.

${L_B}$, ${L_S}$, ${L_T}$, and ${L_C}$ respectively represent the base load, shiftable electric load, transferable electric load, and cutable electric load in the microgrid system.

(2) Constraints on electric power upper and lower limits

$$\begin{gathered} 0 \leqslant {P_{WP}} \leqslant {P_{WP,MAX}} \hfill \\ 0 \leqslant {P_{PV}} \leqslant {P_{PV,MAX}} \hfill \\ {P_{PG,MIN}} \leqslant {P_{PG}} \leqslant {P_{PG,MAX}} \hfill \\ 0 \leqslant {P_{MT}} \leqslant {P_{MT,MAX}} \hfill \\ {P_{BAT,MIN}} \leqslant {P_{BAT}} \leqslant {P_{BAT,MAX}} \hfill \\ \end{gathered}$$

(20)

.

Each generator must satisfy certain upper and lower limits. Where ${P_{WP,MAX}}$ and ${P_{PV,MAX}}$ represent the forecasted output of the wind and PV generators, respectively. ${P_{PG,MIN}}$ and ${P_{PG,MAX}}$ represent the minimum and maximum values of grid power, respectively. ${P_{MT,MAX}}$ represents the rated electrical power of the micro gas turbine. ${P_{BAT,MIN}}$ and ${P_{BAT,MAX}}$ represent the maximum charging and discharging power of the battery energy storage, respectively.

Constraints on the battery energy storage system

The battery energy storage system should have constraints on the state of charge to avoid excessive charging and discharging, which would reduce its lifespan. The expression is as follows:

$$SO{C_{MIN}} \leqslant SOC(t) \leqslant SO{C_{MAX}}$$

(21)

$$SO{C_0}=SO{C_E}$$

(22)

.

In the equation, $SO{C_{MAX}}$ and $SO{C_{MIN}}$ respectively represent the upper and lower limits of the state of charge of the battery energy storage system at time ‘t’. Additionally, during the participation of the battery energy storage system in the operation of the microgrid system, the state of charge should also satisfy the constraint of equal dispatching at the beginning and end. $SO{C_0}$ and $SO{C_E}$ respectively represent the initial and final state of charge of the battery energy storage system.

$$\left\{ {\begin{array}{*{20}{c}} {Ch{a_t}\cdot Disch{a_t}=0,Ch{a_t},Disch{a_t} \in \left\{ {0,1} \right\}} \\ {\sum\limits_{{t=1}}^{T} {\left| {Ch{a_{t+1}} - Ch{a_t}} \right| \leqslant {N_{Cha}}} } \\ {\sum\limits_{{t=1}}^{T} {\left| {Disch{a_{t+1}} - Disch{a_t}} \right| \leqslant {N_{Discha}}} } \end{array}} \right.$$

(23)

.

In addition to the state of charge constraint, there should also be constraints on the charging and discharging states of the battery energy storage system. In the equation, $Ch{a_t}$ represents the battery energy storage system in the charging state, while $Disch{a_t}$ represents it in the discharging state. ${N_{Cha}}$ and ${N_{Discha}}$ respectively denote the limits on the number of charging and discharging cycles of the battery energy storage system.

Constraints on heat power balance

The sum of the heat power from the heat recovery system and the gas boiler should be balanced with the total heat load and the power from the heat storage tank. The calculation formula is as follows:

$$\begin{gathered} {Q_{HRS}}+{Q_{GB}}={Q_L}+{Q_{HS}} \hfill \\ {Q_L}={Q_B}+{Q_S}+{Q_C} \hfill \\ \end{gathered}$$

(24)

.

In the formula, ${Q_{HRS}}$ and ${Q_{GB}}$ respectively represent the heat power from the heat recovery system and the micro gas turbine. ${Q_L}$ represents the total heat load. ${Q_{HS}}$ represents the heat power from the gas boiler. ${Q_B}$, ${Q_S}$, and ${Q_C}$ respectively represent the base heat load, the shiftable heat load, and the cutable heat load.

Constraints on the upper and lower limits of heat power

$$\begin{gathered} 0 \leqslant {Q_{HRS}} \leqslant {Q_{HRS,MAX}} \hfill \\ 0 \leqslant {Q_{GB}} \leqslant {Q_{GB,MAX}} \hfill \\ {Q_{HS,MIN}} \leqslant {Q_{HS}} \leqslant {Q_{HS,MAX}} \hfill \\ \end{gathered}$$

(25)

.

There are also upper and lower limits constraints on heat power. In the equation, ${Q_{HRS,MAX}}$ and ${Q_{GB,MAX}}$ represent the rated heat power of the heat recovery system and the micro gas turbine, respectively. ${Q_{HS,MIN}}$ and ${Q_{HS,MAX}}$ represent the maximum power for heat release and absorption of the heat storage tank.

Constraints on heat storage tank

$$End{o_t}\cdot Exo{t_t}=0$$

(26)

$${W_0}={W_E}$$

(27)

.

The heat storage tank should be constrained by the state of endothermic and exothermic. Here, $End{o_t}$ represents the endothermic state of the heat storage tank, while $Exo{t_t}$ denotes the exothermic state. Additionally, during the participation of the heat storage tank in the operation of the microgrid system, the heat state should also satisfy the constraint of equal dispatching at the beginning and end. ${W_0}$ and ${W_E}$ respectively represent the initial and final heat of the heat storage tank.

Constraints on distribution network

To ensure physical feasibility and compliance with grid interconnection standards, the following distribution network constraints are incorporated:

• (1) Dynamic power exchange limits:

  $$- {P_{net,\hbox{max} }}(t) \leqslant {P_{net}}(t) \leqslant {P_{net,\hbox{max} }}(t),\forall t \in T$$

  (28)

  .

  where ${P_{net,\hbox{max} }}(t)$denotes the time-dependent maximum allowable power exchange, reflecting grid stability requirements during peak/off-peak periods.

• (2) Ramp rate constraints:

  $$- \Delta {P_{\hbox{max} }} \leqslant {P_{net}}(t) - {P_{net}}(t - 1) \leqslant \Delta {P_{\hbox{max} }},\forall t \geqslant 2$$

  (29)

  .

  limiting abrupt power fluctuations to ± 50 kW/h, as mandated by IEEE 1547 for distributed energy integration.

  Aiming at the above model, this paper uses Yalmip toolbox and Cplex solver to solve the example.

Assumptions

To simplify the dynamic complexity of integrated energy systems and reduce computational burden, a series of key assumptions are adopted in constructing the low-carbon resource scheduling optimization model for CHPMS. These assumptions enable the model to focus on core scheduling dynamics while maintaining computational tractability. Detailed explanations of each assumption are provided as follows:

• (1) The electrical and heat loads are assumed to be known and constant throughout the scheduling horizon. Variations caused by stochastic factors such as weather fluctuations, user behavior, or seasonal patterns are disregarded. By treating load demands as deterministic variables, the model avoids the computational complexity associated with stochastic demand modeling, thereby streamlining the optimization process. However, this assumption ignores the inherent variability and unpredictability of actual energy consumption, which may lead to a lower level of model adaptability.

• (2) The power outputs of wind and PV systems are assumed to follow deterministic forecast values. In this model, renewable generation is treated as pre-specified inputs, with scheduling decisions based on predicted generation levels. This assumption eliminates the need to model the uncertainty inherent in renewable energy sources, simplifying the optimization framework.

• (3) The carbon trading system is modeled with a fixed emission allowance and constant carbon credit price. In this model, the total allowable emissions are predetermined, and any excess emissions require purchasing carbon credits at a fixed price. But it may overlook changes in the complexity of carbon markets, such as fluctuations in economic conditions and carbon prices.

• (4) The charging and discharging efficiencies of energy storage systems are assumed to be constant values, ignoring the reality that efficiency decreases with time and usage conditions. This assumption simplifies the calculation of the energy balance, allowing the performance of the energy storage system to be clearly predicted and modeled.

• (5) The operational behaviors of micro gas turbines and gas boilers are approximated as linear functions, ignoring the nonlinear behavior of these devices. Specifically, input-output relationships (e.g., fuel consumption versus power/heat generation) are assumed to follow linear dependencies. This linearization reduces the complexity of nonlinear constraints, thereby enhancing the solvability of the optimization problem.

Collectively, these assumptions establish a structured analytical framework for the CHPMS scheduling model, ensuring computational feasibility by abstracting system uncertainties and focusing on core operational dynamics. By providing a detailed delineation of each assumption, the model’s limitations in specific aspects can be clearly understood, and how these limitations contribute to supporting the realization of optimization objectives. These assumptions serve as a critical foundation for analyzing and interpreting the model’s outcomes.

Case study

Data collection framework and basic parameter setting

The parameterization process for the CHPMS model involved three main phases: data collection, statistical estimation, and field validation. This article focuses on microgrid systems, which include units such as PV generators, wind power, micro gas turbines, gas boilers, heat storage tanks, and battery energy st43eorage systems. Factors considered include carbon trading, time-of-use electricity price, and flexible loads. To ensure transparency and reproducibility, all input data are derived from authoritative sources with rigorous validation. The methodological approach for data collection and parameter estimation in this study was developed through a systematic integration of empirical data, technical specifications, and regulatory frameworks, specifically tailored to China’s energy transition context. Electricity prices were obtained from publicly accessible annual reports published by the Shanxi Provincial Energy Administration, covering industrial time-of-use tariffs. Load profiles and hourly energy generation were sourced from anonymized operational records of industrial parks in the target region, preprocessed to remove commercial identifiers while preserving temporal resolution (Appendix A, Table A1). Hourly weather parameters, including temperature, solar irradiance, and wind speed, were obtained from validated regional meteorological sources and reanalysis datasets of North China Meteorological Consortium dataset (Appendix B, Table B1). Carbon emission factors followed China’s provincial-level lifecycle assessment guidelines, with grid emission intensity calculated from Shanxi’s generation mix (64% coal, 22% gas, 14% renewables). Carbon quotas were aligned with the National Development and Reform Commission’s (NDRC) allocation methodology, incorporating sector-specific benchmarks adjusted for microgrid operational characteristics. Parameter estimation employed Monte Carlo simulations (1,000 iterations) for uncertainty modeling and genetic algorithm optimization (200 generations) for efficiency calibration. Data quality was ensured through isolation forest anomaly detection and multiple imputations for missing values. The dispatching period is set to 24 h, with dispatching intervals of 1 h. Operational parameters of the microgrid system are presented in Table 1. Daily load curves and wind-solar output curves are illustrated in Fig. 3. Time-of-use electricity pricing is depicted in Fig. 4 as well.

Table 1 Operational parameters of the CHPMS.

Fig. 3

Daily load and output curves of wind power and PV.

Fig. 4

The Time-Of-Use Electricity Price.

The study adopts time-of-use price and fixed natural gas price. The electricity price is divided into 6 time periods for analysis. The peak, normal, and valley periods of electricity price are distributed as follows. The peak periods are from 11:00 to 15:00 and from 19:00 to 21:00, with purchase and selling prices of 1.32¥/kWh and 0.82¥/kWh, respectively. The normal periods are from 8:00 to 10:00, 16:00 to 18:00, and 22:00 to 24:00, with purchase and selling prices of 0.83¥/kWh and 0.53¥/kWh, respectively. The valley period is from 0:00 to 7:00, with purchase and selling prices of 0.37¥/kWh and 0.25¥/kWh, respectively. The fixed price of natural gas is 2.5¥/m³.

According to the carbon emission quantification method proposed in this study, the carbon emission coefficients and carbon quotas are set as shown in Table 2.

Table 2 Carbon emission coefficient and carbon quota.

Key parameters were determined through ​system demand and ​empirical calibration. This paper set the PV conversion efficiency to 0.18 ± 0.02, and validated against field data from specific solar plants. Also the battery degradation rate was calibrated to 0.05%/cycle ± 0.005% based on battery manufacturer specifications and cycling tests.

This paper sets the electrical load to include the base electrical load, shiftable electrical load 1 and shiftable electrical load 2, transferable electrical load, and cutable electrical load. The heat load includes the base heat load, shiftable heat load, and cutable heat load. The pre-optimized power distributions of flexible electric load and flexible heat load are illustrated in Figs. 5 and 6, respectively. The parameters of flexible loads are shown in Table 3.

Fig. 5

Pre-optimized distribution of flexible electric loads.

Fig. 6

Pre-optimized distribution of flexible heat loads.

Table 3 Parameters of flexible load.

Analysis of the results of low-carbon economic dispatching

In the case of considering flexible loads, the CHPMS is optimized for dispatching with low carbon economy. The electrical load balance and heat load balance of each energy output are obtained, as shown in Fig. 6. Before and after optimization of electricity load curve and heat load curve as shown in Fig. 7.

In Fig. 7(a), for the electricity load, during the 0:00–7:00 valley period with low demand and a price of 0.25¥/kWh, PV output is absent, while the wind turbine’s operating cost (0.52¥/kWh) exceeds the grid price, leading to its shutdown. The power grid supplies primary power, supplemented by the micro gas turbine, which provides electricity and heat, with excess electricity charging the BES. Throughout the dispatch cycle, the micro gas turbine reduces carbon emissions and alleviates grid pressure during low renewable output and high consumption peaks. In contrast, during 10:00–15:00 and 18:00–21:00, high renewable output (wind/PV) aligns with peak electricity prices, enabling full utilization of distributed energy. The BES discharges during these peak periods and charges during lower-price valley periods, optimizing economic and operational efficiency, while the micro gas turbine’s role diminishes as renewables dominate supply. During the transition periods (7:00–10:00, 15:00–18:00), gradual micro gas turbine ramping balances renewable intermittency, maintaining grid imports < 40 kW. Crucially, the 1-hour BES discharge delay at 18:00 (vs. 17:00 price surge) reflects distribution network ramp-rate constraints, validating the model’s grid-compliant optimization of economic and low-carbon objectives.

Fig. 7

(a) The electrical load balance of each energy output. (b) The heat load balance of each energy output.

The heat load balancing mechanism demonstrates synergistic operation of three heat sources.

For the heat load, the heat recovery system provides most of the heat supply, supplemented by the gas boiler and heat storage tank. Gas boilers play two key roles in heat load balancing. The first is to reduce the output of the gas turbine, and the second is to promote the CHPMS to purchase electricity from the grid, thereby reducing the operation cost of power dispatch. The heat recovery system makes up for power shortages during peak load times. The heat storage tank releases stored heat during peak periods and absorbs excess heat via the gas boiler during low-demand times, effectively implementing peak cutting and valley filling to optimize overall system efficiency through balanced thermal energy distribution.

Figure 8 demonstrates coordinated demand response optimization across electrical and heat systems, with the pre-optimization electric load peaking at 360 kW (20:00) and heat load reaching 230 kW (20:00), exhibiting high volatility. Post-optimization strategies achieve dual-domain peak shaving, reducing electrical and heat peaks by 19.4% (360→290 kW) and 19.6% (230→185 kW) through 60 kW load shifting to valley periods (02:00–06:00) and heat storage dispatch. Simultaneously, the heat load curve shows enhanced stability supported by strategic heat storage charging and discharge. This integrated approach preserves critical industrial baseloads while achieving 19.3% cost reduction and 14.5% gas consumption decrease, validating synergistic electrical-thermal optimization in balancing grid reliability, economic efficiency, and carbon mitigation.

Fig. 8

(a) Electric load curve before and after demand response. (b) Heat load curve before and after demand response.

Fig. 9

The optimized flexible electric load distribution.

By comparing Figs. 5 and 9, it can be seen that the shiftable electric load 1 shifts from 11:00 ~ 13:00 to 4:00 ~ 6:00, and the shiftable electric load 2 shifts from 19:00 ~ 22:00 to 6:00 ~ 9:00. It can be seen that the shiftable electric load is shifted from the peak period of electricity consumption to the lower period of electricity consumption, which is conducive to the consumption of new energy and ensures the economy of dispatch. The transferable electric load is split from the original time period of 12:00 ~ 16:00 to the time period of 3:00 ~ 8:00. In the process of shifting, the duration of the load and the electric power are equal to that before shifting. However, the duration and electric power of the transferable load are changed. The cutable load has different degrees of curtailment in different periods, and the curtailment period is mostly the peak period of electricity consumption.

Fig. 10

The optimized flexible heat load distribution.

By comparing Figs. 6 and 10, the shiftable heat load shifts from 17:00 ~ 20:00 to 7:00 ~ 10:00. It can be seen that the shiftable heat load shifts from the evening peak to the normal period. The cutable heat load can be reduced more during 8:00 ~ 10:00 and 19:00 ~ 22:00. The reduction of heat load also further alleviates the heating pressure at peak times.

Discussion

In order to study the impact of carbon trading on CHPMS scheduling operation and the optimization role of flexible load on low-carbon resource scheduling, five scenarios are set up for comparative analysis.

Scenario 1: Optimal scheduling without considering flexible load and carbon trading.

Scenario 2: Optimal scheduling without considering flexible load.

Scenario 3: Only flexible electric load is considered, and the optimal scheduling of flexible heat load is not considered.

Scenario 4: The optimal scheduling of carbon trading is not considered.

Scenario 5: Comprehensive consideration of CHPMS operating costs and carbon trading costs, including flexible load participation in optimal scheduling.

Scheduling results in five scenarios

In order to compare the economic and environmental benefits under the five scenarios, the low-carbon dispatch results of CHPMS were analyzed respectively from four aspects: operation cost, carbon trading cost, total cost and carbon emission. The specific results are shown in Table 4.

Table 4 The optimal specific results.

Since Scenario 1 and Scenario 4 do not consider carbon trading factors, their carbon trading costs are not considered. As can be seen from Table 4, Scenario 1 and Scenario 4 have the highest carbon emissions. A positive carbon trading cost in Scenario 2 indicates that carbon credits need to be purchased. The carbon emissions of Scenario 2 are significantly lower than those of Scenario 1 and Scenario 4, but the total cost is also increased. Compared with Scenario 2, the carbon trading cost of Scenario 3 is reduced by 39.46%, and the carbon emission is further reduced. Under the comprehensive consideration of flexible load and carbon transaction cost, Scenario 5 has the lowest carbon trading cost and total cost in the scenarios, and the result of carbon emission reduction is also optimistic.

In conclusion, carbon trading can effectively reduce the carbon emissions of CHPMS. The addition of flexible load not only improves the environmental benefit but also reduces the dispatching cost to a certain extent, making the economic and environmental benefit of CHPMS better.

Impact of carbon trading on CHPMS scheduling

In order to verify the impact of carbon trading on CHPMS, Scenario 4 and Scenario 5 are analyzed for low-carbon scheduling, and the output of each energy source before and after the introduction of carbon trading is compared.

Fig. 11

(a) Electric load balancing of Scenario 4. (b) Electric load balancing of Scenario 5.

Compared with Fig. 11, the phenomenon of PV abandonment and wind abandonment in Scenario 4 is more obvious, and the power grid and energy storage jointly make up for the insufficient scene-view output. Scenario 5 increases the consumption of new energy, making the electricity purchased by the grid lower than that in Scenario 4. Battery energy storage system has a certain role in peak cutting and valley filling in both scenarios, but Scenario 5 is more inclined to use new energy to charge the battery energy storage system. It can be seen that carbon trading enables CHPMS to develop in a low-carbon direction and plays an optimization role in the power scheduling of CHPMS.

Impact of flexible load on CHPMS scheduling

To verify the impact of flexible load on CHPMS, Scenario 2, Scenario 3 and Scenario 5 are analyzed for low-carbon scheduling. As can be seen from Table 4, Scenario 2 has the highest total cost of CHPMS without flexible load participating in scheduling. Scenario 3 has the participation of flexible electrical load, and the cost is reduced by 9.66%. Scenario 5 adds a flexible heat load based on Scenario 3 to further reduce the cost.

Fig. 12

Power changes of the power and heat loads in different scenarios.

The changes of electrical and thermal loads in the three scenarios are shown in Fig. 12. Compared with Fig. 12, we can see that Scenario 2 has no flexible load participating in scheduling, while Scenario 3 and Scenario 5 both have flexible electrical load participating in scheduling. Therefore, the power consumption in the peak hours of Scenario 3 and Scenario 5 is reduced, and the peak-valley difference of the power load is also reduced from 230 kW to 155 kW. Since flexible electric loads are involved in scheduling in both Scenario 3 and Scenario 5, the electric loads in Scenario 3 and Scenario 5 appear to coincide. In Scenario 5, the participation of flexible heat load causes the load curve to be clipped and valley filled, and the peak-valley difference of heat load is also reduced from 130 kW to 70.11 kW. Based on the above analysis, the addition of energy storage and flexible load can reduce the total cost and load peak-valley difference of CHPMS to a certain extent.

Fig. 13

Power grid output in Scenario 2 and Scenario 5.

In addition, for electric power scheduling, the main factor that causes the difference in carbon emissions in different scenarios is the difference in power grid output. Therefore, in order to verify the influence of flexible load participation in scheduling, the power grid output of Scenario 2 and Scenario 5 is analyzed, as shown in Fig. 13. Only in a few periods, the power grid output of Scenario 5 will be higher than that of Scenario 2 due to the peak-filling effect of shiftable load and transferable load. Moreover, the addition of cutable load further reduces the power grid output in Scenario 5. The reduction of electricity purchased from the grid makes the utilization of clean energy more fully and rationally, which is also conducive to the reduction of CHPMS carbon emissions and the reduction of carbon trading costs. It can be seen that reasonable scheduling of flexible load is conducive to optimizing the comprehensive benefit of CHPMS.

The proposed optimization model for low-carbon resource dispatching in CHPMS demonstrates significant potential in reducing operational costs and carbon emissions, as validated through the case study. The integration of carbon trading mechanisms and flexible load scheduling enhances the system’s economic and environmental performance, providing a viable framework for sustainable energy management. While the proposed model offers valuable insights into low-carbon resource dispatching, its limitations, stemming from the underlying assumptions, must be acknowledged to contextualize its applicability. Section 4.3 has introduced the assumptions and limitations of model construction in detail. These limitations highlight the need for future research to incorporate stochastic or robust optimization techniques to address uncertainties, develop dynamic carbon trading models to reflect market realities, and integrate more realistic equipment and storage performance models to enhance the model’s alignment with practical scenarios.

Energy dispatching collaborative decision suggestions

In this paper, the cooperative decision model of low-carbon economic scheduling based on CHPMS is constructed, and the operational strategy of promoting system economic scheduling is studied. Based on the analysis of decision results, in order to promote the sustainable development of CHPMS and enhance the practical application value of the system, the following collaborative decision making strategies are proposed.

• (1) In actual operation, CHPMS needs to rely on a variety of resources in the energy environment, and resource clustering of various parties can make CHPMS obtain higher resource utilization efficiency and value. The interaction and feedback between other resources and carbon emission reduction should be fully considered when CHPMS performs economic scheduling. The collaborative decision model proposed in this paper can provide a theoretical basis for the construction of CHPMS project. In other words, in the construction of CHPMS, in addition to the most important clean energy supply system, other available energy systems should be reasonably added on the basis of meeting the energy demand of users, and the joint energy supply behavior of multiple energy systems should be scientifically scheduled, and the optimal economic and environmental benefits should be considered to maximize the allocation of resources. The government should introduce a series of policies to encourage strategic cooperation between energy supply enterprises, such as for CHPMS projects that meet carbon emission requirements, give certain forms of subsidies according to the proportion of clean energy and energy storage utilization level, or other incentive policies.

• (2) From the analysis of decision results, it can be seen that when multiple energy suppliers in the system operate cooperatively, the effective coordination of electricity and heat can be realized, and the effect of peak shifting and valley filling of flexible load can be reflected. Considering carbon trading can also reflect China’s carbon neutrality goal and energy transition needs, to achieve the maximum economic and environmental benefits. Therefore, the government should formulate corresponding incentive and restriction policies according to the possible environmental and social impacts of each link in the CHPMS project to promote the sustainable development of the CHPMS project.

• (3) In the process of realizing value co-creation in CHPMS, the respective operation strategies of the energy supply system will have a great impact on the value co-creation effect. Considering the flexible load participation and carbon trading cost, the economic dispatching of energy supply system can obtain more benefits. The collaborative decision model constructed in this paper can determine a better low-carbon economic scheduling strategy. In the actual operation of CHPMS, the more accurate operation strategies should be formulated considering the utilization of all resources.

Conclusion

China’s energy transformation has exposed such problems as low energy efficiency, high carbon emissions, and obvious abandonment of wind and PV. The economic dispatch optimization management of microgrid system is a solution to absorb abandoned energy and improve energy utilization, and it is also an effective way to realize the value co-creation of the system. Reasonable low-carbon economic dispatching can achieve the optimal economic and environmental benefits of CHPMS, and effective management of energy transfer between energy subjects in CHPMS is another way to achieve system value co-creation. Based on this, this paper establishes CHPMS, which includes two forms of energy, electricity and heat. Considering flexible load participation and carbon transaction cost, a low-carbon economic scheduling collaborative decision model of CHPMS is constructed. Through reasonable scheduling of various energy sources in the system, the synergistic value is added and the value co-creation effect is maximized. Firstly, the value co-creation analysis framework of CHPMS economic scheduling is analyzed. Secondly, a multi-form and multi-characteristic load model is constructed to make full use of the electric and heat flexible loads on the user side, and a low-carbon economic dispatch collaborative decision-making model considering flexible loads and carbon trading costs is established. Then, through example analysis and scenario analysis, the improvement effect of carbon trading on CHPMS structure is verified. It is also found that the cooperative coupling scheduling of CHPMS has more advantages than the independent scheduling of electricity and heat energy. The analysis of scheduling results shows that both the shiftable load and the transferable load have a certain peak shifting effect, and the cutable load can achieve peak cutting. The participation of flexible loads in different scenarios affects the operating cost of CHPMS, and realizes the optimization of the economic and environmental benefits of the system. Finally, the paper puts forward some suggestions for the results of low-carbon economic dispatch, aiming at providing decision support for some related enterprises.

Although this paper provides some ideas for the study of economic scheduling strategies of microgrid systems, the research is mainly applicable to the typical areas of community nature, and there are some limitations. While the proposed model ensures compliance with distribution network constraints (Constraint condition), two key limitations warrant future investigation. First, the deterministic formulation assumes perfect forecasts of renewable generation and loads, whereas real-world uncertainties (e.g., solar irradiance variability) may necessitate stochastic or robust optimization frameworks. Second, the fixed ramp rate threshold could be dynamically adjusted based on real-time grid conditions (e.g., state-of-charge of adjacent energy storage systems) to enhance operational flexibility. Nevertheless, the current constraints—validated under worst-case peak/off-peak scenarios provide a foundational framework for such extensions, balancing simplicity with practical feasibility for grid operators. In the following research, the carbon trading mechanism and the flexible load model will be further analyzed, such as considering more energy forms and broadening the application scope.