Optimal scheduling method for integrated energy system of electric heating and hydrogen based on dynamic pricing of electricity-carbon coupling

Abstract

The application discloses an electricity-heat-hydrogen comprehensive energy system optimization scheduling method based on electricity-carbon coupling dynamic pricing, which comprises the following steps: S1, a deep learning model is used to predict an original carbon transaction price sequence, and a clustering algorithm is combined to generate typical wind-solar output scenes; S2, a physical equipment operation model of the electricity-heat-hydrogen comprehensive energy system is constructed, and the power balance constraints of the wind turbine, the photovoltaic device, the electrolytic cell, the fuel cell and the energy storage device are defined; S3, an electricity-carbon coupling dynamic pricing mechanism is constructed, and based on the real-time net load rate and the predicted carbon price of the system, the comprehensive electricity price containing the carbon emission penalty cost is dynamically calculated; S4, an intelligent optimization algorithm is adopted in combination with the dynamic comprehensive electricity price, and the minimum total operation cost of the system is taken as the target. Through the introduction of the dynamic pricing mechanism coupling the price and the physical equipment and the algorithm closed loop nested architecture, the hydrogen energy equipment actively responds to the carbon market fluctuation, and the low-carbon economic operation level of the comprehensive energy system is improved.

Description

Technical Field

[0001] This invention belongs to the field of integrated energy systems and electricity market technology, specifically involving an optimized scheduling method for an integrated energy system of electricity, heat, and hydrogen with dynamic pricing coupled with electricity and carbon. Background Technology

[0002] Under the "dual carbon" target, renewable energy sources, represented by wind power and photovoltaics, are being connected to the grid on a large scale. Integrated Energy Systems (IES), by coordinating multiple energy forms such as electricity, heat, and hydrogen, have become an important way to promote the consumption of new energy. Among them, the electro-thermal-hydrogen integrated energy system utilizes the energy conversion characteristics of electrolyzers and fuel cells, possessing extremely high regulation flexibility. Existing low-carbon economic dispatch methods for integrated energy systems mainly have the following shortcomings: 1. There is a disconnect between the carbon market and power dispatch mechanisms: Traditional energy management strategies often use fixed carbon prices or static carbon emission factors, which cannot truly reflect the spatiotemporal dynamic changes in the carbon emission intensity of the external power grid, resulting in insufficient robustness of the system in dealing with the risk of high-frequency fluctuations in carbon market prices in the future. 2. The real-time dynamic low-carbon value of hydrogen energy equipment has been overlooked: the existing pricing and cost accounting mechanisms have failed to fully tap the carbon reduction potential of "hydrogen energy" as a cross-source medium. The start-up, shutdown and power status of electrolyzers and fuel cells are difficult to directly map and feed back to the real-time energy purchase cost of the system. 3. Lack of deep closed-loop optimization framework: Existing studies mostly adopt a one-way open-loop mechanism, that is, a fixed electricity price is determined in advance and then scheduling optimization is carried out. This fails to achieve the two-way game and deep coupling of "scheduling decisions change the cleanliness of the system and thus change the electricity price, and dynamic electricity price feeds back to guide equipment scheduling".

[0003] Therefore, there is an urgent need for a new dynamic pricing and scheduling optimization method that can closely integrate the predicted carbon price with the physical operating status of the underlying hydrogen-related equipment in order to improve the low-carbon level and economic benefits of the electrothermal hydrogen system. Summary of the Invention

[0004] The purpose of this invention is to provide an optimized scheduling method for an integrated energy system of electrothermal hydrogen based on dynamic pricing coupled with electricity and carbon. By introducing a dynamic pricing mechanism that couples price and physical properties, and a closed-loop scheduling framework nested within the algorithm, the electrothermal hydrogen system can proactively defend against and optimize carbon market fluctuations, thereby improving the economic efficiency and low-carbon operation of the integrated energy system.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an optimized scheduling method for an integrated electrothermal-hydrogen energy system with dynamic pricing coupled with electricity and carbon, comprising the following steps: S1. Use deep learning models to predict the original carbon trading price series, and combine clustering algorithms to generate typical scenarios of wind and solar power output probability that take into account extreme cases and computational efficiency. S2. Construct a mathematical model of the equipment for the integrated electrothermal hydrogen energy system, and set the power conversion constraints and energy balance equations for the electrolyzer, hydrogen fuel cell, and multi-electrode energy storage device. S3. Construct a dynamic pricing mechanism for electricity-carbon coupling. During the scheduling process, calculate the dynamic carbon emission factor of the system based on the real-time response power of hydrogen-related equipment, and generate a real-time comprehensive electricity price for electricity-carbon coupling. S4. By employing intelligent optimization algorithms combined with predicted environmental parameters and comprehensive electricity prices, the output schemes of each device in the system are iteratively optimized in a closed loop, thereby achieving the goal of minimizing the total operating cost of the system.

[0006] In one embodiment, the generation of typical scenarios for wind and solar power output probabilities related to carbon trading prices in step S1 involves the following steps: S11. Extract historical carbon price data from the carbon market and related energy price series as a feature matrix; S12. Use the CEEMDAN decomposition algorithm model to extract features and perform time series prediction, and output the predicted carbon price for future scheduling cycles. S13. Based on the historical wind speed and illumination data, a probability distribution is fitted, an initial scene set is generated using Monte Carlo simulation, and a clustering algorithm is used to extract a probabilistic typical scene set.

[0007] In one embodiment, the specific steps for constructing the electrocarbon coupling dynamic pricing mechanism in S3 are as follows: S31. Obtain the electrolytic cell power consumption generated by the current iteration of the intelligent optimization algorithm in S4. and fuel cell power generation ; S32. Combining the system's conventional electrical load and predicted wind and solar power output, calculate the system's... t Dynamic net load factor of external power grid at all times ; S33, Combined with the carbon emission intensity constant of the external power grid Calculate the dynamic carbon emission factor that reflects the real-time cleanliness of the system. ; S34. Couple the base electricity price, dynamic carbon emission factor, and predicted carbon price to calculate the combined electricity-carbon price under this iterative state. .

[0008] In one embodiment, the intelligent optimization algorithm in S4 specifically adopts the particle swarm optimization algorithm, and its closed-loop iterative steps are as follows: S41. Initialize the population parameters of the intelligent optimization algorithm and randomly generate a set of initial system scheduling schemes that include the output status of the electrolyzer and fuel cell; S42. Substitute the equipment output of the initial scheme into the electricity-carbon coupling pricing mechanism of S3 to obtain the real-time comprehensive electricity price sequence; S43. Using the minimum day-ahead operating cost of the system as the objective function, calculate the total operating cost and penalty term of the current scheme under the comprehensive electricity price; S44. Determine whether the optimal cost convergence condition or the maximum number of iterations has been met. If not, use the algorithm's own mutation and position update strategy to correct the scheduling scheme and return to S42. If it has been met, output the optimal multi-timescale collaborative scheduling strategy.

[0009] Compared with the prior art, the beneficial effects of the present invention are: 1. This method breaks through the limitations of traditional scheduling that uses fixed carbon prices and static carbon emission factors. It deeply integrates the carbon price predicted by the CEEMDAN algorithm with the real-time operating status of the system, dynamically calculates the carbon emission factor and embeds it into the electricity price system, and truly reflects the spatiotemporal dynamic changes of the carbon emission intensity of the external power grid. This allows the system scheduling decision to accurately match the carbon market price fluctuation pattern, greatly improves the robustness to cope with the risk of high-frequency carbon price fluctuations, and achieves proactive defense against carbon market fluctuations.

[0010] 2. This application, for the first time, directly maps the physical operating states of electrolyzers ("electricity-consuming hydrogen production") and fuel cells ("hydrogen-using electricity") to the real-time energy purchase cost of the system. By dynamically adjusting the electricity price through the real-time response power of hydrogen-related equipment, it fully explores the carbon reduction potential of hydrogen energy as a cross-source medium. The system can utilize green hydrogen to achieve "intertemporal carbon transfer." When carbon prices are high, the hydrogen energy equipment can autonomously adjust to reduce dependence on the external high-carbon power grid, effectively avoiding high electricity prices and carbon emission penalties, and allowing the low-carbon value of hydrogen energy equipment to be directly converted into system economic benefits.

[0011] 3. This scheme changes the existing one-way open-loop model of "pre-setting electricity prices and then optimizing scheduling," by embedding the electricity-carbon coupled pricing module into the closed-loop iterative process of the particle swarm optimization algorithm. This forms a two-way game and deep coupling mechanism where "scheduling decisions change system cleanliness, cleanliness changes dynamically adjust electricity prices, and dynamic electricity prices feed back into and guide scheduling optimization." This achieves coordinated optimization of price response and power scheduling, allowing the scheduling scheme to continuously adapt to dynamic electricity prices and the carbon market environment during iteration, ensuring the global optimum of the system's total operating cost.

[0012] 4. This solution, through the precise generation of typical wind and solar power output scenarios, the scientific construction of equipment mathematical models, and multi-objective intelligent optimization, minimizes system operating costs while simultaneously improving wind and solar power absorption rates and load fulfillment rates. This effectively promotes the absorption of renewable energy sources such as wind and solar power, and enhances the economic efficiency and low-carbon nature of the integrated energy system of electrothermal hydrogen. Furthermore, the technical approach and implementation scheme of this method provide reliable theoretical support and practical technical references for the low-carbon planning, multi-timescale scheduling, and deep integration of the integrated energy system with the electricity market for electrothermal hydrogen systems. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of the framework of this method; Figure 2 A schematic diagram of the CEEMDAN algorithm; Figure 3 A structural diagram of the electrocarbon coupling pricing mechanism; Figure 4 This is a flowchart of the particle swarm optimization algorithm. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] Please see Figures 1-4 This invention provides a technical solution: an optimized scheduling method for an integrated electrothermal-hydrogen energy system with dynamic pricing based on electro-carbon coupling. 1. Data Preprocessing Multiple methods were employed to detect missing power data. First, statistical tools were used to count the number and percentage of missing values ​​in each column, pinpointing the location and extent of the missing values. Second, a missing value matrix was plotted to visually demonstrate the distribution characteristics of missing values ​​across the sample and time series, facilitating the assessment of their clustering and regularity.

[0016] A phased, progressive method is used for power anomaly detection: first, power anomalies exceeding the limit are screened out. The system first analyzes physical anomaly data within a certain range; secondly, it combines meteorological characteristics with power relationship analysis to identify the distribution characteristics of zero-output time, thus distinguishing between normal and abnormal shutdowns; finally, it uses the truncated interquartile range (TQR) method to set dynamic thresholds and identify scattered anomalies caused by interference or measurement errors, thereby balancing detection efficiency and accuracy.

[0017] Upon detecting missing or abnormal data, it is set to null (NULL), and a multi-level repair strategy is employed: Level 1: Single random missing point, filled with the median; Level 2: For consecutive missing values ​​within one day, linear interpolation is used to estimate the missing value based on known points before and after the missing value. Level 3: For continuous absences of more than one day, fill the gaps by referring to the adjacent wind farms according to their operating capacity ratio, and correct for Pearson correlation coefficient. Level 4: Wind farms with more than 25% missing data in a continuous manner are directly removed because the data is not representative, and are then modeled separately before being included in the cluster power prediction.

[0018] 2.CEEMDAN decomposition Please see Figure 2 Initialize parameters, assuming the original power sequence is x(t), the IMF set is empty, and initialize the residual signal: ; noise amplitude is When the number of set iterations is N, and the k-th IMF (k=1,2,…,K) is decomposed, N signals with added noise are first generated: ; In the formula, w i ( t White noise is a random signal whose power spectral density is uniformly distributed across the entire frequency range. This represents the noise amplitude.

[0019] For each signal with added noise Perform EMD decomposition to obtain the first IMF: ; In the formula, EMD1(.) represents the first IMF extracted by EMD decomposition of the signal.

[0020] Calculate the ensemble average of the k-th IMF: ; By adding noise multiple times and performing ensemble averaging, CEEMDAN can effectively suppress mode aliasing, reduce the impact of noise on the decomposition results, and thus improve the accuracy of IMF.

[0021] Update residual signal: ; The noise amplitude is dynamically adjusted based on the characteristics of the residual signal. ,generally Set as a percentage of the original signal standard deviation: ; In the formula, This is the proportionality coefficient, usually taken as... std(x(t)) is the standard deviation of the original signal.

[0022] The noise amplitude gradually decreases with increasing decomposition number, making the addition of noise more adaptable to the local characteristics of the signal. The formula for adjusting the noise amplitude is: ; In the formula, The attenuation factor is usually taken as... .

[0023] When the residual signal Stop the decomposition process when the function is monotonic or cannot be further decomposed.

[0024] Finally, we obtain K IMFs and one residual signal. : ; In the formula, I k ( t ) represents the ensemble average of the k-th IMF. This is the residual signal.

[0025] 3. Dynamic pricing mechanism of electricity-carbon coupling Please see Figure 3 In the traditional model, electricity price is an exogenous given variable. In this embodiment, the electricity price will change in real time according to the scheduling status of the hydrogen-related equipment in the system. First, during any scheduling period... t The degree to which the computing system depends on the external high-carbon power grid, i.e., the net load factor. : ; in, This is the predicted electrical load value. and For wind and solar power generation, The planned power consumption of the electrolytic cell, The planned power generation from fuel cells. If the combined output of wind, solar, and fuel cell power is sufficient, It will approach 0. Secondly, it will be combined with the benchmark carbon emission intensity of the large power grid. Calculate dynamic carbon emission factors : ; Finally, the predicted carbon price will be... By embedding the price system, the final integrated electricity price coupled with carbon emissions is obtained. : ; In the formula, For time-of-use basic electricity pricing, This is an adjustment coefficient. When the system has a high self-sufficiency rate of low-carbon resources (through the generation of more green hydrogen),... Extremely small, effectively immune to the electricity price penalty caused by high predicted carbon prices.

[0026] 4. Particle Swarm Optimization Algorithm Optimization objectives and constraints Please see Figure 4 The algorithm employs a multi-objective optimization problem, whose optimization objectives include: Cost: Minimize the overall operating cost of the system. This cost mainly includes the cost of purchasing electricity from the external power grid based on the dynamic pricing mechanism of electricity-carbon coupling (this cost is constrained by the real-time linkage between the predicted carbon price and the dynamic carbon emission factor of the system), the cost of purchasing natural gas, and the operation and maintenance costs of key equipment such as electrolyzers, hydrogen fuel cells, and multi-energy storage devices.

[0027] Wind and solar power integration rate: Maximize the utilization of wind and solar power generation and reduce wind and solar curtailment.

[0028] Load fulfillment rate: Ensure that the system can effectively meet the various load demands of users, such as electricity, heat, and hydrogen energy.

[0029] During the optimization process, the algorithm needs to follow a series of constraints to ensure the safe and reliable operation of the system. These constraints include system power balance, output limits of power generation devices, power limits of energy conversion devices, and operating state constraints of energy storage devices (such as upper and lower limits of state of charge).

[0030] Implementation process of particle swarm optimization algorithm Population definition: Determine the dimensions of the population, selecting the following six dimensions: power of the electric heating and electrolysis unit, power of the energy storage battery, and power of the three loads: electric heating and hydrogen. Determine the population size, inertia weight, learning factor, and other parameters, and set the boundaries of the variables.

[0031] Initialization: Randomly initialize a group of particles in a given space, including position and velocity information.

[0032] Calculate fitness value: This includes individual and population fitness values. The best fitness value among all individuals is the population fitness value.

[0033] Update Particles: Update the velocity and position of particles, while eliminating particles that do not conform to the constraints.

[0034] Update the individual's best fitness value and the global best fitness value, and determine whether the given termination condition is met. If not, return to the previous step and continue searching until the maximum number of iterations is reached.

[0035] This algorithm yields a set of non-dominated solutions that balance low cost, high renewable energy integration rate, and high load fulfillment rate. It can provide multiple solutions based on different preferences. Through the optimized results, decision-makers can choose schemes that prioritize renewable energy integration or those that prioritize extending energy storage lifespan—this is precisely the advantage of multi-objective optimization algorithms.

[0036] This method breaks through the limitations of conventional integrated energy systems' single static carbon trading mechanism, innovatively embedding predicted carbon prices into real-time electricity prices as a dynamic penalty term. By deeply coupling the physical states of "electricity-consuming hydrogen production" in electrolyzers and "hydrogen-powered electricity generation" in fuel cells, it endows the system with the ability to utilize green hydrogen for "intertemporal carbon transfer" and proactively avoid high electricity prices. Furthermore, by nesting the pricing module within the optimization algorithm closed loop, it achieves coordinated optimization of price response and power scheduling, providing reliable theoretical support and technical solutions for low-carbon planning and multi-timescale scheduling of electrothermal hydrogen systems.

[0037] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An optimized scheduling method for an integrated electrothermal-hydrogen energy system with dynamic pricing coupled with electricity and carbon, characterized in that, Includes the following steps: S1. Use deep learning models to predict the original carbon trading price series, and combine clustering algorithms to generate typical scenarios of wind and solar power output probability that take into account extreme cases and computational efficiency. S2. Construct a mathematical model of the equipment for the integrated electrothermal hydrogen energy system, and set the power conversion constraints and energy balance equations for the electrolyzer, hydrogen fuel cell, and multi-electrode energy storage device. S3. Construct a dynamic pricing mechanism for electricity-carbon coupling. During the scheduling process, calculate the dynamic carbon emission factor of the system based on the real-time response power of hydrogen-related equipment, and generate a real-time comprehensive electricity price for electricity-carbon coupling. S4. By employing intelligent optimization algorithms combined with predicted environmental parameters and comprehensive electricity prices, the output schemes of each device in the system are iteratively optimized in a closed loop, thereby achieving the goal of minimizing the total operating cost of the system.

2. The optimized scheduling method for an integrated electrothermal-hydrogen energy system with dynamic pricing coupled with electricity and carbon as described in claim 1, characterized in that, The generation of typical scenarios for wind and solar power output probabilities related to carbon trading prices in S1 involves the following specific steps: S11. Extract historical carbon price data from the carbon market and related energy price series as a feature matrix; S12. Use the CEEMDAN decomposition algorithm model to extract features and perform time series prediction, and output the predicted carbon price for future scheduling cycles. S13. Based on the historical wind speed and illumination data, a probability distribution is fitted, an initial scene set is generated using Monte Carlo simulation, and a clustering algorithm is used to extract a probabilistic typical scene set.

3. The optimized scheduling method for an integrated electrothermal-hydrogen energy system with dynamic pricing coupled with electricity and carbon as described in claim 1, characterized in that, The specific steps for constructing the electric carbon coupling dynamic pricing mechanism in S3 are as follows: S31. Obtain the electrolytic cell power consumption generated by the current iteration of the intelligent optimization algorithm in S4. and fuel cell power generation ; S32. Combining the system's conventional electrical load and predicted wind and solar power output, calculate the system's... t Dynamic net load factor of external power grid at all times ; S33, Combined with the carbon emission intensity constant of the external power grid Calculate the dynamic carbon emission factor that reflects the real-time cleanliness of the system. ; S34. Couple the base electricity price, dynamic carbon emission factor, and predicted carbon price to calculate the combined electricity-carbon price under this iterative state. .

4. The optimized scheduling method for an integrated electrothermal-hydrogen energy system with dynamic pricing coupled with electricity and carbon as described in claim 1, characterized in that, The intelligent optimization algorithm in S4 specifically adopts the particle swarm optimization algorithm, and its closed-loop iterative steps are as follows: S41. Initialize the population parameters of the intelligent optimization algorithm and randomly generate a set of initial system scheduling schemes that include the output status of the electrolyzer and fuel cell; S42. Substitute the equipment output of the initial scheme into the electricity-carbon coupling pricing mechanism of S3 to obtain the real-time comprehensive electricity price sequence; S43. Using the minimum day-ahead operating cost of the system as the objective function, calculate the total operating cost and penalty term of the current scheme under the comprehensive electricity price; S44. Determine whether the optimal cost convergence condition or the maximum number of iterations has been met. If not, use the algorithm's own mutation and position update strategy to correct the scheduling scheme and return to S42. If it has been met, output the optimal multi-timescale collaborative scheduling strategy.

Images