Micro-grid scheduling method

Abstract

The invention provides a microgrid scheduling method, and the method comprises the steps: constructing a first optimization model which comprises a plurality of objective functions and a plurality of constraint conditions; converting the first optimization model into a second optimization model in a QUBO form; the second optimization model is solved through a mixed quantum approximate optimization algorithm to obtain a micro-grid scheduling strategy, the mixed quantum approximate optimization algorithm comprises a quantum approximate optimization algorithm and a classical algorithm, the quantum approximate optimization algorithm is used for solving the second optimization model through quantum evolution in the multi-round iteration process, and the classical algorithm is used for solving the second optimization model through quantum evolution in the multi-round iteration process. The classical algorithm is used for adjusting optimization parameters of the quantum approximate optimization algorithm based on candidate solutions generated by quantum evolution in each iteration process, parallel exploration of multiple potential solutions is achieved, and therefore the probability of finding an optimal solution is improved within feasible time.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power, and particularly relates to a microgrid scheduling method. Background Art

[0002] The increasing popularity of renewable energy and the wide deployment of microgrids are reshaping the modern power system, especially in rural and remote areas where the central power infrastructure is weak or unreliable. Rural microgrids are characterized by decentralized control and a high penetration of distributed energy resources, and play a key role in improving the energy accessibility, sustainability, and resilience of isolated communities. Such microgrids rely on the combination of renewable energy sources such as solar photovoltaic systems and wind turbines with energy storage systems to maintain a stable and reliable power supply.

[0003] However, the intermittency and randomness of renewable energy and the seasonal fluctuations of electricity demand pose significant challenges to microgrid scheduling. Summary of the Invention

[0004] In view of the problem that the intermittency and randomness of renewable energy and the seasonal fluctuations of electricity demand in the prior art pose significant challenges to microgrid scheduling, the present invention proposes a microgrid scheduling method, which includes:

[0005] Construct a first optimization model, where the first optimization model includes multiple objective functions and multiple constraint conditions;

[0006] Convert the first optimization model into a second optimization model in QUBO form;

[0007] Solve the second optimization model through a hybrid quantum approximate optimization algorithm to obtain a scheduling strategy for the microgrid. The hybrid quantum approximate optimization algorithm includes a quantum approximate optimization algorithm and a classical algorithm. The quantum approximate optimization algorithm is used to solve the second optimization model through quantum evolution in multiple rounds of iteration, and the classical algorithm is used to adjust the optimization parameters of the quantum approximate optimization algorithm based on the candidate solutions generated by quantum evolution in each round of iteration.

[0008] Optionally, solving the second optimization model through a hybrid quantum approximate optimization algorithm to obtain a scheduling strategy for the microgrid includes:

[0009] Map the second optimization model to the Hamiltonian of a quantum system;

[0010] Set all qubits in the quantum system to be in a uniform superposition state;

[0011] In each round of iteration, perform quantum evolution through a quantum circuit composed of a problem Hamiltonian and a hybrid Hamiltonian to obtain candidate solutions for each round of iteration;

[0012] In each iteration process, the expected energy is constructed based on the candidate solution and the problem Hamiltonian, and the optimization parameters of the quantum circuit are adjusted to minimize the expected energy;

[0013] The candidate solution corresponding to the minimum value of the expected energy is used as the scheduling strategy of the microgrid.

[0014] Optionally,

[0015]

[0016] where U QAOA (γ,β) is the quantum evolution, γ and β are the optimization parameters, H cost is the problem Hamiltonian, is the problem Hamiltonian evolution, H mixer is the hybrid Hamiltonian evolution, is the hybrid Hamiltonian evolution.

[0017] Optionally, the problem Hamiltonian includes:

[0018]

[0019] where, is the renewable energy generation cost, is the local power generation, is the energy storage degradation cost, is the discharge energy, is the continuous maintenance cost, is the operating state, τ κ,t ·δ κ,t is the imported energy cost function purchased from the main grid, and are the quantum rotation terms.

[0020] Optionally, the hybrid Hamiltonian includes:

[0021]

[0022] where, is the Pauli X operator.

[0023] Optionally, in each iteration process, the expected energy is constructed based on the candidate solution and the problem Hamiltonian, and the optimization parameters of the quantum circuit are adjusted to minimize the expected energy, including:

[0024]

[0025] where Θ opt is the optimization parameter, H cost |ψ(γ,β) is the problem Hamiltonian, ψ(γ,β)| is the candidate solution.

[0026] Optionally, the multiple objective functions include a cost minimization function:

[0027]

[0028] wherein, is the cost of renewable energy power generation, is the power generation amount, is the energy storage degradation cost, is the generated energy, is the continuous maintenance cost, is the operating state, τ κ,t ·δ κ,t is the imported energy cost function purchased from the main power grid.

[0029] Optionally, the multiple objective functions include an energy self-sufficiency rate maximization function:

[0030]

[0031] wherein, is the local power generation amount, is the energy storage power generation amount, is the imported energy, is the total demand, is the adaptive weight.

[0032] Optionally, the multiple objective functions include a reliability maximization function:

[0033]

[0034] wherein, is the local power generation amount, is the energy storage power generation amount, is the total demand, η l,t and ξ l,t are weights, is the imported energy, is the reserve energy.

[0035] Optionally, the multiple objective functions include a quantum optimization function:

[0036]

[0037] wherein, and are decision variables encoded as quantum states, α p and β q are weights.

[0038] A microgrid scheduling method provided by the present invention has the following technical effects:

[0039] Enhancing the operational efficiency and resilience of microgrids using quantum computing breaks through the limitations of traditional computing methods. The proposed hybrid quantum approximate optimization algorithm transforms the microgrid scheduling problem into a quadratic unconstrained binary optimization (QUBO) problem suitable for quantum computing. This method not only accommodates complex operational constraints but also dynamically adapts to the inherent volatility of renewable energy. The hybrid quantum approximate optimization algorithm enables parallel exploration of multiple potential solutions, thereby increasing the probability of finding the optimal solution within a feasible time.

[0040] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and do not limit the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings, which form a part of this specification, are used to provide a further understanding of the present invention. The schematic embodiments and descriptions thereof of the present invention are used to explain the present invention and do not unduly limit the present invention. In the drawings:

[0042] Figure 1 is a flowchart of a microgrid scheduling method provided for an exemplary embodiment of the present invention;

[0043] Figure 2 is a flowchart of step S103 of a microgrid scheduling method provided for an exemplary embodiment of the present invention;

[0044] Figure 3 is a schematic diagram of energy storage utilization rate and capacity provided for an exemplary embodiment of the present invention;

[0045] Figure 4 is a schematic diagram of quantitative analysis of electricity price and electricity consumption in microgrid management provided for an exemplary embodiment of the present invention;

[0046] Figure 5 is a schematic diagram of the time dynamic change of various renewable energy power demands provided for an exemplary embodiment of the present invention;

[0047] Figure 6 is a schematic diagram of load curve, energy storage utilization rate, and grid stability contour within an hour provided for an exemplary embodiment of the present invention;

[0048] Figure 7 is a schematic diagram of the annual energy storage degradation trend provided for an exemplary embodiment of the present invention;

[0049] Figure 8 is a schematic diagram of the cost distribution of QAOA microgrid operation provided for an exemplary embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0050] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments. It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0051] The following detailed descriptions are all exemplary descriptions, aiming to provide further details of the present invention. Unless otherwise specified, all technical terms adopted in the present invention have the same meaning as commonly understood by those of ordinary skill in the art to which the present invention pertains. The terms used in the present invention are only for describing specific embodiments, and are not intended to limit the exemplary embodiments according to the present invention.

[0052] An exemplary embodiment of the present invention provides a microgrid scheduling method. Refer to Figure 1 , the method includes:

[0053] S101. Construct a first optimization model, where the first optimization model includes multiple objective functions and multiple constraint conditions.

[0054] Among them, the operation constraints of the microgrid are converted into a mathematical model suitable for quantum optimization.

[0055] In one embodiment, the objective functions refer to multiple aspects such as cost minimization, maximum energy self - sufficiency rate, maximum reliability, and quantum optimization. Specifically as follows:

[0056] Cost minimization function:

[0057]

[0058] Among them, is the cost of renewable energy generation, is the power generation amount, is the cost of energy storage degradation, is the generated energy, is the continuous maintenance cost, is the operating state, τ κ,t ·δ κ,t is the cost function of imported energy purchased from the main grid.

[0059] Maximum energy self - sufficiency rate function:

[0060]

[0061] Among them, is the local power generation amount, is the power generation amount of the energy storage, is the imported energy, is the total demand, is the adaptive weight. Improving the energy self - sufficiency rate is crucial for minimizing dependence on external resources and ensuring the resilience of the microgrid.

[0062] Reliability maximization function:

[0063]

[0064] Wherein, is the local power generation, is the energy storage power generation, is the total demand, η l,t and ξ l,t are weights, is the imported energy, is the reserve energy. The reliability of rural microgrids requires a balance between supply and demand and sufficient reserves, and the reliability maximization function ensures the robust reliability of multi-season energy operation.

[0065] Quantum optimization function:

[0066]

[0067] Wherein, and are decision variables encoded as quantum states, α p and β q are weights. Quantum optimization plays a crucial role in improving the energy distribution of microgrids. The quantum optimization function enables the quantum system to explore different configurations and approach the optimal solution through quantum measurement, achieving quantum-enhanced energy scheduling and resource allocation.

[0068] In one embodiment, the constraint conditions refer to multiple aspects, specifically as follows:

[0069] Power balance constraint condition:

[0070]

[0071] Maintaining power balance is the basis for the operation of microgrids to ensure stable system performance. This equation stipulates that the total energy charged into the energy storage system minus the discharged energy imported energy and the demand must be consistent with the sum of while also considering the reserved power In addition, line losses and curtailed energy

[0072] Charge and discharge constraint condition:

[0073]

[0074] The operation of the energy storage system is restricted by its respective charge-discharge limits, ensuring that at any given time, the energy storage system cannot charge and discharge simultaneously. Here, and subject to storage capacity limitations the binary decision variables and are ensured to be mutually exclusive. This guarantees that the microgrid energy storage functions optimally under practical limitations.

[0075] Reserve power constraint:

[0076]

[0077] Ensuring sufficient reserve power is crucial for handling the uncertainties of renewable energy generation and demand fluctuations. This constraint requires that the reserve power must always reach or exceed the minimum threshold to meet emergency requirements. Additionally, an adaptive risk perception term is included: quantifies the uncertainty impact brought by load changes to ensure robustness against unforeseen demand peaks.

[0078] Renewable energy generation constraint:

[0079]

[0080] Renewable energy generation itself is restricted by environmental factors and technical limitations. This equation ensures that the total renewable energy generation does not exceed the maximum potential at any given time. The term considers potential curtailment, enabling the system to dynamically adjust the generation while maintaining grid stability.

[0081] Energy balance constraint:

[0082]

[0083] The energy balance constraint enforces energy balance in the net energy flow dynamics. The total energy generated and discharged by the energy storage must be consistent with the energy of demand and input, ensuring that the system does not accumulate excess energy. The coefficient allows for adjustment according to the energy balance requirements to adapt to the operation of the microgrid.

[0084] Imported energy constraint:

[0085]

[0086] The import of energy from the external power grid must be restricted to prevent over - reliance on external power sources. This constraint ensures that the imported electricity does not exceed the allowable capacity Binary decision variable determines whether imports are allowed at time t.

[0087] Absolute power constraint:

[0088]

[0089] Maintaining grid stability requires that the absolute power imbalance be kept within an acceptable range. This constraint ensures that the sum of local generation and storage does not significantly exceed the total demand and imports, while allowing a certain deviation ∈ stability , to accommodate minor fluctuations.

[0090] Energy imbalance constraint:

[0091]

[0092] Minimizing energy imbalance over the entire planning horizon helps maintain an efficient and resilient microgrid. This quadratic penalty function ensures that deviations from perfect energy balance are penalized, thus suppressing inefficient resource allocation and promoting optimal scheduling.

[0093] Energy balance constraint:

[0094]

[0095] Ensuring energy balance at all times is crucial for system stability. This constraint extends the energy balance formula by adding a weighting coefficient that can be adaptively adjusted according to real - time operating requirements. Additionally, losses and curtailed energy are combined with a dynamic weighting coefficient to ensure robustness to changes. A small deviation ∈ stability is allowed on the right - hand side to accommodate short - term fluctuations without compromising system stability.

[0096] Energy cycle constraint:

[0097]

[0098] This constraint controls the efficiency - adjusted energy flow in the energy storage system. It ensures that the total energy charged into the energy storage system (counted at the charging efficiency minus the energy discharged (counted at the discharging efficiency ), remains within the physical energy storage capacity Within the range. This constraint captures the actual losses during the energy cycle.

[0099] Grid limit constraint:

[0100]

[0101] The grid limit must be strictly adhered to prevent overloading. This constraint ensures that the difference between discharging and charging energy storage, and the net input energy after subtracting the demand, remains within the maximum power exchange limit τgrid,max allowed by the grid. This formula prevents violations that could disrupt the stability of the microgrid.

[0102] Renewable energy generation limit constraint:

[0103]

[0104] The renewable energy generation limit must be strictly enforced to prevent unrealistic scheduling. This equation ensures that the actual deployed renewable energy generation (after curtailment ) does not exceed the theoretical maximum generation This constraint dynamically adapts to the variability of renewable energy generation.

[0105] Operating energy imbalance constraint:

[0106]

[0107] Minimizing the energy imbalance during operation helps maintain an efficient and resilient microgrid. This quadratic penalty function ensures that deviations from perfect energy balance are penalized, thus suppressing inefficient resource allocation and promoting optimal scheduling. The term adapts the severity of the penalty, while ε imbalance,max defines the threshold at which an imbalance is considered unacceptable.

[0108] Battery life degradation constraint:

[0109]

[0110] Battery life degradation is an important consideration for the long-term viability of the microgrid. This constraint ensures that the cumulative charge-discharge cycles (scaled by the life cycle impact factors and ) do not exceed the predetermined total cycle limit This promotes the responsible use of energy storage technologies.

[0111] Microgrid resilience constraint:

[0112]

[0113] The resilience of the microgrid must be maintained at all times. This equation ensures that the weighted energy balance term (including the resilience weighting factor ) always meets the requirement of the minimum resilience threshold θ resilience,min . This avoids excessive dependence on external energy sources and enhances the robustness of the system

[0114] Distributed energy constraint conditions:

[0115]

[0116] This constraint strengthens the demand response mechanism by ensuring that a minimum proportion of the demand is met through stored emissions rather than external supply. The weighting coefficients and reflect the flexibility of demand shifting and energy storage utilization. This constraint ensures that distributed resources can be effectively utilized before resorting to grid power supply.

[0117] Decarbonization constraint conditions:

[0118]

[0119] Decarbonization is the core goal of microgrid operation, ensuring that the dependence on fossil grid power is minimized. This constraint quantifies the net impact of power flow on the environment and is weighted according to their respective emission factors: renewable energy is energy storage emissions are grid imports are The weighted sum must exceed the pre-set carbon neutrality threshold η carbon,min to achieve a sustainable energy transition.

[0120] Voltage stability constraint conditions:

[0121]

[0122] Voltage stability must be maintained within an acceptable range to prevent frequency deviation and operation interruption. This quadratic penalty function ensures that the power imbalance weighted by the voltage sensitivity coefficient does not exceed the tolerable deviation threshold ε voltage,max . This penalty measure discourages serious imbalances between supply and demand that may cause voltage violations.

[0123] Cybersecurity risk constraint conditions:

[0124]

[0125] The cybersecurity risk requirements of the smart microgrid limit the excessive dependence on external grid power, as external grid power is vulnerable to cyberattacks. This equation ensures that according to the network vulnerability coefficient The scaled weighted input power term does not exceed the maximum risk threshold θ cyber,max . Meanwhile, the elastic coefficient will enhance the priority of local power generation and storage rather than external resources.

[0126] Heating limit constraint:

[0127]

[0128] Heating limits must be applied to prevent overheating of the battery storage unit. This constraint ensures that the net storage charge-discharge cycles weighted by the thermal impact factor do not exceed the maximum allowable thermal deviation ∈ thermal,max . This can prevent the degradation of battery performance and life due to excessive thermal stress.

[0129] Demand-side flexibility constraint:

[0130]

[0131] The flexibility of the demand side is crucial for optimizing the operation of the microgrid. This constraint enforces a minimum level of demand-side response, ensuring that a portion of the total demand is met through flexible mechanisms such as energy storage discharge and is weighted by the response coefficients and . This constraint ensures that grid power is used as a last resort, thus encouraging more responsive and adaptable consumption patterns.

[0132] S102. Convert the first optimization model into a second optimization model in QUBO form.

[0133] Among them, the QUBO form is the quadratic unconstrained binary optimization form. Through the QUBO form, complex microgrid scheduling problems can be transformed into problems that can be processed by quantum computers. In the QUBO form, binary variables are used to represent the values of decisions. The decisions are the objects to be solved, and all the decisions of the objective functions constitute the scheduling strategy in this embodiment.

[0134] S103. Solve the second optimization model through a hybrid quantum approximate optimization algorithm to obtain the scheduling strategy of the microgrid.

[0135] Among them, the scheduling strategy is used to schedule the operation of the microgrid. The hybrid quantum approximate optimization algorithm includes the quantum approximate optimization algorithm and classical algorithms. The quantum approximate optimization algorithm (QAOA) is used to solve the second optimization model through quantum evolution in multiple rounds of iteration, and the classical algorithm is used to adjust the optimization parameters of the quantum approximate optimization algorithm based on the candidate solutions generated by quantum evolution in each round of iteration.

[0136] In one embodiment, step S103 specifically includes:

[0137] S1031. Map the second optimization model to the Hamiltonian of the quantum system.

[0138] Specifically:

[0139]

[0140] The hybrid quantum approximate optimization algorithm iteratively adjusts the variable quantum parameters to minimize the energy of the target Hamiltonian. This second optimization model is expressed as a function of the quantum state evolution parameters and and represents the encoding of classical variables into the Hamiltonian. The summation of the cosine and sine terms introduces a dynamic interaction between the quantum rotation and the energy landscape, and the weights α p and β q guide the search trajectory. By iteratively improving these parameters using a classical optimizer, the hybrid quantum approximate optimization algorithm can effectively find an approximate optimal solution to high-dimensional combinatorial problems.

[0141] Among them, each binary quantity corresponds to a qubit, and its value is characterized by |0> or |1>. The quantum state is a superposition of the qubits that are the decisions in the second optimization model. For example, if the number of decisions in the second optimization model is 3, and the values are |0>, |0>, or |1> respectively, then the quantum state is |001>.

[0142] S1032. Set all qubits in the quantum system to be in a uniform superposition state.

[0143] S1033. In each iteration process, perform quantum evolution through the quantum circuit composed of the problem Hamiltonian and the hybrid Hamiltonian to obtain the candidate solution for each iteration process.

[0144] Specifically, the quantum evolution U QAOA (γ,β) includes:

[0145] U QAOA (γ,β) = e -iγHcost e -iβHmixer

[0146] Among them, e -iγH cos t is the problem Hamiltonian evolution, and e -iβH m i xer is the hybrid Hamiltonian evolution. It alternately applies the problem Hamiltonian evolution e -iγH cos t and the hybrid Hamiltonian evolution e -iβH mi xer layers. The parameters γ and β are iteratively adjusted to guide the quantum state towards the direction of the optimal energy configuration.

[0147] The problem Hamiltonian includes:

[0148]

[0149] where is the cost of renewable energy generation, is the local power generation, is the cost of energy storage degradation, is the discharge energy, is the continuous maintenance cost, is the operating state, τ κ,t ·δ κ,t is the cost function of imported energy purchased from the main power grid, and are the quantum rotation terms. The problem Hamiltonian encodes all the cost components of the microgrid operation, including generation cost, storage degradation, operation limits, and quantum variables. The quantum rotation terms and enable the quantum system to explore multiple superpositions, thus promoting the efficient optimization of discrete decision variables.

[0150] The hybrid Hamiltonian includes:

[0151]

[0152] where is the Pauli X operator. The hybrid Hamiltonian H mixer is responsible for driving the transitions between quantum states, ensuring a wide exploration of the solution space. It is expressed as the sum of the Pauli X operators while the Pauli X operator causes state transitions in the quantum register. This enables the optimization algorithm to escape from local minima and ensures that the finally measured solution is globally optimal.

[0153] S1034. During each iteration process, construct the expected energy according to the candidate solution and the problem Hamiltonian, and adjust the optimization parameters of the quantum circuit to minimize the expected energy.

[0154] Specifically, minimizing the expected energy includes:

[0155]

[0156] where the convergence condition of the hybrid quantum approximate optimization algorithm is set to ensure that the energy expectation of the quantum optimization is close to the best classical optimization result Eclassical-opt , thus proving that the hybrid quantum approximate optimization algorithm has competitive or superior performance in microgrid resource allocation.

[0157] The expected energy includes:

[0158]

[0159] Among them, E QAOA The expected energy, calculated according to the measurement probability obtained from repeated quantum state collapses , and this expected value quantifies the effectiveness of quantum-optimized microgrid operation.

[0160] Specifically, the optimization parameters of the quantum circuit include:

[0161]

[0162] Among them, Θ opt is the optimization parameter, H cost |ψ(γ,β) is the problem Hamiltonian, and ψ(γ,β)| is the candidate solution. The final optimization parameter Θ opt is obtained by minimizing the expected energy with respect to the evolving quantum state. This expected value is calculated through repeated quantum measurements, and the classical optimizer refines γ and β to converge to the best possible solution.

[0163] Regarding the candidate solution, since it changes following the quantum evolution process,

[0164]

[0165] Among them, a uniform superposition of all possible binary solutions.

[0166] In one embodiment, the adjustment of the quantum circuit is carried out in a gradient manner to refine the variational parameters:

[0167]

[0168] Among them, the learning rate η controls the update amplitude to ensure stable convergence to the optimal energy allocation strategy

[0169] In one embodiment, after iteration, the final optimization parameter Θ final is obtained. These parameters define the optimal microgrid control strategy, ensuring the best trade-off among cost minimization, system stability, and quantum computing efficiency.

[0170]

[0171] In one embodiment, to ensure the smooth realization of energy balance, a penalty Hamiltonian function H is added during the optimization processpenalty , the function penalizes any behavior that deviates from the energy balance equation:

[0172]

[0173] where the severity of the penalty can be dynamically adjusted. The square term strictly penalizes large mismatches, guiding the quantum optimization process to find a solution that satisfies the microgrid constraints.

[0174] In one embodiment, to improve the satisfaction of the constraints, an additional penalty term H is introduced in the parametric circuit evolution penalty , thus extending the evolution of the hybrid quantum approximate optimization algorithm:

[0175]

[0176] The new parameters are adjusted iteratively to ensure that infeasible solutions are gradually pushed towards the feasible region. This method enhances the robustness of the microgrid optimization based on the hybrid quantum approximate optimization algorithm.

[0177] In one embodiment, in the k-th iteration, the quantum state is obtained by applying the modified hybrid quantum approximate optimization algorithm transformation to the initial superposition state:

[0178]

[0179] This evolution ensures that a feasible optimal solution is explored through quantum-enhanced combinatorial search.

[0180] In one embodiment, in each optimization step, the total cost and the expected value of the penalty Hamiltonian are evaluated:

[0181]

[0182] This expected value guides the parameter update strategy by penalizing violations of the constraints while optimizing the economic objective.

[0183] In one embodiment, the optimization of the optimization parameters follows a gradient-based iterative update rule, and the step size is controlled by the learning rate η:

[0184]

[0185] This adaptive adjustment mechanism ensures the convergence of the quantum optimization process to the optimal microgrid scheduling solution while dynamically enforcing the operating constraints.

[0186] In one embodiment, after sufficient optimization iterations, the optimization parameters converge to the final solution Θ converged :

[0187] Θ converged= l k→∞ (γ k , β k , τ k )

[0188] Θ converged represents the optimal quantum state encoding of the microgrid scheduling problem. The corresponding energy scheduling decision extracted from this solution can be used to formulate the microgrid control strategy in the real world.

[0189] S1035. Take the candidate solution corresponding to the minimum expected energy as the scheduling strategy of the microgrid.

[0190] In one embodiment, to verify the optimization efficiency of the hybrid quantum approximate optimization algorithm, the scheduling strategy of the hybrid quantum approximate optimization algorithm was calculated and the classical optimization benchmark The variance between them is:

[0191]

[0192] A smaller variance indicates that the hybrid quantum approximate optimization algorithm has achieved comparable or better results than the classical method, proving its feasibility in actual deployment.

[0193] In one embodiment, the convergence speed of the hybrid quantum approximate optimization algorithm compared with the classical method is quantified by the acceleration factor κ speedup :

[0194]

[0195] Compared with the classical solution algorithm, the hybrid quantum approximate optimization algorithm is expected to bring computational advantages.. κ speedup > 1 proves that quantum acceleration of microgrid scheduling is superior to traditional methods.

[0196] In one embodiment, the satisfaction of the constraint conditions is a key evaluation index of the quantum optimization effect. The probability P constraint measures the proportion of constraint violations in the scheduling strategy of the optimized hybrid quantum approximate optimization algorithm:

[0197]

[0198] The smaller the value, the more it indicates that the algorithm has successfully achieved feasibility while maintaining optimality.

[0199] In one embodiment, the final expected cost after the best hybrid quantum approximate optimization algorithm iterates k opt is denoted as E final :

[0200]

[0201] This value represents the energy cost of the hybrid quantum approximate optimization algorithm for microgrid operation, benchmarked against classical optimization techniques.

[0202] In one embodiment, to quantify the effectiveness of the hybrid quantum approximate optimization algorithm in optimizing microgrid scheduling, we calculated the relative cost improvement compared to traditional methods:

[0203]

[0204] Metric E improvement Measures the percentage reduction in the total system cost when using the hybrid quantum approximate optimization algorithm compared to traditional solvers. A larger value indicates better quantum-driven optimization.

[0205] In one embodiment, due to its probabilistic nature, the hybrid quantum approximate optimization algorithm can explore a broader solution space by itself. The equation measures the solution diversity of the hybrid quantum approximate optimization algorithm by calculating the deviation between the optimal quantum solution and the average quantum solution over multiple runs:

[0206]

[0207] The higher the diversity, the richer the exploration of other operating strategies.

[0208] In one embodiment, the convergence rate F convergence-rate is used to evaluate the speed at which the hybrid quantum approximate optimization algorithm iterates towards the optimal solution:

[0209]

[0210] It calculates the average change in the expected cost function per iteration, ensuring an efficient and stable optimization process. A lower value indicates smooth convergence, while a larger value indicates oscillatory behavior.

[0211] In one embodiment, quantum entanglement plays a crucial role in improving the computational speed. The entanglement depth C entanglement-depth quantifies the correlation between the variational parameters θ p and λ q of the hybrid quantum approximate optimization algorithm:

[0212]

[0213] The stronger the correlation, the more complex the quantum interactions. A high entanglement depth can improve the quality of the solution but requires careful hardware implementation.

[0214] In one embodiment, real-world quantum devices suffer from decoherence and gate errors, which can affect the optimization results. The hardware noise penalty Phardware-noise The influence of device defects on the solution accuracy is simulated:

[0215]

[0216] Error factor Is proportional to the difference between the quantum and classical solutions, thus allowing mitigation strategies to be incorporated.

[0217] In one embodiment: The quantum circuit must balance computational power and hardware feasibility. The circuit complexity M circuit-complexity Is calculated by adding the gate complexity O gates (g) of all quantum gates g, and then multiplying by the circuit depth D depth (g). Lower complexity enhances the scalability of near-term quantum processors:

[0218]

[0219] In one embodiment, scalability is a key factor in whether the hybrid quantum approximate optimization algorithm can be used for large-scale microgrid scheduling. As the problem size N increases, this equation measures the relative efficiency of the hybrid quantum approximate optimization algorithm and classical solvers:

[0220]

[0221] A stable or decreasing ratio demonstrates that the hybrid quantum approximate optimization algorithm remains competitive for complex systems. Is competitive for complex systems.

[0222] In one embodiment, robustness ensures that the hybrid quantum approximate optimization algorithm solution remains viable under practical conditions. This formula measures robustness by evaluating the interaction between the constraint violation P constraint And the impact of hardware noise P hardware-noise :

[0223]

[0224] The lower the value, the stronger the ability to adapt to errors and uncertainties.

[0225] In one embodiment, the hybrid quantum approximate optimization algorithm typically improves performance by leveraging classical post-processing. The hybrid advantage ratio H hybrid-advantage Quantifies the additional improvement brought by classical fine-tuning to the original hybrid quantum approximate optimization algorithm solution:

[0226]

[0227] A ratio close to 1 indicates quantum dominance, while larger values indicate a greater synergistic effect of hybrid computing.

[0228] In one embodiment, in real-time microgrid operation, execution time is crucial. The total execution time Texecution-time includes the quantum circuit operation time T quantum , the classical pre-optimization time T classical and any additional post-processing time T post-processing :

[0229] T execution-time = T quantum + T classical + T post-processing

[0230] This metric determines the practical feasibility of energy scheduling.

[0231] In one embodiment, even with advanced optimization, the hybrid quantum approximate optimization algorithm may not necessarily find the global minimum. The optimization gap D optimization-gap quantifies the absolute gap between the scheduling strategy E QAOA of the hybrid quantum approximate optimization algorithm and the proven scheduling strategy E optimal :

[0232] D optimization-gap = |E QAOA - E optimal |

[0233] The smaller the gap, the closer the result is to the optimum.

[0234] In one embodiment, the energy workload must be evenly distributed among resources. This equation evaluates the workload distribution by comparing the quantum optimization scheduling decision with the classical benchmark :

[0235]

[0236] An evenly balanced workload can improve resource utilization.

[0237] In one embodiment, the quantum algorithm must be able to adapt to future hardware and algorithm developments. The future compatibility metric X future-compatibility estimates the degree to which the performance of the hybrid quantum approximate optimization algorithm improves as quantum technology gradually improves:

[0238]

[0239] A value close to 1 indicates forward compatibility.

[0240] For the practical application of the microgrid scheduling method proposed in this embodiment, the following results are obtained:

[0241] To verify the effectiveness and computational efficiency of the proposed quantum-based multi-factor resource allocation for the hybrid quantum approximate optimization algorithm 4

[0242] Within this framework, we utilize a rural microgrid. The selected microgrid consists of 35 distributed energy nodes, including 15 photovoltaic (PV) power generation units, 10 wind turbines, and 10 energy storage systems (ESS). The total installed capacity of PV is 2.5 MW, while the wind power generation is 4 MW under optimal wind speed conditions. The microgrid operates independently of the main grid but has a 1 MW capacity backup diesel generator for emergencies. The case study covers a one-year operating range and is divided into four seasonal scenarios to reflect the impact of seasonal variations on renewable energy generation and demand-side behavior. Each season is further divided into weekly scheduling intervals, resulting in a total of 52 optimization instances. The demand-side data is sourced from historical residential and agricultural electricity consumption patterns, with peak summer demand reaching 3.8 MW and relatively lower winter demand of 2.6 MW.

[0243] The dataset includes detailed meteorological records such as solar irradiance levels, wind speed variations, and temperature fluctuations to simulate the uncertainty - reliability of renewable energy generation. The average daily output of solar PV power generation is 4.5 kWh / kW / day. The highest power generation occurs between 10 am and 3 pm, while wind power generation occurs at night due to the prevailing wind patterns in the study area, with peak output between 8 pm and 4 am. The total installed capacity of these energy storage systems is 6 MWh, with a rated capacity of 600 kWh for each battery module and a depth of discharge (DoD) limit of 85% to ensure the battery's service life. The round-trip efficiency of the storage unit is set at 92%, and each storage unit can be charged and discharged up to two times per day.

[0244] Typically, the microgrid includes 50 demand response (DR) participants who can adjust their consumption based on price signals, enabling overall flexibility. During high-demand periods, the load is 400 kW. To incorporate real-world constraints, a 5% contingency reserve requirement is imposed to maintain system stability in the event of unexpected fluctuations in supply or demand.

[0245] The quantum-classical hybrid optimization framework is implemented on a D-Wave Advantage 5000 quantum annealer, using 5000 qubits for quantum state representation and constraint encoding. The optimization based on the hybrid quantum approximate optimization algorithm is executed through PyQuil and PennyLane and interfaced with a classical solver running on a 32-core AMD EPYC 7742 processor (equipped with 1TB of memory). The classical preprocessing and parameter tuning stages use the TensorFlow and Scipy optimization libraries, and the postprocessing is performed through MATLAB and Pandas for result analysis and visualization. Each iteration of the hybrid quantum approximate optimization algorithm contains 20 quantum layers, and the variational parameters are optimized by the gradient-free Nelder-Mead method. The maximum allowed computation time for each scheduling instance is 10 minutes, ensuring near-real-time optimization feasibility. The hybrid framework also includes a quantum noise mitigation strategy that compensates for the hardware-induced decoherence effect, reducing the error rate by 23%, thereby enhancing the robustness of the quantum-derived solutions. The proposed case study provides a real testbed for evaluating the practical applicability of the hybrid quantum approximate optimization algorithm in microgrid scheduling and verifies its performance advantages through comparison with traditional mixed-integer linear programming (MILP) and heuristic particle swarm optimization (PSO) solvers.

[0246]

[0247] Table 1. Summary of Daily Energy Management Metrics

[0248] Table 1 provides a brief overview of the energy dynamics of the microgrid at different times of the day (morning, afternoon, evening), including metrics such as photovoltaic power generation, wind power generation, total energy consumption, stored energy, and released energy. All data are in kilowatt-hours (kWh) and are rounded to one decimal place to ensure accuracy, reflecting the real measurement data that may be observed in an actual operating microgrid.

[0249]

[0250]

[0251] Table 2. Comparison of Classical and Quantum Optimization Techniques for Microgrid Energy Management

[0252] Table 2 aims to quantitatively compare the application results of classical optimization techniques with the quantum approximate optimization algorithm (hybrid quantum approximate optimization algorithm), covering performance metrics such as average energy production, peak load satisfaction, reduction in operating costs, improvement in reliability, and environmental impact, presented in kilowatt-hours (kWh) or percentages, respectively.

[0253] Figure 3Fully demonstrates the management of the energy storage system (ESS) in a rural microgrid over a 24-hour period. The graph shows three key dimensions: the charge, discharge, and remaining capacity of the energy storage system. The charging curve is represented in light blue and follows a sinusoidal pattern, reaching its peak at around the 12th hour (noon), corresponding to the peak solar generation period. This indicates that the energy storage system is mainly charged through photovoltaic solar panels, consistent with the peak solar irradiance period.

[0254] The total charge of the energy storage system can reach 100 kilowatt-hours (kWh), highlighting the system's ability to store excess energy during peak generation periods. In contrast, the discharge curve is shown in gray and exhibits the opposite pattern - reaching its peaks during early morning and evening hours. This indicates that the energy storage system effectively releases stored energy when solar generation is insufficient (usually before sunrise and after sunset). The discharge peak also occurs at the 24th hour (i.e., early morning of the next day), indicating that the system is ready to supply power during the early morning hours when generation capacity cannot immediately meet demand. The observed maximum discharge is approximately 100 kWh, matching the charging capacity, reflecting the balanced strategy of energy inflow and outflow in the daily operation of the microgrid.

[0255] Figure 4 A detailed quantitative analysis of the relationship between electricity price and electricity consumption in the microgrid system was conducted. The dataset in the graph simulates a normal distribution of electricity prices around a mean of 5 and a standard deviation of 2 to reflect actual price fluctuations under market or policy influences. The electricity consumption data is modeled to be directly related to the electricity price and is superimposed with random Gaussian noise to simulate daily operation fluctuations. This realistic simulation provides a basis for analyzing how electricity price changes affect actual electricity consumption patterns. The red regression line shows a significant positive correlation between electricity price and electricity consumption, with a slope of approximately 2.0, indicating that for every 1 unit increase in electricity price, electricity consumption increases by approximately 2 units. The regression equation is further quantified as y = 2.0x + intercept, where the intercept represents the base electricity consumption independent of the electricity price. The R-squared value shown in the upper left corner of the graph is approximately 0.76, meaning that approximately 76% of the electricity consumption changes can be explained by electricity price changes, highlighting the strong linear relationship between the two. The density contour lines deepen the analysis by highlighting the areas with the highest data point density, showing that the most common price - electricity consumption combinations are concentrated around the mean, while the data density is lower in the high and low price ranges. This pattern indicates that although an increase in electricity price generally leads to an increase in electricity consumption, the sensitivity decreases at extreme prices, possibly due to demand elasticity or the upper limit of the microgrid's electricity consumption capacity. This detailed understanding of price sensitivity is crucial for operators and policymakers to optimize energy distribution and pricing strategies within the microgrid. Figure 5 Intuitively shows the association between the operating hours of different renewable energy systems in the microgrid and the fluctuations in electricity demand. The horizontal axis ranges from 0 to 23 hours, reflecting the change in demand from early morning to late night in a day; the vertical axis divides the energy into solar, wind, and hydropower, and the demand patterns for each type of energy are affected by natural conditions and electricity consumption behavior. Specifically:

[0256] Solar energy demand: Peaks between 10:00 and 14:00 (around noon), corresponding to the period of highest solar irradiance. The darkest contour lines in the figure indicate that the demand can reach approximately 50 units (presumably in megawatts), highlighting the strong dependence on solar energy during this period.

[0257] Wind energy demand: Shows a small peak between 18:00 and 23:00 (evening to late night), possibly related to the increase in wind speed at night. The demand fluctuates by about 30 units, contributing steadily but less strongly than solar energy.

[0258] Figure 6 By analyzing three key parameters: the load curve, energy storage utilization rate, and grid stability, the daily dynamic patterns of the microgrid are revealed:

[0259] Load curve: Peak electricity consumption occurs at 8:00 am and 6:00 pm (typical residential and commercial electricity consumption patterns), requiring increased power supply and precautions against grid overload.

[0260] Energy storage utilization rate: Contrary to the load peak, the energy storage system discharges during the peak electricity consumption period to make up for the gap and charges during the low-demand period to balance system costs and stability.

[0261] Grid stability: Reflects the real-time status of voltage / frequency through the 24-hour fluctuation trend. For example, the stability is highest when solar energy is abundant in the afternoon, and energy storage support is required during the morning and evening peaks.

[0262]

[0263] Table 3. Quantum optimization results of the energy system

[0264]

[0265] Table 4. Improving grid performance through quantum optimization

[0266] Table 3 compares the key performance indicators of the energy management system before and after quantum optimization:

[0267] Cost reduction: By optimizing the charging and discharging timing and power purchase strategy, reduce the dependence on high-price periods.

[0268] Efficiency improvement: The quantum algorithm reduces energy transmission losses. For example, it improves the energy storage charging and discharging efficiency from 92% to 95%.

[0269] Enhanced reliability: Dynamically adjust the reserve capacity (such as reserving 5% emergency power) to reduce the risk of power outages.

[0270] Table 4 shows the significant impact of the quantum computing method on improving grid management and operation capabilities. The data compares the performance indicators before and after applying the quantum algorithm:

[0271] Peak demand reduction: Increased from 10.0% to 15.5%, an increase of 5.5%, due to more efficient demand management enabled by quantum technology.

[0272] Renewable energy integration efficiency: Increased from 65.0% to 75.0%, an increase of 10.0%, indicating a significant improvement in the processing and utilization of renewable energy.

[0273] Operational flexibility: Increased from 70.0% to 82.0%, an increase of 12.0%, reflecting a stronger ability to respond to grid demands and sudden changes.

[0274] These data highlight the significant advantages of quantum optimization in the operation of energy systems, emphasizing its role in enhancing grid resilience and efficiency.

[0275] Figure 7 Presents the degradation trend of the energy storage system efficiency over five years with visual impact. This figure is crucial for understanding the losses of energy storage devices during long-term use. The bar chart shows the annual degradation percentage in three different shades of pink: the degradation rate in the first year is less than 3% (light pink, degradation rate 2.8%), indicating that the newly commissioned energy storage system has the best performance and the least wear; over time, the pink gradually deepens, and the degradation rate in the second year rises to 5.6%; by the fifth year, the darkest pink corresponds to a degradation rate of 15.2%, highlighting the need to adopt proactive maintenance strategies or replace equipment to ensure the sustainability and efficiency of the energy storage system. This progressive performance decline is due to the cumulative effect of charge-discharge cycles, environmental factors, and maintenance practices. The color gradient not only facilitates a quick assessment of annual changes but also serves as an analytical tool for predicting future trends and planning maintenance. This figure provides a clear and intuitive visualization of the long-term degradation trajectory for the operation planning and life management of energy storage facilities, which is crucial for relevant decision-makers.

[0276] Figure 8 Details the composition of the operating costs in the quantum-optimized microgrid system. This figure divides the costs into three categories: generation, storage, and maintenance, represented by pink bars of different depths, corresponding to different quantum states or operating scenarios. Specifically:

[0277] Generation cost: Between 40.5 and 45.5 units, which is the highest part of the three categories of costs, reflecting the significant expenditure on energy production in the microgrid operation.

[0278] Storage cost: Between 25.5 and 30.5 units, covering the charge-discharge losses and life management of energy storage devices.

[0279] Maintenance cost: Approximately 15.5 to 20.5 units, involving daily operation and maintenance expenses such as equipment inspection and software updates.

[0280] The visualization results highlight the practical impact of quantum optimization technology on cost management. Through quantum algorithms, the microgrid can dynamically adjust its operating parameters to efficiently reduce costs. For example:

[0281] Dynamically adapt to demand fluctuations: In different scenarios, the changes in storage and maintenance costs indicate that the algorithm can optimize resource allocation based on real-time demand (such as peak electricity consumption) and energy storage efficiency (such as battery aging).

[0282] Prolong the lifespan of energy storage: By reducing unnecessary charge and discharge cycles, the annual battery degradation rate is reduced from 5.6% to 4.2% (combined with Figure 7 data), reducing long-term replacement costs.

[0283] This adaptive cost management significantly improves the economic feasibility and sustainability of the microgrid, ensuring a balance between cost minimization and performance optimization in a complex energy environment.

[0284] In this embodiment, quantum computing is used to improve the operating efficiency and resilience of the microgrid, breaking through the limitations of traditional computing methods. The proposed hybrid quantum approximate optimization algorithm transforms the microgrid scheduling problem into a quadratic unconstrained binary optimization (QUBO) problem suitable for quantum computing. This method not only accommodates complex operating constraints (such as energy conservation, peak load management, cost efficiency), but also dynamically adapts to the inherent volatility of renewable energy. By encoding these constraints as a quantum-friendly Hamiltonian, the hybrid quantum approximate optimization algorithm enables parallel exploration of multiple potential solutions, thereby increasing the probability of finding the optimal solution within a feasible time. We verified the effectiveness of the model through comprehensive simulations based on real data, including a microgrid equipped with a photovoltaic system, wind turbines, and energy storage units. The results show that the hybrid quantum approximate optimization algorithm outperforms traditional optimization techniques in terms of cost reduction, energy efficiency improvement, and system reliability. In addition, this study explores the scalability of quantum algorithms in energy systems, revealing their potential to handle larger-scale and more complex grid architectures in the context of quantum technology advancements. This invention not only confirms the feasibility of quantum algorithms in practical applications, but also lays the foundation for future research on the integration of quantum computing and energy management systems, promoting the construction of a more sustainable, efficient, and resilient energy infrastructure.

[0285] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent replacements can still be made to the specific embodiments of the present invention, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. A microgrid scheduling method, characterized in that, The method includes: Constructing a first optimization model, the first optimization model including a plurality of objective functions and a plurality of constraint conditions; Converting the first optimization model into a second optimization model in QUBO form; Solving the second optimization model through a hybrid quantum approximate optimization algorithm to obtain a scheduling strategy for the microgrid, the hybrid quantum approximate optimization algorithm including a quantum approximate optimization algorithm and a classical algorithm, the quantum approximate optimization algorithm being used to solve the second optimization model through quantum evolution in multiple rounds of iteration, and the classical algorithm being used to adjust the optimization parameters of the quantum approximate optimization algorithm based on the candidate solutions generated by the quantum evolution in each round of iteration.

2. The microgrid scheduling method according to claim 1, wherein The step of solving the second optimization model through the hybrid quantum approximate optimization algorithm to obtain the scheduling strategy for the microgrid includes: Mapping the second optimization model to the Hamiltonian of a quantum system; Setting all qubits in the quantum system to be in a uniform superposition state; In each round of iteration, performing quantum evolution through a quantum circuit composed of the problem Hamiltonian and the hybrid Hamiltonian to obtain candidate solutions for each round of iteration; In each round of iteration, constructing an expected energy according to the candidate solutions and the problem Hamiltonian, and adjusting the optimization parameters of the quantum circuit to minimize the expected energy; Taking the candidate solution corresponding to the minimum value of the expected energy as the scheduling strategy for the microgrid.

3. The microgrid scheduling method according to claim 2, characterized in that, The quantum evolution includes: Among them, U QAOA (γ,β) is a quantum evolution, where γ and β are optimization parameters, and H cost is the problem Hamiltonian, is the problem Hamiltonian evolution, and H mixer is the hybrid Hamiltonian evolution, is the hybrid Hamiltonian evolution.

4. The microgrid scheduling method according to claim 3, wherein The problem Hamiltonian includes: Among them, is the cost of renewable energy power generation, is the local power generation, is the cost of energy storage degradation, is the discharge energy, is the continuous maintenance cost, is the operating state, τ K,t ·δ K,t is the imported energy cost function purchased from the main power grid, and is the quantum rotation term.

5. The microgrid scheduling method according to claim 3, wherein The hybrid Hamiltonian includes: Among them, is the Pauli X operator.

6. The microgrid scheduling method according to claim 2, characterized in that The step of, in each round of iteration, constructing an expected energy according to the candidate solutions and the problem Hamiltonian and adjusting the optimization parameters of the quantum circuit to minimize the expected energy includes: Among them, Θ opt is the optimization parameter, H cost |ψ(γ,β) is the problem Hamiltonian, and ψ(γ,β)| is the candidate solution.

7. The microgrid scheduling method according to claim 1, wherein The plurality of objective functions include a cost minimization function: Among them, is the cost of renewable energy power generation, is the power generation, is the cost of energy storage degradation, is the power generation energy, is the continuous maintenance cost, is the operating state, τ K,t ·δ K,t is the imported energy cost function purchased from the main power grid.

8. The microgrid scheduling method according to claim 1, characterized in that The plurality of objective functions include an energy self-sufficiency rate maximization function: Among them, is the local power generation volume, is the energy storage power generation volume, is the imported energy, is the total demand, is the adaptive weight.

9. The microgrid scheduling method according to claim 1, wherein, The plurality of objective functions include a reliability maximization function: Among them, is the local power generation volume, is the energy storage power generation volume, is the total demand, η l,t and ξ l,t are weights, is the imported energy, is the reserved energy.

10. The microgrid scheduling method according to claim 1, wherein The plurality of objective functions include a quantum optimization function: Among them, and are decision variables encoded as quantum states, and α p and β q are weights.

Images