Public building electric energy substitution equipment and micro-grid interactive frequency modulation method

Abstract

The invention relates to the technical field of power system frequency modulation, and discloses a public building electric energy substitution equipment and micro-grid interactive frequency modulation method, which comprises the following steps: firstly, acquiring equipment operation state data and micro-grid parameters in real time, then constructing a collaborative frequency modulation model based on a dynamic event driving mechanism, defining trigger conditions, and finally, establishing a collaborative frequency modulation model; and describing the coupling relationship between the equipment power response and the power grid frequency. A distributed robust optimization algorithm is designed, a frequency modulation problem is converted into a multi-objective optimization problem, and slack variables are introduced to process uncertainty constraints. And establishing a time-phased frequency modulation strategy, dividing windows according to load prediction, and dynamically adjusting equipment output. And defining physical and communication constraints to construct a mixed integer programming model, and solving by adopting an improved branch and bound algorithm. The method can effectively improve the frequency stability of the micro-grid, optimize the equipment resource configuration, reduce the operation cost, improve the frequency modulation precision and real-time performance, and promote the application of clean energy in the field of public buildings.

Description

Technical Field

[0001] The present invention relates to the technical field of power system frequency regulation, and particularly to an interactive frequency regulation method for electric energy substitution equipment in public buildings and a microgrid. Background Art

[0002] With the increasing global demand for clean energy, the microgrid, as a small power system integrating renewable energy, energy storage devices, and various types of electrical loads, plays an important role in the utilization of distributed energy. As an important place for electric energy consumption, public buildings introduce electric energy substitution equipment, such as electric heating and electric refrigeration equipment, which not only helps to improve energy utilization efficiency but also promotes the consumption of clean energy. However, the coordinated operation of electric energy substitution equipment in public buildings and the microgrid faces many challenges, among which the frequency regulation problem is particularly prominent.

[0003] The frequency stability of the microgrid is crucial for its safe and reliable operation. The intermittency and volatility of renewable energy, such as solar photovoltaic power generation affected by light intensity and wind power generation restricted by wind speed changes, make it difficult to maintain the power balance of the microgrid, resulting in frequency fluctuations. At the same time, the load characteristics of electric energy substitution equipment in public buildings are complex and variable, and their large-scale access further exacerbates the frequency instability problem of the microgrid. During peak electricity consumption periods, electric heating and electric refrigeration equipment operate intensively, and the load increases instantaneously. If not adjusted in time, the microgrid frequency will drop rapidly; while during low electricity consumption periods, the equipment load decreases, which may cause the frequency to rise.

[0004] Traditional frequency regulation methods are difficult to meet the requirements of interactive frequency regulation for electric energy substitution equipment in public buildings and the microgrid. Conventional frequency regulation strategies based on centralized control have problems of heavy communication burden and slow response speed when facing numerous scattered electric energy substitution equipment in the microgrid. Due to the increasing scale and complexity of the microgrid, the centralized control center needs to process a large amount of equipment operation data and grid parameter information, which not only places extremely high requirements on the bandwidth and stability of the communication network, but also the delay in data transmission and processing will lead to the lag in the issuance of frequency regulation instructions and the inability to respond promptly to rapid frequency changes.

[0005] In addition, existing frequency regulation technologies often do not fully consider the characteristics and operation constraints of electric energy substitution equipment. For example, factors such as the charge and discharge efficiency, charge and discharge rate limits, and life loss of energy storage devices are not effectively taken into account in traditional frequency regulation strategies. Unreasonable charge and discharge operations will not only reduce the service life of energy storage devices, increase operating costs, but also may affect their frequency regulation effect. At the same time, there are differences in the response speeds of different types of electric energy substitution equipment, and traditional methods are difficult to perform reasonable priority sorting and coordinated control according to equipment characteristics, resulting in the inability to optimize the allocation of frequency regulation resources.

[0006] In actual application scenarios, the electricity consumption demand of public buildings has obvious periodicity and uncertainty. The electricity consumption patterns on weekdays and holidays, as well as during the day and at night, vary greatly, which increases the difficulty of load forecasting for the microgrid. Inaccurate load forecasting will lead to a mismatch between the frequency regulation strategy and the actual demand, further affecting the frequency stability of the microgrid. Moreover, the problem of communication delay is widespread in the microgrid, which will cause the frequency regulation instructions to not be transmitted to the electric energy substitution equipment in a timely and accurate manner, reducing the accuracy and effectiveness of frequency regulation. Summary of the Invention

[0007] The purpose of the present invention is to provide a method for interactive frequency regulation between electric energy substitution equipment in public buildings and the microgrid to solve the problems raised in the above background technology.

[0008] To achieve the above purpose, the present invention provides the following technical solution: A method for interactive frequency regulation between electric energy substitution equipment in public buildings and the microgrid, the method comprising: Step S1: Real-time collect the operation status data of the electric energy substitution equipment in the public building and the grid parameters of the microgrid. The operation status data at least includes equipment power output, energy storage capacity, charge and discharge efficiency, and the grid parameters at least include frequency deviation, load demand, and renewable energy output; Step S2: Based on the dynamic event-driven mechanism, construct a cooperative frequency regulation model for the electric energy substitution equipment and the microgrid. Describe the coupling relationship between the equipment power response and the grid frequency through discrete-time dynamic equations, and define the trigger conditions for frequency regulation actions; Step S3: Design a distributed robust optimization algorithm to transform the frequency regulation problem into a multi-objective optimization problem. The optimization objectives include minimizing frequency deviation, equalizing equipment life loss, and optimizing the charge and discharge cost of energy storage, and introduce slack variables to handle uncertain constraints; Step S4: Establish a time-division frequency regulation strategy. Divide time windows according to the microgrid load forecasting results. Update the equipment output plan based on rolling horizon control within each window, and generate a dynamic frequency regulation instruction sequence; Step S5: Define the physical constraints and communication constraints of the equipment actions, including power output limits, energy storage charge and discharge rates, and communication delay compensation mechanisms, and construct a mixed-integer programming model to ensure the feasibility of the instructions; Step S6: Use an improved branch and bound algorithm to solve the mixed-integer programming model, combine a heuristic pruning strategy to reduce the search space, and output an optimized equipment output plan that meets the real-time frequency regulation requirements.

[0009] Preferably, in the step S2, the cooperative frequency regulation model further includes: Step S21: Define the trigger threshold for frequency regulation events. When the absolute value of the grid frequency deviation exceeds the preset threshold the equipment frequency regulation action is triggered; Step S22: Establish a discrete-time dynamic equation to describe the relationship between the device output and the frequency deviation. The formula is as follows:

[0010] Wherein, is the frequency deviation at the th moment, is the frequency deviation at the th moment, is the device output adjustment amount at the th moment, is the external disturbance at the th moment, , , are state transition matrices; Step S23: Verify the model convergence through Lyapunov stability analysis to ensure that the steady-state error of the frequency modulation command approaches zero.

[0011] Preferably, in the said step S3, the distributed robust optimization algorithm further includes: Step S31: Divide the devices into master nodes and slave nodes. The master nodes are responsible for global objective optimization, and the slave nodes execute local constraint satisfaction; Step S32: Use the alternating direction method of multipliers to decompose the optimization problem and achieve distributed solution by iteratively updating the dual variables between the master and slave nodes; Step S33: Introduce a robust optimization layer, use an ellipsoidal uncertainty set to describe the fluctuations of renewable energy output, and generate robust feasible solutions.

[0012] Preferably, in the said step S4, the time-division frequency modulation strategy further includes: Step S41: Predict the future time-period microgrid load curve based on a long short-term memory network and divide a rolling optimization window; Step S42: Within each window, dynamically adjust the device output priority according to the real-time frequency deviation. The priority is determined by weighting the device response speed and the remaining energy storage capacity; Step S43: When generating a frequency modulation command sequence, reserve a safety margin to cope with prediction errors. The size of the margin is inversely proportional to the load prediction confidence level.

[0013] Preferably, in the said step S5, the constraint conditions of the mixed-integer programming model include: Step S51: Constraints on the upper and lower limits of device output. The formula is as follows:

[0014] Wherein, is the The minimum allowable output value of a device, is the maximum allowable output value of the th device; Step S52: Energy storage charge and discharge state continuity constraint to ensure that there is no conflict in the charge and discharge modes in adjacent time periods; Step S53: Communication delay compensation constraint, using a time-delay differential equation to model the instruction transmission delay and correcting the delay impact through a prediction-correction mechanism.

[0015] Preferably, in the said step S6, the improved branch and bound algorithm further includes: Step S61: When initializing the branch tree, preferentially select the integer variable that has the greatest impact on the objective function for branching; Step S62: Adopt a tabu search strategy to avoid repeated access to invalid nodes and record the historical optimal solution to accelerate convergence; Step S63: During the pruning process, calculate the lower bound by combining the Lagrangian relaxation method, and prune the branch if the lower bound exceeds the current optimal solution.

[0016] Preferably, in the said step S23, the Lyapunov function is designed as:

[0017] where, is the Lyapunov function, and are positive definite weight matrices, and the matrix parameters are determined by solving a linear matrix inequality, is the transpose of the frequency deviation vector, is the transpose of the device output adjustment amount vector.

[0018] Preferably, in the said step S33, the mathematical expression of the ellipsoidal uncertainty set is:

[0019] where, is the ellipsoidal uncertainty set, is the covariance matrix, which is generated by fitting historical data.

[0020] Preferably, in the said step S41, the input features of the LSTM network include historical load data, weather information, and holiday flags, and the output is the load probability distribution for the future time period.

[0021] Preferably, in the said step S53, the form of the time-delay differential equation is:

[0022] where, is the device time constant, is the actual output value of the th device at time is the communication delay, is the frequency modulation command issued at time After being approximately discretized by Taylor expansion, it is embedded in the optimization model.

[0023] Compared with the prior art, the beneficial effects of the present invention are as follows: In terms of improving the frequency stability of the microgrid, by collecting the operation status data of the electric energy substitution devices in public buildings and the grid parameters of the microgrid in real time, the real-time operation status of the system can be accurately grasped. The cooperative frequency modulation model constructed based on the dynamic event-driven mechanism can quickly respond to the change of the grid frequency. When the absolute value of the grid frequency deviation exceeds the preset threshold, the frequency modulation action of the device is triggered in time, and the coupling relationship between the device power response and the grid frequency is described by the discrete-time dynamic equation, ensuring that the device makes reasonable power adjustment according to the frequency deviation, effectively suppressing the frequency fluctuation, and making the microgrid frequency stable within the allowable range, ensuring the safe and reliable operation of the microgrid. For example, when the output of renewable energy suddenly decreases or the load of public buildings suddenly increases, resulting in a frequency drop, the electric energy substitution device can quickly increase the power output to stabilize the frequency.

[0024] From the perspective of optimizing multiple objectives, the designed distributed robust optimization algorithm transforms the frequency modulation problem into a multi-objective optimization problem, pursuing the minimization of frequency deviation, the equalization of device life loss, and the optimization of energy storage charge and discharge costs at the same time. In terms of minimizing the frequency deviation, by accurately controlling the device output, the adverse effects of frequency fluctuation on the power system and electrical equipment are reduced, and the power quality is improved; for the equalization of device life loss, when the algorithm distributes the frequency modulation tasks, it fully considers the characteristics and operation status of each device, avoids overuse of some devices, makes the life loss of all devices relatively balanced, prolongs the overall service life of the devices, and reduces the device replacement and maintenance costs. For example, in multiple frequency modulation processes, the algorithm reasonably arranges the participation times and output sizes of different devices, avoiding frequent operation or long-term full-load operation of some devices; in terms of optimizing the energy storage charge and discharge costs, by optimizing the charge and discharge strategies of the energy storage device, combining the real-time electricity price information and system requirements, charging during the low electricity price period and discharging during the peak period, the energy storage charge and discharge costs are effectively reduced, and the economic benefits of the microgrid are improved.

[0025] The establishment of the time - segmented frequency modulation strategy also brings many benefits. By dividing the time window according to the micro - grid load prediction results and updating the device output plan based on the rolling - horizon control, it can better adapt to the dynamic changes of the micro - grid load. The long - short - term memory network predicts the micro - grid load curve in the future period, considering various factors such as historical load data, weather information, and holiday flags, improving the accuracy of load prediction. According to the real - time frequency deviation within each window, the priority of device output is dynamically adjusted, and a safety margin is reserved to cope with prediction errors, ensuring efficient frequency modulation under different load conditions. For example, during peak load periods, devices with fast response speed and large remaining energy storage capacity are preferentially arranged to participate in frequency modulation, and at the same time, enough safety margin is reserved to prevent insufficient frequency modulation caused by prediction errors, ensuring the stable operation of the micro - grid under complex load conditions.

[0026] In terms of dealing with constraint conditions, the defined physical constraints and communication constraints of device actions, as well as the constructed mixed - integer programming model, ensure the feasibility of instructions. The upper and lower limits of device output constraints ensure that the device operates within a safe range, avoiding damage to the device due to overload or under - load; the continuity constraint of energy storage charge - discharge status prevents conflicts in the charge - discharge modes of energy storage devices in adjacent periods, ensuring the normal operation and service life of energy storage devices; the communication delay compensation constraint models the instruction transmission delay through a time - delay differential equation and uses a prediction - correction mechanism to correct the delay impact, enabling the device to respond more accurately to frequency - modulation instructions, improving the accuracy and reliability of frequency modulation.

[0027] Finally, the improved branch - and - bound algorithm combined with the heuristic pruning strategy is used to solve the mixed - integer programming model, greatly improving the solution efficiency. When initializing the branch tree, the integer variable that has the greatest impact on the objective function is preferentially selected for branching, which can find the area close to the optimal solution faster; the tabu search strategy avoids repeated access to invalid nodes, records the historical optimal solution to accelerate convergence, reducing the calculation time; the Lagrangian relaxation method calculates the lower bound for pruning, effectively reducing the search space, enabling the algorithm to output an optimized device output plan that meets the real - time frequency - modulation requirements within a short time, meeting the requirements of micro - grid real - time frequency modulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is the working principle diagram of the interaction frequency - modulation method between the electric - energy substitution device of public buildings and the micro - grid described in the present invention; Figure 2 It is the design flow chart of the distributed robust optimization algorithm; Figure 3 It is the flow chart for establishing the time - segmented frequency - modulation strategy; Figure 4 It is the flow chart for constructing the constraints of the mixed - integer programming model. DETAILED DESCRIPTION OF THE INVENTION

[0029] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0030] Please refer to Figures 1-4 , the present invention provides a method for interactive frequency modulation between electric energy substitution equipment in public buildings and a microgrid, and its overall implementation scheme is as follows: Step S1: Data acquisition: Use a dedicated data acquisition device to collect the operation status data of the electric energy substitution equipment in the public building and the grid parameters of the microgrid in real time. For the electric energy substitution equipment, its operation status data at least covers aspects such as equipment power output, energy storage capacity, charge and discharge efficiency, etc. For example, install a power sensor on the equipment to obtain real-time power output data, use an energy storage monitoring system to master the real-time information of the energy storage capacity, and calculate the charge and discharge efficiency based on the charge and discharge records and energy conversion data of the equipment. The grid parameters of the microgrid at least include frequency deviation, load demand, and renewable energy output. The frequency deviation can be obtained by comparing the grid frequency monitor with the standard frequency; the load demand can be estimated based on the total power of various electrical equipment in the microgrid and historical electricity consumption data; the renewable energy output is obtained by monitoring the power generation of renewable energy power generation equipment such as wind turbines and solar panels.

[0031] Step S2: Build a collaborative frequency modulation model: Based on the dynamic event-driven mechanism, build a collaborative frequency modulation model for the electric energy substitution equipment and the microgrid. This model describes the coupling relationship between the equipment power response and the grid frequency through discrete-time dynamic equations, and at the same time defines the trigger conditions for frequency modulation actions. Specifically, when the grid frequency fluctuates abnormally, this model can quickly respond and drive the electric energy substitution equipment to adjust the power output, thereby realizing the regulation of the grid frequency.

[0032] Step S3: Design an optimization algorithm: Design a distributed robust optimization algorithm to transform the frequency modulation problem into a multi-objective optimization problem. In this process, set multiple optimization objectives, including minimizing the frequency deviation to ensure that the grid frequency is stable within a reasonable range; realizing the equalization of equipment life loss to extend the overall service life of the equipment; pursuing the optimization of energy storage charge and discharge costs to reduce the operation cost. And introduce slack variables to handle the uncertain constraints in actual operation, enhancing the adaptability and robustness of the algorithm.

[0033] Step S4: Establish a time - segmented frequency modulation strategy: Divide time windows according to the micro - grid load forecasting results, and formulate the frequency modulation strategy in a time - segmented manner. First, use load forecasting technology to estimate the micro - grid load in the future for a period of time, and then divide different time windows according to the forecasting results. Within each window, based on the method of rolling - horizon control, continuously update the device output plan, and generate a dynamic frequency modulation instruction sequence according to the real - time situation to better adapt to the dynamic changes of the micro - grid load.

[0034] Step S5: Define constraints and build a model: Clarify the physical constraints and communication constraints of device actions. For example, set power output limits to prevent damage to the power grid and the devices themselves caused by excessive or insufficient output of the devices; stipulate the charge - discharge rate of energy storage to ensure the safe and stable operation of energy storage devices; establish a communication delay compensation mechanism to overcome the delay problems generated during the communication process. Through these constraint conditions, build a mixed - integer programming model to ensure the practical feasibility of the generated frequency modulation instructions.

[0035] Step S6: Solve the model to obtain a solution: Use an improved branch - and - bound algorithm to solve the above - mentioned mixed - integer programming model. During the solving process, combine heuristic pruning strategies to reduce unnecessary search spaces and improve the solving efficiency. Finally, output an optimized device output plan that can meet the real - time frequency modulation requirements, and achieve the efficient interactive frequency modulation between the electric energy substitution devices in public buildings and the micro - grid.

[0036] The following further illustrates the implementation of the present invention in combination with Embodiments 1 to 5.

[0037] Embodiment 1: In this embodiment, the construction process of the collaborative frequency modulation model is further elaborated. First, clarify the triggering threshold of the frequency modulation event, and preset a threshold of the absolute value of the power grid frequency deviation . This threshold is determined according to the actual operation of the micro - grid and relevant standards, and it is a key indicator for triggering the device frequency modulation action. When the absolute value of the power grid frequency deviation exceeds the preset threshold , it means that the power grid frequency has fluctuated greatly, and at this time, immediately trigger the device frequency modulation action.

[0038] Then, establish a discrete - time dynamic equation to describe the relationship between device output and frequency deviation. The formula is . In this formula, represents the frequency deviation at the th moment, which reflects the degree to which the power grid frequency deviates from the standard frequency at this moment; is the frequency deviation at the th moment, used to record the frequency deviation state at the previous moment; is the The device output adjustment amount at a moment represents the power output adjustment value made by the device according to the change in the grid frequency; is the external disturbance at the moment, such as the impact of external factors such as sudden changes in renewable energy output and sudden access of high-power loads on the grid frequency; , , are state transition matrices, which are constant matrices determined according to factors such as the system characteristics of the microgrid, device parameters, and grid operating environment, and are used to describe the conversion relationship between system states.

[0039] To ensure that the steady-state error of the frequency modulation command approaches zero, the convergence of the model is verified through Lyapunov stability analysis. The Lyapunov function is designed as . Among them, is the Lyapunov function, which is a function used to measure the stability of the system; and are positive definite weight matrices. The parameters of these two matrices need to be determined by solving linear matrix inequalities. Their role is to weight the frequency deviation vector and the device output adjustment amount vector to reflect the influence degree of different state variables on the system stability. In actual operation, according to the specific parameters and operation requirements of the microgrid, relevant mathematical software or algorithms are used to solve the linear matrix inequalities to obtain appropriate and matrix parameters. Then, according to the Lyapunov stability theory, the model is analyzed and verified. If under certain conditions, the Lyapunov function gradually decreases over time and satisfies the relevant stability conditions, then it can be proved that the coordinated frequency modulation model is stable, that is, the steady-state error of the frequency modulation command will approach zero, thus ensuring the stable regulation of the grid frequency.

[0040] Example 2: This example details the implementation process of the distributed robust optimization algorithm. First, the devices are divided into master nodes and slave nodes. The master node undertakes the important task of global objective optimization in the whole system. It needs to comprehensively consider multiple objectives such as minimizing frequency deviation, equalizing device life loss, and optimizing the charge and discharge cost of energy storage, and coordinate and optimize the operation of the whole system from a macroscopic level. The slave node focuses on implementing local constraint satisfaction. According to the instructions issued by the master node and its own physical constraints and communication constraints, it adjusts the operation state of the device to ensure the normal operation of local devices and the overall stability of the system.

[0041] The alternating direction method of multipliers is used to decompose the optimization problem, and distributed solution is achieved by iteratively updating the dual variables between the master and slave nodes. During the iteration process, the master node calculates and updates the dual variables according to the overall operation condition and optimization objective of the system, and then transmits the relevant information to the slave nodes. After receiving the information from the master node, the slave nodes adjust the operation states of the devices in combination with their own local constraint conditions, and feedback the adjustment results to the master node. The master node updates the dual variables again according to the feedback from the slave nodes, and so on, until certain convergence conditions are met. This distributed solution method can not only make full use of the computing resources of each node to improve the computing efficiency of the algorithm, but also enhance the flexibility and scalability of the system to adapt to microgrid systems of different scales and complexities.

[0042] A robust optimization layer is introduced, and the ellipsoidal uncertainty set is used to describe the fluctuations of renewable energy output. The mathematical expression of the ellipsoidal uncertainty set is . Among them, is the ellipsoidal uncertainty set, which defines the range of fluctuations in renewable energy output; represents the fluctuation amount of renewable energy output at the th moment; is the covariance matrix, which is generated by fitting historical data. In practical applications, a large amount of historical renewable energy output data is collected, and the covariance matrix is calculated using statistical analysis methods. In this way, the uncertainty of renewable energy output can be described more accurately, enabling the optimization algorithm to generate robust feasible solutions, and ensuring the stable operation of the microgrid system and the achievement of the frequency regulation objective even when there are large fluctuations in renewable energy output.

[0043] Embodiment 3: This embodiment deeply explains the specific implementation details of the time-of-use frequency regulation strategy. First, based on the long short-term memory network (LSTM), the microgrid load curve for the future time periods is predicted. The input features of the LSTM network include historical load data, weather information, and holiday flags. Historical load data reflects the past electricity consumption patterns of the microgrid. Through learning historical data, the LSTM network can capture the trends and periodic characteristics of load changes. Weather information, such as temperature, humidity, and sunlight, has a significant impact on the load of the microgrid. For example, in hot summer, the extensive use of cooling equipment such as air conditioners will lead to an increase in load; while during the day with sufficient sunlight, the output of solar power generation equipment increases, and at the same time, the load of some electrical equipment may decrease accordingly. The holiday flag is used to distinguish normal working days from holidays, because during holidays, people's living and working patterns change, and the load characteristics of the microgrid will also be different. Through learning and analyzing these input features, the LSTM network outputs the future The load probability distribution during a time period provides an accurate basis for load forecasting for subsequent time - segmented frequency regulation strategies.

[0044] Divide the rolling optimization window according to the load forecasting results. The division of the rolling optimization window needs to comprehensively consider various factors, such as the frequency of load changes, the accuracy of forecasting, and the limitations of computing resources, etc. Generally speaking, for a micro - grid with relatively frequent load changes, a smaller rolling optimization window can be set to more timely track the load changes; while for a micro - grid with relatively stable load, the size of the rolling optimization window can be appropriately increased to reduce the amount of calculation. Within each window, dynamically adjust the device output priority according to the real - time frequency deviation. The device output priority is determined by weighting the device response speed and the remaining energy storage capacity. The device response speed reflects how quickly the device responds to the frequency regulation command. The faster the response speed of the device, the more quickly it can adjust the power output during frequency regulation, and the greater the role in stabilizing the grid frequency. The remaining energy storage capacity is related to the ability of the device to continuously provide frequency regulation support. The larger the remaining energy storage capacity, the more guaranteed the device is in subsequent frequency regulation. By weighting these two factors, the output priority of the device can be more reasonably determined, improving the frequency regulation efficiency.

[0045] When generating the frequency regulation command sequence, reserve a safety margin to cope with prediction errors. The size of the margin is inversely proportional to the load forecasting confidence level. The load forecasting confidence level reflects the reliability of the forecasting result. When the forecasting confidence level is high, it indicates that the forecasting result is relatively accurate, and at this time, the safety margin can be appropriately reduced; on the contrary, when the forecasting confidence level is low, in order to ensure the stable operation of the micro - grid under large prediction errors, the safety margin needs to be increased. For example, under certain special weather conditions, due to the large uncertainty of weather changes, the confidence level of load forecasting is reduced. At this time, a larger safety margin needs to be reserved to avoid grid frequency instability problems caused by prediction errors.

[0046] Example 4: This example elaborates in detail the constraint conditions of the mixed - integer programming model and its implementation process. First, the upper and lower limits of device output constraint, the formula is . In this formula, is the minimum allowable output value of the th device, which is determined according to the technical parameters of the device and the requirements of safe operation, ensuring that the device can still operate normally at the minimum output without causing damage to the device itself; is the maximum allowable output value of the th device, also determined by the performance and safety limitations of the device, preventing the device from over - outputting and causing device failures or adverse effects on the power grid; is the output value of the th device at the th moment; is a set of devices that includes all the electric energy substitution devices participating in the interactive frequency regulation of the microgrid. During actual operation, the output values of the devices are monitored in real time. Once it is found that they exceed the upper and lower limit ranges, corresponding adjustment measures are immediately taken, such as restricting the power output of the devices or increasing the operating power of the devices, to ensure that the device output is always within the allowable range.

[0047] The continuity constraint of the energy storage charge-discharge state ensures that there are no conflicts in the charge-discharge modes in adjacent time periods. Energy storage devices play an important regulatory role in the microgrid, and the reasonable control of their charge-discharge states is crucial for maintaining grid stability. In actual operation, by establishing the logical relationship and constraint conditions of the energy storage charge-discharge state, it is ensured that there are no contradictions in the charge-discharge modes of the energy storage devices in adjacent time periods. For example, if the energy storage device was in the charging state in the previous time period, then in the current time period, only when certain conditions are met (such as the energy storage capacity not reaching the upper limit, the grid frequency being stable, etc.), is it allowed to continue charging or switch to the discharging state; conversely, if it was in the discharging state in the previous time period, then in the current time period, it is necessary to decide whether to continue discharging or switch to the charging state based on factors such as the remaining energy storage capacity and the grid load demand.

[0048] The communication delay compensation constraint uses a time-delay differential equation to model the instruction transmission delay and corrects the delay impact through a prediction-correction mechanism. The form of the time-delay differential equation is . Among them, is the device time constant, which reflects the dynamic response characteristics of the device itself. Different types of devices have different time constants; is the actual output value of the th device at time ; is the communication delay, which is the time delay generated during the communication network transmission process; is the time when the frequency regulation instruction is issued. In actual applications, first, through the monitoring and analysis of the communication network, the value of the communication delay is obtained. Then, the time-delay differential equation is approximately discretized through Taylor expansion and embedded in the optimization model. After the device receives the frequency regulation instruction, according to the prediction-correction mechanism, combined with the device time constant and the communication delay , the instruction is corrected to compensate for the impact of the communication delay, ensuring that the device can respond to the frequency regulation instruction in a timely and accurate manner, and improving the frequency regulation accuracy and stability of the microgrid.

[0049] Example 5: This embodiment details the implementation process of the improved branch and bound algorithm. When initializing the branch tree, the integer variable that has the greatest impact on the objective function is preferentially selected for branching. In the mixed-integer programming model, there are multiple integer variables, and the values of these variables directly affect the value of the objective function. By analyzing and evaluating each integer variable, the variable that has the greatest impact on the objective function is determined. For example, when considering objectives such as minimizing frequency deviation, equalizing equipment life loss, and optimizing the charge and discharge cost of energy storage, integer variables such as the number of certain equipment put into use and the charge and discharge times of energy storage equipment may have a more significant impact on the objective function. Preferentially selecting these variables for branching can find the region close to the optimal solution faster during the search process, improving the search efficiency of the algorithm.

[0050] The tabu search strategy is adopted to avoid repeated visits to invalid nodes and record the historical optimal solution to accelerate convergence. During the search process, the tabu search strategy records the visited nodes by setting a tabu list. When a new node is searched, first check whether the node is in the tabu list. If it is in the tabu list, skip the node to avoid repeated visits, thereby reducing unnecessary computational effort. At the same time, record the historical optimal solution. During the search process, continuously update the historical optimal solution. Once it is found that the currently searched solution is better than the historical optimal solution, update the historical optimal solution. In this way, during the subsequent search process, the search direction can be guided according to the information of the historical optimal solution, accelerating the convergence speed of the algorithm and finding the optimal equipment output optimization plan that meets the real-time frequency regulation requirements faster.

[0051] During the pruning process, the Lagrangian relaxation method is combined to calculate the lower bound. If the lower bound exceeds the current optimal solution, then prune this branch. The Lagrangian relaxation method is a commonly used method for solving optimization problems. By relaxing the constraint conditions of the original problem, the original problem is transformed into a relaxation problem that is easier to solve. In the improved branch and bound algorithm, the Lagrangian relaxation method is used to calculate the lower bound of each branch. If the calculated lower bound exceeds the currently found optimal solution, it means that this branch cannot produce a better solution. At this time, this branch can be directly pruned without further searching, thus greatly reducing the search space and improving the solution efficiency of the algorithm. In this way, on the premise of ensuring the solution quality, the optimal equipment output optimization plan that meets the real-time frequency regulation requirements can be quickly found, realizing the efficient interactive frequency regulation between the electric energy substitution equipment in public buildings and the microgrid.

[0052] It should be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or elements inherent to such process, method, article or device.

[0053] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for interactive frequency modulation between an electric energy substitution device for public buildings and a microgrid, characterized in that, It includes the following steps: Step S1: Real-time collect the operation status data of the electric energy substitution equipment in the public building and the grid parameters of the microgrid. The operation status data at least includes equipment power output, energy storage capacity, charge-discharge efficiency, and the grid parameters at least include frequency deviation, load demand, and renewable energy output; Step S2: Based on the dynamic event-driven mechanism, construct a coordinated frequency regulation model for the electric energy substitution equipment and the microgrid, describe the coupling relationship between the equipment power response and the grid frequency through discrete-time dynamic equations, and define the trigger conditions for frequency regulation actions; Step S3: Design a distributed robust optimization algorithm, transform the frequency regulation problem into a multi-objective optimization problem, the optimization objectives include minimizing frequency deviation, equalizing equipment life loss, and optimizing the charge-discharge cost of energy storage, and introduce slack variables to handle uncertain constraints; Step S4: Establish a time-segmented frequency regulation strategy, divide time windows according to the microgrid load prediction results, update the equipment output plan based on rolling horizon control within each window, and generate a dynamic frequency regulation instruction sequence; Step S5: Define the physical constraints and communication constraints of equipment actions, including power output limits, energy storage charge-discharge rates, and communication delay compensation mechanisms, and construct a mixed-integer programming model to ensure the feasibility of instructions; Step S6: Use an improved branch and bound algorithm to solve the mixed-integer programming model, combine a heuristic pruning strategy to reduce the search space, and output an optimized equipment output plan that meets the real-time frequency regulation requirements.

2. The method for interactive frequency modulation between the electric energy substitution device of a public building and a microgrid according to claim 1, characterized in that In the said step S2, the coordinated frequency regulation model further includes: Step S21: Define the trigger threshold of the frequency modulation event. When the absolute value of the grid frequency deviation exceeds the preset threshold , trigger the frequency modulation action of the device; Step S22: Establish a discrete-time dynamic equation to describe the relationship between equipment output and frequency deviation, and the formula is: Among them, is the frequency deviation at the moment, is the frequency deviation at the moment, is the equipment output adjustment amount at the moment, is the external disturbance at the moment, , , are state transition matrices; Step S23: Verify the model convergence through Lyapunov stability analysis to ensure that the steady-state error of the frequency regulation instruction approaches zero.

3. The interactive frequency modulation method for the electric energy substitution equipment of public buildings and the microgrid according to claim 1, characterized in that In the said step S3, the distributed robust optimization algorithm further includes: Step S31: Divide the equipment into master nodes and slave nodes. The master nodes are responsible for global objective optimization, and the slave nodes execute local constraint satisfaction; Step S32: Use the alternating direction multiplier method to decompose the optimization problem, and achieve distributed solution by iteratively updating the dual variables between the master and slave nodes; Step S33: Introduce a robust optimization layer, use an ellipsoidal uncertainty set to describe the fluctuation of renewable energy output, and generate a robust feasible solution.

4. The interactive frequency modulation method for the electric energy substitution equipment of public buildings and the microgrid according to claim 1, characterized in that, In the said step S4, the time-segmented frequency regulation strategy further includes: Step S41: Predict the microgrid load curve for the future time period based on the long short-term memory network, and divide the rolling optimization window; Step S42: Within each window, dynamically adjust the equipment output priority according to the real-time frequency deviation, and the priority is determined by the weighted sum of the equipment response speed and the remaining energy storage capacity; Step S43: When generating the frequency regulation instruction sequence, reserve a safety margin to cope with prediction errors, and the size of the margin is inversely proportional to the load prediction confidence level.

5. The interactive frequency modulation method for the electric energy substitution equipment of public buildings and the microgrid according to claim 1, characterized in that, In the said step S5, the constraint conditions of the mixed-integer programming model include: Step S51: Constraints on the upper and lower limits of equipment output, and the formula is: wherein, is the minimum allowable output value of the th device, is the maximum allowable output value of the th device, is a set of devices; Step S52: Continuity constraints on the energy storage charge-discharge state to ensure that there is no conflict in the charge-discharge mode in adjacent periods; Step S53: Communication delay compensation constraints, use a time-delay differential equation to model the instruction transmission delay, and correct the delay impact through a prediction-correction mechanism.

6. The method for interactive frequency modulation between the electric energy substitution equipment of public buildings and the microgrid according to claim 1, characterized in that, In the step S6, the improved branch and bound algorithm further includes: Step S61: When initializing the branch tree, preferentially select the integer variable that has the greatest impact on the objective function for branching; Step S62: Adopt a tabu search strategy to avoid repeated visits to invalid nodes and record the historical optimal solution to accelerate convergence; Step S63: During the pruning process, calculate the lower bound by combining the Lagrangian relaxation method. If the lower bound exceeds the current optimal solution, prune this branch.

7. The method for interactive frequency modulation between the electric energy substitution device for public buildings and the microgrid according to claim 2, characterized in that In the step S23, the Lyapunov function is designed as: Among them, is the Lyapunov function, and are positive definite weight matrices, and the matrix parameters are determined by solving linear matrix inequalities, is the transpose of the frequency deviation vector, is the transpose of the device output adjustment amount vector.

8. The method for interactive frequency modulation between the electric energy substitution equipment of public buildings and the microgrid according to claim 3, characterized in that In the step S33, the mathematical expression of the ellipsoidal uncertainty set is: Among them, is an ellipsoidal uncertainty set, is the covariance matrix, which is generated by fitting historical data.

9. The interactive frequency modulation method for the electric energy substitution equipment of public buildings and the microgrid according to claim 4, characterized in that, In the step S41, the input features of the LSTM network include historical load data, weather information, and holiday flags, and the output is the load probability distribution for the future time period.

10. The method for interactive frequency modulation between the electric energy substitution device of public buildings and the microgrid according to claim 5, wherein In the step S53, the form of the delay differential equation is: Among them, is the device time constant, is the th actual output value of the th device at time is the communication delay, is the frequency modulation command issued at time , which is discretized by Taylor expansion approximation and embedded in the optimization model.

Images