Flexible power peak regulation control method based on air conditioner load resources
By employing a gradient descent optimization algorithm and a temperature boundary constraint-based air conditioning cluster control method, the correlation between air conditioning load control and power grid peak shaving is resolved. This method achieves coordinated optimization of the air conditioning cluster, enabling rapid response to power grid demands and ensuring user comfort and power grid safety.
Patent Information
- Application Number
- CN202511650154.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-12
- Publication Date
- 2026-02-10
AI Technical Summary
Existing air conditioning load control technologies fail to effectively link with grid peak-shaving demands, resulting in rigid control strategies that are difficult to adapt to rapid changes in grid load and lack collaborative optimization in large-scale air conditioning cluster scenarios.
A gradient descent optimization algorithm is used to construct an air conditioning cluster control framework. Combining temperature boundary constraints and dynamic user comfort, the temperature adjustment of the air conditioning cluster is optimized through a loss function to achieve coordinated optimization of air conditioning load and power grid.
It achieves coordinated optimization of air conditioning load and power grid peak shaving, quickly smooths out fluctuations in renewable energy output, ensures user comfort, reduces computing consumption, adapts to real-time power grid demands, and improves cluster collaboration.
Smart Images
Figure CN121507700A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of air conditioner control, more particularly, the present application relates to a flexible power peak shaving control method based on air conditioner load resources. BACKGROUND
[0002] With the proportion of new energy power generation in the power system rising year by year, it has become the core force of energy structure transformation. However, the output of new energy is significantly affected by natural conditions. Photovoltaic power generation depends on sunshine, and wind power depends on wind speed. The strong randomness and intermittency make it difficult to predict the output of new energy and the accuracy of prediction, which not only restricts the large-scale development of new energy, but also forces the power grid to reserve a large amount of standby capacity to deal with output fluctuations, greatly increasing the overall operation cost of the system.
[0003] At the same time, the load side is undergoing profound changes, and new resources such as distributed power, electric vehicles and flexible load are widely connected, which poses two major challenges to power grid operation and dispatching: on the one hand, the pressure of new energy consumption is increasing, and the instability of output makes it difficult to balance power supply and demand; on the other hand, the peak-valley difference of load is becoming increasingly prominent, among which the explosive growth of air conditioning load is particularly critical. In the summer cooling peak period, air conditioning load can cause the instantaneous load of the power grid to increase sharply, further exacerbating the temporal and spatial imbalance of power supply and demand, and posing a severe test to the safe and stable operation of the power grid.
[0004] To solve the above problems, flexible load regulation technology is considered as a key path to realize source-load cooperation, and building air conditioning load has become an ideal regulation object for power grid flexible dispatching due to its obvious advantages: first, air conditioning load accounts for a large part of the total energy consumption of commercial buildings, and has great adjustment potential; second, air conditioning has obvious thermal inertia characteristics, and the room temperature can still be maintained for half an hour after the air conditioner is turned off, providing a buffer space for load regulation. Currently, research on air conditioning load regulation mainly focuses on two directions: one is equipment layer modeling, which analyzes the start-stop characteristics of air conditioner compressors, room temperature dynamic response curves, etc., to build a detailed physical model to predict the power change of a single machine; the other is energy-saving optimization control, which proposes a hierarchical control based on cooling load demand, periodic suspension strategy, etc., aiming to reduce the energy consumption of air conditioning systems.
[0005] However, from the actual needs of power grid dispatching, the existing research has obvious shortcomings: First, the target is limited. The existing scheme pays excessive attention to the energy saving and consumption reduction of the building part, fails to establish direct correlation with the core demand of grid peak filling, new energy fluctuation smoothing and power transmission congestion mitigation, and is out of touch with the actual demand of the grid. Second, the strategy is rigid. The control method mainly depends on the rule of fixed temperature threshold deviation, and cannot be dynamically adjusted according to the real-time dispatching instruction of the grid, so it is difficult to adapt to the rapid change of the grid load. Third, the cluster coordination is missing. The existing model is mainly designed for a single air conditioner, and it is difficult to be extended to a large-scale air conditioner cluster scenario, and it cannot solve the problem of differentiated adjustment and collaborative optimization of different air conditioners in the cluster, resulting in scattered control effect. SUMMARY
[0006] To overcome the existing defects, the embodiments of the present application provide a flexible power peak shaving control method based on air conditioner load resources, which minimizes the power deviation and comfort loss by using gradient descent optimization algorithm, and combines the temperature boundary constraint for air conditioner cluster control, thereby solving the problems of weak correlation with grid peak shaving demand, path dependence on fixed rules for traditional air conditioner load control and lack of cluster adjustment coordination.
[0007] A flexible power peak shaving control method based on air conditioner load resources, comprising the following steps: obtaining air conditioner cluster data, calculating the user acceptable temperature deviation of the i-th air conditioner; constructing the loss function of the whole air conditioner cluster according to the temperature deviation; calculating the partial derivative of the loss function with respect to the set temperature adjustment amount of the i-th air conditioner based on the loss function; iteratively updating the value of the set temperature adjustment amount of the i-th air conditioner according to the partial derivative, wherein the temperature boundary protection rule is used to ensure that the equipment temperature is within a safe range during the iteration process, and the iteration is stopped when the set iteration termination condition is met; calculating the power-temperature linear correlation coefficient of the i-th air conditioner.
[0008] In a preferred embodiment, the air conditioner cluster data is obtained, and the user acceptable temperature deviation of the i-th air conditioner is calculated, which comprises: using the obtained air conditioner cluster data to calculate the user acceptable temperature deviation of the i-th air conditioner by using the normalization method.
[0009] In a preferred embodiment, the loss function of the whole air conditioner cluster is constructed according to the temperature deviation, which comprises: taking the ratio of the set temperature adjustment amount of the i-th air conditioner to the user acceptable temperature deviation as the comfort degree of the air conditioner; according to the preset comfort degree interval where the comfort degree is located, the corresponding comfort degree weight is allocated to the air conditioner; according to the comfort degree, the comfort degree weight and the preset power deviation weight, and combining the deviation between the current whole air conditioner cluster power and the target power, the loss function of the whole air conditioner cluster is constructed.
[0010] In a preferred embodiment, the partial derivative of the loss function with respect to the set temperature adjustment amount of the i-th air conditioner comprises: S1: initializing the set temperature adjustment amount of each air conditioner in the air conditioner cluster; S2: entering an iterative loop, and in each iteration, the partial derivative of the loss function with respect to the set temperature adjustment amount of the i-th air conditioner is calculated; S3: based on the partial derivative value and a preset gradient descent rule, the temperature adjustment amount of each air conditioner is updated synchronously; S4: the total power of the air conditioner cluster in the current iteration is calculated according to the updated temperature adjustment amount, and the deviation between the total power and the peak shaving target power is checked; S5: whether the iteration termination condition is met is judged, the iteration termination condition comprises that the power deviation reaches a preset tolerance range or the iteration number reaches a preset maximum value; if yes, the iteration is stopped and the current set temperature adjustment amount of each air conditioner is output as an optimal solution; otherwise, the next iteration is returned to step S2.
[0011] In a preferred embodiment, the partial derivative of the loss function with respect to the set temperature adjustment amount of the i-th air conditioner further comprises: applying a temperature boundary protection rule to the updated set temperature adjustment amount of each air conditioner to ensure that the set temperature of the air conditioner is within a safe temperature range of the equipment and does not exceed a user acceptable temperature deviation.
[0012] In a preferred embodiment, the power-temperature linear correlation coefficient of the i-th air conditioner is calculated, specifically: the actual running power and the optimal temperature adjustment amount of each air conditioner after iteration optimization are collected; based on the collected multiple sets of actual running power and optimal temperature adjustment amount data, the power-temperature linear correlation coefficient of the i-th air conditioner is recalculated through linear fitting, and is updated to the regulation and control model parameter library.
[0013] The technical effects and advantages of the flexible power peak shaving control method based on air conditioner load resources of the present application are as follows: The present application realizes the collaborative optimization of air conditioner load and power grid peak shaving by constructing an air conditioner cluster regulation and control framework integrating gradient descent optimization algorithm and combining dynamic user comfort constraint mechanism. The technical effects are that different power grid states are adapted relying on dynamic weight coefficients, temperature boundary protection is used to avoid large fluctuations in room temperature, the total power of the cluster is quantitatively matched with the peak shaving target, the system response is accelerated, the cluster collaborative effect is improved, and the calculation consumption is reduced. It can quickly stabilize new energy output fluctuations, support efficient consumption of new energy, guarantee user comfort, and the algorithm is lightweight and can run on edge devices, balancing power grid safe and economic operation and practicality of actual operation. BRIEF DESCRIPTION OF DRAWINGS
[0014] Figure 1 A flowchart of a flexible power peak shaving control method based on air conditioner load resources is provided for the embodiments of the present application. DETAILED DESCRIPTION
[0015] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be clearly and completely described in the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.
[0016] Embodiment 1, Figure 1 A flexible power peak regulation control method based on air conditioner load resources is given, comprising the following steps: S21, acquiring air conditioner cluster data, and calculating a user acceptable temperature offset of an i th air conditioner; S22, constructing a loss function of the whole air conditioner cluster according to the temperature offset; S23, calculating a partial derivative of the loss function to a set temperature adjustment amount of the i th air conditioner based on the loss function; S24, iteratively updating a value of the set temperature adjustment amount of the i th air conditioner according to the partial derivative, wherein a temperature boundary protection rule is used to ensure that the equipment temperature is in a safe range during the iteration process, and the iteration is stopped when a set iteration termination condition is met; S25, calculating a power-temperature linear correlation coefficient of the i th air conditioner.
[0017] The embodiment realizes collaborative optimization of air conditioner load and power grid peak regulation demand by constructing an air conditioner cluster flexible peak regulation framework integrating gradient descent optimization algorithm and combining a dynamic user comfort constraint mechanism. The technical effect lies in adapting to different power grid operation states by relying on a dynamic weight coefficient, avoiding large room temperature fluctuations by using temperature boundary protection logic, and quantitatively matching the total power of the air conditioner cluster and the peak regulation target, thereby solving the problems of response lag, insufficient cluster collaboration and high consumption of computing resources in traditional regulation methods. The method can quickly stabilize new energy output fluctuations, support efficient consumption of new energy, guarantee user comfort, and the lightweight algorithm can run on edge devices such as intelligent gateways, providing a practical solution for safe and economic operation of the power grid, and taking into account regulation accuracy and actual operation feasibility.
[0018] S21, acquiring air conditioner cluster data, and calculating a user acceptable temperature offset of an i th air conditioner.
[0019] In the embodiment, the difference between the ambient temperature and the set temperature in the air conditioner cluster data is determined as follows: The difference between the set temperature and the ambient temperature of the air conditioner is obtained by calculating the difference between the ambient temperature and the set temperature of each air conditioner and taking the absolute value. .
[0020] In the embodiment, the user acceptable temperature offset of the i th air conditioner is specifically calculated as follows:
[0021] wherein, is an offset coefficient, The greater the acceptable offset is also greater, which is conducive to the power regulation of the power grid; The smaller the acceptable offset is also smaller, which is conducive to improving user comfort. Default 0.5, is the basic acceptable offset, which is generally set to 2℃. is the preset maximum possible temperature difference, which is generally set to 10℃ in summer and winter and 8℃ in spring and autumn, is the user acceptable temperature offset of the i-th air conditioner, is the temperature difference between the set temperature of the i-th air conditioner and the ambient temperature, the maximum temperature adjustment range without obvious discomfort of the user, and an association account of a single air conditioner and the acceptable offset is established.
[0022] S22, according to the temperature offset, constructing a loss function of the air conditioner cluster as a whole.
[0023] In this embodiment, the loss function of the air conditioner cluster as a whole is constructed, and the specific formula is:
[0024] In the formula, is the loss function, is the power offset weight, is the comfort weight, is the current total power of the air conditioner cluster, is the target power to be adjusted to, is the set temperature adjustment amount of the i-th air conditioner, is the user acceptable temperature offset of the i-th air conditioner, and n is the total number of air conditioners.
[0025] is a weight coefficient, which can be dynamically adjusted according to the grid consumption; It is biased to meet the grid consumption demand, and when it is increased, the deviation between the total power and the target power will be reduced more quickly, allowing a larger temperature adjustment range (which will sacrifice part of the comfort), and when it is reduced It will reduce the punishment for power deviation, prioritize user comfort, and limit the temperature adjustment range. is the comfort weight, and when it is increased, the set temperature adjustment amount of each air conditioner will be more strictly displayed , ensuring that the temperature fluctuation is within the acceptable range of the user; when When it is reduced, the restriction on temperature adjustment will be relaxed, allowing a larger range of temperature offset to respond to the grid demand. For example, when the grid consumption demand is large: ; flat section time: ; new energy output fluctuation stage: ; office time: .
[0026] S23, based on the loss function, calculate the partial derivative of the loss function to the set temperature adjustment amount of the i-th air conditioner; S24, according to the partial derivative, iteratively update the value of the set temperature adjustment amount of the i-th air conditioner, wherein in the iteration process, the temperature boundary protection rule is used to ensure that the device temperature is in a safe range, and when the set iteration termination condition is met, the iteration is stopped.
[0027] In this embodiment, the partial derivative of the loss function to the set temperature adjustment amount of the i-th air conditioner is calculated, and the specific formula is:
[0028] In the formula, is the partial derivative of the loss function to the set temperature adjustment amount of the i-th air conditioner, is the partial derivative of the target power to the set temperature adjustment amount of the i-th air conditioner, is the power deviation shift weight, is the comfort weight, is the current overall power of the air conditioner cluster, is the target power that needs to be adjusted to, is the set temperature adjustment amount of the i-th air conditioner, is the user acceptable temperature offset of the i-th air conditioner.
[0029] In this embodiment, the temperature boundary protection rule is represented by the following formula:
[0030] Wherein:
[0031] In the formula, is the new set temperature of the i-th air conditioner, is the initial temperature of the i-th air conditioner, is the set temperature adjustment amount of the i-th air conditioner, is the user acceptable temperature offset of the i-th air conditioner, is the minimum design working temperature of the i-th air conditioner, is the maximum design working temperature of the i-th air conditioner.
[0032] In this embodiment, the set iteration termination condition is met when or the maximum iteration number is reached to terminate the iteration, wherein This represents the current overall power of the air conditioning cluster. To achieve the target power that needs to be adjusted, The iteration terminates when the learning rate is reached or the maximum number of iterations (100) is reached.
[0033] S25, calculate the power-temperature linear correlation coefficient of the i-th air conditioner.
[0034] In this embodiment, the calculation of the adjusted total power of the air conditioning cluster needs to be timed: After the temperature adjustment command is issued, wait 10-15 minutes before collecting the current power of the air conditioner. Calculate the total power only after the air conditioner compressor and fan are running stably to avoid power data deviations caused by thermal inertia.
[0035] In this embodiment, the calculation of the power-temperature linear correlation coefficient of the i-th air conditioner specifically involves: The k-value is calculated by taking the temperature difference between the overall temperature of the air conditioning cluster and the temperature before the temperature adjustment command was issued 10-15 minutes later, and the power difference before and after the command was issued within the same time period. The ratio of the actual temperature difference to the actual power difference is the calculated k-value. The specific formula is as follows:
[0036] Let i be the set temperature adjustment amount for the i-th air conditioner. This represents the actual total power difference. Let be the power-temperature linear correlation coefficient of the i-th air conditioner.
[0037] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.
[0038] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.
[0039] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0040] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.
[0041] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
[0042] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A flexible power peak-shaving control method based on air conditioning load resources, characterized in that, Includes the following steps: Acquire air conditioning cluster data and calculate the user-acceptable temperature offset of the i-th air conditioner; Based on the temperature offset, a loss function for the entire air conditioning cluster is constructed; Based on the loss function, calculate the partial derivative of the loss function with respect to the set temperature adjustment amount of the i-th air conditioner; Based on the partial derivative, the value of the set temperature adjustment of the i-th air conditioner is iteratively updated. During the iteration process, the temperature boundary protection rule ensures that the equipment temperature is within a safe range. When the set iteration termination condition is met, the iteration stops. Calculate the power-temperature linear correlation coefficient for the i-th air conditioner.
2. The flexible power peak-shaving control method based on air conditioning load resources according to claim 1, characterized in that, The step of acquiring air conditioning cluster data and calculating the user-acceptable temperature offset of the i-th air conditioner includes: Using the acquired air conditioning cluster data, the user-acceptable temperature offset of the i-th air conditioner is calculated on a per-unit basis using a normalization method.
3. The flexible power peak-shaving control method based on air conditioning load resources according to claim 2, characterized in that, Based on the temperature offset, a loss function for the entire air conditioning cluster is constructed, including: The comfort level of the i-th air conditioner is defined as the ratio of the set temperature adjustment amount to the user's acceptable temperature deviation. Based on the preset comfort range in which the comfort level falls, assign a corresponding comfort weight to the air conditioner; Based on comfort level, comfort weight, and preset power deviation shift weight, and combined with the deviation between the current overall power of the air conditioning cluster and the target power, a loss function for the entire air conditioning cluster is constructed.
4. The flexible power peak-shaving control method based on air conditioning load resources according to claim 3, characterized in that, The calculation of the partial derivative of the loss function with respect to the set temperature adjustment of the i-th air conditioner includes: S1: The set temperature adjustment amount for each air conditioner in the initial air conditioning cluster; S2: Enter the iteration loop. In each iteration, calculate the partial derivative of the loss function with respect to the set temperature adjustment of the i-th air conditioner. S3: Based on the partial derivative values and the preset gradient descent rules, the temperature adjustment amount of each air conditioner is updated synchronously; S4: Based on the updated temperature adjustment, calculate the total power of the air conditioning cluster in the current iteration round, and verify the deviation between the total power and the peak-shaving target power; S5: Determine whether the iteration termination condition is met. The iteration termination condition includes the power deviation reaching a preset tolerance range or the number of iterations reaching a preset maximum value. If met, stop the iteration and output the current set temperature adjustment amount of each air conditioner as the optimal solution. Otherwise, return to step S2 for the next round of iteration.
5. The flexible power peak-shaving control method based on air conditioning load resources according to claim 4, characterized in that, The above also includes: For the updated temperature adjustment settings of each air conditioner, apply temperature boundary protection rules to ensure that the set temperature of the air conditioner is within the safe temperature range of the equipment and does not exceed the user's acceptable temperature deviation.
6. The flexible power peak-shaving control method based on air conditioning load resources according to claim 5, characterized in that, The calculation of the power-temperature linear correlation coefficient for the i-th air conditioner is specifically as follows: Collect the actual operating power and optimal temperature adjustment of each air conditioner after iterative optimization; Based on the collected data of multiple sets of actual operating power and optimal temperature adjustment, the power-temperature linear correlation coefficient of the i-th air conditioner is recalculated through linear fitting and updated to the control model parameter library.
7. The flexible power peak-shaving control method based on air conditioning load resources according to claim 6, characterized in that, The expression for the loss function is: In the formula, For loss function, As the power deviation shift weight, For comfort weighting, This represents the overall power of the current air conditioning cluster. To achieve the target power that needs to be adjusted, Let i be the set temperature adjustment amount for the i-th air conditioner. Let be the user-acceptable temperature offset for the i-th air conditioner, and n be the total number of air conditioners.
8. The flexible power peak-shaving control method based on air conditioning load resources according to claim 7, characterized in that, The partial derivative of the loss function with respect to the set temperature adjustment of the i-th air conditioner is specifically formulated as follows: In the formula, Let be the partial derivative of the loss function with respect to the set temperature adjustment of the i-th air conditioner. Let be the partial derivative of the target power with respect to the set temperature adjustment of the i-th air conditioner. As the power deviation shift weight, For comfort weighting, This represents the overall power of the current air conditioning cluster. To achieve the target power that needs to be adjusted, Let i be the set temperature adjustment amount for the i-th air conditioner. Let be the user-acceptable temperature offset for the i-th air conditioner.
9. The flexible power peak-shaving control method based on air conditioning load resources according to claim 8, characterized in that, The temperature boundary protection rule is expressed by the following formula: in: In the formula, Set the new temperature for the i-th air conditioner. Let be the initial temperature of the i-th air conditioner. Let i be the set temperature adjustment amount for the i-th air conditioner. Let be the user-acceptable temperature offset for the i-th air conditioner. Let i be the minimum design operating temperature for the i-th air conditioner. Let i be the maximum design operating temperature of the i-th air conditioner.
10. The flexible power peak-shaving control method based on air conditioning load resources according to claim 8, characterized in that, The specific iterative termination condition that is satisfied is as follows: when The iteration terminates when the maximum number of iterations is reached, where, This represents the overall power of the current air conditioning cluster. To achieve the target power that needs to be adjusted, The iteration is terminated when the learning rate is reached or the maximum number of iterations is reached.