Urban rail station air conditioner load layered optimization control method oriented to regeneration energy consumption and considering start-stop constraint

Abstract

The invention relates to a regenerated energy consumption-oriented urban rail station air conditioner load layered optimization control method considering start-stop constraints, and belongs to the field of urban rail station air conditioner load regulation and control optimization. The method comprises the steps that a three-layer optimization control framework of an aggregation layer, a cluster layer and an equipment layer is designed so that reliable regulation and control of station air conditioner loads can be achieved under the condition that communication resources are limited, and the framework identifies an air conditioner equivalent operation period through a moving average method on the cluster layer and linearly speculates the normal operation state of the air conditioner; a two-stage optimization method considering the air conditioner shortest start-stop time constraint is provided, in the first stage, in a cluster layer, the response mechanism of each air conditioner is generated by taking maximum absorption of regenerative braking energy as the target; in the second stage, the exclusive regulation and control plans of all the air conditioners are screened out on the aggregation layer with the aim of stabilizing the outsourcing electric power peak-valley difference as the target. According to the method, a reliable, economical and easy-to-implement solution can be provided for participation of the urban rail station air conditioner load in demand response.

Description

Technical Field

[0001] This invention belongs to the field of air conditioning load regulation and optimization in urban rail transit stations, and specifically relates to a hierarchical optimization control method for air conditioning load in urban rail transit stations that takes into account start-stop constraints and is oriented towards the consumption of renewable energy. Background Technology

[0002] The power and lighting system is a high-energy-consuming system in urban rail transit operations, with its energy consumption mainly coming from purchased grid power. It can also absorb regenerative braking energy generated by the traction power supply system. However, due to the lack of real-time optimization of the traction power supply system's power flow, some regenerative braking energy cannot be effectively utilized and is fed back to the grid, resulting in economic waste and potential power quality issues. Air conditioning, as the largest energy-consuming load in the power and lighting system, has the potential to convert surplus regenerative braking energy into heat storage. While ensuring station environmental comfort, it possesses significant flexible adjustment characteristics, making it a high-quality station-level demand response resource. Against this backdrop, how to accurately model and coordinate the station air conditioning load to efficiently absorb regenerative braking energy and reduce the system's purchased power costs and power fluctuations has become an important issue for improving the energy efficiency and operational economy of urban rail transit systems. While existing research has made progress in air conditioning load modeling and control, most studies focus on civil or commercial scenarios and often fail to fully consider actual engineering constraints and communication privacy issues, lacking reliable control schemes for the megawatt-level regenerative braking energy absorption needs of urban rail transit stations. Therefore, developing a hierarchical optimization control method that can take into account air conditioning operation constraints, communication efficiency, and economy is of great engineering significance for promoting the green and low-carbon transformation of urban rail transit.

[0003] The shortcomings of existing technologies are as follows:

[0004] (1) When the air conditioning load aggregation and control is based solely on a fixed mode, the resulting control strategy may fail in actual operation. When the regenerative braking energy fluctuates drastically or passenger flow changes abruptly, it may lead to insufficient response or excessive action of the air conditioning group control, thereby affecting the thermal comfort of the station and causing additional energy consumption. In addition, existing control methods have obvious limitations when dealing with large-scale air conditioning clusters. For example, although the centralized control method can achieve global optimization, it has problems such as high communication pressure and high risk of privacy leakage. While the decentralized control method reduces the communication burden, it is difficult to guarantee the overall optimization effect due to the lack of coordination mechanism, and it cannot effectively balance control accuracy and system reliability.

[0005] (2) In terms of constraint handling and engineering applications, existing research does not adequately consider the key physical constraints in the actual operation of air conditioners, especially ignoring the minimum start-stop time limit of the air conditioner compressor, which makes it impossible to execute control commands in actual equipment. At the same time, the applicability of existing methods in specific urban rail scenarios is flawed. Most studies focus on frequency regulation and peak shaving applications of household or commercial air conditioners, and lack dedicated control strategies for the fluctuation characteristics of megawatt-level regenerative braking energy in urban rail stations, making it difficult for control schemes to achieve economical and reliable energy consumption in actual engineering. Summary of the Invention

[0006] The purpose of this invention is to provide a hierarchical optimization control method for air conditioning load in urban rail transit stations, taking into account start-stop constraints and considering renewable energy consumption, comprising:

[0007] (1) A three-layer optimization control framework of "aggregation layer-cluster layer-equipment layer" is designed to achieve reliable regulation of station air conditioning load under the condition of limited communication resources. In the cluster layer, the equivalent operating cycle of air conditioning is identified by the moving average method, and its normal operating state is linearly predicted, thus laying the foundation for subsequent optimization without the need for continuous information reporting from the equipment layer. Compared with centralized control, the method of this invention significantly reduces communication pressure and privacy risks; compared with decentralized control, it effectively avoids the problems of insufficient or excessive response through the hierarchical coordination mechanism.

[0008] (2) A two-stage optimization method considering the minimum start-up and shutdown time constraint of air conditioners is proposed. The first stage aims to maximize the absorption of regenerative braking energy at the cluster layer and generates the response mechanism for each air conditioner. The second stage aims to smooth the peak-valley difference of purchased power at the aggregation layer and selects the exclusive control plan for each air conditioner. This strategy achieves the dual goals of regenerative energy absorption and economical operation through hierarchical optimization. For this model, two control strategies based on the relationship between trigger time and state transition time are designed and solved using mixed integer linear programming. While strictly satisfying the physical constraints of the equipment, the number of air conditioner switching times is effectively reduced, improving the feasibility and engineering practicality of the control.

[0009] To achieve the above objectives, the technical solution of the present invention is: a hierarchical optimization control method for air conditioning load in urban rail transit stations, considering start-stop constraints and oriented towards renewable energy consumption, comprising:

[0010] A three-layer optimization control framework consisting of aggregation layer, cluster layer, and equipment layer is designed to achieve reliable regulation of station air conditioning load under conditions of limited communication resources. In the cluster layer, the equivalent operating cycle of air conditioning is identified by the moving average method, and its normal operating state is linearly predicted.

[0011] A two-stage optimization method is proposed, taking into account the constraint of the shortest start-stop time of air conditioners. The first stage aims to maximize the absorption of regenerative braking energy at the cluster layer and generate the response mechanism of each air conditioner. The second stage aims to smooth the peak-valley difference of purchased power at the aggregation layer and select the exclusive control plan for each air conditioner.

[0012] Compared to existing technologies, this invention offers the following advantages: The hierarchical optimization control framework constructed in this invention, considering the constraint of the shortest start-stop time for air conditioning, identifies the equivalent cycle of air conditioning and predicts its operating status online using the moving average method. This achieves effective control and coordination of large-scale air conditioning cluster operation information with almost no increase in communication costs. The designed two-stage optimization method aims to maximize the absorption of regenerative braking energy at the cluster level and to smooth fluctuations in purchased power at the aggregation level, simultaneously optimizing energy utilization efficiency and grid interaction friendliness. While strictly meeting equipment physical constraints, it achieves an operating scheme that combines high regenerative energy utilization and low grid power fluctuations. The proposed two control strategies and dedicated control plan generation mechanism can flexibly adjust the triggering time of air conditioning according to actual operating needs, effectively reducing the number of air conditioning on / off cycles while ensuring passenger comfort. This provides a reliable, economical, and easy-to-implement solution for air conditioning load participation in demand response in urban rail stations. Attached Figure Description

[0013] Figure 1 Optimize the control framework for air conditioning load stratification;

[0014] Figure 2 The operating characteristics of a single air conditioner;

[0015] Figure 3 It is the moving average method;

[0016] Figure 4 Inferred air conditioning operation information for the cluster layer;

[0017] Figure 5 To generate a reference adjustment signal;

[0018] Figure 6 The air conditioner's operating time is segmented when it is initially turned on;

[0019] Figure 7 Two control strategies are provided for the initial start-up state of the air conditioner;

[0020] Figure 8 The air conditioner's operating time is segmented when it is initially turned off;

[0021] Figure 9 Two control strategies for the initial off state of the air conditioner

[0022] Figure 10 To regulate the air conditioner's operating status, a reference adjustment signal is tracked.

[0023] Figure 11 The control and operation status of the i-th air conditioner in station j;

[0024] Figure 12 This represents the initial operating state of the air conditioner.

[0025] Figure 13 This is the initial operating temperature of the air conditioner;

[0026] Figure 14 This refers to the operating information of the air conditioner during normal operation.

[0027] Figure 15 This is the power curve of the station's air conditioning load during normal operation;

[0028] Figure 16 Moving average series Standard deviation;

[0029] Figure 17 This is the normal operating temperature of the air conditioner;

[0030] Figure 18 To regulate the operating temperature of the air conditioner;

[0031] Figure 19 To regulate the power curve of the air conditioning load during operation;

[0032] Figure 20 The percentage of air conditioners that are turned on;

[0033] Figure 21 This provides operational information for a single air conditioner before and after control.

[0034] Figure 22 This provides operational information for a single air conditioner before and after load control using different methods.

[0035] Figure 23 This represents the sum of the real-time power of air conditioning loads at all stations.

[0036] Figure 24 The air conditioning load of all stations is supplied by the power grid in real time. Detailed Implementation

[0037] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings.

[0038] This invention proposes a hierarchical optimization control method for air conditioning load in urban rail transit stations, taking into account start-stop constraints and considering renewable energy consumption, comprising:

[0039] A three-layer optimization control framework consisting of aggregation layer, cluster layer, and equipment layer is designed to achieve reliable regulation of station air conditioning load under conditions of limited communication resources. In the cluster layer, the equivalent operating cycle of air conditioning is identified by the moving average method, and its normal operating state is linearly predicted.

[0040] A two-stage optimization method is proposed, taking into account the constraint of the shortest start-stop time of air conditioners. The first stage aims to maximize the absorption of regenerative braking energy at the cluster layer and generate the response mechanism of each air conditioner. The second stage aims to smooth the peak-valley difference of purchased power at the aggregation layer and select the exclusive control plan for each air conditioner.

[0041] The following is a detailed implementation process of the present invention.

[0042] This invention provides a hierarchical optimization control method for air conditioning load in urban rail transit stations, taking into account start-stop constraints and considering renewable energy consumption. The method includes:

[0043] (1) A three-layer optimization control framework of "aggregation layer-cluster layer-equipment layer" is designed to achieve reliable regulation of station air conditioning load under the condition of limited communication resources. In the cluster layer, the equivalent operating cycle of air conditioning is identified by the moving average method, and its normal operating state is linearly predicted, thus laying the foundation for subsequent optimization without the need for continuous information reporting from the equipment layer. Compared with centralized control, this method significantly reduces communication pressure and privacy risks; compared with decentralized control, it effectively avoids the problems of insufficient or excessive response through the hierarchical coordination mechanism.

[0044] (2) A two-stage optimization method considering the minimum start-up and shutdown time constraint of air conditioners is proposed. The first stage aims to maximize the absorption of regenerative braking energy at the cluster layer and generates the response mechanism for each air conditioner. The second stage aims to smooth the peak-valley difference of purchased power at the aggregation layer and selects the exclusive control plan for each air conditioner. This strategy achieves the dual goals of regenerative energy absorption and economical operation through hierarchical optimization. For this model, two control strategies based on the relationship between trigger time and state transition time are designed and solved using mixed integer linear programming. While strictly satisfying the physical constraints of the equipment, the number of air conditioner switching times is effectively reduced, improving the feasibility and engineering practicality of the control.

[0045] 1. Air Conditioning Load Layering Optimization Control Framework

[0046] This invention proposes a three-layer optimized control framework of "aggregation layer - cluster layer - equipment layer" suitable for the air conditioning load of urban rail stations, such as... Figure 1 As shown, the proposed method enhances the reliability and feasibility of real-world scenarios by strengthening the responsibilities of each layer and improving the coordination between adjacent layers.

[0047] The control architecture consists of three layers. The cluster layer (station level) identifies the cyclical characteristics of air conditioning loads and, combined with fluctuating regenerative braking energy, generates and reports multiple demand response mechanisms for each air conditioner. The aggregation layer (system level) optimizes and generates a dedicated control plan for each air conditioner based on the mechanisms of each station, aiming to smooth out the peak-valley difference in total power grid supply, thus avoiding grid impact and penalties. The equipment layer (air conditioner level) strictly executes this plan, only needing to report the initial operating status. Through the collaboration of "station coordination - global optimization - terminal execution," the system maximizes the utilization of regenerative energy while ensuring stable and economical power grid operation.

[0048] 1.1 Cluster Layer: Response Mechanism Based on Regenerative Braking Energy

[0049] Based on the station's total air conditioning load curve, the equivalent cycle parameters of a single air conditioner are identified using the moving average method, and its normal operating status is calculated. Next, regenerative braking energy is converted into a reference control signal. Then, based on the time elapsed between the air conditioner's operating state and the planned switching point, two control strategies are designed: a "segmented dual-control strategy in active state" and a "segmented single-control strategy in closed state." Finally, by controlling the air conditioner's operation to track the regenerative braking energy signal, an air conditioning response mechanism that maximizes the utilization of this energy is constructed.

[0050] 1.2 Speculation on the normal operation of the air conditioner

[0051] The operating characteristics of a single air conditioner on the equipment floor are as follows: Figure 2 As shown.

[0052] In the picture, This indicates the difference between the set maximum and minimum operating temperatures of the air conditioner. Indicates the shortest time the air conditioner will be turned off; Indicates the minimum operating time of the air conditioner; Indicates the rated power of the air conditioning load; This indicates the time the air conditioner is turned off within one operating cycle; This indicates the operating time of the air conditioner within one operating cycle. and They can be calculated using the following formulas:

[0053] (1)

[0054] (2)

[0055] In the formula, Indicates the operating cycle of the air conditioner; This indicates the duty cycle of the air conditioner. Meanwhile, from equation (1), it can be seen that... and They can be represented as:

[0056] (3)

[0057] (4)

[0058] To simplify air conditioning load control, it is assumed that all controlled air conditioners have the same operating cycle and duty cycle. This is reasonable because: the air conditioners involved in the control follow the same temperature control scheme, hence their cycles and duty cycles are consistent; furthermore, the air conditioners in urban rail stations are of uniform model and similar batches, resulting in minimal actual differences. The cluster layer can obtain duty cycle information in advance.

[0059] Since air conditioners have the same operating cycle, the period of the aggregated power curve of multiple air conditioners is the same as the period of the power curve of a single air conditioner. Assume the station... The equivalent operating cycle of the indoor air conditioning is During the air conditioner's running time The memory contains:

[0060] (5)

[0061] In the formula, Indicates the station during normal operation The aggregated power curve of the indoor air conditioning load. To identify the equivalent operating cycle of the air conditioning, a moving average method is used, such as... Figure 3 As shown.

[0062] Figure 3 middle, Indicates power; Indicates time; Indicates station Inner Power curve of an air conditioner during normal operation. (Window size set to...) The window moves during the air conditioner's operating time, and calculations are performed. The average value at which the movement begins The average values ​​at different starting points of the moving average are recorded in the moving average series. As shown in the following formula:

[0063] (6)

[0064] (7)

[0065] As the window size gradually approaches the equivalent operating cycle of the air conditioner, the moving average... The moving average sequence will become consistent. The standard deviation will decrease. Therefore, the window size that minimizes the standard deviation of the moving average series is the equivalent operating period of the air conditioner.

[0066] (8)

[0067] The cluster layer uses the initial operating status of the air conditioner provided by the device layer. and initial operating temperature By combining the equivalent cycle parameters of the air conditioners identified using the moving average method with the known duty cycle information, the station's... Inner The normal operating information of the air conditioner is as follows: when the air conditioner's initial operating status is "on", that is... hour:

[0068] (9)

[0069] (10)

[0070] (11)

[0071] (12)

[0072] (13)

[0073] In the formula, , , , , Indicates the moment of state transition; and They represent the stations respectively. Inner air conditioner The operating status and operating temperature at any given time; and These represent the inferred first number of cluster layers. The normal operating status and normal operating temperature of the air conditioner.

[0074] When the air conditioner is initially in the off state, that is hour, and It is calculated by the following formula:

[0075] (14)

[0076] (15)

[0077] (16)

[0078] Based on the method described above, the cluster layer infers the normal operating information of a certain air conditioner at the device layer, as follows: Figure 4 As shown.

[0079] Figure 4 middle, and These represent the maximum and minimum values ​​within the passenger's comfortable temperature range, respectively. Figure 4 It can be seen that the deviation generated by the linear inference of the cluster layer on the operating information of the air conditioner during normal operation is within an acceptable range, because the operating temperature of the air conditioner changes slowly, and its exponential temperature trajectory can be approximated by a straight line.

[0080] 1.3 Reference Regulation Signal Generation Method Based on Regenerative Braking Energy

[0081] Addressing the megawatt-level regenerative braking energy generated by the traction power supply system, this invention considers the fairness issue of controlled air conditioning load and proposes a reasonable air conditioning load regulation strategy to efficiently absorb the excess capacity of the station's total air conditioning load. This portion of energy. On the one hand, air conditioning is a level 3 load, while lighting systems and other equipment are level 1 loads, so the drastically fluctuating regenerative braking energy is preferentially used for the air conditioning load; on the other hand, some of the reversed regenerative braking energy is several times greater than... When the control of air conditioning load is not considered, it is much greater than The reverse energy feedback not only wastes energy but can also cause severe voltage fluctuations in the power grid supply system, threatening the safe operation of the urban rail system; furthermore, when the regenerative braking energy is less than... At this time, since the regenerative braking energy fed back in is relatively small and absorbing this energy would incur high control costs, the absorption of this energy is not considered. The specific process for generating the reference control signal for formulating the air conditioning load response mechanism is as follows: Figure 5 As shown.

[0082] First, obtain the station information in the traction power supply system. Corresponding traction station The actual regenerative braking energy returned; then, the regenerative braking energy less than [amount missing] is discarded. The part is then used to calculate the regenerative braking energy available for the air conditioning load; finally, the available regenerative braking energy is converted into a square wave signal with an amplitude of 1.

[0083] 1.4 Design methods for two control strategies

[0084] To avoid excessive switching frequency, increased costs, and damage to equipment lifespan caused by air conditioning control, this invention designs two control strategies based on the time between the control trigger moment and the state reversal moment.

[0085] 1) When the air conditioner's initial operating state is "on", that is... hour:

[0086] First, considering the characteristics of the energy fed back from the traction power supply system—namely, its strong fluctuations and regular intermittent nature—the operating time of the air conditioner is determined. Divided into The concept is that within each time period, the air conditioner triggers a control strategy and then returns to its original operating state after the control is activated. Figure 6 As shown.

[0087] Then, in the first Within a certain time range, based on the air conditioning trigger control time Distance to state transition time Length of time Two control strategies are implemented, such as Figure 7 As shown.

[0088] when At that time, the air conditioner was in the first... Within a certain period of time and Two trigger control moments. When At that time, due to The value is relatively small, and the maximum operating temperature of the air conditioner slightly exceeds the set maximum operating temperature, but it will not affect passenger comfort. At that time, the air conditioning load was at the first Only can be executed within a certain time period. Figure 7 (a) shows the control strategy ( Figure 7 (a) indicates the start-up segmented dual-control strategy: at this time, the air conditioner still has a relatively long time before it naturally shuts off (greater than the minimum start-up time constraint). This invention studies the air conditioner's operating time. It is taken from a portion of the long-term operation time of the air conditioner, therefore the period before the first time range and the second After a certain period of time, the minimum start-stop time constraint of the air conditioner is not considered.

[0089] 2) When the air conditioner's initial operating state is off, i.e. hour:

[0090] The method of segmenting the air conditioner's operating time and the air conditioner load control strategy at this time are similar to the method of segmenting the operating time and the control strategy when the air conditioner is initially turned on. See [link to relevant documentation]. Figure 7 (b) shows the control strategy ( Figure 7 (b) represents the closed-state segmented single-control strategy: at this time, the time before the air conditioner naturally shuts off is very short (less than the minimum start time constraint), as shown below. Figure 8 and Figure 9 As shown.

[0091] 1.5 Establish an air conditioning load response mechanism to absorb regenerative braking energy.

[0092] In this invention, when designing the air conditioning load response mechanism at the cluster layer, a single unit independently tracks the reference signal, eliminating the need for information interaction between units and enabling parallel computing. This significantly reduces the computational complexity of optimizing and controlling large-scale air conditioning loads.

[0093] Cluster layer control station Inner The first air conditioner The operating status within a certain time range is configured to track the reference adjustment signal generated by regenerative braking energy, such as... Figure 10 As shown.

[0094] when Figure 10 In the diagram, when the difference between the area of ​​the reference signal within the solid black box and the area within the dashed black box is the largest, the effect of adjusting the air conditioning operation state to track the reference adjustment signal is the best, indicating that in the first... To maximize the utilization of regenerative braking energy for air conditioning load within a certain time frame. When the air conditioner's initial operating state is "on," that is... At that time, the first The specific regulation model within a certain time frame is as follows:

[0095] (17)

[0096] In the formula, This indicates the effect of the cluster layer in controlling the air conditioning operation status by tracking the reference control signal. It is dimensionless, and the larger the value, the better. Indicates the first A set of triggering control moments within a certain time range; Indicate execution Figure 7 The effect of the control strategy in (a) is as follows: ; Indicate execution Figure 7 The effect of the control strategy in (b) is as follows: The specific calculation is as follows:

[0097] (18)

[0098] (19)

[0099] In the formula, This represents the reference adjustment signal. For the first time interval, the minimum on-time constraint does not need to be considered, i.e.:

[0100] (20)

[0101] For the After a certain time frame, the minimum shutdown time constraint no longer needs to be considered, i.e.:

[0102] (twenty one)

[0103] Similarly, when the air conditioner's initial operating state is off, that is... At that time, the first In specific control models within a certain time range and The calculation method is as follows:

[0104] (twenty two)

[0105] (twenty three)

[0106] In summary, for the station Inner The first air conditioner The triggering and control time for a given time period is calculated using equations (17) to (19) or equations (17), (22) to (23). By combining the triggering and control times of the air conditioner within a certain time range, a set of air conditioner triggering and control times is obtained. :

[0107] (twenty four)

[0108] For air conditioning load The clustering layer sends the mechanism for all air conditioning loads at all stations to the aggregation layer to participate in demand response. For the aggregation layer, the clustering layer infers the first... Air conditioner in normal operating condition It is known.

[0109] 2. Polymer layer: Smoothing the peak-to-valley difference in the power curve

[0110] The aggregation layer receives the demand response mechanism from the cluster layer, which involves all air conditioning loads at all stations. Its goal is to smooth out peak-valley differences in the power supply from the grid to all air conditioning loads. Among various air conditioning load response mechanisms, it selects the most suitable one for the equipment layer's air conditioning loads to participate in demand response—that is, a dedicated control plan for the air conditioning loads. The specific process is as follows:

[0111] First, define the aggregation layer optimization scheduling model:

[0112] 1) Objective function

[0113] (25)

[0114] In the formula, This indicates the effect of the cluster layer in smoothing out the peak-valley difference in the energy supplied by the grid to all air conditioning loads; This represents the set of dedicated control plans for all air conditioning loads at all stations in the urban rail transit system. This indicates the number of stations in the urban rail transit system. Expressing the search function Standard deviation; Indicates station The sum of all air conditioning control and operating states within the unit; Indicates station The number of air conditioners installed inside; Indicates station Reference adjustment signal for the indoor air conditioning.

[0115] 2) Constraints of dedicated control plans for air conditioning load

[0116] (26)

[0117] (27)

[0118] (28)

[0119] In the formula, Indicates station A collection of dedicated control plans for all air conditioning loads within the facility; Indicates station Inner A dedicated control plan for Taiwan's air conditioning load; Indicates station Inner Air conditioner The timing for triggering regulation within a short period of time; This indicates the stations sent from the cluster layer to the aggregation layer. Inner Taiwan's air conditioning load response mechanism, by formula Calculated.

[0120] 3) Constraints on the operation status of air conditioning load regulation

[0121] (29)

[0122] In the formula Indicates station Inner The operating status of the air conditioning unit. (Based on the station's...) Inner Initial operating status of the air conditioner And the air conditioner Number of control moments triggered within a short period of time , There are four possible values ​​for , such as Figure 11 As shown.

[0123] Figure 11 middle, Stations inferred from the cluster layer Inner The normal operating status of the air conditioner is determined by the type or Calculated; Indicates station Inner air conditioner The constant control and operation status; and They represent the stations respectively. Inner Air conditioner The first and second trigger control moments within a certain time range; Indicates station Inner Air conditioner State reversal time within a short period of time , by formula or Calculated. Station Inner Air conditioner Triggering time within a short period of time distance Length of time It can be represented as:

[0124] (30)

[0125] Finally, the aggregation layer obtains the station through optimization. Inner Taiwan's dedicated control plan for air conditioning load And send it to the cluster layer and the device layer. All trigger control times in the dedicated control plan are state transition times. After receiving the dedicated control plan for the air conditioning load sent by the aggregation layer, the cluster layer executes the control plan using the methods shown in equations (10), (11) or (15), (16). Figure 7 or Figure 9 The control strategy shown is a linear prediction of the station. Inner Operating status of air conditioner after load regulation and operating temperature .

[0126] 3. Equipment layer: Execute dedicated control plans.

[0127] When the air conditioning load at the equipment layer participates in demand response regulation, it only needs to send the initial operating status and initial operating temperature of each air conditioner to the cluster layer to achieve effective regulation of the air conditioning load without excessively increasing communication costs. When regulating the air conditioning load, the aggregation layer and cluster layer are more concerned with the operating status of the air conditioners than the power of the air conditioning load. To ensure low communication cost operation, the equipment layer does not need to provide the cluster layer with the power information of the air conditioning load.

[0128] After receiving the dedicated control plan for the air conditioning load sent by the aggregation layer, the equipment layer strictly executes the control plan, using the method shown in equation (10) or (15). Figure 7 or Figure 9 The control strategy shown is used to set up stations. Inner The actual operating status of the air conditioner after adjustment And calculate the actual operating temperature after air conditioning control. .

[0129] Case Analysis

[0130] This invention uses a planned subway line as an example to verify the reliability and feasibility of the proposed air conditioning load stratification control method and two regulation strategies. Compared with global optimization based on each air conditioner (involving hundreds of thousands or even millions of variables), the method of this invention optimizes a slightly larger number of variables than the air conditioner's operating cycle, significantly reducing complexity. The established two-stage regulation model is a mixed-integer linear programming problem, which can be efficiently solved using a genetic algorithm in each stage. The example line has a departure interval of 120 s, an operating cycle of 5439 s, and a total of 23 stations. The number of air conditioners at each station is shown in Table 1.

[0131] Table 1 Number of installed units at each station

[0132]

[0133] To cope with varying passenger volumes, stations need to install a corresponding number of air conditioners, and passenger flow differs significantly between stations. Especially at high-traffic transfer stations, a large number of air conditioners are required to ensure a comfortable waiting environment. The physical parameters of each air conditioner are shown in Table 2.

[0134] Table 2 Physical parameters of air conditioners

[0135]

[0136] Considering the heterogeneity of air conditioning, assume the power of all air conditioning loads within the stations of this line. The system follows a normal distribution with a mean of 10kW and a standard deviation of 0.004. Assume that at the initial moment, the number of air conditioners turned on and the ratio of the number of turned-on air conditioners to the total number of air conditioning units installed in the station are as follows: Figure 12 As shown.

[0137] At the initial moment, the operating temperatures of each air conditioner and the temperature distribution range of the air conditioners in each station are as follows: Figure 13 As shown.

[0138] During normal operation, the operating temperature and status of all air conditioners in all stations are as follows: Figure 14 As shown, the power curves of the air conditioning load at each station are as follows: Figure 15 As shown. To facilitate observation of the temperature changes of a specific air conditioner, in Figure 14 The center is indicated by a solid purple-red line.

[0139] According to equations (1) to (4), the actual operating cycle of the air conditioner is 2044 s, of which 583 s is on and 1461 s is off. The moving average method is used to identify... Figure 15 The equivalent period of the air conditioning load power curve at the central station, when the window is set to 2044 s, results in the smallest standard deviation of the moving average series (only 0.0017), see [reference needed]. Figure 16 The results show that the equivalent operating cycle of the air conditioner is indeed 2044 s, and the moving average method can accurately identify this cycle.

[0140] After accurately identifying the air conditioner's operating cycle, the cluster layer, based on the initial state and temperature information provided by the equipment layer, linearly calculates the air conditioner's normal operating information using the methods shown in equations (10), (11) or (15), (16), as follows: Figure 17 As shown.

[0141] For the regenerative braking energy of the traction power supply system, reference control signals for each air conditioner are generated using the method shown. Subsequently, based on a comparison between the time from the air conditioner trigger control moment to the state reversal moment and the shortest start-stop time of the air conditioner, two control strategies are adopted: one at the cluster level aiming to maximize the utilization of regenerative braking energy, and the other at the aggregation level aiming to minimize fluctuations in purchased power. These strategies are used to control the air conditioner load, resulting in the following: Figure 18 The air conditioning control operating temperature curve shown is as follows: Figure 19 The air conditioning load power curves for each station are shown. Figure 20 The comparison shows the actual operating status of the air conditioner at the equipment level and the inferred operating status at the cluster level before and after the air conditioner is regulated.

[0142] Specifically, for the changes in the operating information of a single air conditioner before and after control, one can refer to the changes in the operating information of the 43rd air conditioner in the first station before and after control, such as... Figure 21 As shown.

[0143] Figure 21 In the analysis, the predicted normal operating temperature, operating status, and control operating status of the equipment layer by the cluster layer are consistent with the actual values. However, the predicted control operating temperature deviates significantly from the actual value. This deviation stems from the nonlinear characteristics of the air conditioning operating temperature: as shown in equations (3) and (4), when lowering the same temperature, the higher the initial temperature, the shorter the time required; when raising the same temperature, the higher the initial temperature, the longer the time required. Under conditions of low or no communication, the cluster layer uses a linear method to predict the operating status and temperature of the equipment layer in order to effectively control the air conditioning load. Although the predicted control temperature deviates, the actual air conditioning temperature after control only slightly exceeds the preset limit, which is still within the passenger comfort range and does not affect the travel experience. Therefore, the deviation is within an acceptable range.

[0144] Load regulation can also be achieved by adjusting the air conditioner's highest or lowest operating temperature. This invention provides five temperature adjustment schemes, each adjusting the temperature control range upwards (increasing the highest temperature and decreasing the lowest temperature). , , , , And adjust the air conditioning load according to the plan.

[0145] Taking the 43rd air conditioner at the first station as an example, its operating information before and after adjustment is as follows: Figure 22 As shown in the figure. The results show that all different control methods can meet the premise of minimizing the number of times the air conditioner is switched on and off, but the method of the present invention minimizes the deviation between the controlled operating state and the normal operating state, and has a smaller impact on the service life of the air conditioner.

[0146] After adjusting the air conditioning load, the sum of the real-time power of the air conditioning load of all stations is as follows: Figure 23 As shown, the energy sourced from the grid is as follows: Figure 24 As shown.

[0147] Depend on Figure 23 It can be seen that the method proposed in this invention, compared with the temperature regulation method, is better able to smooth the power curve fluctuation of the station's air conditioning load, which meets the optimization goal of the aggregation layer. Figure 24 As can be seen, the method proposed in this invention can reduce the purchased electricity cost of air conditioning load compared with the temperature regulation method, which meets the optimization goal of the cluster layer. Specific energy information is shown in Table 3.

[0148] Table 3 Energy Information for Each Component During Air Conditioning Load Control

[0149]

[0150] In Table 3, This indicates the total energy consumption of the air conditioning load. ; This indicates that the air conditioning load originates from the power grid. ; This indicates the regenerative braking energy that has been utilized by the air conditioning load. ; This indicates the utilization rate of regenerative braking energy by the air conditioning load; This indicates the rate at which the air conditioning load is reduced from the power source supplied by the grid compared to normal operation. This represents the peak-to-valley difference in the energy supplied by the power grid to the air conditioning load.

[0151] Table 3 shows that when using temperature regulation to control the air conditioning load, the higher the adjusted temperature value, the more fully the air conditioning load utilizes regenerative braking energy, and the flatter the external power curve of the air conditioning load. However, due to the nonlinear change in the operating temperature of the air conditioner, the deviation between the cluster layer and the equipment layer is larger, leading to a maximum increase of 1.06% in the total energy consumption of the air conditioning load. When the adjusted temperature value is 0.8 ℃ or 1 ℃, the increase in the utilization of regenerative braking energy by the air conditioning load is less than the increase in the total energy consumption of the air conditioning load, resulting in an increase in the energy source from the grid after the air conditioning load is regulated. A value less than 0 does not meet the optimization goals of the cluster layer.

[0152] By using the method of this invention to regulate air conditioning load, the energy obtained by the air conditioning load from the grid can be reduced by 4.95% (432 kW·h) with almost no change in total energy consumption, the utilization rate of regenerative braking energy can be increased by 17%, and the standard deviation of grid power supply can be reduced by 33%. The regulation results fully meet the two-stage optimization objectives of the cluster layer and the aggregation layer.

[0153] Compared to the method of this invention, the temperature regulation method has limited control capabilities in demand response, and the communication pressure between the cluster layer and the device layer is relatively high because it requires changing the temperature set range. The method of this invention, even under low communication conditions, can still reliably absorb the regenerative braking energy of the traction system from the air conditioning load and effectively smooth out peak-valley fluctuations in grid power supply.

[0154] The above are preferred embodiments of the present invention. Any changes made to the technical solution of the present invention that do not exceed the scope of the technical solution of the present invention shall fall within the protection scope of the present invention.

Claims

1. A hierarchical optimization control method for air conditioning load in urban rail transit stations, considering start-stop constraints and oriented towards renewable energy consumption, characterized in that... include: A three-layer optimization control framework consisting of aggregation layer, cluster layer, and equipment layer is designed to achieve reliable regulation of station air conditioning load under conditions of limited communication resources. In the cluster layer, the equivalent operating cycle of air conditioning is identified by the moving average method, and its normal operating state is linearly predicted. A two-stage optimization method is proposed, taking into account the constraint of the shortest start-stop time of the air conditioner. In the first stage, the response mechanism of each air conditioner is generated at the cluster level with the goal of maximizing the absorption of regenerative braking energy. The second phase, at the aggregation layer, aims to smooth out the peak-valley difference in purchased power and selects dedicated control plans for each air conditioner.

2. The method for hierarchical optimization control of air conditioning load in urban rail transit stations considering start-stop constraints and oriented towards renewable energy consumption, as described in claim 1, is characterized in that... In the three-layer optimization control framework of aggregation layer-cluster layer-equipment layer, the cluster layer, i.e. the station level, identifies the cyclical characteristics of air conditioning load and, combined with the fluctuating regenerative braking energy, generates and reports multiple demand response mechanisms for each air conditioner. The aggregation layer, or system level, optimizes and generates a dedicated control plan for each air conditioner based on the mechanisms of each station, with the goal of smoothing the peak-valley difference in the total power supply of the power grid, so as to avoid power grid impact and penalties. The equipment layer, or air conditioner level, executes the dedicated control plan for each air conditioner generated by the aggregation layer and reports the initial operating status.

3. The method for hierarchical optimization control of air conditioning load in urban rail transit stations considering start-stop constraints and oriented towards renewable energy consumption, as described in claim 1 or 2, is characterized in that... The cluster layer is implemented as follows: First, based on the station's total air conditioning load curve, the equivalent cycle parameters of individual air conditioners are identified using the moving average method, and their normal operating status is calculated. Specifically, the cluster layer calculates the initial operating status of the air conditioners based on the information provided by the equipment layer. and initial operating temperature , This indicates that the initial operating state of the air conditioner is "on," and vice versa. Combining the equivalent period parameters of the air conditioner obtained through the moving average method with the known duty cycle information, we can infer the station's operating status. Inner The air conditioner is in its initial operating state. Normal operating information below; Secondly, the regenerative braking energy is converted into a reference adjustment signal; specifically, the station's traction power supply system is obtained. Corresponding traction station The actual regenerative braking energy returned; the regenerative braking energy discarded is less than the total installed capacity of the station's air conditioning load. The available regenerative braking energy for the air conditioning load is calculated; the available regenerative braking energy is then converted into a square wave signal with an amplitude of 1. Next, based on the time elapsed between the air conditioner's operating status and the planned switching point, two control strategies were designed: "start-up segmented dual-control strategy" and "closed-up segmented single-control strategy". Ultimately, by regulating the operation of the air conditioner to track the regenerative braking energy signal, an air conditioner response mechanism that maximizes the utilization of the regenerative braking energy signal is constructed.

4. The method for hierarchical optimization control of air conditioning load in urban rail transit stations considering start-stop constraints and oriented towards renewable energy consumption, as described in claim 3, is characterized in that... By regulating the operation of the air conditioner to track the regenerative braking energy signal, an air conditioner response mechanism that maximizes the utilization of the regenerative braking energy signal is constructed, specifically as follows: Cluster layer control station Inner The first air conditioner The operating status within a certain time range is used to track the reference adjustment signal formed by regenerative braking energy; when the initial operating state of the air conditioner is "on", that is... At that time, the first The specific regulation model within a certain time frame is as follows: In the formula, This indicates the effectiveness of the cluster layer in controlling the air conditioning operation status by tracking the reference adjustment signal; a larger value is better. Indicates the first A set of trigger points for regulation within a certain time frame. and For the air conditioner in the first Two triggering control moments exist within a certain time frame. ; This indicates the control effect when the air conditioner's initial operating state is "on" using the segmented dual-control strategy. , Indicates the minimum operating time of the air conditioner; This indicates the control effect when the air conditioner's initial operating state is "on" using a closed-state segmented single-control strategy. , Indicates the first Within a certain time range, the timing of air conditioner trigger control is adjusted. Distance to state transition time The duration is calculated using the following formula: In the formula, Indicates the reference adjustment signal; This indicates the time the air conditioner is turned off within one operating cycle; This indicates the operating time of the air conditioner within one operating cycle; This indicates the shortest time the air conditioner will turn off; for the first time interval, there is no need to consider the shortest time-on constraint, i.e.: For the After a certain time frame, the minimum shutdown time constraint no longer needs to be considered, i.e.: Similarly, when the air conditioner's initial operating state is off, that is... At that time, the first In specific control models within a certain time range and The calculation method is as follows: For the station Inner The first air conditioner Within a certain timeframe, the trigger timing for regulation is calculated based on the above method. By combining the triggering and control times of the air conditioner within a certain time range, a set of air conditioner triggering and control times is obtained. : For air conditioning load The response mechanism involves the cluster layer sending the mechanism for all air conditioning loads at all stations to participate in demand response to the aggregation layer. For the aggregation layer, the cluster layer infers the first... Air conditioner in normal operating condition It is known.

5. The method for hierarchical optimization control of air conditioning load in urban rail transit stations considering start-stop constraints and oriented towards renewable energy consumption, as described in claim 1 or 2, is characterized in that... The aggregation layer receives the mechanism from the cluster layer for all stations and all air conditioning loads to participate in demand response. With the goal of smoothing the peak-valley difference in the power supply from the grid to all air conditioning loads, it selects the most suitable mechanism for the equipment layer air conditioning loads to participate in demand response from various air conditioning load response mechanisms, namely, the dedicated control plan for each air conditioning load.

6. The method for hierarchical optimization control of air conditioning load in urban rail transit stations considering start-stop constraints and oriented towards renewable energy consumption, as described in claim 5, is characterized in that... The aggregation layer is implemented as follows: First, define the aggregation layer optimization scheduling model: 1) Objective function In the formula, This indicates the effect of the cluster layer in smoothing out the peak-valley difference in the energy supplied by the grid to all air conditioning loads; This represents the set of dedicated control plans for all air conditioning loads at all stations in the urban rail transit system. This indicates the number of stations in the urban rail transit system. Expressing the search function Standard deviation; Indicates station The sum of all air conditioning control and operating states within the unit; Indicates station The number of air conditioners installed inside; Indicates station Reference adjustment signal for indoor air conditioning; 2) Constraints of dedicated control plans for air conditioning load In the formula, Indicates station A collection of dedicated control plans for all air conditioning loads within the facility; Indicates station Inner A dedicated control plan for Taiwan's air conditioning load; Indicates station Inner Air conditioner The timing for triggering regulation within a short period of time; This indicates the stations sent from the cluster layer to the aggregation layer. Inner The response mechanism of Taiwan's air conditioning load; 3) Constraints on the operation status of air conditioning load regulation In the formula Indicates station Inner The status of the air conditioning system's operation; according to the station Inner Initial operating status of the air conditioner And the air conditioner Number of control moments triggered within a short period of time , =1 and 0 respectively indicate that the corresponding air conditioner is initially in the on and off state. = And it can take four values: Scenario 1: Scenario 2: Scenario 3: Scenario 4: Stations inferred from the cluster layer Inner The normal operating status of the air conditioner; Indicates station Inner air conditioner The constant control and operation status; and They represent the stations respectively. Inner Air conditioner The first and second trigger control moments within a certain time range; Indicates station Inner Air conditioner State reversal time within a short period of time ;station Inner Air conditioner Triggering time within a short period of time distance Length of time Represented as: Finally, the aggregation layer obtains the station through optimization. Inner Taiwan's dedicated control plan for air conditioning load And send it to the cluster layer and the device layer; In the dedicated control plan, all control trigger moments are state reversal moments; After receiving the dedicated control plan for the air conditioning load from the aggregation layer, the cluster layer, based on whether the air conditioning is initially on or off, employs appropriate methods to execute the corresponding control strategies for that initial on or off state, linearly inferring the station's load. Inner Operating status of air conditioner after load regulation and operating temperature .

7. The method for hierarchical optimization control of air conditioning load in urban rail transit stations considering start-stop constraints and oriented towards renewable energy consumption, as described in claim 6, is characterized in that... Based on whether the air conditioner is initially on or off, the corresponding method is adopted as follows: When the air conditioner is initially in the on state, the following method is used to represent it: When the air conditioner is initially in the off state, the following method is used to represent its operation: In the formula, , ... , , ... Indicates the moment of state transition; This is the initial operating state of the air conditioner. This indicates that the air conditioner is initially running in an "on" state, and vice versa. This is the initial operating temperature; This indicates the maximum operating temperature set for the air conditioner. With minimum operating temperature difference; and They represent the stations respectively. Inner air conditioner The operating status and operating temperature at any given time.

8. The method for hierarchical optimization control of air conditioning load in urban rail transit stations considering start-stop constraints and oriented towards renewable energy consumption, as described in claim 1 or 2, is characterized in that... The device layer is implemented as follows: After receiving the dedicated control plan for the air conditioning load from the aggregation layer, the equipment layer executes the control plan. Based on whether the air conditioning is in an on or off state at the initial operating stage, it adopts the appropriate method to execute the corresponding control strategy for the air conditioning when it is in an on or off state at the initial operating stage. (The last sentence appears to be incomplete and possibly refers to a station.) Inner The actual operating status of the air conditioner after adjustment The actual operating temperature after air conditioning control is calculated. .

9. A hierarchical optimization control system for air conditioning load in urban rail transit stations, considering start-stop constraints and oriented towards renewable energy consumption, characterized in that, It includes a memory, a processor, and computer program instructions stored in the memory and executable by the processor, which, when executed by the processor, enable the implementation of the steps of the method as described in any one of claims 1-8.

10. A computer-readable storage medium having stored thereon computer program instructions executable by a processor, wherein when the processor executes the computer program instructions, it is able to implement the steps of the method as described in any one of claims 1-8.

Images