An air conditioner virtual unit control method based on physical-information fusion

Abstract

The application discloses a kind of based on physical-information fusion air conditioner virtual unit control method, comprising:1, original sample data is collected, and resident air conditioner load is clustered, and air conditioner load virtual unit model is established;2, based on the time that the improved WF 2 Q+ scheduling algorithm is aggregated merchant cloud platform coordination control, and the time that the response state information transmission of each virtual unit air conditioner load is coordinated to aggregated merchant cloud platform;3, determine the response priority order of air conditioner load in each virtual unit under its management and control, and carry out the dynamic scheduling of air conditioner virtual unit;4, by constructing air conditioner user demand curve, meet the flexibility of virtual unit control and user comfort requirement, realize optimal control at scheduling level.The present application can effectively solve the problem of aggregated merchant cloud platform lag response to high response priority unit due to communication delay, by the virtual aggregation management and control of scattered air conditioner load to aggregated merchant, so as to realize the purpose of peak clipping and valley filling, improve power supply efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of air-conditioning virtual machine group control, and in particular to an air-conditioning virtual machine group control method based on physical-information fusion. Background Art

[0002] With the rapid development of the Internet of Things (IoT), the large-scale participation of flexibly controllable load resources, such as air conditioners, in power market ancillary services can effectively reduce grid load during peak hours. Intelligent management of temperature-controlled loads, such as air conditioners, has become a key area of ​​development in demand-side load control technology. To effectively manage and control these demand-side loads, the introduction of load aggregators has become a key solution.

[0003] Existing control strategies for air conditioning loads participating in demand response typically employ optimization approaches. However, when load aggregators are introduced into system regulation, the control problem of clustered air conditioning loads transforms into a collaborative problem among the power grid, the aggregator, and the electricity market. Therefore, it is imperative to establish an information architecture among these three parties and research coordinated control methods for clustered air conditioning virtual machines under aggregator control to collaboratively achieve response objectives. Given the variability in user response preferences, developing individual control strategies for each user's air conditioning load is impractical in practice, making it difficult to achieve coordinated load scheduling and optimization. Summary of the Invention

[0004] In order to overcome the shortcomings of the above-mentioned prior art, the present invention provides an air-conditioning virtual machine group control method based on physical-information fusion, in order to enhance the coordinated scheduling and optimization of loads, effectively solve the problem of delayed response of the aggregator cloud platform to high-response priority units due to communication delays, thereby achieving the purpose of peak shaving and valley filling, and improving the power supply efficiency of the power grid, so as to overcome the problem that it is not practical to formulate a separate control strategy for each user's air-conditioning load.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] The present invention's method for controlling air-conditioning virtual machine groups based on physical-information fusion is characterized in that it is applied to a scenario consisting of a load aggregator, an aggregator cloud platform, and a dispatch center, and is performed in the following steps:

[0007] Step 1: Comprehensive clustering of air conditioning loads considering user attribute similarity and adjustable characteristics;

[0008] Step 1.1: Use the K-means algorithm to cluster user air conditioning loads with similar power usage regularity and adjustable characteristics to obtain user air conditioning load clusters with similar response capabilities. The load aggregator uses user air conditioning load clusters as units to establish a virtual machine group model for user air conditioning load clusters.

[0009] Step 1.2: Divide all the air-conditioning virtual machine groups under the control of the load aggregator into four levels of virtual machine group queues in the order of response capability from strong to weak, where Q1 represents the first group of virtual machine group queues with a response level of one, Q2 represents the second group of virtual machine group queues with a response level of two, Q3 represents the third group of virtual machine group queues with a response level of three, and Q4 represents the fourth group of virtual machine group queues with a response level of four; let the group corresponding to the virtual machine group queue of the i-th level be denoted as Q i ; i=1,2,...,4;

[0010] Step 2: Aggregator cloud platform based on improved WF 2 The Q+ scheduling algorithm coordinates and controls the four levels of air conditioning virtual machine group queues;

[0011] Step 2.1: The group number of the virtual machine group queue that first receives the aggregator cloud platform service is recorded as l; and the virtual service time VT(t+Δt) of the aggregator cloud platform at time t+Δt is obtained using formula (1):

[0012]

[0013] In formula (1), VT(t) represents the virtual service time of the aggregator cloud platform at time t+Δt, Δt represents the time interval; S l The queue Q of the air-conditioning virtual machines in group l that first accept the services of the aggregator cloud platform is l The virtual service start time; A(t)={i|Q i (t-)≠0} represents the group sequence number set of the air-conditioning virtual machine group queue waiting for service in the aggregator cloud platform at time t; Q i (t-) represents the queue Q of the i-th group of air-conditioning virtual machines i The queue length before time t; l = 1, 2, ..., 4;

[0014] Step 2.2: Queue the i-th air-conditioning virtual machine group Q i Virtual service start time S i and virtual service completion time E i Define and schedule by response level;

[0015] Step 3: Dynamic scheduling of air conditioning virtual machine groups;

[0016] Step 3.1: Divide the switch state of the air-conditioning virtual machine group into the switch on upper limit state S on+ , switch on state S on0 , switch open lower limit state S on- , switch closed upper limit state S off+ , switch closed state S off0 , switch closed lower limit state S off- ;

[0017] Initialize the temperature setting value of the air conditioning virtual machine group to T set0 When the operating power of the air-conditioning virtual machine group is arranged in descending order, the upper limit power of the switch is P on+ , switch on power P on0 , Switch on lower limit power P on- , Switch off upper limit power P off+ , switch off power P off0 , Switch off lower limit power P off- ;

[0018] Step 3.2: When the task target of the air conditioning virtual machine group is to reduce the power P at time t t,off At (t), the number of units N that need to be shut down in all air conditioning virtual units is determined by formula (4): close :

[0019]

[0020] In formula (4), P r,ik (t) is the queue Q of the i-th group of air-conditioning virtual machines at time t i The operating power of the jth air conditioner in the kth air conditioner virtual machine group;

[0021] Step 3.3: When the task target of the air conditioning virtual machine group is to increase the power P at time t t,on At (t), the number of units N that need to be turned on in all air-conditioning virtual units is determined by formula (5): open :

[0022]

[0023] Step 3.4: When the task goal is to reduce power, the operating power of the air-conditioning virtual machine group is reduced, and the queue Q of the i-th air-conditioning virtual machine group is determined by formula (6). i The shutdown coefficient D of the kth air conditioning virtual machine group off,ik :

[0024]

[0025] In formula (6): D t,ik The i-th group of air-conditioning virtual machine queue Qi The weighted value of the room temperature of the kth air-conditioning virtual machine group at time t, T r,ik (t) is the queue Q of the i-th group of air-conditioning virtual machines i The room temperature of the kth air-conditioning virtual unit at time t; D ct,ik The i-th group of air-conditioning virtual machine queue Q i The weighted value of the number of times the kth air-conditioning virtual machine group is controlled, N t,ik (t) is the number of air-conditioning virtual machine group queues Q of group i up to time t. i The cumulative number of times the load in the kth air-conditioning virtual machine group is controlled;

[0026] Step 3.5: When the task goal is to increase power, the operating power of the air-conditioning virtual machine group is increased, and the i-th virtual machine group queue Q is determined by formula (7). i The opening coefficient D of the kth air conditioning virtual machine group on,ik :

[0027]

[0028] Step 3.6: When the power grid assigns the dispatching task to the load aggregator, the load aggregator adjusts the number of response units of the air conditioning virtual machine groups according to the task target and the response level of each air conditioning virtual machine group;

[0029] Step 4: By constructing the air-conditioning user demand curve, the flexibility of the air-conditioning virtual group control and the user comfort requirements are met, and the optimal control at the scheduling level is achieved;

[0030] Step 4.1: Use the i-th group of air-conditioning virtual machine queue Q i The temperature adjustment amplitude Vr of the jth air conditioner in the kth air conditioner virtual machine group i,kj As a control signal issued by load aggregators;

[0031] Step 4.2, the load aggregator groups the i-th group of air-conditioning virtual machines into queue Q i The power function value p of the jth air conditioner in the kth air conditioner virtual machine group i,kj =D i,kj (Vr i,kj ) is synthesized, where D i,kj represents the queue Q of the i-th group of air-conditioning virtual machines i The power-temperature adjustment amplitude function of the jth air conditioner in the kth air conditioner virtual machine group is used; thus, the i-th air conditioner virtual machine group queue Q is obtained using formula (8): i The total power function value P of all air conditioners in the kth air conditioner virtual machine group i,k :

[0032]

[0033] In formula (8), N k Represents the i-th group of air-conditioning virtual machine queue Q i The number of air conditioners in the kth air conditioner virtual machine group;

[0034] Step 4.3: The load aggregator uses formula (9) to obtain the i-th group of air-conditioning virtual machine group queue Q i The temperature adjustment amplitude Vr of the kth air conditioning virtual machine group i,k :

[0035]

[0036] In formula (9), D i,k represents the queue Q of the i-th group of air-conditioning virtual machines i The total power-temperature adjustment amplitude function of all air conditioners in the kth air conditioner virtual machine group; represents the queue Q of the i-th group of air-conditioning virtual machines i The inverse function of the total power-temperature adjustment amplitude function of all air conditioners in the kth air conditioner virtual machine group;

[0037] Step 4.4, the dispatch center adjusts the temperature amplitude Vr i,k It is sent to the load aggregator as a control signal, so that the required power is distributed to each air-conditioning virtual machine group through the aggregator's cloud platform.

[0038] The air-conditioning virtual machine group control method based on physical-information fusion according to the present invention is also characterized in that step 2.2 includes:

[0039] Step 2.2.1, starting state;

[0040] The initial value of the virtual service time VT(t0) of the aggregator cloud platform at time t0, the i-th air-conditioning virtual machine group queue Q i The initial value of the virtual service start time S0 and the virtual service completion time E0 are both initialized to 0;

[0041] Step 2.2.2, response status;

[0042] The aggregator cloud platform uses formula (2) to calculate the i-th air-conditioning virtual machine group queue Q i Virtual service start time S i , use formula (3) to calculate the i-th air-conditioning virtual machine group queue Q i The virtual service completion time E i ; Thus, the virtual service start time and virtual service completion time of all air-conditioning virtual machine group queues are obtained, and the min-th group of air-conditioning virtual machine group queues Q with the smallest virtual service start time are found. min, thereby determining the priority of each air-conditioning virtual machine group queue in the ascending order of the virtual service start time; min = 1, 2, ..., 4;

[0043]

[0044]

[0045] In formula (2): t i represents the i-th air-conditioning virtual machine group queue Q i The time of accepting the aggregator cloud platform service; VT(t i ) represents t i The virtual service time of the aggregator cloud platform corresponding to the moment; Q i (t i -) represents the i-th air-conditioning virtual machine group queue Q i In t i The queue length before time; E i-1 represents the i-1th air conditioning virtual machine group queue Q i-1 The virtual service completion time;

[0046] In formula (3): L i Represents the i-th virtual machine group queue Q i The message length, including Q i The operating status information of each air conditioning virtual machine group, user-set temperature adjustment amount and adjustable capacity; i Indicates that the aggregator cloud platform is the i-th air-conditioning virtual machine group queue Q i The rate of service provided.

[0047] The step 3.6 includes:

[0048] Step 3.6.1: Arrange the highest priority group f of air conditioning virtual machine queue Q f Priority participation in response, the response priority relationship of air conditioners in different operating states is: Closing coefficient D under the same operating state off,ik Or opening coefficient D on,ik The smaller the air conditioner virtual machine group, the higher the priority. The fth group of air-conditioning virtual machine queues Q has the highest priority f The switch state is the upper limit state S on+ Air conditioning virtual machine group, Q f,Son0 The fth group of air-conditioning virtual machine queues Q has the highest priority f The switch state is the switch open state S on0 Air conditioning virtual machine group, The fth group of air-conditioning virtual machine queues Q has the highest priority fThe middle switch state is the switch open lower limit state S on- Air conditioning virtual machine group;

[0049] Step 3.6.2: If the highest priority air conditioning virtual machine group still cannot meet the scheduling task target after responding, the next priority group s air conditioning virtual machine group queue Q s As a secondary air conditioning virtual machine group to join the control, the air conditioning response priority relationship of different operating states within the secondary air conditioning virtual machine group is: Closing coefficient D under the same operating state off,ik Or opening coefficient D on,ik The smaller the air conditioner virtual machine group, the higher the priority. The sth group of air-conditioning virtual machine queue Q is the second highest priority. s The switch state is the upper limit state S on+ Air conditioning virtual machine group, The sth group of air-conditioning virtual machine queue Q is the second highest priority. s The switch state is the switch open state S on0 Air conditioning virtual machine group, The sth group of air-conditioning virtual machine queue Q is the second highest priority. s The middle switch state is the switch open lower limit state S on- Air conditioning virtual machine group;

[0050] Step 3.6.3: If the first two priority levels of the virtual machine group queues still cannot meet the scheduling task target after the response, the third priority level t group of virtual machine group queue Q t As a three-level air conditioning virtual machine group to join the control, the air conditioning response priority relationship of different operating states within the three-level air conditioning virtual machine group is: Closing coefficient D under the same operating state off,ik Or opening coefficient D on,ik The smaller the air conditioner virtual machine group, the higher the priority. The tth virtual machine group queue Q represents the third-level priority t The switch state is the upper limit state S on+ Air conditioning virtual machine group, The tth virtual machine group queue Q represents the third-level priority t The switch state is the switch open state S on0 Air conditioning virtual machine group, The tth virtual machine group queue Q represents the third-level priority t The middle switch state is the switch open lower limit state S on- Air conditioning virtual machine group;

[0051] In step 3.6.4, regardless of whether the air conditioning virtual machine group queues of the first, second, and third priority levels meet the scheduling task objectives, the air conditioning virtual machine group queues of the fourth priority level do not participate in the load aggregator's control task.

[0052] An electronic device of the present invention includes a memory and a processor, and is characterized in that the memory is used to store a program that supports the processor to execute the air conditioning virtual machine group control method, and the processor is configured to execute the program stored in the memory.

[0053] The present invention provides a computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and the computer program executes the steps of the air-conditioning virtual machine group control method when the computer program is executed by a processor.

[0054] Compared with the prior art, the beneficial effects of the present invention are embodied in:

[0055] Based on the modeling of air-conditioning load virtual machine groups and their ability to participate in demand response, this paper proposes an optimization control strategy for virtual machine groups with different response priorities. By dividing the response levels and using the market equilibrium mechanism, the paper uses virtual electricity prices to represent the control signals issued by the aggregator to the air-conditioning users in the virtual machine group. The paper constructs the demand curve of the air-conditioning load of each user in the group according to their respective electricity comfort requirements and load control flexibility, so as to ensure that the flexibility of the virtual machine group control and the user comfort requirements are met when updating the task objectives.

[0056] 2 The present invention classifies the air-conditioning operating status within the virtual machine group and sorts the air-conditioning load response order within the virtual machine group to avoid large power fluctuations caused by centralized control. It uses an improved weighted fair queue scheduling algorithm to solve the problem of delayed response of the aggregator cloud platform to high-priority units due to communication delays.

[0057] 3 Based on the data connection between the aggregator's cloud platform and users, the present invention addresses the "over-regulation" problem of virtual machine groups in the sub-priority queue generated by the improved weighted fair queue scheduling algorithm being unable to obtain services from the cloud platform. This paper proposes an optimal control method for the aggregator at the scheduling level, solving the problems of inoperable control strategies formulated for each user's air-conditioning load and difficulty in coordinated scheduling and optimization of air-conditioning loads. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 Flowchart of the air-conditioning virtual machine group control method of the present invention;

[0059] Figure 2 This is the air conditioning load demand curve diagram of the present invention.

[0060] Specific implementation party

[0061] In this embodiment, a control method for air-conditioning virtual machine groups based on physical-information fusion is applied to a scenario consisting of a load aggregator, an aggregator cloud platform, and a dispatch center. The K-means algorithm is used to cluster residential air-conditioning loads with similar power consumption regularity and adjustable characteristics to achieve the purpose of building a virtual machine group model. Secondly, the air-conditioning load demand curve is constructed based on the market equilibrium mechanism, and the improved WF 2 The Q+ scheduling algorithm coordinates and controls the aggregator cloud platform, ensuring the transmission delay of high-priority virtual machine groups and the fairness of responses between priority queues. By performing optimal control at the scheduling level, the aggregator effectively realizes the virtual aggregation control of distributed air-conditioning loads. Specifically, Figure 1 As shown, the method is carried out in the following steps:

[0062] Step 1: Comprehensive clustering of air conditioning loads considering user attribute similarity and adjustable characteristics;

[0063] Step 1.1: Use the K-means algorithm to cluster user air-conditioning loads with similar power consumption regularity and adjustable characteristics to obtain user air-conditioning load clusters with similar response capabilities. The load aggregator uses user air-conditioning load clusters as units to establish a virtual machine group model for user air-conditioning load clusters.

[0064] Step 1.2: To avoid the phenomenon of "over-regulation" during the response process of the aggregator, all air-conditioning virtual machine groups under the control of the load aggregator are divided into four levels of virtual machine group queues in the order of response capability from strong to weak. Among them, Q1 represents the first virtual machine group queue with a response level of one. The virtual machine group ranked as one has a strong response willingness and a large controllable capacity, and its response level is the highest; Q2 represents the second group of virtual machine group queues with a response level of two. Its response capability is relatively weak. It participates in the response only when and only when the virtual machine group with a response level of one does not meet the regulation requirements; Q3 represents the third group of virtual machine group queues with a response level of three. Its response capability is the weakest. It participates in the response when the virtual machine groups with response levels of one and two cannot meet the response requirements; Q4 represents the fourth group of virtual machine group queues with a response level of four; let the group corresponding to the virtual machine group queue of the i-th level be recorded as Q i ; It includes users whose air-conditioning load does not meet the regularity of electricity consumption during the clustering process and whose response willingness is extremely weak; i=1,2,...,4.

[0065] Step 2: Aggregator cloud platform based on improved WF 2 The Q+ scheduling algorithm coordinates and controls the four levels of air conditioning virtual machine group queues;

[0066] Step 2.1: The group number of the virtual machine group queue that first receives the aggregator cloud platform service is recorded as l; and the virtual service time VT(t+Δt) of the aggregator cloud platform at time t+Δt is obtained using formula (1):

[0067]

[0068] In formula (1), VT(t) represents the virtual service time of the aggregator cloud platform at time t+Δt, Δt represents the time interval; S l The queue Q of the air-conditioning virtual machines in group l that first accept the services of the aggregator cloud platform is l The virtual service start time; A(t)={i|Q i (t-)≠0} represents the group sequence number set of the air-conditioning virtual machine group queue waiting for service in the aggregator cloud platform at time t; Q i (t-) represents the queue Q of the i-th group of air-conditioning virtual machines i The queue length before time t; l = 1, 2, ..., 4.

[0069] Step 2.2: Queue the i-th air-conditioning virtual machine group Q i Virtual service start time S i and virtual service completion time E i Define and schedule by response level;

[0070] Step 2.2.1, starting state;

[0071] The initial value of the virtual service time VT(t0) of the aggregator cloud platform at time t0, the i-th air-conditioning virtual machine group queue Q i The initial value of the virtual service start time S0 and the virtual service completion time E0 are both initialized to 0.

[0072] Step 2.2.2, response status;

[0073] The aggregator cloud platform uses formula (2) to calculate the i-th air-conditioning virtual machine group queue Q i Virtual service start time S i , use formula (3) to calculate the i-th air-conditioning virtual machine group queue Q i The virtual service completion time E i ; Thus, the virtual service start time and virtual service completion time of all air-conditioning virtual machine group queues are obtained, and the min-th group of air-conditioning virtual machine group queues Q with the smallest virtual service start time are found. min , thereby determining the priority of each air-conditioning virtual machine group queue in the ascending order of the virtual service start time; min = 1, 2, ..., 4;

[0074]

[0075]

[0076] In formula (2): t i represents the i-th air-conditioning virtual machine group queue Q i The time when the aggregator cloud platform service is accepted; VT(t i ) represents t i The virtual service time of the aggregator cloud platform corresponding to the moment; Q i (t i -) represents the i-th air-conditioning virtual machine group queue Q i In t i The queue length before time; E i-1 represents the i-1th air conditioning virtual machine group queue Q i-1 The virtual service completion time;

[0077] In formula (3): L i Represents the i-th virtual machine group queue Q i The message length, including Q i The operating status information of each air conditioning virtual machine group, user-set temperature adjustment amount and adjustable capacity; i Indicates that the aggregator cloud platform is the i-th air-conditioning virtual machine group queue Q i The rate of service provided.

[0078] Step 3: Dynamic scheduling of air conditioning virtual machine groups;

[0079] Step 3.1: To prevent the aggregator's synchronous control from damaging the diversity of the user's air conditioning load operation, causing fluctuations in the air conditioning operation power and temporary synchronization problems in the regional load, the on / off state of the air conditioning virtual machine group is divided into the switch on upper limit state S on+ , switch on state S on0 , switch open lower limit state S on- , switch closed upper limit state S off+ , switch closed state S off0 , switch closed lower limit state S off- ;

[0080] Initialize the temperature setting value of the air conditioning virtual machine group to T set0 When the operating power of the air-conditioning virtual machine group is arranged in descending order, the upper limit power of the switch is P on+ , switch on power P on0 , Switch on lower limit power P on- , Switch off upper limit power P off+ , switch off power P off0 , Switch off lower limit power P off-; When the system task objectives need to participate in the regulation of air conditioning load adjustment T set0 T set When the load is high, the operating status of the air conditioning load will change.

[0081] Step 3.2: When the task target of the air conditioning virtual machine group is to reduce the power P at time t t,off At (t), the number of units N that need to be shut down in all air conditioning virtual units is determined by formula (4): close :

[0082]

[0083] In formula (4), P r,ik (t) is the queue Q of the i-th group of air-conditioning virtual machines at time t i The operating power of the jth air conditioner in the kth air conditioner virtual machine group.

[0084] Step 3.3: When the task target of the air conditioning virtual machine group is to increase the power P at time t t,on At (t), the number of units N that need to be turned on in all air-conditioning virtual units is determined by formula (5): open :

[0085]

[0086] Step 3.4: When the task goal is to reduce power, the operating power of the air-conditioning virtual machine group is reduced, and the queue Q of the i-th air-conditioning virtual machine group is determined by formula (6). i The shutdown coefficient D of the kth air conditioning virtual machine group off,ik :

[0087]

[0088] In formula (6): D t,ik The i-th group of air-conditioning virtual machine queue Q i The weighted value of the room temperature of the kth air-conditioning virtual machine group at time t, T r,ik (t) is the queue Q of the i-th group of air-conditioning virtual machines i The room temperature of the kth air-conditioning virtual unit at time t; D ct,ik The i-th group of air-conditioning virtual machine queue Q i The weighted value of the number of times the kth air-conditioning virtual machine group is controlled, N t,ik (t) is the number of air-conditioning virtual machine group queues Q of group i up to time t. i The cumulative number of times the load in the kth air-conditioning virtual machine group is controlled;

[0089] D t,ik With D ct,ikThe value of is determined by the weight of room temperature and control times. t,ik The larger the value, the lower the user's electricity comfort level, and the larger the temperature adjustment range of the user within the same control level; D ct,ik The larger it is, the better the fairness of the control of the user's air-conditioning load is, and the smaller the difference in the number of controlled times of the user's air-conditioning load within the same control level is.

[0090] Step 3.5: When the task goal is to increase power, the operating power of the air-conditioning virtual machine group is increased, and the i-th virtual machine group queue Q is determined by formula (7). i The opening coefficient D of the kth air conditioning virtual machine group on,ik :

[0091]

[0092] When determining the response priority of the air conditioning virtual machine group with the same response level, the weight coefficient D on,ik Arranged in order from low to high, D on,ik Small groups of air-conditioning virtual machines are given priority to be turned on or off;

[0093] Step 3.6: When the power grid assigns the dispatching task to the load aggregator, the load aggregator adjusts the number of response units of the air conditioner virtual machine groups according to the task target and the response level of each air conditioner virtual machine group as follows:

[0094] Step 3.6.1: Arrange the highest priority group f of air conditioning virtual machine queue Q f Priority participation in response, the response priority relationship of air conditioners in different operating states is: Closing coefficient D under the same operating state off,ik Or opening coefficient D on,ik The smaller the air conditioner virtual machine group, the higher the priority. The fth group of air-conditioning virtual machine queues Q has the highest priority f The switch state is the upper limit state S on+ Air conditioning virtual machine group, The fth group of air-conditioning virtual machine queues Q has the highest priority f The switch state is the switch open state S on0 Air conditioning virtual machine group, The fth group of air-conditioning virtual machine queues Q has the highest priority f The middle switch state is the switch open lower limit state S on- Air-conditioned virtual machine group.

[0095] Step 3.6.2: If the highest priority air conditioning virtual machine group still cannot meet the scheduling task target after responding, the next priority group s air conditioning virtual machine group queue Qs As a secondary air conditioning virtual machine group to join the control, the air conditioning response priority relationship of different operating states within the secondary air conditioning virtual machine group is: Closing coefficient D under the same operating state off,ik Or opening coefficient D on,ik The smaller the air conditioner virtual machine group, the higher the priority. The sth group of air-conditioning virtual machine queue Q is the second highest priority. s The switch state is the upper limit state S on+ Air conditioning virtual machine group, The sth group of air-conditioning virtual machine queue Q is the second highest priority. s The switch state is the switch open state S on0 Air conditioning virtual machine group, The sth group of air-conditioning virtual machine queue Q is the second highest priority. s The middle switch state is the switch open lower limit state S on- Air-conditioned virtual machine group.

[0096] Step 3.6.3: If the first two priority levels of the virtual machine group queues still cannot meet the scheduling task target after the response, the third priority level t group of virtual machine group queue Q t As a three-level air conditioning virtual machine group to join the control, the air conditioning response priority relationship of different operating states within the three-level air conditioning virtual machine group is: Closing coefficient D under the same operating state off,ik Or opening coefficient D on,ik The smaller the air conditioner virtual machine group, the higher the priority. The tth virtual machine group queue Q represents the third-level priority t The switch state is the upper limit state S on+ Air conditioning virtual machine group, The tth virtual machine group queue Q represents the third-level priority t The switch state is the switch open state S on0 Air conditioning virtual machine group, The tth virtual machine group queue Q represents the third-level priority t The middle switch state is the switch open lower limit state S on- Air-conditioned virtual machine group.

[0097] In step 3.6.4, regardless of whether the air conditioning virtual machine group queues of the first, second, and third priority levels meet the scheduling task objectives, the air conditioning virtual machine group queues of the fourth priority level do not participate in the load aggregator's control task.

[0098] Step 4: By constructing the air-conditioning user demand curve, the flexibility of the air-conditioning virtual group control and the user comfort requirements are met, and the optimal control at the scheduling level is achieved;

[0099] Step 4.1: Use the i-th group of air-conditioning virtual machine queue Q i The temperature adjustment amplitude Vr of the jth air conditioner in the kth air conditioner virtual machine group i,kj As the control signal issued by the load aggregator; the i-th group of air-conditioning virtual machine queue Q i The jth air conditioner in the kth air conditioner virtual machine group is constructed according to its respective power comfort requirements and load control flexibility as follows: Figure 2 The demand curve shown in the figure, where the horizontal axis temperature adjustment amplitude represents the control signal, and its range is limited to [-1,1]; the vertical axis represents the i-th group of air-conditioning virtual machine queues Q in the control period i The real-time power of the jth air conditioner in the kth air conditioner virtual machine group and the queue Q of the i-th air conditioner virtual machine group i The real-time power of the kth air-conditioning virtual machine group.

[0100] Step 4.2, the load aggregator groups the i-th group of air-conditioning virtual machines into queue Q i The power function value p of the jth air conditioner in the kth air conditioner virtual machine group i,kj =D i,kj (Vr i,kj ) is synthesized, where D i,kj represents the queue Q of the i-th group of air-conditioning virtual machines i The power-temperature adjustment amplitude function of the jth air conditioner in the kth air conditioner virtual machine group is used; thus, the i-th air conditioner virtual machine group queue Q is obtained using formula (8): i The total power function value P of all air conditioners in the kth air conditioner virtual machine group i,k :

[0101]

[0102] In formula (8), N k Represents the i-th group of air-conditioning virtual machine queue Q i The number of air conditioners in the kth air conditioner virtual machine group.

[0103] Step 4.3: The load aggregator uses formula (9) to obtain the i-th group of air-conditioning virtual machine group queue Q i The temperature adjustment amplitude Vr of the kth air conditioning virtual machine group i,k :

[0104]

[0105] In formula (9), D i,k represents the queue Q of the i-th group of air-conditioning virtual machinesi The total power-temperature adjustment amplitude function of all air conditioners in the kth air conditioner virtual machine group; represents the queue Q of the i-th group of air-conditioning virtual machines i The inverse function of the total power-temperature adjustment amplitude function of all air conditioners in the kth air conditioner virtual machine group;

[0106] In each control cycle, the cloud platform needs to update the demand curve in real time according to the response status of the virtual machine group to ensure that the flexibility of virtual machine group control and user comfort requirements are met when updating the task objectives.

[0107] Step 4.4, the dispatch center adjusts the temperature amplitude Vr i,k It is sent to the load aggregator as a control signal, so that the required power is distributed to each air-conditioning virtual machine group through the aggregator's cloud platform.

[0108] In this embodiment, an electronic device includes a memory and a processor, wherein the memory is used to store a program that supports the processor to execute the above method, and the processor is configured to execute the program stored in the memory.

[0109] In this embodiment, a computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method are executed.

[0110] In summary, the present invention considers the comprehensive clustering results of air-conditioning loads based on the improved WF 2 The Q+ algorithm is used to control the aggregator cloud platform, coordinate the time for transmitting the air-conditioning load response status information of each virtual machine group to the aggregator cloud platform, reduce the time difference caused by information transmission delay when the cloud platform receives the header group of information uploaded by each virtual machine group, and propose a dynamic scheduling method for virtual machine groups. Finally, based on the air-conditioning user demand curve of the market equilibrium mechanism, optimal control at the scheduling level is achieved.

Claims

1. A method for controlling air-conditioning virtual machine groups based on physical-information fusion, characterized in that: It is applied in a scenario consisting of a load aggregator, an aggregator cloud platform, and a dispatch center, and is carried out as follows: Step 1: Comprehensive clustering of air conditioning loads considering user attribute similarity and adjustable characteristics; Step 1.1: Use the K-means algorithm to cluster user air conditioning loads with similar power usage regularity and adjustable characteristics to obtain user air conditioning load clusters with similar response capabilities. The load aggregator uses user air conditioning load clusters as units to establish a virtual machine group model for user air conditioning load clusters. Step 1.2: Divide all the air-conditioning virtual machine groups under the control of the load aggregator into four levels of virtual machine group queues in the order of response capability from strong to weak, where Q1 represents the first group of virtual machine group queues with a response level of one, Q2 represents the second group of virtual machine group queues with a response level of two, Q3 represents the third group of virtual machine group queues with a response level of three, and Q4 represents the fourth group of virtual machine group queues with a response level of four; let the group corresponding to the virtual machine group queue of the i-th level be denoted as Q i ; i=1,2,...,4; Step 2: Aggregator cloud platform based on improved WF 2 The Q+ scheduling algorithm coordinates and controls the four levels of air conditioning virtual machine group queues; Step 2.1: The group number of the virtual machine group queue that first receives the aggregator cloud platform service is recorded as l; and the virtual service time VT(t+Δt) of the aggregator cloud platform at time t+Δt is obtained using formula (1): In formula (1), VT(t) represents the virtual service time of the aggregator cloud platform at time t+Δt, Δt represents the time interval; S l The queue Q of the air-conditioning virtual machines in group l that first accept the services of the aggregator cloud platform is l The virtual service start time; A(t)={i|Q i (t-)≠0} represents the group sequence number set of the air-conditioning virtual machine group queue waiting for service in the aggregator cloud platform at time t; Q i (t-) represents the queue Q of the i-th group of air-conditioning virtual machines i The queue length before time t; l = 1, 2, ..., 4; Step 2.2: Queue the i-th air-conditioning virtual machine group Q i Virtual service start time S i and virtual service completion time E i Define and schedule by response level; Step 3: Dynamic scheduling of air conditioning virtual machine groups; Step 3.1: Divide the switch state of the air-conditioning virtual machine group into the switch on upper limit state S on+ , switch on state S on0 , switch open lower limit state S on- , switch closed upper limit state S off+ , switch closed state S off0 , switch closed lower limit state S off- ; Initialize the temperature setting value of the air conditioning virtual machine group to T set0 When the operating power of the air-conditioning virtual machine group is arranged in descending order, the upper limit power of the switch is P on+ , switch on power P on0 , Switch on lower limit power P on- , Switch off upper limit power P off+ , switch off power P off0 , Switch off lower limit power P off- ; Step 3.2: When the task target of the air conditioning virtual machine group is to reduce the power P at time t t,off At (t), the number of units N that need to be shut down in all air conditioning virtual units is determined by formula (4): close : In formula (4), P r,ik (t) is the queue Q of the i-th group of air-conditioning virtual machines at time t i The operating power of the jth air conditioner in the kth air conditioner virtual machine group; Step 3.3: When the task target of the air conditioning virtual machine group is to increase the power P at time t t,on At (t), the number of units N that need to be turned on in all air-conditioning virtual units is determined by formula (5): open : Step 3.4: When the task goal is to reduce power, the operating power of the air-conditioning virtual machine group is reduced, and the queue Q of the i-th air-conditioning virtual machine group is determined by formula (6). i The shutdown coefficient D of the kth air conditioning virtual machine group off,ik : In formula (6): D t,ik The i-th group of air-conditioning virtual machine queue Q i The weighted value of the room temperature of the kth air-conditioning virtual machine group at time t, T r,ik (t) is the queue Q of the i-th group of air-conditioning virtual machines i The room temperature of the kth air-conditioning virtual unit at time t; D ct,ik The i-th group of air-conditioning virtual machine queue Q i The weighted value of the number of times the kth air-conditioning virtual machine group is controlled, N t,ik (t) is the number of air-conditioning virtual machine group queues Q of group i up to time t. i The cumulative number of times the load in the kth air-conditioning virtual machine group is controlled; Step 3.5: When the task goal is to increase power, the operating power of the air-conditioning virtual machine group is increased, and the i-th virtual machine group queue Q is determined by formula (7). i The opening coefficient D of the kth air conditioning virtual machine group on,ik : Step 3.6: When the power grid assigns the dispatching task to the load aggregator, the load aggregator adjusts the number of response units of the air conditioning virtual machine groups according to the task target and the response level of each air conditioning virtual machine group; Step 4: By constructing the air-conditioning user demand curve, the flexibility of the air-conditioning virtual group control and the user comfort requirements are met, and the optimal control at the scheduling level is achieved; Step 4.1: Use the i-th group of air-conditioning virtual machine queue Q i The temperature adjustment amplitude Vr of the jth air conditioner in the kth air conditioner virtual machine group i,kj As a control signal issued by load aggregators; Step 4.2, the load aggregator groups the i-th group of air-conditioning virtual machines into queue Q i The power function value p of the jth air conditioner in the kth air conditioner virtual machine group i,kj =D i,kj (Vr i,kj ) is synthesized, where D i,kj represents the queue Q of the i-th group of air-conditioning virtual machines i The power-temperature adjustment amplitude function of the jth air conditioner in the kth air conditioner virtual machine group is used; thus, the i-th air conditioner virtual machine group queue Q is obtained using formula (8): i The total power function value P of all air conditioners in the kth air conditioner virtual machine group i,k : In formula (8), N k Represents the i-th group of air-conditioning virtual machine queue Q i The number of air conditioners in the kth air conditioner virtual machine group; Step 4.3: The load aggregator uses formula (9) to obtain the i-th group of air-conditioning virtual machine group queue Q i The temperature adjustment amplitude Vr of the kth air conditioning virtual machine group i,k : In formula (9), D i,k represents the queue Q of the i-th group of air-conditioning virtual machines i The total power-temperature adjustment amplitude function of all air conditioners in the kth air conditioner virtual machine group; represents the queue Q of the i-th group of air-conditioning virtual machines i The inverse function of the total power-temperature adjustment amplitude function of all air conditioners in the kth air conditioner virtual machine group; Step 4.4, the dispatch center adjusts the temperature amplitude Vr i,k It is sent to the load aggregator as a control signal, so that the required power is distributed to each air-conditioning virtual machine group through the aggregator's cloud platform.

2. The air-conditioning virtual machine group control method based on physical-information fusion according to claim 1 is characterized in that: The step 2.2 includes: Step 2.2.1, starting state; The initial value of the virtual service time VT(t0) of the aggregator cloud platform at time t0, the i-th air-conditioning virtual machine group queue Q i The initial value of the virtual service start time S0 and the virtual service completion time E0 are both initialized to 0; Step 2.2.2, response status; The aggregator cloud platform uses formula (2) to calculate the i-th air-conditioning virtual machine group queue Q i Virtual service start time S i , use formula (3) to calculate the i-th air-conditioning virtual machine group queue Q i The virtual service completion time E i ; Thus, the virtual service start time and virtual service completion time of all air-conditioning virtual machine group queues are obtained, and the min-th group of air-conditioning virtual machine group queues Q with the smallest virtual service start time are found. min , thereby determining the priority of each air-conditioning virtual machine group queue in the ascending order of the virtual service start time; min = 1, 2, ..., 4; In formula (2): t i represents the i-th air-conditioning virtual machine group queue Q i The time of accepting the aggregator cloud platform service; VT(t i ) represents t i The virtual service time of the aggregator cloud platform corresponding to the moment; Q i (t i -) represents the i-th air-conditioning virtual machine group queue Q i In t i The queue length before time; E i-1 represents the i-1th air conditioning virtual machine group queue Q i-1 The virtual service completion time; In formula (3): L i Represents the i-th virtual machine group queue Q i The message length, including Q i The operating status information of each air conditioning virtual machine group, user-set temperature adjustment amount and adjustable capacity; i Indicates that the aggregator cloud platform is the i-th air-conditioning virtual machine group queue Q i The rate of service provided.

3. The air-conditioning virtual machine group control method based on physical-information fusion according to claim 2 is characterized in that: The step 3.6 includes: Step 3.6.1: Arrange the highest priority group f of air conditioning virtual machine queue Q f Priority participation in response, the response priority relationship of air conditioners in different operating states is: Closing coefficient D under the same operating state off,ik Or opening coefficient D on,ik The smaller the air conditioner virtual machine group, the higher the priority. The fth group of air-conditioning virtual machine queues Q has the highest priority f The switch state is the upper limit state S on+ Air conditioning virtual machine group, The fth group of air-conditioning virtual machine queues Q has the highest priority f The switch state is the switch open state S on0 Air conditioning virtual machine group, The fth group of air-conditioning virtual machine queues Q has the highest priority f The middle switch state is the switch open lower limit state S on- Air conditioning virtual machine group; Step 3.6.2: If the highest priority air conditioning virtual machine group still cannot meet the scheduling task target after responding, the next priority group s air conditioning virtual machine group queue Q s As a secondary air conditioning virtual machine group to join the control, the air conditioning response priority relationship of different operating states within the secondary air conditioning virtual machine group is: Closing coefficient D under the same operating state off,ik Or opening coefficient D on,ik The smaller the air conditioner virtual machine group, the higher the priority. The sth group of air-conditioning virtual machine queue Q is the second highest priority. s The switch state is the upper limit state S on+ Air conditioning virtual machine group, The sth group of air-conditioning virtual machine queue Q is the second highest priority. s The switch state is the switch open state S on0 Air conditioning virtual machine group, The sth group of air-conditioning virtual machine queue Q is the second highest priority. s The middle switch state is the switch open lower limit state S on- Air conditioning virtual machine group; Step 3.6.3: If the first two priority levels of the virtual machine group queues still cannot meet the scheduling task target after the response, the third priority level t group of virtual machine group queue Q t As a three-level air conditioning virtual machine group to join the control, the air conditioning response priority relationship of different operating states within the three-level air conditioning virtual machine group is: Closing coefficient D under the same operating state off,ik Or opening coefficient D on,ik The smaller the air conditioner virtual machine group, the higher the priority. The tth virtual machine group queue Q represents the third-level priority t The switch state is the upper limit state S on+ Air conditioning virtual machine group, The tth virtual machine group queue Q represents the third-level priority t The switch state is the switch open state S on0 Air conditioning virtual machine group, The tth virtual machine group queue Q represents the third-level priority t The middle switch state is the switch open lower limit state S on- Air conditioning virtual machine group; In step 3.6.4, regardless of whether the air conditioning virtual machine group queues of the first, second, and third priority levels meet the scheduling task objectives, the air conditioning virtual machine group queues of the fourth priority level do not participate in the load aggregator's control task.

4. An electronic device comprising a memory and a processor, characterized in that: The memory is used to store a program that supports the processor to execute the air conditioning virtual machine group control method according to any one of claims 1 to 3, and the processor is configured to execute the program stored in the memory.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the air conditioning virtual machine group control method according to any one of claims 1 to 3 are executed.

Images