Distributed power control method for building temperature control load cluster

Abstract

The invention discloses a distributed power control method for a building temperature control load cluster. The method comprises the steps that a state space model of a temperature control load is constructed based on thermodynamic characteristics of the temperature control load; an expandable communication network oriented to a large number of distributed temperature control loads is constructed; based on a state space model of a communication network topology and a temperature control load, the temperature control load and an adjacent load exchange variable information through a leader consistency protocol to serve as a control variable to update a power variable and a temperature variable of the temperature control load; and constructing a target function and a constraint condition to obtain an auxiliary control variable, and adding the auxiliary control variable to an original control variable to update a power variable and a temperature variable of the temperature control load again. According to the control strategy, the huge flexibility provided by the temperature control load cluster in the building is fully utilized to provide auxiliary service for the power grid, the fair distribution of the output and the comfort demand between the temperature control loads is ensured in the process of dynamically regulating and controlling the temperature control load cluster, and meanwhile, the privacy of the user is protected.

Description

Technical Field

[0001] The present invention belongs to the field of electrical engineering, and more specifically, relates to a distributed power control method for a building temperature control load cluster. Background Art

[0002] Integrating large-scale renewable energy into the power grid is the future dominant direction of the energy and power industries. The main challenge in integrating renewable energy into the power system is to balance supply and demand amidst the fluctuations of renewable energy generation. Due to the inherent randomness and volatility of clean energy, the deployment of large-scale clean energy poses a huge challenge to the flexibility and stability of the power grid. In response to the dilemmas brought about by the uncertainties of renewable energy, the concept of energy storage has been proposed. Energy storage plays a key role as an efficient source of flexibility. However, the initial investment and operating costs of energy storage devices are relatively high. The large amount of investment required for large-scale energy storage devices may undermine the economic viability of microgrids. Therefore, recent research has proposed to effectively utilize the generalized energy storage resources on the demand side to provide grid ancillary services. The growth of distributed and non-dispatchable generation, such as rooftop photovoltaic (PV) solar energy, increasingly relies on non-interruptive control of demand response to support the balanced operation of the power grid.

[0003] Buildings account for more than 40% of the global energy consumption. Large load-side adjustable resources have become a research hotspot due to their huge adjustable potential. Temperature control loads have become an important control resource in demand-side response due to their unique thermal inertia and ease of adjustment, and play an important role in maintaining the power balance of the power grid. In recent years, high-temperature weather has occurred frequently, resulting in a gradual increase in the proportion of indoor air-conditioning loads in microgrid communities. These indoor air-conditioning loads have gradually become an important part of temperature control loads.

[0004] The power of a single temperature control load is relatively small. In order to effectively provide grid regulation services and solve the problem of power instability in the power grid, a large number of temperature control loads need to be aggregated to jointly achieve power regulation. Facing a large number of distributed temperature control load units, an effective control strategy is needed to achieve unified control of these loads. Early research work often relied on a temperature-based priority list control method to manage large-scale temperature control load clusters. However, in the temperature sequence control algorithm, temperature control loads usually adopt binary 0-1 switching control, which is more suitable for fixed-frequency air-conditioning control and will cause the occurrence of load life reduction and power rebound phenomena. With the increasing popularity of integrated circuits, more and more attention has been paid to the use of consensus algorithms for adaptive and unified control of large-scale continuously adjustable loads (including integrated circuits). Since the temperature control load is closely related to the user's usage situation, the user experience of the temperature control load must be considered during the regulation process. Currently, the continuous regulation method considering user comfort is not yet mature. Summary of the Invention

[0005] In view of the existing technical gaps, the present invention proposes a distributed power control method for a building temperature control load cluster, which makes full use of the great flexibility provided by the temperature control load cluster in the building to provide auxiliary services for the power grid, and ensures the fair distribution of the output and comfort requirements among the temperature control loads during the dynamic regulation of the temperature control load cluster.

[0006] The technical solution of the present invention to solve the above technical problems is as follows: A distributed power control method for a building temperature control load cluster, comprising: S1. Construct a state space model based on the thermodynamic characteristics of the temperature control load; S2. Construct an expandable communication network for a large number of distributed temperature control loads; S3. Based on the communication network topology and the state space model of the temperature control load, the temperature control load and adjacent loads exchange variable information through the leader consensus protocol as control variables to update their own power variables and temperature variables; S4. Construct an objective function and constraint conditions to obtain auxiliary control variables and add them to the original control variables to update the power variables and temperature variables of the temperature control load again.

[0007] Further, the S1 includes: According to the distributed power control method for a building temperature control load cluster described in claim 1, wherein the S1 includes: Based on the thermodynamic characteristics of the temperature control load, a dynamic model of a single temperature control load unit described by a differential equation is obtained. In order to easily and uniformly control the temperature of the temperature control load, the temperature variable y(t) is introduced T in (t) is the indoor temperature of each temperature control load at time t, and are respectively the lowest and highest acceptable temperatures of the indoor temperature, and where δ is the comfort interval, T set is the set temperature value; when the temperature variable y(t) is 0.5, the indoor temperature is equal to the set temperature; when 0 ≤ y(t) < 0.5, the indoor temperature is lower than the set temperature but within the acceptable comfort range; when 0.5 < y(t) ≤ 1, the indoor temperature is higher than the set temperature but within the acceptable comfort range; when y(t) < 0 ∪ y(t) > 1, the indoor temperature exceeds the comfort range.

[0008] Substituting the dynamic model of the temperature control load into its temperature variable, the state space model of the temperature control load can be obtained: where, T out(t) is the ambient temperature at time t, η is the energy efficiency ratio, the energy efficiency ratio of the cooling-type temperature control load is positive, and the energy efficiency ratio of the heating-type temperature control load is negative. P r is the rated power of the temperature control load, with the unit of kW, C tcl and R tcl are the equivalent capacitance and resistance of the room, x i (t) is the power variable of the i-th temperature control load, that is, the output power percentage at time t; y i (t) is the temperature variable of the i-th temperature control load. In this state space model, the power variable and temperature variable of the temperature control load are used as state variables to describe the change of the indoor temperature under the combined influence of the power input and the outdoor temperature of the temperature control load.

[0009] The state space model matrix form of the temperature control load is: where p = [x y] T is the state variable, A is the system matrix, representing the connection of the internal state of the system, B is the control matrix, and is specifically expressed as: Furthermore, the S2 includes: For the cluster communication network composed of N temperature control load units in a residential building, in addition, each building is equipped with a reference temperature control load, which is responsible for uniformly regulating the temperature control load cluster in the building and establishing communication between floors. The communication topology of the entire building is represented by the Laplacian matrix where is the symbol of the matrix Kronecker product, n f is the number of floors participating in the regulation in the residential building, and I is the identity matrix. The four matrices L1, L2, L3, Γ1, Γ2 are defined as follows: The Laplacian matrix of the topological graph Γ1 represents the local interaction of the temperature control load units in each household. The Laplacian matrix of the topological graph Γ2 represents the overall interconnection between the households on each floor. The Laplacian matrix of the topological graph Γ3 represents the relationship between the floors. The Laplacian matrix represents the dominant temperature control load in each household, indicating that the load actively participates in the load communication on its respective floor. The Laplacian matrix represents the reference temperature control load in each building, which is responsible for the macroscopic regulation of the temperature control load cluster of the entire building.

[0010] Furthermore, the S3 includes: S3.1. In order to ensure the fair distribution of the power regulation task in the load cluster while protecting user privacy, the leader consensus algorithm is introduced. The control variable u of the i-th temperature control load i can be expressed as: where a ij is the communication weight between the i-th and j-th temperature-controlled load units, and a i0 is the communication weight between each temperature-controlled load and the reference temperature-controlled load. a ij = 1 indicates that there is a communication link between two temperature-controlled load units, and a ij = 0 indicates that there is no communication link between two temperature-controlled load units. x0 and y0 are the power variable and temperature variable values given by the reference temperature-controlled load respectively. κ1, κ2, κ3, and κ4 are the communication weight coefficients, representing the influence magnitudes of the power variable and temperature variable of the temperature-controlled load on its control variable.

[0011] Its matrix form is: u = -κ α L c p - κ β B r p + E r where u = [u0, u1, u2... u N T , κ α = [κ1 κ2], κ β = [κ3 κ4] E r = κ β B r [x0 y0] T , B r = diag{a i0} (i = 0, 1, 2..., N) is the adjacency matrix of the reference temperature-controlled load. Substituting the discretized control protocol into the state-space model of the temperature-controlled load, we can obtain p(k + 1) = Gp(k) + E rr where G = I + A - κ α BL c - κ β BB r , E rr = BE r + c.

[0012] S3.2. Under the communication network and the control protocol of S3.1, the state variable values of the temperature-controlled load cluster reach consistency. The steady-state value x f of the power variable is the power variable value x0 of the reference temperature-controlled load, and the steady-state value y f of the temperature variable can be deduced as To ensure the user comfort experience, the steady-state value y f of the temperature variable of the temperature-controlled load should be controlled within the range of (0, 1). Therefore, under the constraint conditions of the user comfort experience, regulation is taken according to the outdoor temperature T​out and the temperature set value T set Given reference temperature control load power variable value wherein, the temperature set value T set is determined by the power dispatching instruction and is calculated by , where m is the number of temperature control load units participating in the regulation. The final output power value of the temperature control load cluster is

[0013] Furthermore, the said S4 includes: S4.1. In order to obtain the optimal control variable, an auxiliary control variable u L (k) is added to the original control variable, and the original control variable is updated to u p = -(κ α + κ μ )L c p - κ β B r p + E r , wherein, κ μ = [μ μ] is the coefficient of the auxiliary control variable.

[0014] S4.2. In order to obtain the coefficient of the auxiliary control variable, the temperature control load state variable is optimized to obtain the objective function wherein, the given Q L = diag(q l , q l , … q l ) ∈ R Np×Np , R L = diag(r l , r l , … r l ) ∈ R Nc×Nc are real, symmetric and positive definite weighting matrices, N c and N p are respectively the step lengths of control and prediction in the optimization iteration process. Considering the comfort requirements of the temperature control load, constraint conditions are added, wherein, are the minimum and maximum control amounts of N temperature control load clusters.

[0015] S4.3 adopts a distributed optimization algorithm to solve the optimization problem in S4.2, and further iteratively optimizes to obtain the coefficient of the auxiliary control variable and the optimal control variable of the temperature control load unit. Description of the Drawings

[0016] Figure 1 It is the example three - layer building inter - layer and intra - layer communication topology diagram used in the present invention.

[0017] Figure 2 Schematic diagram of the distributed power signal regulation structure proposed by the present invention.

[0018] Figure 3 Power variable and temperature variable changes of each load in the simulation system of the present invention when the temperature control load cluster does not add a reference temperature control load.

[0019] Figure 4 Power variable and temperature variable changes of each load in the simulation system of the present invention when the temperature control load cluster adds a reference temperature control load.

[0020] Figure 5 Power variable and temperature variable changes of each load in the simulation system of the present invention when the temperature control load cluster adds a reference temperature control load and performs optimization iteration.

[0021] Figure 6 Schematic diagram of the process of the method of the present invention. Detailed implementation manners

[0022] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0023] Embodiment 1: Based on Figure 1 The shown communication network, in the application scenario of a given power command value, the proposed control method is verified. The power variable and temperature value of the given reference temperature control load. The power variable of the reference temperature control load and its initial temperature variable. The temperature control load model parameters for simulation verification are shown in Table 1. There are a total of temperature control loads in the building. In order to introduce randomness elements in the simulation, the power and temperature variables of the temperature control loads are uniformly distributed in the range of 0.3 to 0.7.

[0024] A distributed power control method for a building temperature control load cluster, the steps include: S1. Construct a state space model based on the thermodynamic characteristics of the temperature control load; S2. Construct an expandable communication network for a large number of distributed temperature control loads; S3. Based on the communication network topology and the state space model of the temperature control load, the temperature control load and adjacent loads exchange variable information through the leader consensus protocol as control variables to update their own power variables and temperature variables; S4. Construct an objective function and constraint conditions to obtain auxiliary control variables and add them to the original control variables to update the power variables and temperature variables of the temperature control load again.

[0025] The temperature control loads and controller parameters used in the simulation are shown in Table 1. Table 1 Temperature control loads and controller parameters Parameter Value Parameter Value <![CDATA[Room equivalent thermal resistance R tcl > 2 <![CDATA[The number of temperature-controlled loads per household, m in > 5 <![CDATA[Room equivalent heat capacity C tcl > 2 <![CDATA[Number of building floors n floor > 3 <![CDATA[Rated power P r > 5 <![CDATA[Communication weight κ1]]> 0.3 Coefficient of performance η 2.5 <![CDATA[Communication weight κ2]]> 0.005 Comfort range δ 2 <![CDATA[Communication weight κ3]]> 0.8 Number of households per floor m 3 <![CDATA[Communication weight κ4]]> 0.05

[0026] Figure 3 For the power variable and temperature variable changes of each load in the temperature control load cluster of the simulation system when the reference temperature control load is not added. When the initial values of the specified power variable and comfort variable are set, the temperature control loads gradually reach the same power and comfort level after 100 seconds under Control Strategy 100. However, as Figure 3 can be clearly seen, the comfort variable of the temperature control load cluster is significantly lower than zero, indicating that the indoor temperature of the building is not within the specified comfort range. Figure 4 A reference temperature control load is introduced in the building to synchronize the power variable and comfort variable of all temperature control loads with the reference temperature control load. The power variable of the temperature control load is synchronized at t f = 15 s, and the comfort variable is synchronized at around t f = 25 s. Compared with Figure 3 , the regulation stabilization time of the entire building is significantly shortened. However, as can be seen from the Figure 4 magnified view, under this control method, when t < 7 s, there are still some loads outside the comfort range.

[0027] To strictly ensure that the room temperature of the entire building remains within the comfort range, the optimized control strategy proposed in S4 of the present invention is introduced, and the simulation results are as Figure 5 shown. The results show that under this control strategy, the convergence time of the power variable and comfort variable is significantly shortened. The power variable is synchronized at around t = 5 s, and the comfort variable reaches stability at around t f = 15 s. During the control process, strict measures are taken to ensure that the comfort variable remains within the range of (0.1). This is achieved by incorporating the constraints into the control strategy optimization conditions and optimizing the control step size in each iteration.

[0028] From the comparison of several control strategies, it can be seen that adding a reference temperature control load and optimizing the iterative control can significantly improve the time for the temperature control load regulation of the entire building to reach stability, strictly ensure that the indoor temperature remains within the specified comfort range, protect the user's privacy, and further utilize the flexibility of the building to provide ancillary services for the power grid.

Claims

1. A distributed power control method for a building temperature control load cluster, characterized in that, Including: S1. Construct the state - space model based on the thermodynamic characteristics of the temperature - controlled load; S2. Construct an expandable communication network for a large number of distributed temperature - controlled loads; S3. Based on the communication network topology and the state - space model of the temperature - controlled load, the temperature - controlled load and adjacent loads exchange variable information through the leader - following consensus protocol as control variables to update their own power variables and temperature variables; S4. Construct the objective function and constraint conditions to obtain the auxiliary control variables, add them to the original control variables, and update the power variables and temperature variables of the temperature - controlled load again.

2. The distributed power control method for a building temperature control load cluster according to claim 1, wherein The S1 includes: S1.

1. Based on the thermodynamic characteristics of the temperature - controlled load, obtain the dynamic model of a single temperature - controlled load: Among them, T in (t) and T out (t) are the indoor temperature and the ambient temperature of each temperature-controlled load at time t, η is the energy efficiency ratio, the energy efficiency ratio of the cooling-type temperature-controlled load is positive, and the energy efficiency ratio of the heating-type temperature-controlled load is negative. P r is the rated power of the temperature-controlled load, with the unit of kW, C tcl and R tcl are the equivalent capacitance and resistance of the room, x(t) is the power variable of the temperature-controlled load, that is, the output power percentage at time t; y(t) is the temperature variable of the temperature-controlled load, and are respectively the lowest and highest acceptable indoor temperatures, and where δ is the comfort interval, T set is the set temperature value; When the temperature variable y(t) is 0.5, the indoor temperature is equal to the set temperature; when 0 ≤ y(t) < 0.5, the indoor temperature is lower than the set temperature but within the acceptable comfort range; when 0.5 < y(t) ≤ 1, the indoor temperature is higher than the set temperature but within the acceptable comfort range; when y(t) < 0 ∪ y(t) > 1, the indoor temperature exceeds the comfort range; S1.

2. Substitute the dynamic model of the temperature - controlled load into the temperature variable to obtain: Obtain the state - space model of the temperature - controlled load as: where x i and y i respectively represent the power variable and the temperature variable of the i-th thermostatic load. In this state space model, the power variable and the temperature variable of the thermostatic load are used as state variables to describe the change of the indoor temperature under the combined influence of the power input and the outdoor temperature; The matrix form of the state - space model is: where p = [x y] T is a state variable, A is a system matrix representing the connection of the internal state of the system, B is a control matrix, and c is an additional matrix, specifically expressed as:

3. A distributed power control method for a building temperature control load cluster according to claim 1, characterized in that The S2 includes: S2.

1. For a cluster communication network composed of N temperature-controlled load units in a residential building, it consists of N nodes, each node representing a temperature-controlled load, and the connecting lines represent two-way communication links; the temperature-controlled loads on each floor can communicate with each other, and the residential building has a three-layer layout; assume that there are m = 3 families on each floor, and each family is equipped with m in = 5 temperature-controlled load devices; each building is equipped with a central control center, which is responsible for managing the temperature-controlled load cluster in the building and establishing communication between floors; the communication topology Laplacian matrix L c is described by the following formula: where is the symbol of the matrix Kronecker product, and n f is the number of floors participating in the regulation in the residential building. The four matrices L1, L2, L3, Γ1, Γ2 are defined as follows: The Laplacian matrix of the topological graph Γ1 represents the local interaction of the temperature control load units in each household, and the Laplacian matrix of the topological graph Γ2 represents the overall interconnection between the households on each floor; the Laplacian matrix of the topological graph Γ3 represents the relationship between the floors; the Laplacian matrix represents the dominant temperature control load in each household, indicating that the load actively participates in the load communication on its respective floor; the Laplacian matrix represents the reference temperature control load in each building, which is responsible for the macro-control of the temperature control load cluster of the whole building.

4. A distributed power control method for a building temperature control load cluster according to claim 1, characterized in that, The S3 includes: S3.

1. To ensure the fair distribution of power regulation tasks in the load cluster while protecting user privacy, the leader consensus algorithm is introduced, and the control variable u of the i-th thermostatic load is i expressed as: where a ij is the communication weight between the i-th and j-th temperature-controlled load units, and a i0 is the communication weight between each temperature-controlled load and the reference temperature-controlled load. a ij = 1 indicates that there is a communication link between two temperature-controlled load units, and a ij = 0 indicates that there is no communication link between two temperature-controlled load units. x0 and y0 are the power variable and temperature variable values given by the reference temperature-controlled load respectively; κ1, κ2, κ3, and κ4 are communication weight coefficients respectively, characterizing the influence of the power variable and temperature variable of the temperature-controlled load on its control variable. The control variable u of its temperature-controlled load cluster is in matrix form as follows: u = -κ α L c p - κ β B r p + E r where \(u = [u_0, u_1, u_2... u N T , \kappa α = [\kappa_1 \kappa_2], \kappa β = [\kappa_3 \kappa_4]E r = \kappa β B r [x_0 y_0] T , B r = diag\{a i0 \}(i = 0, 1, 2..., N) is the adjacency matrix of the reference temperature control load. Substituting the discretized control protocol into the state space model of the temperature control load, we get:​ p(k + 1) = Gp(k) + E rr where G = I + A - κ α BL c -κ β BB r ,E rr = BE r + c; S3.

2. Under the communication network and the control protocol of S3.1, the state variable values of the temperature - controlled load cluster reach consistency, and the steady - state values of the power variable and temperature variable are as follows. where x f and y f are respectively the steady-state values of the power variable and the temperature variable of the temperature-controlled load cluster. To ensure the user's comfort experience, a temperature variable constraint is introduced: To meet the user comfort experience constraint conditions, take regulation and give the power variable value of the temperature - controlled load as follows: Among them, the temperature set value T set is given by the power dispatching instruction and m is the number of temperature control load units participating in regulation; the final output power value P of the temperature control load cluster tt is as follows:

5. A distributed power control method for a building temperature control load cluster according to claim 1, characterized in that The S4 includes: S4.

1. To obtain the optimal control variable, an auxiliary control variable u L (k) is added to the original control variable: p(k + 1)=Gp(k)+E rr +Bu L (k) = G'p(k) + E rr u L B(k) = -κ μ L c Bp(k) It should be noted that there seems to be an error in the original text where "](k)=-κ" should probably be "B(k) = -κ". The translation is based on the corrected understanding. where k is the number of iteration steps, and κ μ = [μμ] is the coefficient of the auxiliary control variable; at this time, the optimal control variable u p (k) is u p = -(κ α + κ μ )L c p - κ β B r p + E r S4.

2. To obtain the optimal control variables, set the objective function as follows: where, given Q L = diag(q l , q l , … q l ) ∈ R Np×Np , R L = diag(r l , r l , … r l ) ∈ R Nc×Nc are real, symmetric and positive definite weighting matrices, N c and N p are the step sizes of control and prediction respectively in the optimization iteration process; Considering the comfort requirements of the temperature - controlled load, add constraint conditions: Among them, are the minimum and maximum control amounts of N temperature-controlled load clusters; S4.

3. Use the distributed optimization algorithm to solve the optimization problem in S4.2, further iteratively optimize to obtain the optimal control variables of the temperature - controlled load unit, and update the control variable u of S3.1.

Images