Energy-storage-free type photovoltaic air conditioner dynamic power supply method and system

Abstract

The invention provides an energy-storage-free dynamic power supply method and system for a photovoltaic air conditioner. The method belongs to the technical field of building energy conservation and photovoltaic air conditioner control and comprises the steps that a photovoltaic output prediction value and an air conditioner load prediction value are obtained; time-of-use electricity price data serves as input, a linear programming model with the operation cost minimization as the target is constructed, and power is dynamically supplied to the air conditioner according to the power supply strategy and the time-of-use electricity price response strategy based on the linear programming model; the system can dynamically supply power to the air conditioner through photovoltaic in a non-energy-storage scene, so that the operation cost is minimized, the heat energy storage potential of a building envelope structure is developed, load cross-time transfer is achieved through pre-cooling / pre-heating, and the operation cost is further reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of building energy conservation and photovoltaic air conditioner control, and more particularly, to a dynamic power supply method and system for a photovoltaic air conditioner without energy storage. Background Art

[0002] The photovoltaic direct-drive air conditioner system can effectively reduce building energy consumption by directly using photovoltaic power generation to drive the air conditioner load. In the prior art, the invention patent with the publication number of CN115473435A proposes a photovoltaic air conditioner system based on voltage matching, which realizes power adaptation with the load by adjusting the output voltage of the photovoltaic array. However, such a solution has the following defects:

[0003] The adopted power supply control strategy only switches the power supply mode according to the real-time matching degree of photovoltaic output and load, resulting in operating costs. For example, during the peak noon period with sufficient photovoltaic output, if the air conditioner load suddenly increases and power needs to be purchased from the grid, the operating cost will increase by more than 30%. In order to balance the photovoltaic output fluctuation, most solutions rely on batteries or supercapacitors. For example, the invention patent with the publication number of CN119362536A, but adding batteries or supercapacitors will increase the system cost by 50% - 80%. Moreover, the batteries or supercapacitors have a short lifespan and complex maintenance, which limits the technology promotion. Summary of the Invention

[0004] The purpose of the present invention is to provide a dynamic power supply method and system for a photovoltaic air conditioner without energy storage, which can minimize the operating cost by dynamically supplying power to the air conditioner through photovoltaic in the energy storage-free scenario, and develop the thermal energy storage potential of the building envelope structure to realize load transfer across time periods through precooling / preheating, further reducing the operating cost.

[0005] The present invention is implemented as follows:

[0006] In a first aspect, the present invention provides a dynamic power supply method for a photovoltaic air conditioner without energy storage, including:

[0007] Obtaining the predicted value of photovoltaic output and the predicted value of air conditioner load

[0008] Taking the time-of-use electricity price data as the input, constructing a linear programming model with the goal of minimizing the operating cost, and the constraint conditions of the linear programming model are:

[0009]

[0010] Wherein, P grid (t) is the power consumption from the grid at time t, and T(t) is the indoor temperature;

[0011] Based on a linear programming model, dynamically supply power to the air conditioner according to the power supply strategy and the time-of-use electricity price response strategy;

[0012] The power supply strategy includes: the priority of photovoltaic power supply is greater than that of grid power intake, and when the photovoltaic output is insufficient, switch to grid power intake;

[0013] The time-of-use electricity price response strategy includes: defining the electricity price as a dynamic switching threshold. If the electricity price is in the peak period, increase the set temperature of the air conditioner by 1 - 3 °C to reduce the grid power intake; if the electricity price is in the low period, reduce the air conditioner load demand the next day by precooling the building envelope.

[0014] Based on the first aspect, obtaining the predicted value of photovoltaic output The steps include:

[0015] Adopt the LSTM model as the photovoltaic output prediction model, and define the input sequence X t =[G(t - τ), T a (t - τ), P pv (t - τ)], where τ = 0, 1, …, 23 represents the data of the past 24 hours, X t represents the input feature vector at time τ, G(τ) represents the solar irradiance at time τ; T a (τ) represents the ambient temperature at time τ; P pv (τ) represents the historical actual photovoltaic output at time τ; The hidden state update equation of the LSTM model is:

[0016] h t =tanh(W h ·[h t-1 , x t +b h );

[0017] Among them, h t represents the hidden state vector at the current time t, W h represents the hidden state weight matrix, b h represents the hidden state bias term, h t-1 represents the hidden state vector at the previous time, and xt represents the input data at the current time;

[0018] The output layer is mapped to the predicted value for the next 24 hours through a fully connected network:

[0019]

[0020] Among them, represents the predicted value of photovoltaic output from hour 1 to 24 starting from time t = 0, W o represents the output layer weight matrix; h t represents the hidden state vector at the current time; bo Represents the output layer bias term, which adjusts the baseline offset of the predicted value.

[0021] Based on the first aspect, obtaining the air-conditioning load prediction value The steps include:

[0022] Adopt the random forest model as the air-conditioning load prediction model. Let the input feature set F = {N(t), T out (t), H(t), D(t)}, where D(t) ∈ {0, 1} represents the date type, N(t) is the number of people at time t, and T out (t) is the temperature at time t; select the splitting node by minimizing the Gini impurity to generate a set of decision trees; the final air-conditioning load prediction value is the weighted average of the outputs of each tree, specifically:

[0023]

[0024] Among them, is the predicted air-conditioning load demand for time period t, M is the number of decision trees, and f m is the prediction function of the m-th tree.

[0025] Based on the first aspect, the linear programming model is:

[0026]

[0027] Among them, C total is the daily average total operating cost of the system, in yuan; P grid (t) is the power taken from the power grid at time period t, in kW; π(t) is the time-of-use electricity price at time period t, in yuan / kWh; λ is the temperature deviation penalty coefficient, in yuan / °C, with a default value of 10; T(t) is the actual indoor temperature at time period t, in °C; Tset(t) is the air-conditioning set temperature at time period t, in °C; max(x, 0) represents taking the larger value of x and 0, which is used to quantify the penalty cost for the part where the temperature exceeds the standard.

[0028] Based on the first aspect, the electricity price is defined as a piecewise linear function of the dynamic switching threshold:

[0029]

[0030] Among them, η th (t) is the photovoltaic power supply switching threshold at time period t, indicating the percentage of photovoltaic output that needs to reach the load demand to allow switching to photovoltaic power supply; π(t) is the electricity price at time period t, in yuan / kWh.

[0031] Based on the first aspect, if the electricity price is in the low valley period, the steps to reduce the air-conditioning load demand of the next day by precooling the building envelope include:

[0032] Define the Lyapunov function:

[0033]

[0034] By designing the switching logic, make V < 0 to ensure the asymptotic stability of the system during the mode switching process;

[0035] Based on the unsteady heat transfer theory, the wall temperature distribution T(x, t) satisfies:

[0036]

[0037] where T(x, t) is the temperature distribution at the location of the wall and time t, and α is the thermal diffusivity of the material, with the unit of m 2 / s. The calculation formula of α is where: k is the wall thermal conductivity, with the unit of W / (m·K), ρ is the wall density, with the unit of kg / m 3 ; c p is the specific heat capacity of the wall, with the unit of J / (kg·K);

[0038] The boundary conditions are:

[0039]

[0040] By solving through the method of separation of variables, the relationship between the precooling duration τ and the temperature drop ΔT the next day is obtained:

[0041]

[0042] where β n is the root of the characteristic equation, and C n is the coefficient determined by the initial temperature distribution; accordingly, the precooling duration is formulated to make the room temperature rise controllable during the peak period the next day.

[0043] In a second aspect, the present invention provides a dynamic power supply system for a photovoltaic air conditioner without energy storage, including:

[0044] A prediction value acquisition module for acquiring the predicted value of photovoltaic power output and the predicted value of air conditioner load

[0045] A linear programming module for constructing a linear programming model with the goal of minimizing the operating cost, taking the time-of-use electricity price data as the input. The constraint conditions of the linear programming model are:

[0046]

[0047] where P grid (t) is the grid power consumption at time t, and T(t) is the indoor temperature;

[0048] A dynamic power supply module, which is used to dynamically supply power to an air conditioner based on a linear programming model according to a power supply strategy and a time-of-use electricity price response strategy;

[0049] The power supply strategy includes: the priority of photovoltaic power supply is greater than that of grid power taking, and when the photovoltaic output is insufficient, it switches to grid power taking;

[0050] The time-of-use electricity price response strategy includes: defining the electricity price as a dynamic switching threshold. If the electricity price is in the peak period, the set temperature of the air conditioner is increased by 1 - 3 °C to reduce the grid power consumption; if the electricity price is in the low valley period, the building envelope is precooled to reduce the air conditioner load demand of the next day.

[0051] Based on the second aspect, the dynamic power supply module includes a grid switch, and the grid switch includes:

[0052] The first contactor K1 in the photovoltaic power supply circuit;

[0053] The second contactor K2 in the grid power supply circuit;

[0054] A PLC controller, which is used to send an interlock control signal to K1 and K2 to ensure that the two circuits are not conducted simultaneously;

[0055] A discharge resistor is also connected in parallel in the photovoltaic power supply circuit, which is used to consume the surplus electric energy when the photovoltaic output exceeds the air conditioner load demand and prevent it from flowing back to the grid.

[0056] In the third aspect, the present invention provides an electronic device, including:

[0057] A memory, which is used to store one or more programs;

[0058] A processor;

[0059] When the one or more programs are executed by the processor, the above-mentioned method is implemented.

[0060] In the fourth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by the processor, the above-mentioned method is implemented.

[0061] Compared with the prior art, the present invention has at least the following advantages or beneficial effects:

[0062] The present invention provides a dynamic power supply method for a photovoltaic air conditioner without energy storage. By constructing a linear programming model with the goal of minimizing the operating cost, it can dynamically adjust the photovoltaic power supply switching threshold and the temperature setting value. The electricity price is defined as the dynamic switching threshold. During the peak electricity price period, the set temperature of the air conditioner is increased by 1-3°C to reduce the electricity consumption from the power grid. It ensures that photovoltaic power is preferentially used during high-price periods, reduces the dependence on the power grid, and minimizes the operating cost by dynamically powering the air conditioner. During the low electricity price period, the building envelope is precooled to reduce the air conditioning load demand the next day. The thermal energy storage potential of the building envelope is developed, and the load is transferred across time periods through precooling / preheating, further reducing the operating cost. Through the collaborative optimization of dynamic threshold adjustment and building thermal inertia, the operating cost is reduced without the need for energy storage equipment, and it can ensure that the temperature fluctuation is still controllable within ±1.5°C under the condition of ±20% fluctuation of the photovoltaic output. Compared with the prior art, the initial investment is reduced by more than 50%, and it is applicable to the photovoltaic air conditioner transformation of public buildings such as libraries and office buildings. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required in the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0064] Figure 1 It is a schematic flowchart of a dynamic power supply method for a photovoltaic air conditioner without energy storage according to the present invention;

[0065] Figure 2 It is a schematic structural diagram of a dynamic power supply system for a photovoltaic air conditioner without energy storage according to the present invention;

[0066] Figure 3 It is a schematic structural diagram of an electronic device according to the present invention.

[0067] Reference Signs:

[0068] 1. Predicted value acquisition module; 2. Linear programming module; 3. Dynamic power supply module; 4. Processor; 5. Memory; 6. Communication interface; DETAILED DESCRIPTION OF THE EMBODIMENTS

[0069] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0070] The following will describe in detail some embodiments of the present invention with reference to the accompanying drawings. Without conflict, the following various embodiments and the various features in the embodiments can be combined with each other.

[0071] Embodiment

[0072] Through long-term research and practice, the applicant found that the existing technology for powering air conditioners with photovoltaic power only switches the power supply mode according to the real-time matching of photovoltaic output and load, resulting in the forced purchase of electricity at high prices during peak hours; at the same time, relying on fixed thresholds for switching causes frequent operation of equipment, accelerating hardware wear and causing temperature fluctuations to exceed the limit.

[0073] In view of this, the embodiments of the present invention provide a dynamic power supply method for a photovoltaic air conditioner without energy storage, which can minimize the operating cost by dynamically powering the air conditioner with photovoltaic power in a scenario without energy storage, and develop the thermal energy storage potential of the building envelope structure to achieve load transfer across time periods through precooling / preheating, further reducing the operating cost.

[0074] Please refer to Figure 1 , and this dynamic power supply method for a photovoltaic air conditioner without energy storage includes the following steps:

[0075] S1: Obtain the predicted value of photovoltaic output and the predicted value of air conditioner load

[0076] Specifically, the steps for obtaining the predicted value of photovoltaic output include: Using the LSTM model as the photovoltaic output prediction model, defining the input sequence X t = [G(t - τ), T a (t - τ), P pv (t - τ)], where τ = 0, 1,..., 23 represents the data of the past 24 hours, X t represents the input feature vector at time τ, G(τ) represents the solar irradiance at time τ; T a (τ) represents the ambient temperature at time τ; P pv (τ) represents the historical actual photovoltaic output at time τ; the hidden state update equation of the LSTM model is:

[0077] h t = tanh(W h ·[h t-1 , x t + b h );

[0078] where, h t represents the hidden state vector at the current time t, W h represents the hidden state weight matrix, b h represents the hidden state bias term, ht-1 represents the hidden state vector at the previous moment, and \(x_t\) represents the input data at the current moment;

[0079] The output layer is mapped to the predicted values for the next 24 hours through a fully connected network:

[0080]

[0081] Among them, represents the predicted photovoltaic output values from 0 to 24 hours in period \(t\), \(W\) o represents the weight matrix of the output layer; \(h\) t represents the hidden state vector at the current moment; \(b\) o represents the bias term of the output layer, which adjusts the baseline offset of the predicted value.

[0082] Obtaining the predicted value of air-conditioning load The steps include: using a random forest model as the air-conditioning load prediction model, setting the input feature set \(F = \{N(t), T\) out (t), H(t), D(t)\}, where \(D(t)\in\{0, 1\}\) represents the date type, \(N(t)\) is the number of people at time \(t\), and \(T\) out (t) is the temperature at time \(t\); selecting split nodes by minimizing the Gini impurity to generate a set of decision trees; the final predicted value of air-conditioning load is the weighted average of the outputs of each tree, specifically:

[0083]

[0084] Among them, is the predicted air-conditioning load demand in period \(t\), \(M\) is the number of decision trees, and \(f\) m is the prediction function of the \(m\)-th tree. By evaluating the importance of feature permutation, features with importance lower than the threshold \(\theta\) are removed, and the default \(\theta = 0.1\).

[0085] Modeling the time series features through the LSTM model and further optimizing the feature weights by the random forest model, the two models cooperate to improve the refined prediction accuracy, robustness and generalization of photovoltaic output and air-conditioning load.

[0086] It should be noted that in order to improve the model prediction accuracy, the LSTM model and the random forest model can be pre-trained with historical data. After verifying that the prediction error MAPE ≤ 10%, the trained LSTM model and random forest model are obtained. Input the collected synchronous meteorological, photovoltaic output, load demand and electricity price signals into the trained LSTM model and random forest model to obtain the prediction curve for input into the linear programming model.

[0087] S2: Using the time-of-use electricity price data as the input, construct a linear programming model with the goal of minimizing the operating cost. The constraint conditions of the linear programming model are:

[0088]

[0089] Among them, P grid (t) is the power consumption of the power grid in the t period, and T(t) is the indoor temperature;

[0090] Specifically, the linear programming model is:

[0091]

[0092] Among them, C total is the total daily operating cost of the system, with the unit of yuan; P grid (t) is the power taken from the power grid in the t period, with the unit of kW; π(t) is the time-of-use electricity price in the t period, with the unit of yuan / kWh; λ is the temperature deviation penalty coefficient, with the unit of yuan / °C, and the default value is 10; T(t) is the actual indoor temperature in the t period, with the unit of °C; Tset(t) is the air conditioner set temperature in the t period, with the unit of °C; max(x, 0) represents taking the larger value of x and 0, which is used to quantify the penalty cost of the temperature exceeding the standard.

[0093] By creating a linear programming model, resource allocation and cost control can be optimized. The linear programming model adjusts the energy storage charge and discharge strategy in real time according to the volatility of solar power generation and load demand, reduces redundant energy waste, and reduces the dependence on the external power grid, thereby significantly reducing the comprehensive operating cost of the microgrid.

[0094] Furthermore, after the linear programming model generates the optimal power supply strategy, the power grid switcher executes the power supply loop switching and simultaneously rewrites the air conditioner set temperature parameter. Specifically, the power grid switcher includes:

[0095] The first contactor (K1) in the photovoltaic power supply loop;

[0096] The second contactor (K2) in the power grid power supply loop;

[0097] The PLC controller is used to send interlock control signals to K1 and K2 to ensure that the two loops are not conducted simultaneously.

[0098] Furthermore, a dump resistor is also connected in parallel in the photovoltaic power supply loop, which is used to consume the surplus electric energy when the photovoltaic output exceeds the air conditioner load demand and prevent reverse flow to the power grid.

[0099] S3: Based on the linear programming model, supply power to the air conditioner dynamically according to the power supply strategy and the time-of-use electricity price response strategy;

[0100] The power supply strategy includes: the priority of photovoltaic power supply is greater than taking power from the power grid, and switch to taking power from the power grid when the photovoltaic output is insufficient; by setting the power supply strategy, ensure that photovoltaic power is preferentially used during high-price periods to reduce the dependence on the power grid.

[0101] The time-of-use electricity price response strategy includes: defining the electricity price as a dynamic switching threshold. If the electricity price is in the peak period, the set temperature of the air conditioner is increased by 1 - 3°C to reduce the power consumption from the power grid; if the electricity price is in the valley period, the air-conditioning load demand for the next day is reduced by precooling the building envelope. For example, when the electricity price is in the peak period and the photovoltaic output is sufficient, the set temperature of the air conditioner is increased to 27 - 28°C; when the electricity price is in the peak period and the photovoltaic output is insufficient, the power consumption from the power grid is limited to no more than 80% of the predicted load demand. In the valley period of the electricity price, the air-conditioning system is controlled to operate at least 4 hours at a temperature 2 - 3°C lower than the conventional set temperature; the operating cost is minimized by dynamically powering the air conditioner. The purpose of such a setting is that in the prior art, during the peak period at noon when the photovoltaic output is sufficient, if the air-conditioning load suddenly increases and additional power needs to be purchased from the power grid, the operating cost will increase by more than 30%. Therefore, in this embodiment, a linkage mechanism between electricity price data and load dispatching is set up to actively reduce the electricity demand during high-price periods when the photovoltaic output fluctuates.

[0102] The piecewise linear function that defines the electricity price as a dynamic switching threshold is:

[0103]

[0104] Among them, η th (t) is the photovoltaic power supply switching threshold at time t, indicating the percentage of photovoltaic output that needs to reach the load demand to allow switching to photovoltaic power supply; π(t) is the electricity price at time t, with the unit of yuan / kWh.

[0105] Furthermore, if the electricity price is in the valley period, the steps of reducing the air-conditioning load demand for the next day by precooling the building envelope include:

[0106] Define the Lyapunov function:

[0107]

[0108] Ensure the asymptotic stability of the system during mode switching by designing the switching logic such that V < 0;

[0109] Based on the unsteady heat transfer theory, the wall temperature distribution T(x, t) satisfies:

[0110]

[0111] Among them, T(x, t) is the temperature distribution at the position of the wall and time t, α is the thermal diffusivity of the material, with the unit of m 2 / s, and the calculation formula of α is Among them: k is the thermal conductivity of the wall, with the unit of W / (m·K), ρ is the density of the wall, with the unit of kg / m 3 ; c pis the specific heat capacity of the wall, with the unit of J / (kg·K);

[0112] The boundary conditions are as follows:

[0113]

[0114] By using the method of separation of variables to solve, the relationship between the pre-cooling duration τ and the temperature drop ΔT the next day is obtained:

[0115]

[0116] where β n is the root of the characteristic equation, and C n is the coefficient determined by the initial temperature distribution; based on this, the pre-cooling duration is formulated to make the temperature rise amplitude during the peak period the next day controllable.

[0117] Specifically, the inventor has found through research that the thermal delay time of the concrete wall of the library can reach 4 - 6 hours. However, when the existing technology uses photovoltaic power to supply power to the air conditioner, the thermal energy storage characteristics of the building envelope structure are not fully utilized. Therefore, in this embodiment, the thermal energy storage potential of the building envelope structure is developed, and the load is transferred across time periods through pre-cooling / pre-heating to further reduce the operating cost.

[0118] Furthermore, this embodiment can also set a multiple fault tolerance mechanism:

[0119] When the LSTM model and the random forest model fail to predict, switch to the historical mean mode (that is, use the mean of the historical photovoltaic output data and air conditioner load data as the photovoltaic output prediction value and air conditioner load prediction value). When the constraint conditions of the linear programming model conflict, gradually relax the temperature limit, and trigger an emergency stop when the switching timeout occurs. In addition, considering that the fixed threshold switching strategy in the existing technology (such as switching when the photovoltaic output ≥ 80% of the load) is difficult to adapt to sudden weather changes or prediction errors. Frequent switching of the power supply mode will accelerate the wear of the contactor and cause excessive temperature fluctuations indoors. This embodiment can also suppress the frequent switching of the power supply mode through a hysteresis interval, and compensate for the prediction error by combining adaptive threshold correction (gain coefficient γ = 0.2) to ensure that the temperature fluctuation can still be controlled within ±1.5°C under the condition of ±20% fluctuation of the photovoltaic output.

[0120] Please refer to Figure 2 , in the second aspect, the present invention provides a non-energy storage type photovoltaic air conditioner dynamic power supply system, including:

[0121] A predicted value acquisition module 1 for acquiring the predicted value of photovoltaic output and the predicted value of air conditioner load

[0122] The linear programming module 2 is used to construct a linear programming model with the goal of minimizing the operating cost, taking the time-of-use electricity price data as the input. The constraint conditions of the linear programming model are as follows:

[0123]

[0124] Among them, P grid (t) is the power grid power consumption at time t, and T(t) is the indoor temperature;

[0125] The dynamic power supply module 3 is used to dynamically supply power to the air conditioner based on the linear programming model according to the power supply strategy and the time-of-use electricity price response strategy;

[0126] The power supply strategy includes: the priority of photovoltaic power supply is greater than that of power grid power consumption, and when the photovoltaic output is insufficient, it switches to power grid power consumption;

[0127] The time-of-use electricity price response strategy includes: defining the electricity price as a dynamic switching threshold. If the electricity price is in the peak period, the set temperature of the air conditioner is increased by 1 - 3 °C to reduce the power grid power consumption; if the electricity price is in the low period, the air-conditioning load demand of the next day is reduced by precooling the building envelope.

[0128] The dynamic power supply module includes a power grid switch, and the power grid switch includes:

[0129] The first contactor K1 in the photovoltaic power supply circuit;

[0130] The second contactor K2 in the power grid power supply circuit;

[0131] The PLC controller is used to send interlock control signals to K1 and K2 to ensure that the two circuits are not turned on at the same time;

[0132] A unloading resistor is also connected in parallel in the photovoltaic power supply circuit, which is used to consume the surplus electric energy when the photovoltaic output exceeds the air-conditioning load demand and prevent reverse flow to the power grid.

[0133] Specifically, for the specific implementation manner of the photovoltaic air-conditioning dynamic power supply system without energy storage, please refer to the specific implementation manner of the photovoltaic air-conditioning dynamic power supply method without energy storage, and no more details will be elaborated here.

[0134] Please refer to Figure 3 , Thirdly, the present invention provides an electronic device, including:

[0135] The memory 5 is used to store one or more programs;

[0136] The processor 4; the processor 4 is connected to the memory 5 through the communication interface 6;

[0137] When one or more programs are executed by the processor 4, the above-mentioned all or part of the methods are implemented.

[0138] Fourthly, the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor 4, the above-mentioned all or part of the methods are implemented.

[0139] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above-mentioned exemplary embodiments, and without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced within the present invention. Any reference signs in the claims should not be construed as limiting the claims concerned.

Claims

1. A dynamic power supply method for a photovoltaic air conditioner without energy storage, characterized in that, including: Obtain the predicted value of photovoltaic power output and the predicted value of air-conditioning load Taking time-of-use electricity price data as input, a linear programming model is constructed with the goal of minimizing operating costs. The constraint conditions of the linear programming model are: Among them, P grid (t) is the power consumption of the power grid in the t period, and T(t) is the indoor temperature; Based on the linear programming model, the air conditioner is dynamically powered according to the power supply strategy and the time-of-use electricity price response strategy; The power supply strategy includes: the priority of photovoltaic power supply is greater than that of grid power supply, and when the photovoltaic output is insufficient, it switches to grid power supply; The time-of-use electricity price response strategy includes: defining the electricity price as a dynamic switching threshold. If the electricity price is in the peak period, the set temperature of the air conditioner is increased by 1 - 3 °C to reduce the grid power consumption; if the electricity price is in the valley period, the air-conditioning load demand of the next day is reduced by precooling the building envelope.

2. The dynamic power supply method of a non-energy storage type photovoltaic air conditioner according to claim 1, characterized in that, Obtaining the predicted value of photovoltaic output The steps include: Using the LSTM model as the photovoltaic output prediction model, define the input sequence X t = [G(t - τ), T a (t - τ), P pv (t - τ)], where τ = 0, 1, …, 23 represents the data of the past 24 hours, and X t represents the input feature vector at time τ, G(τ) represents the solar irradiance at time τ; T a (τ) represents the ambient temperature at time τ; P pv (τ) represents the historical actual photovoltaic output at time τ; the hidden state update equation of the LSTM model is: h t =tanh(W h ·[h t-1 ,x t +b h ); where h t represents the hidden state vector at the current time t, W h represents the hidden state weight matrix, b h represents the hidden state bias term, h t-1 represents the hidden state vector at the previous time, and xt represents the input data at the current time; The output layer is mapped to the predicted value for the next 24 hours through a fully connected network: Among them, represents the predicted value of PV output power from 0 to 24 hours when the time period t starts from 0, W o represents the weight matrix of the output layer; h t represents the hidden state vector at the current moment; b o represents the bias term of the output layer, which adjusts the reference offset of the predicted value.

3. A dynamic power supply method for a non-energy storage type photovoltaic air conditioner according to claim 2, characterized in that Obtaining the predicted value of air-conditioning load The steps include: The random forest model is adopted as the air - conditioner load prediction model. Let the input feature set F = {N(t), T out (t), H(t), D(t)}, where D(t) ∈ {0, 1} represents the date type, N(t) is the number of people at time t, and T out (t) is the temperature at time t. The splitting nodes are selected by minimizing the Gini impurity to generate a set of decision trees. The final air - conditioner load prediction value is the weighted average of the outputs of each tree, specifically: Among them, is the predicted air-conditioning load demand at time period t, M is the number of decision trees, and f m is the prediction function of the m-th tree.

4. A dynamic power supply method for a photovoltaic air conditioner without energy storage according to claim 3, characterized in that The linear programming model is: Among them, C total is the daily average total operating cost of the system, in yuan; P grid (t) is the electricity consumption taken from the power grid during period t, in kW; π(t) is the time-of-use electricity price during period t, in yuan / kWh; λ is the temperature deviation penalty coefficient, in yuan / °C, with a default value of 10; T(t) is the actual indoor temperature during period t, in °C; Tset(t) is the air-conditioning set temperature during period t, in °C; max(x, 0) represents taking the larger value between x and 0, and is used to quantify the penalty cost for the part where the temperature exceeds the standard.

5. A dynamic power supply method for a non-energy storage type photovoltaic air conditioner according to claim 4, characterized in that The piecewise linear function that defines the electricity price as a dynamic switching threshold is: Among them, η th (t) is the photovoltaic power supply switching threshold in the t period, indicating the percentage of photovoltaic output that needs to reach the load demand to allow switching to photovoltaic power supply; π(t) is the electricity price in the t period, with the unit of yuan / kWh.

6. A dynamic power supply method for a photovoltaic air conditioner without energy storage according to claim 4, characterized in that If the electricity price is in the valley period, the steps to reduce the air-conditioning load demand of the next day by precooling the building envelope include: Defining a Lyapunov function: By designing the switching logic to make V < 0, it is ensured that the system is asymptotically stable during the mode switching process; Based on the unsteady heat transfer theory, the wall temperature distribution T(x, t) satisfies: Among them, T(x,t) is the temperature distribution at the location of the wall and time t, and α is the thermal diffusivity of the material, with the unit of m 2 / s. The calculation formula of α is Among them: k is the thermal conductivity of the wall, with the unit of W / (m·K), ρ is the density of the wall, with the unit of kg / m 3 ; c p is the specific heat capacity of the wall, with the unit of J / (kg·K); The boundary conditions are: Solving by the method of separation of variables, the relationship between the precooling duration τ and the temperature drop ΔT of the next day is obtained: where β n is the root of the characteristic equation, and C n is the coefficient determined by the initial temperature distribution; based on this, the pre-cooling duration is formulated to make the room temperature rise controllable during the peak period of the next day.

7. A photovoltaic air-conditioning dynamic power supply system without energy storage, characterized in that, including A predicted value acquisition module, configured to obtain a predicted value of photovoltaic power output and a predicted value of air-conditioning load A linear programming module, configured to use time-of-use electricity price data as input to construct a linear programming model with the objective of minimizing the operating cost. The constraint conditions of the linear programming model are as follows: Among them, P grid (t) is the power consumption of the power grid in the t period, and T(t) is the indoor temperature; A dynamic power supply module for dynamically powering the air conditioner based on the linear programming model according to the power supply strategy and the time-of-use electricity price response strategy; The power supply strategy includes: the priority of photovoltaic power supply is greater than that of grid power supply, and when the photovoltaic output is insufficient, it switches to grid power supply; The time-of-use electricity price response strategy includes: defining the electricity price as a dynamic switching threshold. If the electricity price is in the peak period, the set temperature of the air conditioner is increased by 1 - 3 °C to reduce the grid power consumption; if the electricity price is in the valley period, the air-conditioning load demand of the next day is reduced by precooling the building envelope.

8. A dynamic power supply system for a photovoltaic air conditioner without energy storage according to claim 7, characterized in that The dynamic power supply module includes a grid switch, and the grid switch includes: The first contactor K1 in the photovoltaic power supply circuit; The second contactor K2 in the grid power supply circuit; A PLC controller for sending an interlock control signal to K1 and K2 to ensure that the two circuits are not conducted simultaneously; The photovoltaic power supply circuit is also connected in parallel with a load dump resistor for consuming surplus electric energy when the photovoltaic output exceeds the air-conditioning load demand to prevent backflow to the grid.

9. An electronic device, characterized in that, including: A memory for storing one or more programs; A processor; When the one or more programs are executed by the processor, the method described in any one of claims 1 - 6 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the method described in any one of claims 1 - 6 is implemented.

Images