Coupling mechanism and intelligent control method of energy production and consumption for zero-carbon building operation in tropical zone

Abstract

The application relates to the field of building energy saving and intelligent control technology, and discloses a hot-tropical zero-carbon building operation energy production and consumption coupling mechanism and intelligent regulation and control method, which comprises the following steps: S1, building related data acquisition and preprocessing; S2, constructing a photovoltaic power generation quantification model, constructing an air conditioner refrigeration load and electric power quantification model, and constructing other electric load prediction models; S3, constructing a photovoltaic-air conditioner refrigeration-other electric load coupling mechanism model; S4, constructing a multi-objective optimization model; and S5, rolling time domain intelligent regulation and control. The hot-tropical zero-carbon building operation energy production and consumption coupling mechanism and intelligent regulation and control method, under the premise of only considering the refrigeration working condition, unifies photovoltaic power generation power, air conditioner refrigeration electric power and other electric loads into an electric energy balance equation, quantitatively describes the quantitative relationship between the electric loads and the grid power purchase power through a power boundary and a power fluctuation penalty term, and forms a coupling mechanism model which can be directly used for engineering design and simulation.

Description

Technical Field

[0001] This invention relates to the field of building energy conservation and intelligent control technology, specifically to a coupling mechanism and intelligent control method for energy production and consumption in tropical zero-carbon buildings. Background Technology

[0002] Due to the characteristics of high temperatures, high radiation, and long cooling load periods throughout the year in tropical regions, the proportion of air conditioning in building energy consumption is significantly higher than in temperate regions. In order to reduce operating energy consumption and carbon emissions, more and more buildings are integrating solar photovoltaic systems on rooftops or facades to partially offset the electricity demand of air conditioning and other electrical loads through on-site power generation, thereby achieving "zero carbon" or "near-zero carbon" operation.

[0003] In existing technologies, on the one hand, photovoltaic systems are mostly based on "generating as much as possible and connecting it to the grid" or simply for local self-consumption, without fully considering the short-term fluctuations in photovoltaic output, intraday peak-valley differences, and time-scale matching relationships with air conditioning cooling loads and other electrical loads; on the other hand, air conditioning control mostly adopts fixed temperature settings or simple time-period control, lacking a comprehensive response to time-of-use changes in electricity prices, grid carbon emission factors, and photovoltaic power generation, making it difficult to achieve coordination between prioritizing local photovoltaic consumption and peak shaving and valley filling.

[0004] Meanwhile, existing research on zero-carbon buildings in tropical regions mostly focuses on optimizing building envelopes, improving equipment energy efficiency, or individual demand-side response strategies. It lacks a systematic characterization of the quantitative coupling mechanism of the entire system—"photovoltaic power generation capacity – integrated energy consumption primarily for cooling"—during the operational phase. Furthermore, it lacks multi-objective intelligent control methods that can be directly applied to engineering implementation, balancing operating costs, carbon emissions, and indoor comfort. In engineering practice, the following problems often arise:

[0005] 1. The mismatch between the peak output of photovoltaic power and the peak cooling load means that a large amount of electricity still needs to be purchased from the grid under high radiation and high temperature conditions, which weakens the emission reduction effect of zero-carbon buildings;

[0006] 2. Control strategies are mostly oriented towards a single objective (such as energy consumption or comfort), failing to fully consider electricity prices and grid carbon emission factors, making it difficult to quantify and optimize operating costs and carbon emissions;

[0007] 3. Existing control methods are not very adaptable to weather fluctuations and load uncertainties in tropical regions, and lack rolling optimization strategies that combine prediction models with self-learning mechanisms;

[0008] 4. The lack of an interpretable mathematical model at the system level that unifies photovoltaic power generation models, cooling load models, other electrical load models, and grid power fluctuation constraints poses difficulties for engineering design, commissioning, and operation and maintenance.

[0009] Therefore, it is necessary to propose a coupling mechanism and intelligent control method and system for photovoltaic power generation and integrated energy consumption mainly based on air conditioning in the operation phase of zero-carbon buildings in tropical regions. Under the premise of considering only the cooling operation, it is necessary to achieve coordinated optimization and intelligent operation of photovoltaic power generation, cooling-dominated energy consumption and other electrical energy consumption. Summary of the Invention

[0010] To address the shortcomings of existing technologies, this invention provides a coupling mechanism and intelligent control method for the operation of zero-carbon buildings in tropical regions, which has the advantages of achieving efficient synergy between photovoltaic power generation and energy consumption, primarily for cooling, through a clear mathematical model, executable optimization control steps, and a modular system architecture. It solves the problems of unclear coupling mechanism, difficulty in balancing multiple conflicting objectives, insufficient adaptability to uncertainties in tropical conditions, and insufficient engineering feasibility in existing technologies.

[0011] To achieve the aforementioned goal of efficient synergy between photovoltaic power generation and energy consumption primarily for cooling through a clear mathematical model, executable optimization control steps, and a modular system architecture, this invention provides the following technical solution: a coupling mechanism and intelligent control method for power generation and energy consumption in tropical zero-carbon buildings. The power generation side is a solar photovoltaic power generation system integrating dynamic facade shading, and the energy consumption side is the building's comprehensive electrical load primarily for air conditioning and cooling, excluding heating operations. The method includes the following steps:

[0012] S1. Acquire outdoor weather and building-related data at different times in the tropical region and perform preprocessing. The building-related data includes discrete time data. Horizontal total solar radiation Outdoor air temperature Indoor air temperature Relative humidity, photovoltaic array backsheet temperature, air conditioning power Other electrical load power Electricity price and power grid carbon emission factors The time step is ;

[0013] S2, the total horizontal solar radiation The equivalent irradiance of the photovoltaic array surface is converted using a conversion formula. Construct a photovoltaic power generation quantification model to calculate photovoltaic power generation output. ;

[0014] Calculate building cooling load based on building-related data. Based on this, by introducing the cooling energy efficiency ratio, a quantitative model of air conditioning cooling load and electrical power is constructed to calculate the air conditioning cooling power. ;

[0015] Based on building-related data, a time-segmentation and data-driven approach is used to construct a data structure that includes the power of other electrical loads. The predictive model predicts the power of other electrical loads at the next moment;

[0016] S3. Based on the results obtained in steps S1-S2, establish the power balance relationship, form a coupling mechanism model of photovoltaic-air conditioning-other loads in the operation phase, and calculate the power purchased by the power grid.

[0017] S4. Based on building-related data and power grid purchase power, and using the air conditioning set temperature sequence as the decision variable, while simultaneously satisfying multiple constraints of indoor temperature, air conditioning power and air conditioning set temperature, a final multi-objective optimization model is constructed. An intelligent optimization strategy combining reinforcement learning and evolutionary algorithm is adopted to obtain the optimal control solution that satisfies the constraints.

[0018] S5. Based on steps S1-S4, a rolling time-domain predictive control strategy is adopted to update the photovoltaic power generation prediction value for the next time step using irradiance, outdoor temperature, and other electrical loads. Predicted cooling load power Other electrical load power forecasts Through a final multi-objective optimization model, the optimal air conditioning set temperature and corresponding air conditioning control quantity at the current moment are obtained, and the photovoltaic control quantity corresponding to the photovoltaic side control strategy is determined simultaneously. The air conditioning control quantity is converted into a set temperature command for the air conditioning controller, and the photovoltaic control quantity is converted into an output control command for the photovoltaic inverter, and then sent to the corresponding equipment for execution, prioritizing the use of photovoltaic power and suppressing... Fluctuations are controlled only at the current moment; the process of repeated updates and optimizations enables intelligent and coordinated regulation of energy production and consumption in tropical zero-carbon buildings under cooling-only conditions.

[0019] Preferably, in step S2, a photovoltaic power generation quantification model is constructed to calculate the photovoltaic power generation, as shown in formula (1):

[0020] (1)

[0021] In formula (1), For photovoltaics Photovoltaic power generation at any given time For reference photovoltaic efficiency, For effective light-receiving area, The temperature power coefficient, This refers to the actual temperature of the photovoltaic module. This is a reference temperature.

[0022] Preferably, in step S2, a quantitative model of air conditioning cooling load and electrical power is constructed, and the air conditioning cooling power is calculated, as follows:

[0023] Calculate the building cooling load under cooling conditions. Its calculation is shown in formula (2):

[0024] (2)

[0025] In formula (2), The equivalent heat transfer coefficient of the building envelope. For equivalent enclosure area, To allow solar heat to enter the interior through the building envelope. For heat gain inside personnel and equipment;

[0026] Calculate the cooling power of the air conditioner As shown in formula (3):

[0027] (3)

[0028] In formula (3), This refers to the cooling energy efficiency ratio.

[0029] Preferably, in step S2, the calculation of the prediction model is as shown in formula (4):

[0030] (4)

[0031] In formula (4), The time period to which the time belongs. For weekday / weekend type, To fit the obtained nonlinear function, For other electrical loads at the current moment, For updatable parameters, This is the predicted power value of other electrical loads at the next moment.

[0032] Preferably, in step S3, an energy balance relationship is established, forming a photovoltaic-air conditioning-other load coupling mechanism model for the operation phase, and the power purchased by the power grid is calculated, as follows:

[0033] The calculation of the electrical energy balance relationship is shown in formula (5):

[0034] (5)

[0035] In formula (5), Power purchased from the power grid;

[0036] Apply power boundary constraints, as shown in equation (6):

[0037] (6)

[0038] In formula (6), Minimum power purchase capacity, Maximum power purchase capacity;

[0039] In this process, formula (5) is transformed into a vector form in the prediction time domain, as shown in formula (7):

[0040] (7)

[0041] In formula (7), Forecast sequence of power purchased from the power grid, This is a predicted sequence of air conditioning cooling power. For other electrical load power prediction sequences, This is a photovoltaic power generation prediction sequence;

[0042] In the formula,

[0043] ,

[0044] ,

[0045] ,

[0046] ;

[0047] Add grid power fluctuation penalty item Its calculation is shown in formula (8):

[0048] (8)

[0049] In formula (8), The fluctuation penalty coefficient, For the first Time and the The difference in power purchase capacity of the power grid at any given time This represents the total number of moments.

[0050] Preferably, in step S4, the final multi-objective optimization model is constructed as follows:

[0051] ① Set the air conditioner to a specific temperature sequence. Using optional photovoltaic side control strategies as decision variables, and with electricity purchase cost, carbon emissions, and comfort deviation as objectives, an initial multi-objective optimization model is constructed, the calculation of which is shown in formula (9):

[0052] (9)

[0053] In formula (9), As weight, For the target comfortable temperature, To minimize the overall cost;

[0054] The constraints on indoor temperature and air conditioning power are applied, as shown in formula (10):

[0055] (10)

[0056] In formula (10), This represents the lower limit of indoor temperature. This represents the upper limit of indoor temperature. This refers to the rated maximum cooling capacity of the air conditioner.

[0057] Apply the air conditioning set temperature boundary, the boundary of which is shown in formula (11):

[0058] (11)

[0059] In formula (11), Set the lower limit of the temperature for the air conditioner. Set the upper limit of the temperature for the air conditioner;

[0060] ② In formula (9), the target comfort temperature has a comfort deviation, which is based not only on the square of the temperature deviation, but also on a simplified comfort index that includes humidity. , i.e., quantification The degree of deviation between the indoor environment at any given time and the comfort state expected by a person is calculated as shown in formula (12): (12)

[0061] In formula (12), This refers to indoor relative humidity. For the target relative humidity, These are the weighting coefficients;

[0062] Calculate comfort target items The comfort index at each moment is summed up, and its calculation is shown in formula (13):

[0063] (13)

[0064] ③ Operating costs With carbon emissions The calculations are shown in formulas (14) and (15) respectively:

[0065] (14)

[0066] (15)

[0067] In formulas (14) and (15), For time step;

[0068] ④ The final multi-objective optimization model is constructed using a normalized weighted form, and the calculation is shown in formula (16):

[0069] (16)

[0070] In formula (16), For reference only. For normalized weights.

[0071] Preferably, in step S4, an intelligent optimization strategy combining reinforcement learning and evolutionary algorithms is used to obtain the optimal control solution that satisfies the constraints, as follows:

[0072] ① Construct an instantaneous cost function using the air conditioner set temperature change and the photovoltaic side control strategy as the action space. The calculation is shown in formula (17):

[0073] (17)

[0074] By using reinforcement learning algorithms to learn near-optimal policies, blind search can be avoided;

[0075] ②The air conditioning set temperature sequence and photovoltaic side control strategy in the entire prediction time domain are used as individual codes. The final multi-objective optimization model of formula (16) is used as the fitness. The evolutionary algorithm is used to perform global optimization. Under the premise of satisfying the constraints, the optimal control solution that satisfies the constraints is obtained.

[0076] Preferably, in step S5, a rolling time-domain predictive control strategy is used to predict photovoltaic power generation, cooling load, and other electrical loads, and the calculation is shown in formulas (18)-(20):

[0077] (18)

[0078] (19)

[0079] (20)

[0080] In the formula, This is the predicted value of photovoltaic power generation at the next moment. This is the predicted total cooling load of the building at the next moment. Forecast values ​​of other building loads at the next moment. , , These are internally adjustable parameters. , , For prediction functions;

[0081] By minimizing the prediction error, online self-learning correction of key parameters in the coupling mechanism model is achieved, and its calculation is shown in formula (21):

[0082] (twenty one)

[0083] Preferably, the optimization decision module includes:

[0084] Target normalization unit, used for Normalization is performed according to formula (16);

[0085] The constraint processing unit is used to transform the constraints shown in formulas (6), (10), and (11) into penalty functions or projection operators to ensure a feasible solution.

[0086] The rolling optimization unit is used to solve the finite-time domain optimization problem based on the updated prediction sequence and coupling mechanism model at each control moment, and only outputs the control quantity of the current step, thereby realizing the coordinated optimization of photovoltaic power generation, air conditioning cooling energy consumption and other electrical energy consumption under cooling-only conditions.

[0087] Compared with existing technologies, this invention provides a coupling mechanism and intelligent control method for energy production capacity and energy consumption in tropical zero-carbon buildings, which has the following beneficial effects:

[0088] 1. The coupling mechanism and intelligent control method of the operating capacity and energy consumption of the tropical zero-carbon building, under the premise of only considering the cooling condition, integrates the photovoltaic power generation, air conditioning cooling power and other electrical loads into the power balance equation. The quantitative relationship between the power boundary and the power purchased by the grid is clearly described through the power boundary and power fluctuation penalty term, forming a coupling mechanism model that can be directly used for engineering design and simulation, overcoming the problem of "unclear system-level relationship" in traditional methods.

[0089] 2. The coupling mechanism and intelligent control method of the operating capacity and energy consumption of the tropical zero-carbon building constructs a multi-objective optimization function that includes operating costs, equivalent carbon emissions and comfort indicators. The weights can be flexibly configured according to different scenarios to achieve a quantitative trade-off between "low cost, low carbon and high comfort". Compared with the control method that only takes energy consumption or indoor temperature as a single objective, it is more in line with the comprehensive operating needs of tropical zero-carbon buildings.

[0090] 3. The coupling mechanism of energy production capacity and energy consumption in the operation of the tropical zero-carbon building and the intelligent control method, through short-term prediction models of photovoltaic output, cooling load and other electrical loads and online learning error correction, enable the present invention to maintain the consistency between the model and actual operation under typical tropical conditions such as high radiation, rapid changes in cloud cover and outdoor temperature fluctuations; combined with the intelligent solution mechanism of rolling time-domain optimization and reinforcement learning-evolutionary algorithm, the control strategy can be continuously optimized to improve the robustness and economy of the system in uncertain environments.

[0091] 4. The coupling mechanism and intelligent control method of energy production capacity and energy consumption in the operation of this tropical zero-carbon building are achieved by explicitly considering the grid power fluctuation penalty term in the objective function. Furthermore, by prioritizing the use of photovoltaic power generation in control decisions, this invention can effectively reduce grid-side power peaks and fluctuations, promote local consumption of photovoltaic power, reduce distribution-side capacity and transmission-side pressure, and improve the overall benefits of zero-carbon buildings from a system-level perspective.

[0092] 5. The coupling mechanism of energy production and consumption in the operation of the tropical zero-carbon building and the intelligent control method propose a complete system module division and functional definition: data acquisition module, modeling and prediction module, optimization decision module, control execution module and strategy update module. It can be integrated into the existing building automation system (BAS) or energy management system (EMS) through software upgrade or modular expansion. It has clear interfaces and implementation paths and strong engineering feasibility.

[0093] 6. The coupling mechanism and intelligent control method of the operating capacity and energy consumption of the tropical zero-carbon building do not model the heating operation. Instead, the modeling and control focus on the cooling-dominated energy consumption scenario in tropical regions, making the model and algorithm more in line with actual operation and avoiding unnecessary complexity. Under the premise of ensuring comfort constraints, the building achieves efficient operation under typical cooling load-dominated conditions by optimizing the cooling load and coordinating with other electrical loads. Attached Figure Description

[0094] Figure 1 This is a flowchart illustrating the coupling mechanism and intelligent control method of building energy production capacity and energy consumption in this invention.

[0095] Figure 2 This is a schematic diagram of the coupling mechanism between building capacity and energy consumption and the intelligent control system structure of the present invention;

[0096] Figure 3 This is a schematic diagram of the photovoltaic-air conditioning-grid power coupling mechanism of the present invention. Detailed Implementation

[0097] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments and accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0098] Please see Figure 1A mechanism and intelligent control method for coupling energy production and consumption in tropical zero-carbon buildings are disclosed. The energy production side is a solar photovoltaic power generation system with integrated dynamic facade shading, and the energy consumption side is the building's comprehensive electrical load, mainly for air conditioning and cooling, without involving heating conditions. The method includes the following steps:

[0099] S1. Acquire outdoor weather and building-related data at different times in tropical regions, and perform time alignment, outlier removal, and interpolation completion; the building-related data includes at least discrete time intervals. Horizontal total solar radiation Outdoor air temperature Indoor air temperature Relative humidity, photovoltaic array backsheet temperature, air conditioning power Other electrical load power (Electric power consumed by all other electrical equipment in the building, excluding the energy consumption of the air conditioning system and the photovoltaic array itself), electricity price and power grid carbon emission factors The time step is ;

[0100] S2, the total horizontal solar radiation The equivalent irradiance of the photovoltaic array surface is converted using a conversion formula. Construct a photovoltaic power generation quantification model to calculate photovoltaic power generation output. ;

[0101] Calculate building cooling load based on building-related data. Based on this, by introducing the cooling energy efficiency ratio, a quantitative model of air conditioning cooling load and electrical power is constructed to calculate the air conditioning cooling power. ;

[0102] Based on building-related data, a time-segmentation and data-driven approach is used to construct a data structure that includes the power of other electrical loads. The predictive model predicts the power of other electrical loads at the next moment;

[0103] Specifically, in step S2, a photovoltaic power generation quantification model is constructed, and its calculation is shown in formula (1):

[0104] (1)

[0105] In formula (1), For photovoltaics Photovoltaic power generation at any given time For reference operating conditions (standard operating conditions are 25℃ and 1000W / m²), the photovoltaic efficiency is... The effective light-receiving area is expressed in m². The temperature power coefficient, The actual temperature of the photovoltaic module (collected by a sensor). Reference temperature (general 25℃);

[0106] Specifically, in step S2, a quantitative model of air conditioning cooling load and electrical power is constructed to calculate the air conditioning cooling power. The details are as follows:

[0107] Calculate the building cooling load considering only the cooling operation. The total heat removed from the room by the air conditioner per unit time is a key parameter that determines the energy consumption of the air conditioner. This cooling load consists of three parts: heat transferred from the building envelope, solar radiation heat, and internal heat dissipated by people and equipment in the room. Its calculation is shown in formula (2).

[0108] (2)

[0109] In formula (2), The equivalent heat transfer coefficient of the building envelope is expressed in W / (m²·℃), reflecting the overall thermal insulation performance of the building envelope, including walls and windows. The equivalent enclosure area is expressed in m². Heat transferred to the building envelope; The solar radiation heat entering the room through the building envelope is measured in W. The internal heat dissipated by people and equipment inside the room, measured in W;

[0110] Calculate the cooling power of the air conditioner As shown in formula (3):

[0111] (3)

[0112] In formula (3), For cooling energy efficiency ratio (e.g., That is, consuming 1 unit of electricity to generate the heat equivalent to 4 units of electricity. (Higher efficiency means greater energy saving; does not involve heating.)

[0113] Specifically, in step S2, the calculation of the prediction model is as shown in formula (4):

[0114] (4)

[0115] In formula (4), The time period to which the time belongs. For weekday / weekend type, To fit the obtained nonlinear function, For other electrical loads at the current moment, For updatable parameters, For the predicted power values ​​of other electrical loads at the next time step;

[0116] The specific calculation process of formula (4) is as follows:

[0117] ①The source and quantification definition of input data

[0118] Time period Based on time step The 24 hours of a natural day are divided into equal time periods, and quantified using digital coding, for example, by... The 24 hours are divided into 96 time periods. The value range is 1 to 96; The value of comes from discrete time. Time attribute matching results;

[0119] Weekday / Weekend Type Binary encoding is used for quantization, with a value of 1 for weekdays and 0 for weekends and public holidays. The value is determined by the public calendar table and discrete time. The date information was matched to obtain the result;

[0120] Other electrical load power at the current moment The building-related data also includes at least the total power value of all electrical equipment within the building, excluding the air conditioning system and the photovoltaic array's own operating energy consumption. This data is directly obtained after outlier removal and filtering, with appropriate acquisition accuracy and time step. Maintain consistency.

[0121] ② Updatable parameters Acquisition and Update

[0122] Initial parameter acquisition: Select at least 30 days of building-related data as the training set. The training set data dimensions are consistent with the input data, and preprocessing including time alignment, outlier removal, and linear interpolation completion is performed. Based on a time segmentation strategy, ... and Different combinations of sub-training sets are constructed, and data-driven methods such as multinomial fitting, gradient boosting trees, or shallow BP neural networks are used to fit each sub-training set to obtain the initial parameters of each model in the system. , These are the model parameters for the corresponding fitting methods, such as the coefficients of a polynomial fit, the weights and biases of a neural network, and the node splitting parameters of a decision tree.

[0123] Online parameter updates: A sliding window mechanism is used, with a fixed update window (e.g., 7 days). Building-related data generated during building operation is continuously added to the training set, while historical building-related data outside the window is removed. The models are then refitted and updated. The value of is chosen to ensure the model's adaptability to changes in building electricity consumption patterns.

[0124] ③Nonlinear functions Construction

[0125] Construction based on a two-dimensional modeling strategy of time segmentation and data-driven approach Firstly, according to Time period coding and The binary encoding is used to match the corresponding segmented sub-models; each model is a non-linear mapping function, obtained by fitting the training set data, and is used to achieve the transition from... arrive The nonlinear mapping and fitting process take the minimum mean square error between the predicted and measured values ​​as the objective function.

[0126] ④ Specific calculation steps

[0127] At discrete time Extract from preprocessed building-related data , , The specific quantitative value;

[0128] according to and The value of is matched with the corresponding models and currently valid updatable parameters in formula (4). ;

[0129] Will , , Substitute into valid nonlinear functions In the process, the next moment is calculated using a function. Other electrical load power forecasts .

[0130] S3. Based on the results obtained in steps S1-S2, establish the power balance relationship, form a coupling mechanism model of photovoltaic-air conditioning-other loads during the operation phase, and calculate the power purchased by the power grid. ;

[0131] Specifically, in step S3, an energy balance relationship is established, a photovoltaic-air conditioning-other load coupling mechanism model is formed during the operation phase, and the power purchased by the power grid is calculated, as follows:

[0132] (5)

[0133] In formula (5), Power purchased from the power grid;

[0134] To ensure the safe and compliant operation of the system, power boundary constraints are applied, as shown in formula (6):

[0135] (6)

[0136] In formula (6), Minimum power purchase capacity, Maximum power purchase capacity;

[0137] In the prediction time domain, formula (5) is transformed into vector form, as shown in formula (7):

[0138] (7)

[0139] In formula (7), Forecast sequence of power purchased from the power grid, This is a predicted sequence of air conditioning cooling power. For other electrical load power prediction sequences, This is a predicted sequence of active power generation.

[0140] In the formula,

[0141] ,

[0142] ,

[0143] ,

[0144] ;

[0145] Add grid power fluctuation penalty item Its calculation is shown in formula (8):

[0146] (8)

[0147] In formula (8), The fluctuation penalty coefficient, For the first Time and the The difference in power purchase capacity of the power grid at any given time This represents the total number of moments.

[0148] S4. Based on building-related data and power grid purchase capacity, and using the air conditioning set temperature sequence as the decision variable, while simultaneously satisfying multiple constraints such as indoor temperature, air conditioning power, and air conditioning set temperature, a final multi-objective optimization model is constructed. An intelligent optimization strategy combining reinforcement learning and evolutionary algorithms is employed to obtain the optimal control sequence that satisfies the constraints.

[0149] Specifically, in step S4, the final multi-objective optimization model is constructed as follows:

[0150] ① Set the air conditioner to a specific temperature sequence. Using optional photovoltaic side control strategies as decision variables, and with electricity purchase cost, carbon emissions, and comfort deviation as objectives, a multi-objective optimization function is constructed, reflecting the synergistic requirements of economy, environmental protection, and comfort. The calculation of the multi-objective optimization function is shown in formula (9):

[0151] (9)

[0152] In formula (9), As weight, For the target comfortable temperature, To minimize the overall cost;

[0153] Indoor temperature and air conditioning power constraints are applied, as shown in formula (10):

[0154] (10)

[0155] In formula (10), This represents the lower limit of indoor temperature. This represents the upper limit of indoor temperature. This refers to the rated maximum cooling capacity of the air conditioner.

[0156] Apply the air conditioning set temperature boundary, the boundary of which is shown in formula (11):

[0157] (11)

[0158] In formula (11), Set the lower limit of the temperature for the air conditioner. Set the upper limit of the temperature for the air conditioner;

[0159] ② In formula (9), the target comfort temperature has a comfort deviation, which is based not only on the square of the temperature deviation, but also on a simplified comfort index that takes humidity into account. , i.e., quantification The degree of deviation between the indoor environment at any given time and the comfort state expected by a person is calculated as shown in formula (12):

[0160] (12)

[0161] In formula (12), This refers to indoor relative humidity. For the target relative humidity, These are the weighting coefficients;

[0162] Calculate comfort target items The comfort index at each moment is summed up, and its calculation is shown in formula (13):

[0163] (13)

[0164] ③ Operating costs With carbon emissions The calculations are shown in formulas (14) and (15) respectively:

[0165] (14)

[0166] (15)

[0167] In formulas (14) and (15), For time step;

[0168] ④ The total objective function is constructed using a normalized weighted form, and its calculation is shown in formula (16):

[0169] (16)

[0170] In formula (16), For reference only. For normalized weights;

[0171] Specifically, in step S4, an intelligent optimization strategy combining reinforcement learning and evolutionary algorithms is used to obtain the optimal control solution that satisfies the constraints, as follows:

[0172] ① Construct an instantaneous cost function using the air conditioner set temperature change and the photovoltaic side control strategy as the action space. The calculation is shown in formula (17):

[0173] (17)

[0174] By using reinforcement learning algorithms to learn near-optimal policies, blind search can be avoided;

[0175] ②The air conditioning set temperature sequence and photovoltaic side control strategy in the entire prediction time domain are used as individual codes. The final multi-objective optimization model of formula (16) is used as the fitness. The evolutionary algorithm is used to perform global optimization. Under the premise of satisfying the constraints, the optimal control solution that satisfies the constraints is obtained.

[0176] S5. Based on steps S1-S4, a rolling time-domain predictive control strategy is adopted, using irradiance, outdoor temperature, and other electrical loads to update the predicted photovoltaic power generation value for the next time step. Predicted cooling load power Other electrical load power forecasts Through a final multi-objective optimization model, the optimal air conditioning set temperature and corresponding air conditioning control quantity at the current moment are obtained, and the photovoltaic control quantity corresponding to the photovoltaic side control strategy is determined simultaneously. The air conditioning control quantity is converted into a set temperature command for the air conditioning controller, and the photovoltaic control quantity is converted into an output control command for the photovoltaic inverter, and then sent to the corresponding equipment for execution, prioritizing the use of photovoltaic power and suppressing... Fluctuations are controlled only at the current moment; the process of repeated updates and optimizations enables intelligent and coordinated regulation of energy production and consumption in tropical zero-carbon buildings under cooling-only conditions.

[0177] Specifically, in step S5, a rolling time-domain predictive control strategy is used to predict photovoltaic output, cooling load, and other electrical loads, and the calculation is shown in formulas (18)-(20):

[0178] (18)

[0179] (19)

[0180] (20)

[0181] In the formula, This is the predicted value of photovoltaic power generation at the next moment. This is the predicted total cooling load of the building at the next moment. Forecast values ​​of other building loads at the next moment. , , These are internally adjustable parameters. , , For prediction functions;

[0182] By minimizing the prediction error, online self-learning correction of key parameters in the coupling mechanism model is achieved, and its calculation is shown in formula (21):

[0183] (twenty one)

[0184] A mechanism and intelligent control system for coupling energy production and consumption during the operation of tropical zero-carbon buildings, used to realize the mechanism and intelligent control method for coupling energy production and consumption during the operation of tropical zero-carbon buildings, including:

[0185] The data acquisition module is used to collect operating data of photovoltaic arrays, air conditioning equipment and other electrical loads, as well as indoor and outdoor environmental data.

[0186] The modeling and prediction module is used to construct and maintain the photovoltaic power generation model, air conditioning cooling model and other electrical load models described in formulas (1)-(5), and to make predictions for future periods based on formulas (18)-(21);

[0187] The optimization decision module is used to solve the optimal control sequence of air conditioning set temperature and photovoltaic side control strategy according to the multi-objective optimization model and intelligent optimization solution strategy of formula (9) or formula (16);

[0188] The control execution module is used to convert optimization decisions into control commands for the air conditioner and photovoltaic inverter and issue them for execution; when the system is configured with a tilt adjustment mechanism / dynamic facade shading mechanism, the control execution module further generates and issues tilt control commands / shading angle control commands to the corresponding mechanisms for execution;

[0189] The strategy update module is used to update the prediction model parameters and target weights based on operational feedback, so as to achieve precise and intelligent control of photovoltaic power generation and comprehensive energy consumption, mainly cooling, during the operation of tropical zero-carbon buildings.

[0190] Example 1:

[0191] This embodiment takes a 2000㎡ tropical office zero-carbon building in Sanya City as an example to carry out the coupling and intelligent control of energy production capacity and energy consumption. The south-facing facade of the building integrates a dynamic facade shading photovoltaic system, and the air conditioner only operates in cooling mode. The control is achieved according to the coupling mechanism and intelligent control method of energy production capacity and energy consumption of tropical zero-carbon building proposed in this invention.

[0192] (I) Specific steps:

[0193] Step S1: Data Acquisition and Preprocessing

[0194] by Using a time step, 30 consecutive days of operational data for the building were collected, and time alignment, outlier removal, and interpolation were performed.

[0195] Data collected: Total horizontal solar radiation : 0~1200W / ㎡; outdoor air temperature 25~35℃; Indoor air temperature Temperature range: 24~28℃; Relative humidity: 60%~85%; Photovoltaic array backsheet temperature: 28~65℃; Air conditioning power consumption. : 0~150kW; other electrical load power 20~50kW; Electricity price Peak hours: 0.8 yuan / kWh; Normal hours: 0.5 yuan / kWh; Off-peak hours: 0.3 yuan / kWh; Grid carbon emission factor 0.5~0.6 kg CO2 / kWh;

[0196] Preprocessing operations: using 12 were eliminated according to principle Abnormal data points are filled by linear interpolation to complete the missing data, for example, at a certain time. The previous value was 800W / ㎡, the subsequent value was 750W / ㎡, and the total value is 775W / ㎡.

[0197] Step S2:

[0198] ① Construct a quantitative model for photovoltaic power generation

[0199] Equivalent irradiance conversion: The real-time tilt angle of the photovoltaic array is dynamically adjusted from 0° to 60°. The formula for converting the total horizontal solar radiation into equivalent irradiance is as follows:

[0200]

[0201] in For real-time tilt angle;

[0202] when , When the equivalent irradiance is calculated, the formula is:

[0203]

[0204] Constructing a photovoltaic power generation model: using formulas The parameters are set as follows:

[0205] ,

[0206] ,

[0207] ,

[0208] ,

[0209]

[0210] when , hour, Substituting, we get:

[0211]

[0212] ② Construct a quantitative model of air conditioning cooling load and electrical power

[0213] The calculation formula for constructing the cooling load model is as follows:

[0214]

[0215] When setting , , , The power is negatively correlated with the shading angle, with 25kW at 0° and 8kW at 60°.

[0216] when , A shading angle of 30° corresponds to hour, ;

[0217] The calculation formula for constructing the air conditioner's electrical power model is as follows:

[0218]

[0219] The value varies with outdoor temperature: 3.5 at 30℃ and 3.0 at 35℃. , Insertion is worthwhile hour, .

[0220] ③ Construct prediction models for other electrical loads

[0221] Using a time-segmented and data-driven approach, the calculation formula for constructing a prediction model for other electrical loads is as follows:

[0222]

[0223] in, The time period to which the time belongs. For weekday / weekend type, To fit the obtained nonlinear function, For updatable parameters; the fitted result is For example, during weekday afternoons , hour, The error is 1.8% compared to the measured value of 27.5kW.

[0224] Step S5: Construct a coupling mechanism model of photovoltaic-air conditioning-other electrical loads

[0225] The formula for establishing the electrical energy balance relationship is as follows:

[0226]

[0227] The formula for calculating the applied power boundary constraints is as follows:

[0228] ,

[0229] when , , hour:

[0230]

[0231] That is, 8.11kW of electricity was sold to the grid, which meets the constraints.

[0232] Step S4: Construct the final multi-objective optimization model

[0233] The decision variable is set as the air conditioner set temperature series. Based on the temperature range of 24~28℃ and photovoltaic side control strategies, a multi-objective optimization function is constructed:

[0234]

[0235] set up , , , At the same time, apply , constraint.

[0236] Step S5: Scrolling Time Domain Intelligent Control

[0237] A rolling time-domain strategy with a 1-hour prediction time domain and a 15-minute control step size is adopted. The prediction model is updated using an LSTM irradiance prediction model with an error of ±5% and an outdoor temperature prediction model with an error of ±0.5℃. , , Solve the optimization problem every 15 minutes and execute only the current control variable;

[0238] If the next 15 minutes are predicted If the power output is reduced from 800W / ㎡ to 700W / ㎡, the photovoltaic tilt angle will be adjusted to 40° and the air conditioner temperature will be set to 27℃ to prioritize the absorption of photovoltaic power.

[0239] (II) Quantitative Case Settlement

[0240] The control effects of the system without intelligent control and the system under this scheme were compared on July 15th. The specific settlement process is as follows:

[0241] (1) Basic parameters

[0242] There were a total of 96 15-minute time steps on the day, with 24 steps during peak hours, 24 steps during normal hours, and 48 steps during off-peak hours; the average carbon emission factor of the power grid was 0.55 kgCO2 / kWh; the photovoltaic reference power and load power were obtained according to the above steps.

[0243] (2) Calculation of no control scheme

[0244] Photovoltaic power generation: Fixed tilt angle 20°, average equivalent irradiance 380W / ㎡, total power generation ;

[0245] Total electrical load: Total power consumption of air conditioner Other loads Total load ;

[0246] Grid Interaction: Electricity Purchase There is no electricity sales, with peak hours at 120 kWh, normal hours at 80 kWh, and off-peak hours at 148.17 kWh.

[0247] Electricity purchase cost: Yuan;

[0248] Carbon emissions: ;

[0249] Comfort deviation: Average indoor temperature 25.8℃, sum of squared deviations .

[0250] (3) Calculation of the control scheme in this embodiment

[0251] Photovoltaic power generation: With dynamic tilt angle adjustment and an average equivalent irradiance of 450W / ㎡, the total power generation is... This represents a year-on-year increase of 19.8%.

[0252] Total electrical load: Air conditioner set temperature dynamically adjusted, total power consumption , decreased by 14.7%, other loads Total load ;

[0253] Grid Interaction: Electricity Purchase Electricity sold to the grid during peak hours Net purchased electricity ;

[0254] Electricity purchase cost: Peak hour purchase of 95 kWh, normal hour purchase of 75 kWh, and off-peak hour purchase of 167.63 kWh. Yuan, a year-on-year decrease of 11.07%;

[0255] Carbon emissions: This represents a year-on-year decrease of 3.02%.

[0256] Comfort deviation: Average indoor temperature 26.1℃, sum of squared deviations This represents a year-on-year decrease of 34.1%.

[0257] (4) Settlement conclusion

[0258] This solution achieves a 11.07% reduction in electricity purchase costs, a 3.02% reduction in carbon emissions, and a 34.1% reduction in comfort deviation, while increasing photovoltaic power generation by 19.8%, thus achieving the synergistic optimization goal of "low cost, low carbon, and high comfort" for tropical zero-carbon buildings.

[0259] Example 2:

[0260] To verify the inventiveness and rationality of this invention, a complete control experiment was designed. Data was obtained through standardized experimental preparation, scheme, and process, and the performance differences between the two strategies were compared and analyzed, as follows:

[0261] (I) Experimental Preparation

[0262] 1. Experimental Site and Model:

[0263] The experiment was conducted using a 2000㎡ tropical office building model of the same size in the same area of ​​Sanya City. It was divided into an experimental group and a control group. The building envelope parameters, photovoltaic module specifications, and air conditioning equipment models of the experimental group and the control group were completely identical. The effective light-receiving area of ​​the photovoltaic modules was 500㎡ with a reference efficiency of 18%, the air conditioning cooling energy efficiency ratio ranged from 3.0 to 3.5, and the basic power of other electrical loads was 20 to 50kW.

[0264] 2. Experimental Apparatus

[0265] Equipped with a meteorological monitoring instrument, with monitoring accuracy of: solar radiation ±5W / ㎡, temperature ±0.2℃;

[0266] Power monitoring instrument, monitoring accuracy: power ±0.5%, energy consumption ±0.1kW·h;

[0267] Environmental monitoring instrument, monitoring accuracy: indoor temperature ±0.1℃, relative humidity ±2%;

[0268] Data acquisition instrument, acquisition frequency 15 minutes / time;

[0269] Ensure data synchronization.

[0270] 3. Basic parameter calibration

[0271] Before the experiment, the photovoltaic module power generation efficiency and air conditioning cooling efficiency of the experimental and control models were uniformly calibrated. The photovoltaic module generated 90kW under standard irradiance of 1000W / ㎡ and 25℃, and the air conditioning cooling efficiency ratio was 3.3 under an outdoor temperature of 32℃, thus eliminating the influence of equipment performance differences on the experiment.

[0272] (II) Experimental Design

[0273] 1. Control Group Strategy

[0274] The control group adopted a conventional fixed control scheme, as detailed below:

[0275] Photovoltaic system: fixed tilt angle of 20°, not adjusted with changes in solar radiation, implementing a grid connection strategy of "generating as much as it connects";

[0276] Air conditioning system: The temperature is set at a fixed 26℃ all year round and does not respond to changes in photovoltaic output or electricity prices;

[0277] Load management: Other electrical loads do not participate in coordinated control and operate according to their inherent electricity consumption patterns.

[0278] 2. Experimental Group Strategy

[0279] The experimental group adopted the intelligent control strategy of the present invention, namely the technical solution of Example 1, to dynamically adjust the photovoltaic tilt angle and the air conditioning set temperature, prioritize the absorption of photovoltaic power and suppress grid power fluctuations.

[0280] 3. Experimental Period and Variable Control

[0281] The experiment lasted for 30 days, from July 1 to July 30, covering the typical high temperature and high radiation period in tropical regions. The outdoor meteorological conditions and building foundation power load of the two groups of experiments were kept exactly the same, and only the control strategy was changed to ensure the comparability of the experimental results.

[0282] (III) Experimental Procedure

[0283] 1. Data Collection Phase (Days 1-5)

[0284] Baseline data were collected for the experimental and control groups, including total horizontal solar radiation, outdoor temperature, indoor temperature, photovoltaic power generation, air conditioning power consumption, other electrical loads, grid interaction power, and electricity price. The consistency of the initial states of the two groups was verified, and the collected data showed that the deviation of the baseline data of the two groups was less than 2%.

[0285] 2. Strategy Implementation Phase (Days 6-25)

[0286] The control group started the traditional fixed control strategy, while the experimental group started the intelligent regulation strategy of this invention. The group recorded various operating data in 15-minute time steps every day, including hourly photovoltaic power generation, air conditioning power consumption, grid electricity purchase and sale, indoor temperature, electricity purchase cost, carbon emissions, etc.

[0287] 3. Data verification phase (days 26-30)

[0288] Outlier removal and completion are performed on the collected data. Invalid data is removed using the 3σ principle, and missing data is completed using linear interpolation to ensure data integrity and accuracy.

[0289] (iv) Experimental Data

[0290] 1. Hourly Key Indicator Data

[0291]

[0292] 2. Comparison of 30-day cumulative indicators

[0293]

[0294] 3. Economic and environmental benefit data for the entire lifecycle and different time periods

[0295]

[0296] (V) Experimental Conclusions

[0297] Traditional fixed strategies do not consider the time matching between photovoltaic (PV) output and cooling load. 30-day experimental data show that its local PV consumption rate is only 8.2%, the maximum grid power fluctuation reaches 24.9kW, and the proportion of electricity purchased during high electricity price periods reaches 35%, resulting in high electricity purchase costs and carbon emissions. This invention achieves dynamic coordination between PV production capacity and cooling energy consumption through coupled mechanism modeling and rolling time-domain control, increasing the local PV consumption rate to 10.1% and reducing the proportion of electricity purchased during high electricity price periods to 28%. It solves the technical pain points of traditional strategies such as "unclear system-level relationships and difficulty in balancing multiple conflicting objectives" and has outstanding creativity.

[0298] In a 30-day experiment, this invention achieved a 19.8% increase in photovoltaic power generation, a 14.7% reduction in air conditioning power consumption, an 11.07% reduction in electricity purchase costs, a 3.02% reduction in carbon emissions, a 34.1% reduction in indoor comfort deviation, and a 79.5% reduction in the maximum fluctuation of grid power. This verifies the engineering feasibility and practical application value of the technical solution. Compared with traditional strategies, it has significant performance advantages and can stably achieve the synergistic optimization goal of "low cost, low carbon, and high comfort" in tropical zero-carbon buildings.

[0299] In summary, the coupling mechanism and intelligent control method of the operating capacity and energy consumption of the tropical zero-carbon building, under the premise of only considering the cooling condition, integrates photovoltaic power generation, air conditioning cooling power and other electrical loads into the power balance equation. It clearly describes the quantitative relationship between the power generation and the power purchased by the grid through the power boundary and power fluctuation penalty term, forming a coupling mechanism model that can be directly used for engineering design and simulation, thus overcoming the problem of "unclear system-level relationship" in traditional methods.

[0300] Furthermore, the coupling mechanism of the operating capacity and energy consumption of this tropical zero-carbon building and the intelligent control method have constructed a multi-objective optimization function that includes operating costs, equivalent carbon emissions and comfort indicators. The weights can be flexibly configured according to different scenarios to achieve a quantitative trade-off between "low cost, low carbon and high comfort". Compared with the control method that only takes energy consumption or indoor temperature as a single objective, it is more in line with the comprehensive operating needs of tropical zero-carbon buildings.

[0301] Furthermore, the coupling mechanism of energy production capacity and energy consumption in this tropical zero-carbon building operation and the intelligent control method, through short-term prediction models of photovoltaic output, cooling load and other electrical loads and online learning error correction, enable the invention to maintain consistency between the model and actual operation under typical tropical conditions such as high radiation, rapid changes in cloud cover, and outdoor temperature fluctuations. Combined with the intelligent solution mechanism of rolling time-domain optimization and reinforcement learning-evolutionary algorithm, the control strategy can be continuously optimized to improve the robustness and economy of the system in uncertain environments.

[0302] Furthermore, the coupling mechanism between the operating capacity and energy consumption of this tropical zero-carbon building, as well as its intelligent control method, explicitly considers a grid power fluctuation penalty term in the objective function. Furthermore, by prioritizing the use of photovoltaic power generation in control decisions, this invention can effectively reduce the peak and fluctuation of grid-side power, promote the local consumption of photovoltaic power, reduce the capacity of the distribution side and the pressure on the transmission side, and improve the overall benefits of zero-carbon buildings from a system level.

[0303] Furthermore, the coupling mechanism of energy production capacity and energy consumption in the operation of the tropical zero-carbon building and the intelligent control method propose a complete system module division and functional definition: data acquisition module, modeling and prediction module, optimization decision module, control execution module and strategy update module. It can be integrated into the existing building automation system (BAS) or energy management system (EMS) through software upgrade or modular expansion. It has clear interfaces and implementation paths and strong engineering feasibility.

[0304] Furthermore, the coupling mechanism and intelligent control method for the operating capacity and energy consumption of this tropical zero-carbon building do not model the heating conditions, but instead focus the modeling and control on the cooling-dominated energy consumption scenario in tropical regions. This makes the model and algorithm more in line with actual operation and avoids unnecessary complexity. Under the premise of ensuring comfort constraints, the building achieves efficient operation under typical cooling-dominated conditions by optimizing the cooling load and coordinating it with other electrical loads. This solves the problems of unclear coupling mechanism, difficulty in balancing multiple conflicting objectives, insufficient adaptability to the uncertainty of tropical conditions, and insufficient engineering feasibility in existing technologies.

[0305] The relevant modules involved in this system are all hardware system modules or functional modules that combine computer software programs or protocols with hardware in the prior art. The computer software programs or protocols involved in these functional modules are technologies known to those skilled in the art and are not improvements to this system. The improvement of this system lies in the interaction or connection between the modules, that is, in improving the overall structure of the system to solve the corresponding technical problems that this system aims to address.

[0306] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A coupling mechanism and intelligent control method for energy production and consumption in tropical zero-carbon buildings, characterized in that, Includes the following steps: S1. Acquire outdoor weather and building-related data at different times in the tropical region and perform preprocessing. The building-related data includes discrete time data. Horizontal total solar radiation Outdoor air temperature Indoor air temperature Relative humidity, photovoltaic array backsheet temperature, air conditioning power Other electrical load power Electricity price and power grid carbon emission factors The time step is ; S2, the total horizontal solar radiation The equivalent irradiance of the photovoltaic array surface is converted using a conversion formula. Construct a photovoltaic power generation quantification model to calculate photovoltaic power generation output. ; Calculate building cooling load based on building-related data. Based on this, by introducing the cooling energy efficiency ratio, a quantitative model of air conditioning cooling load and electrical power is constructed to calculate the air conditioning cooling power. ; Based on building-related data, a time-segmentation and data-driven approach is used to construct a data structure that includes the power of other electrical loads. The predictive model predicts the power of other electrical loads at the next moment; S3. Based on the results obtained in steps S1-S2, establish the power balance relationship, form a coupling mechanism model of photovoltaic-air conditioning-other loads during the operation phase, and calculate the power purchased by the power grid. ; Establish a power balance relationship, form a coupling mechanism model of photovoltaic-air conditioning-other loads during the operation phase, and calculate the power purchased by the power grid, as follows: The calculation of the electrical energy balance relationship is shown in formula (5): （5） In formula (5), Power purchased from the power grid; Apply power boundary constraints, as shown in equation (6): （6） In formula (6), Minimum power purchase capacity, Maximum power purchase capacity; In this process, formula (5) is transformed into a vector form in the prediction time domain, as shown in formula (7): （7） In formula (7), Forecast sequence of power purchased from the power grid, This is a predicted sequence of air conditioning cooling power. For other electrical load power prediction sequences, This is a photovoltaic power generation prediction sequence; In the formula, 、 、 、 ； Add grid power fluctuation penalty item Its calculation is shown in formula (8): （8） In formula (8), The fluctuation penalty coefficient, For the first Time and the The difference in power purchase capacity of the power grid at any given time This represents the total number of time points. S4. Based on building-related data and power grid purchase power, the air conditioning set temperature sequence at different times is used as the decision variable. The model satisfies multiple constraints of indoor temperature, air conditioning power and air conditioning set temperature. The final multi-objective optimization model is constructed, and an intelligent optimization strategy combining reinforcement learning and evolutionary algorithm is adopted to obtain the optimal control solution that satisfies the constraints. S5. Based on steps S1-S4, a rolling time-domain predictive control strategy is adopted, using irradiance, outdoor temperature, and other electrical loads to update the predicted photovoltaic power generation value for the next time step. Predicted cooling load power Other electrical load power forecasts Through a final multi-objective optimization model, the optimal air conditioning set temperature and corresponding air conditioning control quantity at the current moment are obtained, and the photovoltaic control quantity corresponding to the photovoltaic side control strategy is determined simultaneously. The air conditioning control quantity is converted into a set temperature command for the air conditioning controller, and the photovoltaic control quantity is converted into an output control command for the photovoltaic inverter, and then sent to the corresponding equipment for execution, prioritizing the use of photovoltaic power generation and suppressing... Fluctuations are controlled only at the current moment; the process of repeated updates and optimizations enables intelligent and coordinated regulation of energy production and consumption in tropical zero-carbon buildings under cooling-only conditions.

2. The coupling mechanism and intelligent control method for the operational capacity and energy consumption of a tropical zero-carbon building according to claim 1, characterized in that, In step S2, a photovoltaic power generation quantification model is constructed to calculate the photovoltaic power generation, as shown in formula (1): （1） In formula (1), For photovoltaics Photovoltaic power generation at any given time For reference photovoltaic efficiency, For effective light-receiving area, The temperature power coefficient, This refers to the actual temperature of the photovoltaic module. This is a reference temperature.

3. The coupling mechanism and intelligent control method for the operational capacity and energy consumption of a tropical zero-carbon building according to claim 1, characterized in that, In step S2, a quantitative model of air conditioning cooling load and power is constructed to calculate the air conditioning cooling power. The details are as follows: Calculate the building cooling load under cooling conditions. Its calculation is shown in formula (2): （2） In formula (2), The equivalent heat transfer coefficient of the building envelope. For equivalent enclosure area, To allow solar heat to enter the interior through the building envelope. For heat gain inside personnel and equipment; Calculate the cooling power of the air conditioner As shown in formula (3): （3） In formula (3), This refers to the cooling energy efficiency ratio.

4. The coupling mechanism and intelligent control method for the operational capacity and energy consumption of a tropical zero-carbon building according to claim 1, characterized in that, In step 2, a system is constructed that includes other electrical load power. The prediction model predicts the power of other electrical loads at the next moment, and its calculation is shown in formula (4): （4） In formula (4), The time period to which the time belongs. For weekday / weekend type, To fit the obtained nonlinear function, For the current electrical load power, For updatable parameters, This is the predicted power value of other electrical loads at the next moment.

5. The coupling mechanism and intelligent control method for the operational capacity and energy consumption of a tropical zero-carbon building according to claim 1, characterized in that, In step S4, the final multi-objective optimization model is constructed as follows: ① Set the air conditioner to a specific temperature sequence. Using photovoltaic side control strategy as decision variables, and with electricity purchase cost, carbon emissions and comfort deviation as objectives, an initial multi-objective optimization model is constructed, the calculation of which is shown in formula (9): （9） In formula (9), As weight, For the target comfortable temperature, To minimize the overall cost; The constraints on indoor temperature and air conditioning power are applied, as shown in formula (10): （10） In formula (10), This represents the lower limit of indoor temperature. This represents the upper limit of indoor temperature. This refers to the rated maximum cooling capacity of the air conditioner. Apply the air conditioning set temperature boundary, the boundary of which is shown in formula (11): （11） In formula (11), Set the lower limit of the temperature for the air conditioner. Set the upper limit of the temperature for the air conditioner; ② In formula (9), the target comfort temperature has a comfort deviation. Based on the square of the temperature deviation, a simplified comfort index including humidity is adopted. Its calculation is shown in formula (12): （12） In formula (12), This refers to indoor relative humidity. For the target relative humidity, These are the weighting coefficients; Calculate comfort target items The comfort index at each moment is summed up, and its calculation is shown in formula (13): （13） ③ Operating costs With carbon emissions The calculations are shown in formulas (14) and (15) respectively: （14） （15） In formulas (14) and (15), For time step; ④ The final multi-objective optimization model is constructed using a normalized weighted form, and the calculation is shown in formula (16): （16） In formula (16), For reference only. For normalized weights.

6. The coupling mechanism and intelligent control method for the operational capacity and energy consumption of a tropical zero-carbon building according to claim 5, characterized in that, In step S4, an intelligent optimization strategy combining reinforcement learning and evolutionary algorithms is used to obtain the optimal control solution that satisfies the constraints, as detailed below: ① Construct an instantaneous cost function using the air conditioner set temperature change and the photovoltaic side control strategy as the action space. The calculation is shown in formula (17): （17） By using reinforcement learning algorithms to learn near-optimal policies, blind search can be avoided; ②The air conditioning set temperature sequence and photovoltaic side control strategy in the entire prediction time domain are used as individual codes. The final multi-objective optimization model of formula (16) is used as the fitness. The evolutionary algorithm is used to perform global optimization. Under the premise of satisfying the constraints, the optimal control solution that satisfies the constraints is obtained.

7. The coupling mechanism and intelligent control method for the operation capacity and energy consumption of a tropical zero-carbon building according to claim 4, characterized in that, In step S5, a rolling time-domain predictive control strategy is used to predict the photovoltaic power generation, other electrical load power and cooling load power, and the calculation is shown in formulas (18)-(20): （18） （19） （20） In the formula, This is the predicted value of photovoltaic power generation at the next moment. This is the predicted total cooling load of the building at the next moment. Forecast values ​​of other building loads at the next moment. These are internally adjustable parameters. For prediction functions; By minimizing the prediction error, online self-learning correction of key parameters in the coupling mechanism model is achieved, and its calculation is shown in formula (21): （21）。

Images