AI-enabled building energy consumption real-time monitoring and energy-saving recommendation method and system

Abstract

The invention relates to an AI-enabled building energy consumption real-time monitoring and energy-saving recommendation method and system. The method comprises the following steps: acquiring real-time operation data of a building, and standardizing the real-time operation data to obtain a standardized data set; wherein the building real-time operation data comprises equipment state data and environment data; real-time energy consumption of the building is obtained through equipment state data in the real-time operation data of the building; when the real-time energy consumption of the building exceeds a preset threshold value, modeling based on the coupling relationship among the standardized data set, the building physical parameters and the equipment operation parameters to obtain a parameterized coupling matrix; inputting the standardized data set and the parameterized coupling matrix into an MCA-MPC model to obtain an energy-saving optimization instruction; the energy-saving optimization instruction is used for indicating the equipment to adjust the running state. By adopting the method, the real-time influence of environment change and the diversity of user behavior modes can be fully considered, and the energy consumption saving scheme can be recommended in real time.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of energy consumption management, and particularly relates to an AI-enabled building energy consumption real-time monitoring and energy-saving recommendation method and system. BACKGROUND

[0002] The building industry is one of the main fields of global energy consumption, and its energy consumption management problem is increasingly prominent, so it is urgent to realize building energy saving and consumption reduction. In the traditional technology, energy consumption management is realized through static benchmark model and independent control of each device, which relies on preset scenes and sensor feedback, cannot fully consider the real-time influence of environmental changes and the diversity of user behavior patterns, and ignores the operation relationship between each energy consumption system, resulting in energy waste. SUMMARY

[0003] Therefore, it is necessary to provide an AI-enabled building energy consumption real-time monitoring and energy-saving recommendation method and system which can fully consider the real-time influence of environmental changes and the diversity of user behavior patterns and recommend energy-saving solutions in real time.

[0004] In a first aspect, the application provides an AI-enabled building energy consumption real-time monitoring and energy-saving recommendation method, comprising:

[0005] Obtaining building real-time operation data and standardizing the real-time operation data to obtain a standardized data set; wherein the building real-time operation data includes device state data and environmental data;

[0006] Obtaining the real-time energy consumption of the building through the device state data in the building real-time operation data;

[0007] When the real-time energy consumption of the building exceeds a preset threshold, modeling based on the coupling relationship between the standardized data set, the building physical parameters and the device operation parameters to obtain a parameterized coupling matrix; wherein the parameterized coupling matrix is used to represent the relationship between the device state data and the environmental data;

[0008] Inputting the standardized data set and the parameterized coupling matrix into an MCA-MPC model to obtain an energy-saving optimization instruction; the energy-saving optimization instruction is used to instruct the device to adjust the operating state.

[0009] Further, the MCA-MPC model comprises:

[0010] A prediction subnetwork is used to predict the personnel activity state based on the historical personnel flow heat map; the personnel activity state includes personnel distribution and activity intensity;

[0011] A target subnetwork is used to determine an optimization proposition based on the standardized data set, the parameterized coupling matrix and the personnel activity state; the optimization proposition includes a target function and a constraint set;

[0012] An optimized sub-network is used to generate energy-saving optimization instructions based on personnel activity status and optimization propositions.

[0013] Furthermore, based on historical population movement heatmaps, the activity status of people is predicted, including:

[0014] Based on historical population flow heat maps, the periodic flow of people is predicted to obtain the regional baseline population density.

[0015] Based on real-time public travel data, the population distribution of the regional baseline population density is adjusted to obtain a population migration probability matrix; real-time public travel data is used to characterize the planned travel of people.

[0016] The adjusted population density is obtained by adjusting the regional baseline population density based on the population migration probability matrix.

[0017] Based on historical population flow heat maps, the impact of special historical events on the adjusted population density is quantified to obtain a correction term for population density.

[0018] The regional baseline population density, the adjusted population density, and the correction terms are fused together based on pre-set weights to obtain the population activity status.

[0019] Furthermore, based on the standardized dataset, parameterized coupling matrix, and personnel activity status, optimization propositions are determined, including:

[0020] An energy consumption prediction model is constructed based on the parameterized coupling matrix and device status data in the standardized dataset.

[0021] Based on the activity status of personnel and the preset comfort temperature, a temperature constraint set is constructed using the following formula:

[0022] T min (j,i)=T comfort -ΔT deadband ·(1-o j (i))

[0023] T max (j,i)=T comfort +ΔT deadband ·(1-o j (i))

[0024] Among them, T min (j,i) represents the minimum temperature, T max (j,i) represents the maximum temperature, T comfort For the preset comfortable temperature, o j (i) represents the population density in the region, ΔT deadband For temperature control dead zone bandwidth;

[0025] Based on the energy consumption prediction model and the temperature constraint set, a target function is constructed;

[0026] Based on the temperature constraint set and the device physical constraint, a constraint set is constructed, and the constraint set and the target function are integrated to obtain an optimization proposition.

[0027] Further, the expression formula of the target function is:

[0028]

[0029] Wherein, J(u) is the target function, u is the control variable vector, i is the time step number in the prediction time domain, N p is the prediction time domain length, is the predicted energy consumption of the i-th step, w i is the time weight factor, λ is the adjustment cost weight, k is the time step number in the control time domain, N c is the control time domain length, Δu k is the device state change vector, ρ is the penalty weight, j is the building area number, T j (i) is the predicted temperature of area j at the i-th step, T max (j,i) is the time-varying upper limit of temperature of area j at the i-th step, T min (j,i) is the time-varying lower limit of temperature of area j at the i-th step.

[0030] Further, based on the personnel activity state and the optimization proposition, energy-saving optimization instructions are generated, including:

[0031] Based on the standardized data set, an initialization vector is calculated, which is used to represent the initial state of the device operation;

[0032] Based on the personnel activity state and the initialization vector, the solution of the optimization proposition is calculated, and the solution is sorted according to the time sequence to obtain control instruction data;

[0033] Based on the constraint set in the optimization proposition, the feasibility of the control instruction data is verified to obtain a feasibility report;

[0034] Based on the feasibility report, the control instruction sequence is optimized to obtain energy-saving optimization instructions.

[0035] Further, based on the coupling relationship modeling of the standardized data set, the building physical parameters and the device operation parameters, a parameterized coupling matrix is obtained, including:

[0036] Based on the rated parameters of the device and the indoor and outdoor humidity, the air conditioning-fresh air coupling coefficient matrix is calculated by the following formula:

[0037]

[0038] Wherein, mac-fan is the air conditioning-new air coupling coefficient matrix, representing the change of the total power of the air conditioning system when the air volume changes, V air is the new air volume flow rate, k is the power coefficient of the new air fan, n is the power index of the fan, p is the air density, h fg is the latent heat of water vaporization, c p is the specific heat capacity of air, W out is the external humidity, W in is the internal humidity, T out is the external temperature, T in is the internal temperature, and COP is the refrigeration coefficient.

[0039] Based on the building physical parameters and external temperature data, the air conditioning-weather coupling coefficient matrix is calculated by the following formula:

[0040]

[0041] wherein, m ac-temp is the air conditioning-weather coupling coefficient matrix, representing the change of the total power of the air conditioning system when the external temperature changes, U is the overall heat transfer coefficient of the building envelope, A is the total area of the building envelope, p sys is the total efficiency of the system, p is the air density, V air is the new air volume flow rate, c p is the specific heat capacity of air, and COP is the refrigeration coefficient.

[0042] Based on the building physical parameters and solar radiation data, the air conditioning-radiation coupling coefficient matrix is calculated by the following formula:

[0043]

[0044] wherein, m ac-solar is the air conditioning-radiation coupling coefficient matrix, representing the change of the total power of the air conditioning system when the solar radiation changes, A is the total area of the building envelope, r win is the window-wall ratio, SHGC is the solar heat gain coefficient, p sys is the total efficiency of the system, f vent is the radiation heat absorption coefficient of the new air, and COP is the refrigeration coefficient.

[0045] Based on the per capita heat dissipation, the air conditioning-human flow coupling coefficient matrix is calculated by the following formula:

[0046]

[0047] wherein, m ac-people is the air conditioning-human flow coupling coefficient matrix, q person is the per capita heat dissipation, p sys is the total efficiency of the system.

[0048] The fusion air conditioner-fresh air coupling coefficient matrix, the air conditioner-weather coupling coefficient matrix, the air conditioner-radiation coupling coefficient matrix and the air conditioner-human flow coupling coefficient matrix to obtain a parameterized coupling matrix.

[0049] In a second aspect, the present application further provides an AI-enabled building energy consumption real-time monitoring and energy saving recommendation system, comprising:

[0050] The source module is configured to obtain building real-time operation data, and perform standardization processing on the real-time operation data to obtain a standardized data set; wherein the building real-time operation data comprises equipment state data and environment data;

[0051] The energy consumption module is configured to obtain the real-time energy consumption of the building by the equipment state data in the building real-time operation data.

[0052] The coupling module is configured to, when the real-time energy consumption of the building exceeds a preset threshold, model a coupling relationship based on the standardized data set, the building physical parameters and the equipment operation parameters to obtain a parameterized coupling matrix; wherein the parameterized coupling matrix is used to represent the relationship between the equipment state data and the environment data.

[0053] The instruction module is configured to input the standardized data set and the parameterized coupling matrix into an MCA-MPC model to obtain an energy saving optimization instruction; the energy saving optimization instruction is used to instruct the equipment to adjust the operation state.

[0054] In a third aspect, the present application further provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method provided in the first aspect of the present application when executing the computer program.

[0055] In a fourth aspect, the present application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps of the method provided in the first aspect of the present application.

[0056] The AI-enabled building energy consumption real-time monitoring and energy-saving recommendation method and system described above, by obtaining building real-time operation data, and standardizing the real-time operation data to obtain a standardized data set, wherein the building real-time operation data includes equipment state data and environmental data; obtaining the real-time energy consumption of the building through the equipment state data in the building real-time operation data; when the real-time energy consumption of the building exceeds a preset threshold, modeling based on the coupling relationship of the standardized data set, building physical parameters and equipment operation parameters to obtain a parameterized coupling matrix; wherein the parameterized coupling matrix is used to represent the relationship between the equipment state data and the environmental data; inputting the standardized data set and the parameterized coupling matrix into the MCA-MPC model to obtain energy-saving optimization instructions; the energy-saving optimization instructions are used to instruct the technical means of adjusting the running state of the equipment, which can realize the technical effect of real-time recommendation of energy-saving schemes considering the real-time influence of environmental changes and the diversity of user behavior patterns. BRIEF DESCRIPTION OF DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments or the related art, the drawings needed to be used in the embodiments or the related art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0058] Figure 1 A process flow diagram of an AI-enabled building energy consumption real-time monitoring and energy-saving recommendation method provided by the present application;

[0059] Figure 2 A structure diagram of an MCA-MPC model provided by the present application;

[0060] Figure 3 A structure diagram of an AI-enabled building energy consumption real-time monitoring and energy-saving recommendation system provided by the present application. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solutions and advantages of the present application more clear, the present application will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application.

[0062] In one embodiment, as shown in Figure 1 An AI-enabled building energy consumption real-time monitoring and energy-saving recommendation method is provided, and the present embodiment takes the method applied to a terminal as an example. It should be understood that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is realized through the interaction of the terminal and the server. In the present embodiment, the method includes the following steps:

[0063] Step 101, obtaining building real-time operation data, and standardizing the real-time operation data to obtain a standardized data set; wherein the building real-time operation data includes equipment state data and environment data.

[0064] Specifically, the equipment state data can include air conditioning power, fresh air damper opening, lighting brightness and other equipment real-time parameters; the environment data can include indoor and outdoor temperature and humidity, CO2 concentration, solar radiation intensity, and people flow density sensor readings. The terminal obtains equipment and environment data streams in real time through sensors in the building automation system, and uniformly maps the data in the data stream to the [0, 1] interval to obtain a standardized data set.

[0065] Step 102, obtaining the real-time energy consumption of the building through the equipment state data in the building real-time operation data.

[0066] The real-time energy consumption is the current total power of the building, which is obtained by directly summing the equipment state data by the terminal. For example, the real-time energy consumption of the building can include the energy consumption of the air conditioning system and the lighting equipment, and the air conditioning system can include the fresh air system and the refrigeration system.

[0067] Step 103, when the real-time energy consumption of the building exceeds a preset threshold, modeling based on the coupling relationship between the standardized data set, the building physical parameters and the equipment operation parameters to obtain a parameterized coupling matrix; wherein the parameterized coupling matrix is used to represent the relationship between the equipment state data and the environment data.

[0068] Specifically, the preset threshold is set in advance based on historical data and building energy efficiency benchmarks; the parameterized coupling matrix includes at least four dimensions, which can optionally include the relationship between the air conditioner and the fresh air, the relationship between the air conditioner and the air temperature, the relationship between the air conditioner and the solar radiation, and the relationship between the air conditioner and the environment number; the building physical parameters can include fixed properties such as building envelope heat transfer coefficient and window-to-wall ratio; the equipment operation parameters can include equipment inherent parameters such as air conditioning energy efficiency ratio and fan power index. The terminal compares the size relationship between the real-time energy consumption and the preset threshold, and if the real-time energy consumption exceeds the threshold, calculates each coupling coefficient based on real-time equipment state, environment state and other data, and splices each coupling coefficient into a unified coupling matrix.

[0069] Step 104, inputting the standardized data set and the parameterized coupling matrix into the MCA-MPC model to obtain energy-saving optimization instructions; the energy-saving optimization instructions are used to instruct the equipment to adjust the operating state.

[0070] Specifically, the MCA-MPC (Multivariable Coupling Adaptive Model Predictive Control) model has at least a three-layer structure, which can include a prediction subnetwork, an objective subnetwork, and an optimization subnetwork. It is used to predict future states, determine the objective function corresponding to the future states, solve the objective function, and finally obtain energy-saving optimization instructions.

[0071] This embodiment provides an AI-enabled method for real-time monitoring and energy-saving recommendations of building energy consumption. It acquires real-time building operation data and standardizes this data to obtain a standardized dataset. The real-time building operation data includes equipment status data and environmental data. The real-time energy consumption of the building is obtained from the equipment status data. When the real-time energy consumption exceeds a preset threshold, a parametric coupling matrix is ​​obtained based on the coupling relationship between the standardized dataset, building physical parameters, and equipment operation parameters. This parametric coupling matrix characterizes the relationship between equipment status data and environmental data. The standardized dataset and the parametric coupling matrix are input into an MCA-MPC model to obtain energy-saving optimization instructions. These instructions instruct the equipment to adjust its operating status. Through these techniques, dynamic energy consumption optimization is achieved while ensuring user comfort. The method fully considers the impact of real-time environmental changes and the diversity of user behavior patterns on building energy consumption and recommends energy-saving solutions in real time.

[0072] In one embodiment, such as Figure 2 As shown, the MCA-MPC model 200 includes:

[0073] Predictive subnetwork 201 is used to predict the activity status of people based on historical population flow heat maps; the activity status of people includes population distribution and activity intensity.

[0074] The historical population flow heatmap is a spatiotemporal matrix data that records the historical population density in each area of ​​the building. The population activity status includes two dimensions: population distribution (the predicted population density vector for each area) and activity intensity (the metabolic equivalent of the population, determining the average heat generated by each person). The prediction subnetwork 201 can extract fixed population flow patterns using an LSTM (Long Short-Term Memory) model, analyze the impact of historical events using a Gaussian regression algorithm, and thus obtain the population activity status.

[0075] The objective subnetwork 202 is used to determine the optimization proposition based on the standardized dataset, parameterized coupling matrix, and personnel activity status; the optimization proposition includes the objective function and constraint set.

[0076] Specifically, the optimization proposition contains at least two core elements, wherein the objective function is a comprehensive cost function that needs to be minimized, and the constraint set is a set of device physical limitations and comfort requirements. The target subnetwork 202 constructs temperature constraints based on human comfort temperature and device physical limitations based on physical device parameters, etc., to minimize energy consumption as the optimization goal, determines the objective function, and obtains the optimization proposition.

[0077] The optimization subnetwork 203 is configured to generate energy-saving optimization instructions based on the personnel activity state and the optimization proposition.

[0078] Specifically, the energy-saving optimization instructions are a set of operation instruction sequences for instructing the device to adjust the operating state, thereby achieving the minimization of building energy consumption while maintaining the comfort of personnel. The optimization subnetwork 203 iteratively solves the objective function in the optimization proposition based on the relationship between the device and the environment and the personnel, wherein each solution is an operation instruction, and the feasibility of the solution is verified based on the constraint set, the infeasible solution is eliminated, and the remaining solutions are integrated to obtain the energy-saving optimization instructions.

[0079] The core of the traditional MPC model is to repeatedly solve a finite time domain optimization problem online. The model can include a prediction model and a rolling optimization process. The prediction model usually uses a simple unit impulse response curve for prediction, and the rolling optimization uses a QP solver (Quadratic Programming, quadratic programming) for optimization. The model has problems such as excessive simplification of the model, only focusing on the air conditioning system itself, fixed constraints, single optimization target, and ignoring the coupling relationship between devices. Therefore, under the architecture of the traditional MPC model, the LSTM model and the Gaussian regression algorithm are introduced to predict personnel behavior by fusing periodic rules and sudden disturbances, and the construction of a multi-objective optimization proposition is improved, that is, the minimization of the final energy consumption is focused on, while the comfort of personnel and the operation life of the device are also focused on, and the requirements of the three are dynamically balanced. The SQP (Sequential Quadratic Programming, sequential quadratic programming) algorithm is used for rolling optimization to realize nonlinear optimization. The MCA-MPC model upgrades the MPC model comprehensively through environmental coupling quantization, dynamic prediction and correction to multi-objective decision making under the core of repeatedly solving a finite time domain optimization problem.

[0080] The embodiment realizes the balance between building energy consumption and comfort by constructing the MCA-MPC model to predict and optimize building energy consumption, prevents user discomfort caused by energy saving, and improves the rationality and feasibility of energy-saving recommendations.

[0081] In one embodiment, the personnel activity state is predicted based on a historical personnel flow heat map, including:

[0082] Step 301, predicting periodic flow of personnel based on historical personnel flow heat map, obtaining regional baseline personnel density.

[0083] Specifically, the historical personnel flow heat map is a historical human flow density space-time matrix of each region of the building; the periodic flow is a repetitive personnel activity mode; the regional baseline personnel density is a normal flow distribution vector excluding special events. Exemplarily, the periodic flow can include the morning peak on weekdays and the lunch restaurant peak. The terminal predicts the future normalized personnel distribution situation through the LSTM network based on the personnel distribution in the past period.

[0084] Step 302, adjusting the personnel distribution situation of the regional baseline personnel density based on real-time public travel data, obtaining a personnel migration probability matrix; the real-time public travel data is used to represent the scheduled travel of personnel.

[0085] Specifically, the real-time public travel data is the scheduled information of personnel; each element in the personnel migration probability matrix represents the migration probability from one region to another region. Exemplarily, the real-time public travel data can include: scheduled meetings, restaurant reservation rates, scheduled activities, etc. Based on the public travel, the migration situation of personnel between regions in the building before the travel is calculated, and the personnel migration probability matrix is obtained.

[0086] Step 303, adjusting the regional baseline personnel density based on the personnel migration probability matrix, obtaining the adjusted personnel density.

[0087] Wherein, the adjusted personnel density is a density vector corrected by travel, and the redistribution of density can be realized by matrix multiplication.

[0088] Step 304, quantifying the influence of historical special events on the adjusted personnel density based on the historical personnel flow heat map, obtaining a correction term of the personnel density.

[0089] Specifically, the historical special event is a marked abnormal activity, which can include: fire drill, star signing, etc. The correction term of the density is the flow change vector caused by the event. The terminal quantifies the influence of the abnormal event on the personnel density in the building through the Gaussian process regression algorithm, and obtains the correction term.

[0090] Step 305, fusing the regional baseline personnel density, the adjusted personnel density and the correction term based on the pre-set weight, obtaining the personnel activity state.

[0091] Specifically, the personnel activity state is the final prediction result, which can include personnel distribution density and personnel activity intensity. Optionally, the pre-weight setting is that the weight of the regional baseline personnel density is 0.4, the weight of the adjusted personnel density is 0.5, and the weight of the correction term is 0.1.

[0092] The embodiment provides accurate personnel activity state input for building energy saving by multi-model fusion prediction, supports subsequent energy saving optimization, and improves the accuracy of energy saving recommendation.

[0093] In one of the embodiments, based on the standardized data set, the parameterized coupling matrix and the personnel activity state, an optimization proposition is determined, including:

[0094] In step 401, based on the parameterized coupling matrix and the equipment state data in the standardized data set, an energy consumption prediction model is constructed.

[0095] The energy consumption prediction model is a linear regression model, which is used to predict energy consumption. The terminal constructs a change matrix of equipment state change and multiplies it with the parameterized coupling matrix to obtain energy consumption prediction.

[0096] In step 402, based on the personnel activity state and the preset comfortable temperature, a temperature constraint set is constructed by the following formula:

[0097] T min (j,i)=T comfort -ΔT deadband ·(1-o j (i))

[0098] T max (j,i)=T comfort +ΔT deadband ·(1-o j (i))

[0099] Wherein, T min (j,i) is the minimum temperature, T max (j,i) is the maximum temperature, T comfort is the preset comfortable temperature, o j (i) is the personnel density in the area, and ΔT deadband is the temperature control dead zone bandwidth.

[0100] Specifically, the comfortable temperature is the most suitable temperature for the human body; the temperature control dead zone bandwidth is the range of temperature fluctuation; the temperature constraint set is the time-varying function aggregation of the upper and lower limits of the temperature in each area, and the more dense the personnel in the area, the more strict the temperature control. For example, when the personnel density in the area is less than 0.2, the upper limit of the temperature is automatically covered to 28℃, and the lower limit of the temperature is 20℃ to avoid invalid energy consumption.

[0101] In step 403, based on the energy consumption prediction model and the temperature constraint set, a target function is constructed.

[0102] The target function is a comprehensive cost function to be minimized, which has three objectives, namely, minimum energy consumption, minimum equipment operation and maximum personnel comfort.

[0103] At step 404, a constraint set is constructed based on the temperature constraint set and the device physical constraint, and the constraint set and the objective function are integrated to obtain an optimization proposition.

[0104] The device physical constraint is a hard limit for device operation; the constraint set is a set of all constraint conditions; and the optimization proposition is a complete optimization problem including the objective function and the constraint set. The constraint set can include temperature constraints, minimum start-stop time of the device, power limit of the device, carbon dioxide concentration constraints, and the like. The terminal is configured to

[0105] The embodiment constructs an optimization proposition related to energy saving, device protection and personnel comfort by physical rule constraints and multi-objective optimization, and provides a basis for subsequent energy saving recommendation.

[0106] In one embodiment, the expression formula of the objective function is as follows:

[0107]

[0108] wherein J(u) is the objective function, u is a control variable vector, i is a time step number in a prediction time domain, N p is a prediction time domain length, is a predicted energy consumption at the i-th step, w i is a time weight factor, λ is an adjustment cost weight, k is a time step number in a control time domain, N c is a control time domain length, Δu k is a device state change vector, ρ is a penalty weight, j is a building area number, T j (i) is a predicted temperature of the area j at the i-th step, T max (j,i) is a time-varying upper limit of the temperature of the area j at the i-th step, T min (j,i) is a time-varying lower limit of the temperature of the area j at the i-th step.

[0109] Specifically, the objective function includes three objectives, which are minimum energy consumption, minimum device adjustment times, and most suitable temperature for human activities, respectively. The time weight factor is often bound with time period electricity price, so as to introduce energy consumption cost into the consideration range of the energy saving scheme. Different adjustment strategies are switched by changing the time weight factor, the adjustment cost weight and the penalty weight.

[0110] In one embodiment, based on the personnel activity state and the optimization proposition, an energy saving optimization instruction is generated, including:

[0111] At step 601, an initialization vector is calculated based on a standardized data set, and the initialization vector is used to represent an initial state of device operation.

[0112] Specifically, the initialization vector is the starting point of the optimization calculation of each device, containing the initial setting value of the device. The terminal extracts the latest device standardization parameters from the standardized data set, and arranges the device parameters in a fixed order into a vector.

[0113] Step 602, based on the personnel activity state and the initialization vector, calculate the solution of the optimization proposition, and sort the solution in time sequence to obtain the control instruction data.

[0114] Specifically, the control instruction data is a set of time-sequenced device operation instructions. The terminal adjusts the temperature constraint according to the personnel activity state, uses the initialization vector as the initial value, and iteratively calculates the solution of the optimization proposition by the SQP algorithm to obtain the control instruction data.

[0115] Step 603, based on the constraint set in the optimization proposition, verify the feasibility of the control instruction data to obtain a feasibility report.

[0116] Among them, the feasibility report is a diagnosis document quantifying the violation situation. The terminal compares whether the device running state indicated by the control instruction data conforms to the pre-defined constraint condition through the rule engine, and then obtains the violation item and the violation intensity.

[0117] Step 604, based on the feasibility report, optimize the control instruction sequence to obtain energy-saving optimization instructions.

[0118] Specifically, the energy-saving optimization instructions are executable and safe instructions for instructing the device to adjust its running state. For example, when the power exceeds the limit, reduce the power of non-critical devices; when the device running changes suddenly, insert an overshoot point; when the device running exceeds the upper limit, such as when the air conditioner temperature setting is lower than the minimum air conditioner temperature, run at the maximum device running power.

[0119] This embodiment forces the safety requirements in device running by optimization solution and feasibility verification, and improves the reliability of energy-saving recommendations.

[0120] In one of the embodiments, it is characterized in that, based on the coupling relationship modeling of the standardized data set, the building physical parameters and the device running parameters, a parameterized coupling matrix is obtained, including:

[0121] Step 701, based on the device rated parameters and indoor and outdoor humidity, calculate the air conditioner-fresh air coupling coefficient matrix by the following formula:

[0122]

[0123] Among them, m ac-fan is the air conditioner-fresh air coupling coefficient matrix, representing the change amount of the total power of the air conditioning system when the fresh air volume changes, V airis the fresh air volume flow rate, k is the fresh air fan power coefficient, n is the fan power index, p is the air density, h fg is the water vapor latent heat, c p is the air specific heat capacity, W out is the outside humidity, W in is the inside humidity, T out is the outside temperature, T in is the inside temperature, COP is the refrigeration coefficient.

[0124] Specifically, the air conditioner-fresh air coupling coefficient matrix represents the instantaneous power change of the air conditioning system (including the compressor and the fan) when the fresh air volume changes by 1 cubic meter per second. When the power of the fresh air system rises, the power of the air conditioning system rises synchronously.

[0125] Step 702, based on the building physical parameters and the external temperature data, the air conditioner-weather coupling coefficient matrix is calculated by the following formula:

[0126]

[0127] wherein, m ac-temp is the air conditioner-weather coupling coefficient matrix, representing the change of the total power of the air conditioning system when the external temperature changes, U is the overall heat transfer coefficient of the envelope structure, A is the total area of the envelope structure, η sys is the total system efficiency, p is the air density, V air is the fresh air volume flow rate, c p is the air specific heat capacity, COP is the refrigeration coefficient.

[0128] Specifically, the air conditioner-weather coupling coefficient matrix represents the total power increment of the air conditioning system required to maintain the indoor set temperature when the outdoor temperature rises by 1℃, which includes two kinds of loads, namely the heat conduction load of the envelope structure and the outdoor air heat load brought by the fresh air system.

[0129] Step 703, based on the building physical parameters and the solar radiation data, the air conditioner-radiation coupling coefficient matrix is calculated by the following formula:

[0130]

[0131] wherein, m ac-solar is the air conditioner-radiation coupling coefficient matrix, representing the change of the total power of the air conditioning system when the solar radiation changes, A is the total area of the envelope structure, r win is the window-wall ratio, SHGC is the solar heat gain coefficient, η sys is the total system efficiency, f vent is the fresh air radiation heat absorption coefficient, COP is the refrigeration coefficient.

[0132] Specifically, the air conditioner-radiation coupling coefficient matrix represents the power that the air conditioning system needs to increase when the solar radiation intensity increases by 1 W / m 2 Specifically, the air conditioner-radiation coupling coefficient matrix represents the power that the air conditioning system needs to increase when the solar radiation intensity increases by 1 W / m

[0133] In step 704, based on the heat dissipation per capita, the air conditioner-people flow coupling coefficient matrix is calculated by the following formula:

[0134]

[0135] wherein, m ac-people is the air conditioner-people flow coupling coefficient matrix, q person is the heat dissipation per capita, and η sys is the total efficiency of the system.

[0136] Specifically, the heat dissipation per capita is affected by the activity intensity, and the air conditioner-people flow coupling coefficient matrix represents the power that the air conditioning system needs to increase when one more person is added.

[0137] In step 705, the parameterized coupling matrix is obtained by fusing the air conditioner-fresh air coupling coefficient matrix, the air conditioner-weather coupling coefficient matrix, the air conditioner-radiation coupling coefficient matrix, and the air conditioner-people flow coupling coefficient matrix.

[0138] Specifically, the parameterized coupling matrix represents the response relationship of the total power of the air conditioner when the four parameters of fresh air, temperature, radiation, and people flow change.

[0139] The embodiment introduces the influence of fresh air, temperature, radiation, and people flow on the air conditioning system into the energy-saving recommendation through the coupling matrix modeling of the four dimensions, thereby improving the accuracy of the energy-saving recommendation.

[0140] It should be understood that although each step in the flowchart involved in each embodiment as described above is displayed in sequence according to the arrow, these steps are not necessarily executed in sequence according to the arrow. Unless explicitly stated herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other sequences. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or steps or stages in other steps.

[0141] Based on the same inventive concept, the embodiments of the present application also provide an AI-enabled building energy consumption real-time monitoring and energy saving recommendation system for implementing the AI-enabled building energy consumption real-time monitoring and energy saving recommendation method described above. The implementation scheme for solving the problem provided by the system is similar to the implementation scheme described in the above method, so the specific limitations in one or more AI-enabled building energy consumption real-time monitoring and energy saving recommendation system embodiments provided below can be referred to the limitations of the AI-enabled building energy consumption real-time monitoring and energy saving recommendation method described above, which will not be repeated here.

[0142] In one exemplary embodiment, as shown in Figure 3 An AI-enabled building energy consumption real-time monitoring and energy saving recommendation system 800 is provided, comprising:

[0143] A source module 801 is configured to obtain building real-time operation data and perform standardization processing on the real-time operation data to obtain a standardized data set; wherein the building real-time operation data comprises equipment state data and environment data;

[0144] An energy consumption module 802 is configured to obtain real-time energy consumption of the building by using the equipment state data in the building real-time operation data;

[0145] A coupling module 803 is configured to, when the real-time energy consumption of the building exceeds a preset threshold, model a coupling relationship based on the standardized data set, building physical parameters and equipment operation parameters to obtain a parameterized coupling matrix; wherein the parameterized coupling matrix is used to represent the relationship between the equipment state data and the environment data;

[0146] An instruction module 804 is configured to input the standardized data set and the parameterized coupling matrix into an MCA-MPC model to obtain energy saving optimization instructions; the energy saving optimization instructions are used to instruct the equipment to adjust the operation state.

[0147] Further, the MCA-MPC model comprises:

[0148] A prediction sub-network is configured to predict a personnel activity state based on a historical personnel flow heat map; the personnel activity state comprises personnel distribution and activity intensity;

[0149] A target sub-network is configured to determine an optimization proposition based on the standardized data set, the parameterized coupling matrix and the personnel activity state; the optimization proposition comprises a target function and a constraint set;

[0150] An optimization sub-network is configured to generate energy saving optimization instructions based on the personnel activity state and the optimization proposition.

[0151] Further, the prediction sub-network is further configured to:

[0152] Based on the historical personnel flow heat map, the periodic flow of personnel is predicted, and the regional baseline personnel density is obtained;

[0153] Based on real-time public travel data, the personnel distribution of the regional baseline personnel density is adjusted, and a personnel migration probability matrix is obtained. The real-time public travel data is used to represent the scheduled travel of personnel;

[0154] Based on the personnel migration probability matrix, the regional baseline personnel density is adjusted, and an adjusted personnel density is obtained;

[0155] Based on the historical personnel flow heat map, the influence of historical special events on the adjusted personnel density is quantified, and a correction term of the personnel density is obtained;

[0156] The regional baseline personnel density, the adjusted personnel density, and the correction term are fused based on a pre-set weight to obtain a personnel activity state.

[0157] Further, the target sub-network is also used for:

[0158] Based on the parameterized coupling matrix and the device state data in the standardized data set, an energy consumption prediction model is constructed;

[0159] Based on the personnel activity state and the pre-set comfortable temperature, a temperature constraint set is constructed by the following formula:

[0160] T min (j,i)=T comfort -ΔT deadband ·oj(i)

[0161] T max (j,i)=T comfort +ΔT deadband ·o j (i)

[0162] Wherein, T min (j,i) is the minimum temperature, T max (j,i) is the maximum temperature, T comfort is the pre-set comfortable temperature, o j (i) is the personnel density in the region, ΔT deadband is the temperature control dead zone bandwidth;

[0163] Based on the energy consumption prediction model and the temperature constraint set, a target function is constructed;

[0164] Based on the temperature constraint set and the device physical constraint, a constraint set is constructed, and the constraint set and the target function are integrated to obtain an optimization proposition.

[0165] Further, the expression formula of the target function is:

[0166]

[0167] where J(u) is the objective function, u is the control variable vector, i is the time step index in the prediction horizon, N p is the prediction horizon length, is the predicted energy consumption at the i-th step, w i is the time weight factor, λ is the regulation cost weight, k is the time step index in the control horizon, N c is the control horizon length, Δu k is the device state change vector, ρ is the penalty weight, j is the building zone index, T j (i) is the predicted temperature of zone j at the i-th step, T max (j, i) is the time-varying upper limit of temperature of zone j at the i-th step, T min (j, i) is the time-varying lower limit of temperature of zone j at the i-th step.

[0168] Further, the optimization sub-network comprises:

[0169] calculating an initialization vector based on the standardized dataset, the initialization vector being used to represent an initial state of the device operation;

[0170] calculating a solution of the optimization proposition based on the personnel activity state and the initialization vector, and sorting the solution in a time sequence to obtain control instruction data;

[0171] performing feasibility verification on the control instruction data based on the constraint set in the optimization proposition to obtain a feasibility report;

[0172] optimizing the control instruction sequence based on the feasibility report to obtain energy-saving optimization instructions.

[0173] Further, the coupling module 803 is further used for:

[0174] calculating an air conditioner-fresh air coupling coefficient matrix based on the device rated parameters and the indoor and outdoor humidity through the following formula:

[0175]

[0176] where m ac-fan is the air conditioner-fresh air coupling coefficient matrix, representing the change amount of the total power of the air conditioning system when the fresh air volume changes, V air is the fresh air volume flow rate, k is the fan power coefficient, n is the fan power index, ρ is the air density, h fg is the water vapor latent heat, c p is the air specific heat capacity, W out is the external humidity, W in is the internal humidity, T out is the external temperature, T in is the internal temperature, and COP is the refrigeration coefficient.

[0177] Based on the building physical parameters and external temperature data, the air conditioner-weather coupling coefficient matrix is calculated by the following formula:

[0178]

[0179] Wherein, m ac-temp is the air conditioner-weather coupling coefficient matrix, representing the change of the total power of the air conditioning system when the external temperature changes, U is the comprehensive heat transfer coefficient of the building envelope, A is the total area of the building envelope, η sys is the total efficiency of the system, p is the air density, V air is the volume flow rate of fresh air, c p is the specific heat capacity of air, and COP is the refrigeration coefficient.

[0180] Based on the building physical parameters and solar radiation data, the air conditioner-radiation coupling coefficient matrix is calculated by the following formula:

[0181]

[0182] Wherein, m ac-solar is the air conditioner-radiation coupling coefficient matrix, representing the change of the total power of the air conditioning system when the solar radiation changes, A is the total area of the building envelope, r win is the window-wall ratio, SHGC is the solar heat gain coefficient, η sys is the total efficiency of the system, f vent is the radiation heat absorption coefficient of fresh air, and COP is the refrigeration coefficient.

[0183] Based on the per capita heat dissipation, the air conditioner-human flow coupling coefficient matrix is calculated by the following formula:

[0184]

[0185] Wherein, m ac-people is the air conditioner-human flow coupling coefficient matrix, q person is the per capita heat dissipation, η sys is the total efficiency of the system.

[0186] The parameterized coupling matrix is obtained by integrating the air conditioner-fresh air coupling coefficient matrix, the air conditioner-weather coupling coefficient matrix, the air conditioner-radiation coupling coefficient matrix and the air conditioner-human flow coupling coefficient matrix.

[0187] In one embodiment, a computer device is provided, comprising a memory and a processor, the memory storing a computer program, and the processor implementing the steps of the AI-enabled building energy consumption real-time monitoring and energy saving recommendation method as described above when executing the computer program.

[0188] In one embodiment, a computer readable storage medium is provided, having stored thereon a computer program, which when executed by a processor implements the steps of any of the above method embodiments.

[0189] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts are described in the part of the method embodiments. The above described device embodiments are only illustrative, wherein the components described as separate components can or can not be physically separated, the components displayed as units can or can not be physical units, i.e. can be located in one place or can be distributed to multiple network units. Part or all of the modules can be selected to achieve the purpose of the present disclosure according to actual needs. Those skilled in the art can understand and implement it without creative labor.

[0190] The above described embodiments only express several implementation manners of the present application, the description is more specific and detailed, but it should not be understood as a limitation to the patent scope of the application. It should be pointed out that, for those skilled in the art, without departing from the concept of the present application, several modifications and improvements can be made, which are all within the protection scope of the present application.

Claims

1. An AI-enabled building energy consumption real-time monitoring and energy saving recommendation method, characterized in that, The method comprises: obtaining building real-time operation data, and standardizing the real-time operation data to obtain a standardized data set; wherein the building real-time operation data comprises equipment state data and environment data; obtaining real-time energy consumption of the building through the equipment state data in the building real-time operation data; when the real-time energy consumption of the building exceeds a preset threshold, modeling based on a coupling relationship of the standardized data set, building physical parameters and equipment operation parameters to obtain a parameterized coupling matrix; wherein the parameterized coupling matrix is used to represent the relationship between the equipment state data and the environment data; inputting the standardized data set and the parameterized coupling matrix into an MCA-MPC model to obtain energy-saving optimization instructions; the energy-saving optimization instructions are used to instruct the equipment to adjust the operating state.

2. The AI-enabled building energy consumption real-time monitoring and energy saving recommendation method of claim 1, wherein The MCA-MPC model comprises: a prediction sub-network for predicting a personnel activity state based on a historical personnel flow heat map; the personnel activity state comprises personnel distribution and activity intensity; a target sub-network for determining an optimization proposition based on the standardized data set, the parameterized coupling matrix and the personnel activity state; the optimization proposition comprises an objective function and a constraint set; an optimization sub-network for generating the energy-saving optimization instructions based on the personnel activity state and the optimization proposition.

3. The AI-enabled building energy consumption real-time monitoring and energy saving recommendation method of claim 2, wherein The prediction of the personnel activity state based on the historical personnel flow heat map comprises: predicting the periodic flow of personnel based on the historical personnel flow heat map to obtain a regional reference personnel density; adjusting the personnel distribution of the regional reference personnel density based on real-time public travel data to obtain a personnel migration probability matrix; the real-time public travel data is used to represent the scheduled travel of personnel; adjusting the regional reference personnel density based on the personnel migration probability matrix to obtain an adjusted personnel density; quantifying the influence of historical special events on the adjusted personnel density based on the historical personnel flow heat map to obtain a correction term of the personnel density; fusing the regional reference personnel density, the adjusted personnel density and the correction term based on a pre-set weight to obtain the personnel activity state.

4. The AI-enabled building energy consumption real-time monitoring and energy saving recommendation method of claim 2, wherein The determination of the optimization proposition based on the standardized data set, the parameterized coupling matrix and the personnel activity state comprises: constructing an energy consumption prediction model based on the parameterized coupling matrix and the equipment state data in the standardized data set; constructing a temperature constraint set based on the personnel activity state and a preset comfortable temperature through the following formula: T min (j,i) = T comfort - ΔT deadband • (1 - o j (i)) T max (j,i) = T comfort + ΔT deadband • (1 - o j (i)) where T min (j,i) is the minimum temperature, T max (j,i) is the maximum temperature, T comfort is a preset comfort temperature, o j (i) is the density of people in the area, ΔT deadband is the temperature control deadband width; constructing an objective function based on the energy consumption prediction model and the temperature constraint set; constructing a constraint set based on the temperature constraint set and equipment physical constraints, and integrating the constraint set and the objective function to obtain the optimization proposition.

5. The AI-enabled building energy consumption real-time monitoring and energy saving recommendation method of claim 4, wherein, The expression formula of the objective function is: where J(u) is the objective function, u is the vector of control variables, i is the time step index in the prediction horizon, N p is the prediction horizon length, is the predicted energy consumption at step i, w i is the time weighting factor, λ is the tuning cost weight, k is the time step index in the control horizon, N c is the control horizon length, Δu k is the device state change vector, ρ is the penalty weight, j is the building zone index, T j (i) is the predicted temperature of zone j at step i, T max (j, i) is the time-varying upper temperature limit of zone j at step i, T min (j, i) is the time-varying lower temperature limit of zone j at step i.

6. The AI-enabled building energy consumption real-time monitoring and energy saving recommendation method of claim 2, wherein The generation of the energy-saving optimization instructions based on the personnel activity state and the optimization proposition comprises: calculating an initialization vector based on the standardized data set, the initialization vector being used to represent the initial state of equipment operation; Based on the personnel activity state and the initialization vector, a solution of the optimization proposition is calculated, and the solution is sorted in time sequence to obtain control instruction data; Based on the constraint set in the optimization proposition, the control instruction data is subjected to feasibility verification to obtain a feasibility report; Based on the feasibility report, the control instruction sequence is optimized to obtain the energy-saving optimization instruction.

7. The AI-enabled building energy consumption real-time monitoring and energy saving recommendation method according to any one of claims 1 to 6, characterized in that, The parameterized coupling matrix is obtained based on the coupling relationship modeling of the standardized data set, building physical parameters and equipment operation parameters, including: Based on the equipment rated parameters and indoor and outdoor humidity, the air conditioner-fresh air coupling coefficient matrix is calculated by the following formula: wherein, m ac-fan is an air conditioner-fresh air coupling coefficient matrix, representing the change in the total power of the air conditioning system when the fresh air volume changes, V air is the fresh air volume flow, k is the fresh air fan power coefficient, n is the fan power index, p is the air density, h fg is the water vapor latent heat, c p is the air specific heat capacity, W out is the outside humidity, W in is the inside humidity, T out is the outside temperature, T in is the inside temperature, and COP is the refrigeration coefficient. Based on the building physical parameters and external temperature data, the air conditioner-weather coupling coefficient matrix is calculated by the following formula: wherein m ac-temp is the air conditioning-weather coupling coefficient matrix, representing the change in total power of the air conditioning system when the external temperature changes, U is the comprehensive heat transfer coefficient of the building envelope, A is the total area of the building envelope, η sys is the total efficiency of the system, ρ is the air density, V air is the volume flow rate of fresh air, c p is the specific heat capacity of air, and COP is the refrigeration coefficient; Based on the building physical parameters and solar radiation data, the air conditioner-radiation coupling coefficient matrix is calculated by the following formula: Wherein, m ac-solar is the air conditioning-radiation coupling coefficient matrix, representing the change in total power of the air conditioning system when the solar radiation changes, A is the total area of the building envelope, r win is the window-wall ratio, SHGC is the solar heat gain coefficient, η sys is the total efficiency of the system, f vent is the new air radiation heat absorption coefficient, COP is the refrigeration coefficient; Based on the per capita heat dissipation, the air conditioner-human flow coupling coefficient matrix is calculated by the following formula: wherein m ac-people is the air-conditioning-human flow coupling coefficient matrix, q person is the per capita heat dissipation, η sys is the total system efficiency; The parameterized coupling matrix is obtained by fusing the air conditioner-fresh air coupling coefficient matrix, the air conditioner-weather coupling coefficient matrix, the air conditioner-radiation coupling coefficient matrix and the air conditioner-human flow coupling coefficient matrix.

8. An AI-enabled building energy consumption real-time monitoring and energy saving recommendation system, characterized in that, The system comprises: A source module is configured to obtain building real-time operation data and perform standardized processing on the real-time operation data to obtain a standardized data set; wherein the building real-time operation data comprises equipment state data and environment data; An energy consumption module is configured to obtain real-time energy consumption of a building by the equipment state data in the building real-time operation data; A coupling module is configured to, when the real-time energy consumption of the building exceeds a preset threshold, perform coupling relationship modeling based on the standardized data set, building physical parameters and equipment operation parameters to obtain a parameterized coupling matrix; wherein the parameterized coupling matrix is used to represent the relationship between the equipment state data and the environment data; An instruction module is configured to input the standardized data set and the parameterized coupling matrix into an MCA-MPC model to obtain an energy-saving optimization instruction; the energy-saving optimization instruction is used to instruct the equipment to adjust the operation state. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor executes the computer program to implement the steps of the method of any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 7.

Images