PVT heat pump multi-source energy management method and system based on AI intelligence

Abstract

The embodiment of the invention discloses a PVT heat pump multi-source energy management method and system based on AI intelligence, belongs to the technical field of photovoltaic thermoelectricity, and solves the problem that an existing PVT heat pump energy adjusting system is low in energy comprehensive utilization rate. Collecting environment temperature data and energy demand information; determining a predicted energy demand value based on a multiple linear regression model; predicting the output energy of the PVT assembly and the thermoelectric module in a future period of time through a preset time sequence prediction model to obtain a predicted energy output value; according to the current environment temperature data and the predicted energy demand value, priority coefficients corresponding to the PVT assembly and the thermoelectric module are determined; determining a difference value between the predicted energy demand value and the predicted energy output value, and obtaining to-be-adjusted energy information based on the difference value and a priority coefficient; and matching the to-be-adjusted energy information with a preset heat storage / electricity storage integrated buffer device to realize dynamic adjustment of output energy.

Description

Technical Field

[0001] The present application relates to the field of photovoltaic thermoelectric technology, and in particular to a multi-source energy management method and system for PVT heat pumps based on AI intelligence. Background Art

[0002] In the energy sector, as user demands for electricity and heat become increasingly diverse and dynamic, efficient and precise energy regulation systems are becoming crucial. Currently, PVT (photovoltaic thermal) heat pumps, as integrated energy devices combining photovoltaic power generation and solar thermal utilization, face numerous challenges in energy regulation.

[0003] With technological advancements and rising living standards, user-side electricity and heat demand is experiencing complex fluctuations. On the one hand, users' daily productivity and daily life, such as office equipment and household appliances, experience significant fluctuations in usage time and power requirements, leading to significant peaks and valleys in electricity demand. On the other hand, user heat demands vary significantly across seasons, time periods, and usage scenarios, such as winter heating, summer cooling, and daily hot water supply.

[0004] Existing energy regulation systems based on PVT heat pumps have difficulty in rationally allocating resources when regulating energy in response to these complex and changing situations, resulting in overuse of some components and underuse of others, reducing the overall energy utilization efficiency. Summary of the Invention

[0005] The embodiments of the present application provide an AI-based intelligent PVT heat pump multi-source energy management method and system for solving the following technical problems: the existing energy regulation system based on PVT heat pumps has difficulty in reasonably allocating resources during the energy regulation process, resulting in overuse of some components and underutilization of some components, thereby reducing the comprehensive energy utilization efficiency.

[0006] The embodiments of this application adopt the following technical solutions:

[0007] The embodiment of the present application provides a multi-source energy management method for a PVT heat pump based on AI intelligence. The method includes: collecting ambient temperature data sent by an ambient temperature sensor in real time; and obtaining energy demand information fed back by a user end in real time; wherein the energy demand information includes power demand information and heat demand information; constructing a linear relationship between ambient temperature data, energy demand information, and the output energy of the PVT component and the thermoelectric module based on a multivariate linear regression model, so as to determine a predicted energy demand value based on the linear relationship; predicting the output energy of the PVT component and the thermoelectric module in a future period of time through a preset time series prediction model to obtain a predicted energy output value; determining the priority coefficients corresponding to the PVT component and the thermoelectric module respectively based on the current ambient temperature data and the predicted energy demand value; determining the difference between the predicted energy demand value and the predicted energy output value, and obtaining the energy information to be adjusted based on the difference and the priority coefficient; matching the energy information to be adjusted with a preset integrated heat storage / electricity storage buffer device to achieve dynamic adjustment of the output energy.

[0008] The embodiment of the present application constructs a multivariate linear regression model to analyze key factors such as ambient temperature and energy demand, and constructs a linear relationship between the output energy of the PVT component and the thermoelectric module. It can comprehensively analyze the impact of various factors on energy demand, and improves the accuracy of the prediction compared to a single factor prediction. The priority coefficient is determined based on the current ambient temperature data and the predicted energy demand value, and energy resources can be allocated according to actual conditions. By dynamically adjusting the priority coefficient, the performance and stability of the entire energy system are improved, and the complex and changing energy demand and environmental conditions are better coped with. By determining the difference between the predicted energy demand value and the predicted energy output value, and combining the priority coefficient to obtain the energy information to be adjusted, a clear and accurate basis is provided for energy regulation. The energy information to be adjusted is matched with the preset heat storage / electricity storage integrated buffer device, and the output energy is dynamically adjusted. The distribution and use of energy can be flexibly adjusted according to the real-time energy supply and demand situation, avoiding energy waste during transmission and conversion.

[0009] In one implementation of the present application, a linear relationship between ambient temperature data, energy demand information, PVT components, and thermoelectric module output energy is constructed based on a multivariate linear regression model, specifically including:

[0010] Based on the multiple linear regression model:

[0011]

[0012] Determine a linear relationship;

[0013] Where Y is the predicted energy demand value; β0 is the first regression coefficient; β1 is the second regression coefficient; β2 is the third regression coefficient; Te is the ambient temperature data; γ is the weight coefficient of solar radiation intensity to ambient temperature; Gs is the solar radiation intensity; E PVT Energy required for PVT components; E T is the energy required by the thermoelectric module; α1 is the first adjustment parameter; α2 is the second adjustment parameter; ε is the error term.

[0014] In one implementation of the present application, the output energy of the PVT component and the thermoelectric module in the future period is predicted by a preset time series prediction model to obtain a predicted energy output value, specifically including: sorting the historical output energy values of the PVT component based on chronological order to construct an electricity output graph; sorting the historical output energy values of the thermoelectric module based on chronological order to construct a heat output graph; dividing the electricity output graph and the heat output graph into multiple regions according to the changing state of the ambient temperature data, and determining the energy change difference corresponding to adjacent regions; inputting the future ambient temperature prediction information, the electricity output graph, the heat output graph and the energy change difference into the preset time series prediction model to obtain the predicted energy output value.

[0015] In one implementation of the present application, priority coefficients corresponding to the PVT components and thermoelectric modules are determined based on the current ambient temperature data and the predicted energy demand value, specifically including: expanding the predicted energy demand value by segment values to obtain energy demand segment values; matching the energy demand segment values with the preset information database to obtain a first reference priority coefficient set; comparing the current ambient temperature data with the environmental feature labels corresponding to each coefficient in the first reference priority coefficient set, and determining the second reference priority coefficient that best matches the current ambient temperature data feature through the feature similarity obtained by comparison, and forming a second reference priority coefficient set; constructing a geographically weighted regression model based on the correlation between the geographical location of the PVT heat pump and the preset priority coefficient set, so as to obtain the priority coefficient through the geographically weighted regression model and the second reference priority coefficient set.

[0016] In one implementation of the present application, the difference between the predicted energy demand value and the predicted energy output value is determined, and the energy information to be adjusted is obtained based on the difference and the priority coefficient, specifically including: performing difference decomposition of different scales on the difference between the predicted energy demand value and the predicted energy output value through wavelet transform to obtain difference fluctuation and difference trend; constructing an input set based on difference fluctuation, difference trend, priority coefficient, ambient temperature data and energy demand information; inputting the input set into a dual-depth Q network to adjust the priority coefficient based on different energy state data through the dual-depth Q network, and outputting the priority coefficients corresponding to different energy states; obtaining the energy information to be adjusted based on the difference and the priority coefficients corresponding to different energy states.

[0017] In one implementation of the present application, the energy information to be adjusted is obtained based on the difference and the priority coefficients corresponding to different energy states, specifically including: taking the difference, priority coefficient and other relevant factors as nodes to determine the joint probability distribution between different nodes; wherein, other relevant factors include at least one of seasonal factors, user type and weather; based on the difference, the priority coefficients corresponding to different energy states, the joint probability distribution and the preset energy output constraints, the adjustment value and adjustment direction in the energy information to be adjusted are determined by a linear programming optimization algorithm.

[0018] In one implementation of the present application, the adjustment value and adjustment direction of the energy information to be adjusted are determined by a linear programming optimization algorithm, specifically including: constructing a comprehensive objective function:

[0019]

[0020] (C d (x PVT ,x TE )+C e (x PVT ,x TE ))-ω4·V ar E;

[0021] Build constraints:

[0022] (1-l)x PVT +(1-l)x TE ≥D;

[0023] -x PVT -max≤x PVT ≤x PVT -max;

[0024] -x TE -max≤x TE ≤x TE -max;

[0025] The comprehensive objective function and constraint conditions are input into the linear programming solver, and the adjustment value and adjustment direction of the energy information to be adjusted are output through the linear programming solver; wherein Z is the comprehensive objective function; P PVT is the priority coefficient of the PVT component; P TE is the priority coefficient of the thermoelectric module; x PVT is the energy regulation of the PVT component; x TE is the energy regulation amount of the thermoelectric module; D is the difference; λ is the discount factor corresponding to the joint probability distribution; T is time; ω1 is the first weight coefficient; ω2 is the second weight coefficient; ω3 is the third weight coefficient; ω4 is the fourth weight coefficient; R is the priority coefficient; Q is other factors; C d is the adjustment cost function; C e is the environmental emission cost; V ar E is the variance of energy supply and demand balance; l is the energy transmission loss rate; θ is a positive number.

[0026] In one implementation of the present application, the energy information to be adjusted is matched with a preset integrated heat storage / electricity storage buffer device to achieve dynamic adjustment of the output energy, specifically including: determining the state of charge of the battery based on the rated capacity and remaining capacity of the battery corresponding to the preset integrated heat storage / electricity storage buffer device, and obtaining the electricity storage adjustment space based on the state of charge and the charge-discharge characteristic curve; determining the current heat storage amount based on the temperature and specific heat capacity of the heat storage medium corresponding to the preset integrated heat storage / electricity storage buffer device, and determining the heat storage adjustment space based on the current heat storage amount and the geometric parameters of the heat storage device; matching the adjustment value and adjustment direction with the electricity storage adjustment space and the heat storage adjustment space, respectively, and dynamically adjusting the preset integrated heat storage / electricity storage buffer device based on the matching results.

[0027] In one implementation of the present application, the adjustment value and the adjustment direction are matched with the electricity storage adjustment space and the heat storage adjustment space, and based on the matching result, the preset heat storage / electricity storage integrated buffer device is dynamically adjusted, specifically including: determining the corresponding adjustment space based on the adjustment direction; comparing the adjustment value with the adjustment space to determine the adjustment difference; when the adjustment difference is not greater than the adjustment space, responding to the adjustment instruction, dynamically adjusting the preset heat storage / electricity storage integrated buffer device; when the adjustment difference is greater than the adjustment space, issuing a dynamic adjustment alarm.

[0028] An embodiment of the present application provides an AI-based intelligent PVT heat pump multi-source energy management system, comprising: a data acquisition unit, which collects ambient temperature data sent by an ambient temperature sensor in real time, and obtains energy demand information fed back by a user end in real time; wherein the energy demand information includes power demand information and heat demand information; an energy demand value prediction unit, which constructs a linear relationship between ambient temperature data, energy demand information, and the output energy of the PVT component and the thermoelectric module based on a multivariate linear regression model, so as to determine a predicted energy demand value based on the linear relationship; an energy output value prediction unit, which predicts the output energy of the PVT component and the thermoelectric module in a future period of time through a preset time series prediction model to obtain a predicted energy output value; a priority coefficient determination unit, which determines the priority coefficients corresponding to the PVT component and the thermoelectric module respectively based on current ambient temperature data and the predicted energy demand value; a to-be-adjusted energy determination unit, which determines the difference between the predicted energy demand value and the predicted energy output value, and obtains the to-be-adjusted energy information based on the difference and the priority coefficient; and a dynamic adjustment unit, which matches the to-be-adjusted energy information with a preset heat storage / electricity storage integrated buffer device to achieve dynamic adjustment of the output energy.

[0029] At least one of the above-mentioned technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects: By constructing a multivariate linear regression model, the embodiments of the present application analyze key factors such as ambient temperature and energy demand, establishing a linear relationship between the output energy of the PVT component and the thermoelectric module, and comprehensively analyzing the impact of various factors on energy demand, the prediction accuracy is improved compared to single-factor predictions. Priority coefficients are determined based on current ambient temperature data and predicted energy demand values, enabling energy resources to be allocated based on actual conditions. By dynamically adjusting the priority coefficients, the performance and stability of the entire energy system are improved, better coping with complex and changing energy demands and environmental conditions. By determining the difference between the predicted energy demand value and the predicted energy output value and combining it with the priority coefficients to obtain the energy information to be adjusted, a clear and accurate basis for energy regulation is provided. The energy information to be adjusted is matched with the pre-set integrated heat / electricity storage buffer device, and the output energy is dynamically adjusted. This allows for flexible adjustment of energy distribution and use based on real-time energy supply and demand conditions, avoiding energy waste during transmission and conversion. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments described in the present application. For those skilled in the art, other drawings can be obtained based on these drawings without inventive work. In the drawings:

[0031] Figure 1A flow chart of a multi-source energy management method for a PVT heat pump based on AI intelligence provided in an embodiment of the present application;

[0032] Figure 2 A schematic structural diagram of an AI-based intelligent PVT heat pump multi-source energy management system provided in an embodiment of the present application.

[0033] Reference numerals:

[0034] 200: AI-based intelligent PVT heat pump multi-source energy management system, 201: data acquisition unit, 202: energy demand value prediction unit, 203: energy output value prediction unit, 204: priority coefficient determination unit, 205: energy to be adjusted determination unit, 206: dynamic adjustment unit. DETAILED DESCRIPTION

[0035] The embodiments of the present application provide a PVT heat pump multi-source energy management method and system based on AI intelligence.

[0036] In order to enable those skilled in the art to better understand the technical solutions in this application, the following will clearly and completely describe the technical solutions in the embodiments of this application in conjunction with the drawings in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0037] The technical solutions proposed in the embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0038] Figure 1 This is a flow chart of a multi-source energy management method for a PVT heat pump based on AI intelligence provided in an embodiment of the present application. Figure 1 As shown, the AI-based multi-source energy management method for PVT heat pumps includes the following steps:

[0039] Step 101: collect ambient temperature data sent by an ambient temperature sensor in real time, and obtain energy demand information fed back by a user end in real time; wherein the energy demand information includes power demand information and heat demand information.

[0040] In one implementation of this application, multiple ambient temperature sensors are provided to obtain real-time ambient temperature data. Furthermore, a smart electricity meter collects real-time user electricity consumption data, while a smart heat meter measures user calorie consumption data. The acquired electricity and calorie consumption data are used to obtain energy demand information from the user.

[0041] Step 102: Based on a multiple linear regression model, a linear relationship is constructed among the ambient temperature data, the energy demand information, the PVT component, and the output energy of the thermoelectric module to determine a predicted energy demand value based on the linear relationship.

[0042] In one implementation of the present application, a multivariate linear regression model is constructed based on the acquired historical data, including ambient temperature data, energy demand information, and output energy of PVT components and thermoelectric modules.

[0043] Among them, the constructed multiple linear regression model is as follows:

[0044]

[0045] A linear relationship was determined through the model;

[0046] Where Y is the predicted energy demand value; β0 is the first regression coefficient; β1 is the second regression coefficient; β2 is the third regression coefficient; Te is the ambient temperature data; γ is the weight coefficient of solar radiation intensity to ambient temperature; Gs is the solar radiation intensity; E PVT Energy required for PVT components; E T is the energy required by the thermoelectric module; α1 is the first adjustment parameter; α2 is the second adjustment parameter; ε is the error term.

[0047] Furthermore, after the model training and optimization are completed, the real-time collected ambient temperature data and the real-time output energy data of the PVT components and thermoelectric modules are input, and the input data are substituted into the trained multivariate linear regression model to calculate the predicted energy demand value.

[0048] Step 103: Predict the output energy of the PVT component and the thermoelectric module in the future by using a preset time series prediction model to obtain a predicted energy output value.

[0049] In one implementation of the present application, the historical output energy values of PVT components are sorted chronologically to construct an electrical output graph. The historical output energy values of thermoelectric modules are sorted chronologically to construct a heat output graph. Based on the changing state of ambient temperature data, the electrical output graph and the heat output graph are each divided into multiple regions, and the energy change differences corresponding to adjacent regions are determined. Future ambient temperature forecast information, the electrical output graph, the heat output graph, and the energy change differences are input into a preset time series prediction model to obtain predicted energy output values.

[0050] Specifically, collect the electrical output power data of the PVT component over the past period of time. For example, record the electrical output power value of the PVT component every 15 minutes over the past month, and sort these data in chronological order to ensure the temporal continuity of the data. Similarly, collect the thermal output power data of the thermoelectric module over the same period of time, also recording it every 15 minutes, and sort these data in chronological order. With time as the horizontal axis and the electrical output power of the PVT component as the vertical axis, plot the sorted data into a line graph. Also, plot the heat output graph of the thermoelectric module with time as the horizontal axis and the thermal output power of the thermoelectric module as the vertical axis.

[0051] Furthermore, ambient temperature data within the same time period is collected and accurately matched with the timestamps of the PVT component and thermoelectric module data. The power output diagram is divided into multiple regions based on the changing state of the ambient temperature. For example, when the ambient temperature rises, the corresponding PVT component power output may also change. The portion of the power output diagram corresponding to this temperature rise phase is divided into one region; when the temperature fluctuates, the corresponding power output portion is divided into another region. The ambient temperature changes within each region have similar characteristics. The same operation is performed on the heat output diagram of the thermoelectric module. For example, when the ambient temperature is low and stable, the heat output of the thermoelectric module may be relatively stable in order to meet the heating demand. The portion of the heat output diagram corresponding to this phase is divided into one region.

[0052] Furthermore, for adjacent areas of the power output graph, the energy change difference is calculated, and similar calculations are performed for adjacent areas of the heat output graph. The ambient temperature forecast information for a period of time in the future is obtained through the relevant meteorological data platform. The future ambient temperature forecast information, the constructed power output graph, the heat output graph, and the calculated energy change difference are integrated together as input data for the preset time series prediction model. The prepared data is input into the preset time series prediction model. This model is a long short-term memory network trained based on historical data. The model will predict the future energy output of PVT components and thermoelectric modules based on the input ambient temperature forecast information, combined with the historical power output graph, heat output graph, and the laws reflected by the energy change difference.

[0053] Step 104 : Determine the priority coefficients corresponding to the PVT components and the thermoelectric modules respectively according to the current ambient temperature data and the predicted energy demand value.

[0054] In one implementation of the present application, the predicted energy demand value is numerically expanded to obtain an energy demand segment value. The energy demand segment value is matched with a preset information database to obtain a first reference priority coefficient set. The current ambient temperature data is compared with the environmental feature label corresponding to each coefficient in the first reference priority coefficient set. By comparing the feature similarity obtained, a second reference priority coefficient that best matches the current ambient temperature data feature is determined and constitutes a second reference priority coefficient set. Based on the correlation between the geographical location of the PVT heat pump and the preset priority coefficient set, a geographically weighted regression model is constructed to obtain a priority coefficient through the geographically weighted regression model and the second reference priority coefficient set.

[0055] Specifically, for each predicted energy demand value, determine the segment to which it belongs, and expand it to the representative value or range value of the segment. For example, if the power demand value predicted at a certain moment is 120kW, it falls within the [100-150kW) segment. Match the energy demand segment value obtained in the previous step with the energy demand segment field in the database. For example, for the expanded power demand segment value [100-150kW), search the database for all records whose energy demand segment fields contain this value. Multiple records that meet the conditions can be found, and these records correspond to different environmental feature labels and priority coefficients. Collect these found priority coefficients to form a first reference priority coefficient set. Among them, the preset information database in the embodiment of the present application stores a large amount of information related to energy demand, environmental features and priority coefficients. The structure of the database may include multiple fields, such as energy demand segment fields, environmental feature label fields, including temperature ranges, seasons, weather conditions, etc., and corresponding priority coefficient fields.

[0056] Furthermore, for each coefficient in the first reference priority coefficient set, its corresponding environmental feature label is extracted. For example, for a coefficient of 0.6, its corresponding environmental feature label is "summer, sunny, temperature 25-30°C." Using cosine similarity, the similarity value between each environmental feature label and the current ambient temperature data is obtained. The priority coefficient corresponding to the environmental feature label with the highest similarity value is selected to form the second reference priority coefficient set.

[0057] Furthermore, the geographical location information of the PVT heat pump, such as longitude and latitude coordinates, is collected. At the same time, the potential correlation between the preset priority coefficient set and the geographical location is analyzed. The geographically weighted regression model is used to determine the correlation. The geographically weighted regression model can take into account the spatial heterogeneity of the geographical location, that is, the variable relationship in different locations may be different. When constructing the model, the geographical location of the PVT heat pump is used as the independent variable, the priority coefficient is used as the dependent variable, and other relevant factors such as population density and energy supply infrastructure are considered as control variables. The second reference priority coefficient set is input into the geographically weighted regression model as the initial value or constraint condition. After model calculation and adjustment, the priority coefficient of the PVT heat pump in the area is finally obtained.

[0058] Step 105: Determine the difference between the predicted energy demand value and the predicted energy output value, and obtain the energy information to be adjusted based on the difference and the priority coefficient.

[0059] In one implementation of the present application, the difference between the predicted energy demand value and the predicted energy output value is decomposed at different scales using a wavelet transform to obtain difference fluctuations and difference trends. An input set is constructed based on the difference fluctuations, difference trends, priority coefficients, ambient temperature data, and energy demand information. The input set is input into a dual-depth Q network, which adjusts the priority coefficients based on different energy state data and outputs the priority coefficients corresponding to different energy states. Based on the difference and the priority coefficients corresponding to different energy states, the energy information to be adjusted is obtained.

[0060] Specifically, for the difference sequence between the predicted energy demand value and the predicted energy output value, the difference signal is decomposed into different frequency components using wavelet basis functions. The low-frequency components correspond to the long-term trend of the signal, while the high-frequency components correspond to the short-term fluctuations of the signal.

[0061] The input set consists of difference fluctuations, difference trends, priority coefficients, ambient temperature data, and energy demand information. This input set is fed into a dual-depth Q-network, which evaluates different priority coefficient adjustments based on previously learned experience. The dual-depth Q-network is trained by collecting a large amount of historical data related to the energy system, including predicted energy demand values, predicted energy output values, ambient temperature data, energy demand information, and corresponding priority coefficients for different time periods. The dual-depth Q-network consists of two deep neural networks: an online network and a target network. This network is trained in a loop, with the parameters of the online network copied to the target network at a certain number of training steps to ensure a certain degree of synchronization between the target and online networks. Training stops when the maximum number of training steps is reached or the loss function falls within a certain number of steps.

[0062] For example, the network might evaluate the expected reward for increasing the priority coefficient from 0.6 to 0.7, and then decreasing it to 0.5. The reward value is often related to indicators such as the degree of energy supply and demand balance and energy efficiency. If adjusting the priority coefficient reduces the difference between energy supply and demand and improves energy efficiency, a positive reward is given; otherwise, a negative reward is given. After calculation and comparison, the network selects the priority coefficient corresponding to the action with the highest expected reward as the output.

[0063] In one implementation of this application, the difference, priority coefficient, and other relevant factors are used as nodes to determine the joint probability distribution between different nodes; the other relevant factors include at least one of seasonal factors, user type, and weather. Based on the difference, the priority coefficient corresponding to different energy states, the joint probability distribution, and preset energy output constraints, a linear programming optimization algorithm is used to determine the adjustment value and adjustment direction of the energy information to be adjusted.

[0064] Specifically, the causal relationships between different nodes are determined and directed edges are established. For example, seasonal factors affect a user's energy demand, which in turn affects the energy difference, so a directed edge is established from the seasonal factor node to the difference node. Weather conditions affect the energy output of PVT components, which in turn affects the energy difference, so a directed edge is established from the weather node to the difference node. User type affects energy demand patterns and influences the determination of the reference priority coefficient, so a directed edge is established from the user type node to the reference priority coefficient node. Through these directed edges, a network structure is constructed that reflects the interrelationships between various factors.

[0065] Furthermore, a large amount of historical data related to each node is collected. Based on this data, the conditional probability of each node given the state of its parent node is calculated. Using these conditional probabilities, the joint probability distribution of different node state combinations in the entire Bayesian network is determined using the joint probability formula of the Bayesian network. Based on the difference, the priority coefficients corresponding to different energy states, the joint probability distribution, and the preset energy output constraints, a linear programming optimization algorithm is used to determine the adjustment value and adjustment direction of the energy information to be adjusted.

[0066] Specifically, construct a comprehensive objective function:

[0067]

[0068] (C d (x PVT ,x TE )+C e (x PVT ,x TE ))-ω4·V ar E;

[0069] Build constraints:

[0070] (1-l)x PVT +(1-l)x TE ≥D;

[0071] -x PVT -max≤x PVT ≤x PVT -max;

[0072] -x TE -max≤x TE ≤x TE -max;

[0073] Input the comprehensive objective function and constraint conditions into the linear programming solver, and output the adjustment value and adjustment direction in the energy information to be adjusted through the linear programming solver;

[0074] Among them, Z is the comprehensive objective function; P PVT is the priority coefficient of the PVT component; P TE is the priority coefficient of the thermoelectric module; x PVT is the energy regulation of the PVT component; x TE is the energy regulation amount of the thermoelectric module; D is the difference; λ is the discount factor corresponding to the joint probability distribution; T is time; ω1 is the first weight coefficient; ω2 is the second weight coefficient; ω3 is the third weight coefficient; ω4 is the fourth weight coefficient; R is the priority coefficient; Q is other factors; C d is the adjustment cost function; C e is the environmental emission cost; V ar E is the variance of energy supply and demand balance; l is the energy transmission loss rate; θ is a positive number.

[0075] Step 106: Match the energy information to be adjusted with the preset integrated heat / electricity storage buffer device to achieve dynamic adjustment of output energy.

[0076] In one implementation of the present application, the battery's state of charge (SOC) is determined based on the rated capacity and remaining capacity of the battery corresponding to a pre-installed integrated heat / electricity storage buffer device. The SOC and charge / discharge characteristic curves provide a storage adjustment space. The current heat storage capacity is determined based on the temperature and specific heat capacity of the heat storage medium corresponding to the pre-installed integrated heat / electricity storage buffer device. The heat storage adjustment space is determined based on the current heat storage capacity and the geometric parameters of the heat storage device. The adjustment value and adjustment direction are matched to the storage adjustment space and heat storage adjustment space, respectively. Based on the matching results, the pre-installed integrated heat / electricity storage buffer device is dynamically adjusted.

[0077] Specifically, the battery's state of charge (SOC) is determined based on the ratio of its remaining capacity to its rated capacity. Furthermore, the energy storage adjustment range is determined based on the SOC and the charge-discharge characteristic curve. The charge-discharge characteristic curve describes the current and voltage ranges within which the battery can safely and efficiently charge and discharge at different SOCs, as well as the corresponding capacity changes. The energy storage adjustment range includes the maximum charge capacity and the maximum discharge capacity.

[0078] Furthermore, the current heat storage capacity is calculated using the heat calculation formula: Heat = Mass × Specific Heat Capacity × Temperature Change. The geometric parameters of the heat storage device include volume and effective heat storage space height. These parameters, along with the density of the heat storage medium, determine the maximum heat storage capacity of the heat storage device. The heat storage adjustment space includes the maximum heat storage difference and the maximum heat release difference. The received adjustment value is compared with the power storage adjustment space. If the charge adjustment is within the power storage adjustment space range, the match is successful. If it is not within this range, the discharge adjustment exceeds the power storage adjustment space range and the match fails.

[0079] In one implementation of the present application, a corresponding adjustment space is determined based on the adjustment direction. The adjustment value is compared with the adjustment space to determine an adjustment difference. If the adjustment difference is not greater than the adjustment space, the preset integrated heat / electricity storage buffer device is dynamically adjusted in response to an adjustment instruction. If the adjustment difference is greater than the adjustment space, a dynamic adjustment alarm is issued.

[0080] Furthermore, when the power storage regulation is successfully matched, such as when charging is successfully matched, the battery management system controls the charging circuit to charge the battery at an appropriate current and voltage, ensuring that the charge is completed within a safe charging parameter range. If the match fails, such as when the discharge exceeds the range, the system can issue an alarm or adjust the regulation strategy, such as reducing the discharge request and re-matching and regulation. When the heat storage regulation is successfully matched, the heating operation is performed according to the preset heating power and time. If the heat release regulation fails, such as when the heat release exceeds the range, the system can adjust the heat release strategy based on actual conditions, such as reducing the heat release power to ensure that the heat release operation is carried out within the heat storage regulation range, thus achieving dynamic adjustment of the pre-installed integrated heat storage / power storage buffer device.

[0081] Figure 2 This is a schematic diagram of the structure of a PVT heat pump multi-source energy management system based on AI intelligence provided in the embodiment of this application. Figure 2As shown, the AI-based PVT heat pump multi-source energy management system 200 includes: a data acquisition unit 201, which collects ambient temperature data sent by the ambient temperature sensor in real time, and obtains energy demand information fed back by the user in real time; the energy demand information includes power demand information and heat demand information. An energy demand value prediction unit 202, which uses a multivariate linear regression model to construct a linear relationship between ambient temperature data, energy demand information, and the output energy of the PVT component and thermoelectric module, and determines the predicted energy demand value based on the linear relationship. An energy output value prediction unit 203, which uses a preset time series prediction model to predict the output energy of the PVT component and thermoelectric module over a period of time in the future, obtains a predicted energy output value. A priority coefficient determination unit 204, which determines the priority coefficients corresponding to the PVT component and thermoelectric module respectively based on the current ambient temperature data and the predicted energy demand value. A to-be-adjusted energy determination unit 205, which determines the difference between the predicted energy demand value and the predicted energy output value, and obtains the to-be-adjusted energy information based on the difference and the priority coefficients. The dynamic adjustment unit 206 matches the energy information to be adjusted with the preset integrated heat storage / electricity storage buffer device to achieve dynamic adjustment of the output energy.

[0082] The various embodiments in this application are described in a progressive manner. Similar portions between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from the other embodiments. In particular, the device, apparatus, and non-volatile computer storage medium embodiments are generally similar to the method embodiments, so their descriptions are relatively simple. For relevant portions, refer to the descriptions of the method embodiments.

[0083] The foregoing is merely an embodiment of the present application and is not intended to limit the present application. It will be apparent to those skilled in the art that various modifications and variations may be made to the embodiments of the present application. However, such modifications or substitutions do not deviate from the spirit and scope of the technical solutions of the various embodiments of the present application.

Claims

1. AI-based intelligent PVT heat pump multi-source energy management method, characterized by: The method comprises: Real-time collection of ambient temperature data sent by the ambient temperature sensor, and real-time acquisition of energy demand information fed back by the user end; wherein the energy demand information includes power demand information and heat demand information; Constructing a linear relationship between the ambient temperature data, the energy demand information, the PVT component, and the output energy of the thermoelectric module based on a multiple linear regression model to determine a predicted energy demand value based on the linear relationship; The output energy of the PVT component and the thermoelectric module in the future is predicted by a preset time series prediction model to obtain a predicted energy output value; Determining priority coefficients corresponding to the PVT component and the thermoelectric module, respectively, based on current ambient temperature data and the predicted energy demand value; Determining a difference between the predicted energy demand value and the predicted energy output value, and obtaining energy information to be adjusted based on the difference and the priority coefficient; The energy information to be adjusted is matched with a preset heat storage / electricity storage integrated buffer device to achieve dynamic adjustment of output energy.

2. The AI-based intelligent PVT heat pump multi-source energy management method according to claim 1 is characterized in that: The multivariate linear regression model is used to construct a linear relationship between the ambient temperature data, energy demand information, PVT components, and thermoelectric module output energy, specifically including: Based on the multiple linear regression model: determining the linear relationship; Where Y is the predicted energy demand value; β0 is the first regression coefficient; β1 is the second regression coefficient; β2 is the third regression coefficient; Te is the ambient temperature data; γ is the weight coefficient of solar radiation intensity to ambient temperature; Gs is the solar radiation intensity; EPVT is the energy demand of the PVT component; E T is the energy required by the thermoelectric module; α1 is the first adjustment parameter; α2 is the second adjustment parameter; ε is the error term.

3. The AI-based intelligent PVT heat pump multi-source energy management method according to claim 1 is characterized in that: The output energy of the PVT component and the thermoelectric module in a future period of time is predicted by using a preset time series prediction model to obtain a predicted energy output value, specifically including: Sorting the historical output energy values of the PVT components based on chronological order to construct an energy output graph; Sorting the historical output energy values of the thermoelectric modules based on chronological order to construct a heat output graph; Dividing the power output graph and the heat output graph into a plurality of regions according to the change state of the ambient temperature data, and determining energy change differences corresponding to adjacent regions; The future ambient temperature prediction information, the power output graph, the heat output graph, and the energy change difference are input into a preset time series prediction model to obtain a predicted energy output value.

4. The AI-based intelligent PVT heat pump multi-source energy management method according to claim 1 is characterized in that: The determining of priority coefficients corresponding to the PVT component and the thermoelectric module, respectively, based on the current ambient temperature data and the predicted energy demand value, specifically includes: Expanding the predicted energy demand value by segment value to obtain an energy demand segment value; Matching the energy demand segment value with a preset information database to obtain a first reference priority coefficient set; Comparing the current ambient temperature data with the environmental feature labels corresponding to each coefficient in the first reference priority coefficient set, and determining a second reference priority coefficient that best matches the current ambient temperature data feature based on feature similarity obtained from the comparison, and forming a second reference priority coefficient set; Based on the correlation between the geographical location of the PVT heat pump and the preset priority coefficient set, a geographically weighted regression model is constructed to obtain the priority coefficient through the geographically weighted regression model and the second reference priority coefficient set.

5. The AI-based intelligent PVT heat pump multi-source energy management method according to claim 1 is characterized in that: The determining a difference between the predicted energy demand value and the predicted energy output value, and obtaining the energy information to be adjusted based on the difference and the priority coefficient, specifically includes: By wavelet transform, the difference between the predicted energy demand value and the predicted energy output value is decomposed into difference values of different scales to obtain difference fluctuation and difference trend; constructing an input set based on the difference fluctuation, the difference trend, the priority coefficient, ambient temperature data, and energy demand information; Inputting the input set into a dual-depth Q network, so that the dual-depth Q network adjusts the priority coefficient based on different energy state data and outputs priority coefficients corresponding to different energy states; The energy information to be adjusted is obtained based on the difference and the priority coefficients respectively corresponding to the different energy states.

6. The AI-based intelligent PVT heat pump multi-source energy management method according to claim 5 is characterized in that: The obtaining of the energy information to be adjusted based on the difference and the priority coefficients corresponding to the different energy states specifically includes: The difference, the priority coefficient, and other relevant factors are used as nodes to determine a joint probability distribution between different nodes; wherein the other relevant factors include at least one of a seasonal factor, a user type, and weather; Based on the difference, the priority coefficients corresponding to the different energy states, the joint probability distribution and the preset energy output constraint conditions, the adjustment value and adjustment direction in the energy information to be adjusted are determined by a linear programming optimization algorithm.

7. The AI-based intelligent PVT heat pump multi-source energy management method according to claim 6 is characterized in that: The step of determining the adjustment value and adjustment direction of the energy information to be adjusted by using a linear programming optimization algorithm specifically includes: Construct a comprehensive objective function: (C d (x PVT ,x TE )+C e (x PVT ,x TE ))-ω4·V ar E; Build constraints: (1-l)x PVT +(1-l)x TE ≥D; -x PVT -max≤x PVT ≤x PVT -max; -x TE -max≤x TE ≤x TE -max; Inputting the comprehensive objective function and the constraint conditions into a linear programming solver, and outputting the adjustment value and adjustment direction of the energy information to be adjusted through the linear programming solver; Among them, Z is the comprehensive objective function; P PVT is the priority coefficient of the PVT component; P TE is the priority coefficient of the thermoelectric module; x PVT is the energy regulation of the PVT component; x TE is the energy regulation amount of the thermoelectric module; D is the difference; λ is the discount factor corresponding to the joint probability distribution; T is time; ω1 is the first weight coefficient; ω2 is the second weight coefficient; ω3 is the third weight coefficient; ω4 is the fourth weight coefficient; R is the priority coefficient; Q is other factors; C d is the adjustment cost function; C e is the environmental emission cost; V ar E is the variance of energy supply and demand balance; l is the energy transmission loss rate; θ is a positive number.

8. The AI-based intelligent PVT heat pump multi-source energy management method according to claim 1 is characterized in that: The step of matching the energy information to be adjusted with the preset integrated heat storage / electricity storage buffer device to achieve dynamic adjustment of output energy specifically includes: Determining the state of charge of the battery based on the rated capacity and remaining capacity of the battery corresponding to the preset integrated heat storage / electricity storage buffer device, and obtaining the power storage adjustment space based on the state of charge and the charge-discharge characteristic curve; Determining a current heat storage amount based on the temperature and specific heat capacity of the heat storage medium corresponding to the preset integrated heat storage / electricity storage buffer device, and determining a heat storage adjustment space based on the current heat storage amount and geometric parameters of the heat storage device; The adjustment value and the adjustment direction are matched with the electricity storage adjustment space and the heat storage adjustment space respectively, and based on the matching results, the preset heat storage / electricity storage integrated buffer device is dynamically adjusted.

9. The AI-based intelligent PVT heat pump multi-source energy management method according to claim 8 is characterized in that: The step of matching the adjustment value and the adjustment direction with the electricity storage adjustment space and the heat storage adjustment space, and dynamically adjusting the preset heat storage / electricity storage integrated buffer device based on the matching result, specifically includes: Based on the adjustment direction, a corresponding adjustment space is determined; Comparing the adjustment value with the adjustment space to determine an adjustment difference; In the case where the adjustment difference is not greater than the adjustment space, responding to an adjustment instruction, dynamically adjusting the preset heat storage / electricity storage integrated buffer device; When the adjustment difference is greater than the adjustment space, a dynamic adjustment alarm is performed.

10. AI-based intelligent PVT heat pump multi-source energy management system, characterized by: The system comprises: A data acquisition unit collects ambient temperature data sent by the ambient temperature sensor in real time, and acquires energy demand information fed back by the user end in real time; wherein the energy demand information includes power demand information and heat demand information; an energy demand value prediction unit, which constructs a linear relationship between the ambient temperature data, the energy demand information, the PVT component, and the output energy of the thermoelectric module based on a multiple linear regression model, so as to determine a predicted energy demand value based on the linear relationship; An energy output value prediction unit, which predicts the output energy of the PVT component and the thermoelectric module in a future period of time by using a preset time series prediction model to obtain a predicted energy output value; a priority coefficient determination unit, which determines the priority coefficients corresponding to the PVT component and the thermoelectric module respectively according to the current ambient temperature data and the predicted energy demand value; an energy-to-be-adjusted determining unit, configured to determine a difference between the predicted energy demand value and the predicted energy output value, and obtain information on the energy to be adjusted based on the difference and the priority coefficient; The dynamic adjustment unit matches the energy information to be adjusted with the preset heat storage / electricity storage integrated buffer device to achieve dynamic adjustment of the output energy.

Images