Heating scheduling method and device and storage medium

Abstract

The application relates to a heat supply scheduling method and device and a storage medium, wherein the heat supply scheduling method comprises the following steps: acquiring multi-source sensing data of multiple regions, and converting the multi-source sensing data into a multi-source fusion feature vector; inputting the multi-source fusion feature vector into a preset time sequence prediction model to generate time sequence prediction results of the regions; storing preset energy according to the time sequence prediction results and a preset photovoltaic preferential consumption preset strategy; respectively calculating preheating starting durations of the regions based on the multi-source sensing data of the regions; and respectively performing heat supply on the regions based on the preheating starting durations of the regions and the preset energy. The problems that it is difficult to guarantee the effects of multi-region heat supply and freezing prevention in the related art are solved, and intelligent distribution and self-adaptive regulation and control of multi-region heat are achieved.

Description

Technical Field

[0001] This application relates to the field of energy management, and in particular to heating dispatching methods, devices and storage media. Background Technology

[0002] In rural and remote areas of high-altitude and cold regions, schools and public toilets are often located back-to-back. Extreme low temperatures in winter easily lead to problems such as frozen and cracked pipes in these toilets, poor indoor environments, and high pressure on heating and frost prevention systems. Currently, energy self-sufficiency is mainly achieved through off-grid integrated solar-storage-thermal systems. However, existing dispatching methods rely primarily on fixed rules and thresholds, making it impossible to accurately allocate heat according to the energy needs of different areas, thus failing to guarantee heating and frost prevention effects across multiple regions.

[0003] Currently, no effective solution has been proposed to address the problem of insufficient heating and frost protection in multiple areas in related technologies. Summary of the Invention

[0004] This application provides a heating scheduling method, apparatus, and storage medium to at least solve the problem in related technologies that it is difficult to guarantee heating and antifreeze effects in multiple areas.

[0005] In a first aspect, embodiments of this application provide a heating scheduling method, the method comprising:

[0006] Acquire multi-source sensing data from multiple regions and convert the multi-source sensing data into a multi-source fusion feature vector;

[0007] The multi-source fusion feature vector is input into a preset time-series prediction model to generate time-series prediction results for each region;

[0008] Based on the time-series prediction results and the preset photovoltaic priority consumption and storage strategy, pre-stored energy is stored;

[0009] Based on the multi-source sensing data of each region, the preheating start-up time of each region is calculated respectively;

[0010] Heating is provided to each region based on the preheating start-up time and the pre-stored energy.

[0011] In some embodiments, storing pre-stored energy based on the time-series prediction results and a preset photovoltaic priority consumption and storage strategy includes:

[0012] Based on the time-series prediction results, determine whether there will be peak energy consumption events in each region within a preset time period in the future;

[0013] If a peak energy consumption event is determined to occur in any region, energy storage status parameters are collected; when the energy storage status parameters are lower than a preset threshold, pre-stored energy is stored according to a preset photovoltaic priority consumption tiered pre-storage strategy.

[0014] In some embodiments, the heating of each region based on the preheating start-up time and the pre-stored energy includes:

[0015] Based on the time-series prediction results for each region, the peak energy consumption arrival time for each region is determined.

[0016] When the remaining time before the peak energy consumption time is less than or equal to the preheating start time, the pre-stored energy is used to heat each area respectively.

[0017] In some embodiments, the method of using the pre-stored energy to heat each area includes:

[0018] Based on the multi-source sensing data, the heating operation mode is determined;

[0019] Based on the aforementioned heating operation mode, the pre-stored energy is used to provide heating to each area.

[0020] In some embodiments, the step of using the pre-stored energy to provide heating to each area based on the heating operation mode includes:

[0021] Based on historical data and a preset multi-objective optimization function, control commands are generated.

[0022] Based on the heating operation mode and the control command, the pre-stored energy is used to heat each area.

[0023] In some embodiments, the multi-source sensing data includes external sensing data and internal operational data; the step of converting the multi-source sensing data into a multi-source fusion feature vector includes:

[0024] The external sensing data and the internal operating data are converted into external feature vectors and internal feature vectors, respectively.

[0025] The external feature vector and the internal feature vector are concatenated and linearly mapped to generate the multi-source fusion feature vector.

[0026] In some embodiments, the external sensing data includes the building heating rate, initial outdoor temperature, and initial indoor temperature; the calculation of the preheating start-up time for each region based on the multi-source sensing data for each region includes:

[0027] Obtain the target temperature for each region;

[0028] Based on the building heating rate, indoor initial temperature, outdoor initial temperature, target temperature, and preset self-learning correction coefficient for each region, the preheating start-up time for each region is calculated.

[0029] In some embodiments, the time-series prediction results include a pedestrian flow prediction curve, a heat load prediction curve, and a photovoltaic power generation prediction curve.

[0030] Secondly, embodiments of this application provide a heating dispatching device, the device comprising:

[0031] The fusion module is used to acquire multi-source sensing data from multiple regions and convert the multi-source sensing data into a multi-source fusion feature vector.

[0032] The prediction module is used to input the multi-source fusion feature vector into a preset time-series prediction model to generate time-series prediction results for each region.

[0033] The pre-storage module is used to store pre-stored energy based on the time-series prediction results and the preset photovoltaic priority consumption pre-storage strategy;

[0034] The preheating module is used to calculate the preheating start-up time for each region based on the multi-source sensing data of each region.

[0035] The heating module is used to provide heating to each area based on the preheating start-up time and the pre-stored energy.

[0036] Thirdly, embodiments of this application provide a storage medium storing a computer program thereon, which, when executed by a processor, implements the heating scheduling method as described in the first aspect above.

[0037] Compared to related technologies, the heating scheduling method, apparatus, and storage medium provided in this application involve acquiring multi-source sensing data from multiple regions and converting it into a multi-source fusion feature vector. The multi-source fusion feature vector is then input into a preset time-series prediction model to generate time-series prediction results for each region. Based on the time-series prediction results and a preset photovoltaic priority consumption and storage strategy, pre-stored energy is collected. The preheating start-up time for each region is calculated based on the multi-source sensing data. Finally, heating is provided to each region based on its preheating start-up time and pre-stored energy. This solves the problem in related technologies of difficulty in ensuring heating and antifreeze effects across multiple regions, achieving intelligent allocation and adaptive control of heat across multiple regions.

[0038] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0039] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0040] Figure 1 This is a schematic diagram of the architecture of a heating dispatching system according to an embodiment of this application;

[0041] Figure 2 This is a flowchart of a heating scheduling method according to an embodiment of this application;

[0042] Figure 3 This is a flowchart of step S230;

[0043] Figure 4 This is a schematic diagram of the overall process of the heating scheduling method according to the embodiments of this application;

[0044] Figure 5 This is a schematic diagram illustrating the verification effect under extreme no-light conditions according to an embodiment of this application;

[0045] Figure 6 This is a structural block diagram of a heating dispatching device according to an embodiment of this application. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application. Furthermore, it is understood that although the efforts made in such a development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, modifications to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.

[0047] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0048] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “a,” “an,” “an,” “the,” and similar words used in this application do not indicate quantity limitation and may indicate singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. The terms “connected,” “linked,” “coupled,” and similar words used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. “Multiple” used in this application means two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. The terms “first,” “second,” “third,” etc., used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects.

[0049] Public toilets in high-altitude rural and remote areas (rural public toilets) generally face difficulties in winter frost protection, heating, and hot water supply, especially in regions such as Qinghai, Tibet, and Inner Mongolia, where extreme winter temperatures can reach below -20°C. Public toilets without heating facilities suffer from severe problems such as frozen and cracked pipes, poor toilet conditions, and even closure. Meanwhile, these areas have abundant solar resources, providing a natural advantage for developing off-grid integrated solar-energy storage systems. In the typical application scenario where schools and public toilets are located back-to-back, the two types of facilities have naturally complementary characteristics in terms of energy consumption time and demand: schools have high energy demand during weekdays and are largely idle during holidays / summer / winter breaks; while rural public toilets have frost protection and basic heating needs all day, with demand being relatively more prominent at night and on holidays. However, existing energy storage and heating system scheduling schemes still have many shortcomings. First, they lack scene perception capabilities and cannot automatically identify the school's operational status. Fixed scheduling strategies can easily lead to energy waste. Second, the accuracy of load forecasting is insufficient, making it difficult to fit the nonlinear relationship between multi-dimensional parameters, which can easily lead to problems such as curtailment of solar power or insufficient heating. Third, cross-scene collaborative scheduling has not been achieved. The energy systems of schools and public toilets are independent of each other, resulting in low overall energy utilization efficiency. Fourth, it is difficult to balance operational resilience and construction economy. Large-capacity energy storage configurations increase initial investment, while reducing energy storage capacity can lead to antifreeze failure under extreme operating conditions. Fifth, the preheating timing control method is crude, relying solely on timed or fixed temperature thresholds to start the equipment. It cannot dynamically adjust based on the environment, building, and pedestrian flow, resulting in both energy loss and difficulty in ensuring user comfort.

[0050] To address the shortcomings of the existing technology, this application proposes a heating dispatching system applicable to back-to-back scenarios in multiple regions. Figure 1 This is a schematic diagram of the architecture of a heating dispatching system according to an embodiment of this application. Taking a scenario where a school and a rural public toilet are back-to-back as an example, the system is divided into three layers: a field perception layer, a DC microgrid main controller layer, and a cloud management layer. The field perception layer consists of a data acquisition and protection module and multi-source sensors deployed in the school area, and a power distribution and protection module, rural public toilet monitoring sensors, and indoor environmental detection module deployed in the rural public toilet area. It is responsible for the underlying multi-source data acquisition and millisecond-level local electrical protection. The DC microgrid main controller layer is deployed in the field computer room and serves as the local control hub of the system. On the one hand, it receives dispatching instructions from the cloud management layer and accurately controls the power and direction of the cross-scenario transmission lines to achieve smooth transmission of surplus power from the school to the public toilet. In extreme cases, it can provide reverse backup power. On the other hand, it aggregates the operating data of all field devices and connects to the cloud management layer, while also receiving remote dispatching instructions to ensure the real-time performance and stability of the system. The cloud-based management layer consists of an energy management system, a cloud platform and remote monitoring, a local control panel and a human-machine interaction module. Deployed on a cloud server, it is responsible for non-real-time tasks such as historical data storage, big data analysis, unified management of multiple sites, remote parameter setting and report generation. It also accesses meteorological data and external information to provide data support for system scheduling.

[0051] In terms of hardware layout, the DC bus power supply side of the school area connects to the DC power output from the photovoltaic array module via the combiner and protection module, and the energy storage power from the energy storage battery module. The power consumption side of the DC bus connects to the air source heat pump module, solid thermal storage module, lighting socket power consumption module, and data acquisition and protection module, respectively driving the indoor heating system, thermal storage medium circulation system, on-campus electrical loads, and connecting to on-campus multi-source sensors, achieving priority guarantee for on-campus heating and power consumption, and on-site storage of surplus energy. The power distribution and protection module in the rural public toilet area receives surplus power from the school via cross-scene transmission lines. On the one hand, it powers the phase change thermal storage module and electric heat tracing module to form an anti-freeze heating system; on the other hand, it powers the public toilet ventilation module and public toilet electrical load module. At the same time, it collects public toilet environment and equipment operation data through indoor environmental detection module and rural public toilet monitoring sensors to support anti-freeze and heating control.

[0052] During system operation, the energy management system dynamically allocates energy flow based on real-time photovoltaic output, battery state of charge (SOC), thermal storage temperature, indoor temperature, and meteorological data. It determines whether photovoltaic power should be prioritized for supplying on-campus loads, charging batteries, storing thermal energy, or transmitting it to public restrooms via cross-scenario transmission lines. Simultaneously, it receives fault signals from the acquisition and protection modules and quickly performs fault isolation, system restart, and backup power switching. This system, through a three-layer architecture, is adapted to back-to-back school and public restroom layouts in cold regions, achieving cross-scenario energy sharing, layered energy storage management, and refined anti-freeze heating control, while balancing system stability, economy, and anti-freeze resilience.

[0053] Based on the aforementioned heating dispatching system, this embodiment provides a heating dispatching method. Figure 2 This is a flowchart of a heating scheduling method according to an embodiment of this application, such as... Figure 2 As shown, the process includes the following steps:

[0054] Step S210: Acquire multi-source sensing data from multiple regions and convert the multi-source sensing data into multi-source fusion feature vectors;

[0055] Specifically, multi-source sensing data is collected simultaneously from two major areas: schools and rural public toilets. The data types include time status labels, meteorological environmental parameters, pedestrian traffic and heating / electricity load, equipment operating parameters, and system configuration parameters. The collected raw data undergoes preprocessing such as outlier removal and normalization. The four-dimensional time labels are converted into embedded vectors, and various feature data are then concatenated and linearly mapped to integrate and convert them into a unified format of multi-source fusion feature vectors, which serve as the input basis for the intelligent prediction model.

[0056] Step S220: Input the multi-source fusion feature vector into the preset time series prediction model to generate time series prediction results for each region;

[0057] Specifically, the preprocessed multi-source fusion feature vector is input into a pre-deployed and parameter-configured temporal Transformer prediction model. This model employs a three-layer stacked encoder structure and is configured with parameters such as multi-head attention mechanism and adaptive learning rate. The model performs multi-head self-attention calculation and position-by-position feedforward network operation on the input feature vector, and completes deep feature mining by combining residual connections and layer normalization. Based on the multi-dimensional feature information of the input, the model continuously extrapolates and generates time-series prediction results for two major areas—schools and rural public toilets—over the next 72 hours, specifically including regional pedestrian flow prediction curves, heat load prediction curves, and photovoltaic power generation prediction curves. Simultaneously, the system introduces a real-time data feedback loop, feeding back subsequently collected on-site operational data (such as pedestrian flow data) to the temporal prediction model. Through an online self-learning mechanism, the model weight parameters are dynamically updated to continuously optimize the model's prediction accuracy, ensuring that the photovoltaic power generation prediction error, pedestrian flow prediction accuracy, and heat load correction error are all maintained within the set acceptable range.

[0058] Step S230: Store pre-stored energy based on the time-series forecast results and the preset photovoltaic priority consumption and storage strategy;

[0059] Specifically, based on the time-series forecasts of pedestrian traffic, heat load, and photovoltaic power generation in each area, combined with real-time parameters such as battery state of charge and thermal storage temperature, a photovoltaic priority consumption and pre-storage strategy is triggered and executed. This strategy follows a fixed power allocation priority, first meeting the real-time electricity and heating needs on site, and then utilizing surplus photovoltaic power to charge and store energy for the energy storage batteries and thermal storage devices, thus achieving energy pre-storage. The system adopts a zoned management mode for the batteries and thermal storage devices, dividing them into normal operation zones and resilience reserve zones. While maximizing the consumption of photovoltaic power generation and reducing energy waste, it also reserves sufficient energy for special operating conditions such as continuous no sunlight and extreme low temperatures, balancing energy utilization efficiency and system operational reliability.

[0060] Step S240: Calculate the preheating start-up time for each region based on the multi-source sensing data of each region.

[0061] Specifically, real-time multi-source sensing data from various areas is retrieved to obtain environmental parameters such as initial indoor temperature, target comfort temperature, and outdoor temperature. Combined with the average building heating rate and benchmark outdoor temperature calibrated on-site, the preheating start-up time for school areas and rural public toilet areas is calculated to determine the optimal preheating trigger node.

[0062] Step S250: Based on the preheating start-up time and pre-stored energy of each area, heat is supplied to each area respectively.

[0063] Specifically, based on the preheating start-up time calculated for each area and the pre-stored energy in the batteries and thermal storage devices, heating and antifreeze control are implemented in different areas. The system uses the user arrival time as the judgment node, and starts heating after the preheating trigger condition is met. It prioritizes the use of previously stored heat energy for heating, and supplements the heat with air source heat pumps when the heat energy is insufficient. In extreme low-temperature environments, electric heat tracing is used as a backup antifreeze measure. The system dynamically adjusts the output power of various heating devices according to the measured indoor temperature, controls the indoor environment according to the preset temperature range, ensures user comfort during the day, and strictly adheres to the antifreeze temperature limit at night. With the help of pre-stored electrical energy, heat energy, and adaptive preheating strategies, continuous and stable heating and pipeline antifreeze are achieved in each area, giving full play to the peak-shaving and emergency guarantee role of pre-stored energy.

[0064] In the above steps, by sequentially completing a series of processes such as multi-source sensing data acquisition and feature fusion, high-precision time-series prediction based on time-series models, priority photovoltaic consumption and hierarchical and zoned energy pre-storage, adaptive calculation of preheating start-up time for each region, and regional coordinated heating and anti-freezing, different operating scenarios can be automatically identified and energy allocation strategies can be dynamically adjusted. This effectively captures the nonlinear correlation between multiple parameters to improve prediction accuracy, achieving energy complementarity and mutual assistance between schools and rural public toilets. While significantly reducing the photovoltaic curtailment rate and improving overall energy utilization efficiency, the hierarchical energy storage mechanism balances the economics of system construction with the resilience of anti-freezing operation under extreme low temperatures and continuous no-light conditions. Combined with adaptive preheating control, the indoor temperature is precisely regulated, which not only improves the usage environment of rural public toilets in high-altitude and cold regions and eliminates the problem of pipe freezing and cracking, but also reduces the workload of manual operation and maintenance. The overall solution has strong practicality and scalability.

[0065] In some of these embodiments, such as Figure 3 As shown, step S230 stores pre-stored energy based on the time-series forecast results and the preset photovoltaic priority consumption and storage strategy, including the following steps:

[0066] Step S231: Based on the time series prediction results, determine whether there will be a peak energy consumption event in each region within the future preset time.

[0067] Step S232: If it is determined that there is a peak energy consumption event in any area, the energy storage status parameters are collected; when the energy storage status parameters are lower than the preset threshold, the pre-stored energy is stored in accordance with the preset photovoltaic priority consumption hierarchical pre-storage strategy.

[0068] Specifically, based on the generated time-series predictions of pedestrian traffic, heat load, and photovoltaic power generation, the system analyzes two major areas—schools and rural public toilets—to determine whether various energy consumption peaks will occur within a preset time period. Once an energy consumption peak is detected in either area, the system collects real-time energy storage status parameters such as battery state of charge and thermal storage temperature, comparing them to preset thresholds. When the energy storage status parameters fall below the set thresholds, the system activates a photovoltaic priority consumption and tiered pre-storage strategy, consuming photovoltaic power in a tiered manner according to the order of "real-time load priority, battery charging second, and thermal storage charging last," completing the pre-storage of electrical and thermal energy. Simultaneously, the system manages the operating status of energy storage devices by region, reserving emergency energy to cope with special operating conditions such as continuous lack of sunlight and extreme cold, while enhancing photovoltaic consumption capacity.

[0069] In the above steps, by first predicting peak energy demand based on time-series forecasts and then triggering a tiered pre-storage linkage mechanism based on the real-time status of energy storage, the timing of energy storage can be accurately grasped, avoiding energy loss and redundant equipment operation caused by blind energy storage. According to the allocation rules of photovoltaic tiered consumption, the utilization of local photovoltaic clean energy can be maximized, effectively reducing the annual curtailment rate of the system and improving the overall energy utilization efficiency. At the same time, with the help of the energy storage zoning management mode, without increasing the capacity of energy storage equipment or the initial investment in the control system, the pre-stored energy can be fully utilized to form an emergency reserve, effectively balancing the economic efficiency of system operation and the anti-freezing safety and resilience under extreme conditions, ensuring that schools and rural public toilets in high-altitude and cold regions can still operate stably in environments without sunlight and in extreme cold.

[0070] In some embodiments, heating is provided to each region based on its preheating start-up time and pre-stored energy, including:

[0071] Based on the time-series forecast results for each region, the peak energy consumption arrival time for each region is determined;

[0072] When the remaining time before the peak energy demand arrives is less than or equal to the preheating start-up time, the pre-stored energy is used to heat each area separately.

[0073] Specifically, based on the time-series forecasts of pedestrian traffic and heat load in each area, the peak energy consumption times for schools and rural public toilets are determined. The remaining time before the peak energy consumption time is calculated in real time. When this remaining time is less than or equal to the preheating start-up time for the corresponding area, heating is officially started. The system prioritizes using pre-stored heat energy from thermal storage devices for heating. When heat is insufficient, air-source heat pumps provide supplementary heating. In extreme low-temperature environments, electric heat tracing is used as a backup measure against freezing. Simultaneously, the output power of each heating device is dynamically adjusted based on the measured indoor temperature, and the indoor environment is controlled according to preset temperature ranges. While ensuring user comfort, the system relies on pre-stored energy to prevent pipes in high-altitude and cold regions from freezing and cracking.

[0074] In the above steps, the peak energy consumption time of each area is accurately determined based on the time-series prediction results. The heating trigger node is intelligently determined by combining the calculated preheating start-up time. At the appropriate time, the pre-stored energy is called to carry out the heating operation. This abandons the traditional extensive control mode of timed start-stop and fixed threshold triggering. It can dynamically complete the preheating according to the flow of people and the ambient temperature. This not only effectively reduces ineffective energy consumption and improves energy utilization efficiency, but also ensures that the indoor temperature is adjusted to a comfortable range before users use it. At the same time, with the help of the layered pre-stored energy reserves and the multi-equipment collaborative heating mechanism, the indoor temperature is continuously maintained to meet the standard, and the problem of pipe freezing and cracking in low temperature environment is strictly prevented. It takes into account the heating comfort, energy saving and system anti-freeze reliability.

[0075] In some embodiments, pre-stored energy is used to heat each area separately, including:

[0076] Based on multi-source sensing data, determine the heating operation mode;

[0077] Based on the heating operation model, pre-stored energy is used to provide heating to each area.

[0078] Specifically, based on the four-dimensional time status labels in the multi-source sensing data and combined with preset judgment rules, the system automatically identifies and switches the corresponding heating operation mode, which is divided into four categories: weekday mode, weekend / holiday mode, winter / summer vacation mode, and extreme low temperature mode. After determining the current heating operation mode, the system calls on the pre-stored energy such as electrical energy and heat energy stored in the early stage to carry out heating operations in different areas according to the energy allocation, power scheduling, and temperature control rules corresponding to each mode.

[0079] During weekdays, priority is given to ensuring heating needs in school areas. Surplus photovoltaic energy and pre-stored energy are supplied to school-supported public toilets first, and the remaining energy is then transmitted to rural public toilets. During weekends and holidays, the energy priority for schools is lowered, and most of the pre-stored energy is transmitted to rural public toilets to meet their all-weather anti-freezing and heating needs. During winter and summer vacations, schools only maintain a minimum operating load, and all pre-stored energy is concentrated on heat storage and anti-freezing heating for rural public toilets. When the outdoor temperature reaches the extreme cold standard, the system switches to extreme low temperature mode, prioritizing the use of pre-stored energy in the energy storage resilience reserve area to fully protect the pipes of rural public toilets from freezing, while the room temperature control standards in school areas can be appropriately relaxed. During the heating process, the energy usage sequence is always followed: priority is given to heat storage and release, supplemented by air source heat pumps, and electric heat tracing is used as a backup. The output power of each device is dynamically adjusted to maintain the temperature of public toilets in schools at the first preset temperature (e.g., 10℃~12℃) and the temperature of public toilets in rural areas at the second preset temperature (e.g., 8℃~10℃). The indoor temperature throughout the area at night is not lower than the 6℃ antifreeze threshold. The pre-stored energy is precisely allocated and efficiently utilized according to different operating modes.

[0080] The above steps automatically identify the current heating operation mode through multi-source sensing data, and allocate pre-stored energy differently according to four different modes: weekdays, weekends and holidays, winter and summer vacations, and extreme low temperatures. This can dynamically adjust the energy allocation priority according to the energy consumption characteristics of schools and rural public toilets at different times, breaking the limitations of traditional fixed scheduling strategies, giving full play to the complementary advantages of dual-scenario energy use, and maximizing the activation of pre-stored electrical and thermal energy. At the same time, under extreme low temperature conditions, resilient reserve energy can be used first to ensure anti-freezing safety. This not only effectively improves the utilization efficiency of pre-stored energy and reduces energy loss, but also stably meets the heating and anti-freezing needs of different areas under various operating scenarios, reduces manual mode switching and operation and maintenance intervention, and further improves the overall intelligence level and environmental adaptability of the system.

[0081] In some embodiments, based on the heating operation mode, pre-stored energy is used to heat each area separately, including:

[0082] Based on historical data and a preset multi-objective optimization function, control commands are generated.

[0083] Based on the heating operation mode and control commands, pre-stored energy is used to provide heating to each area.

[0084] Specifically, historical operating data, equipment parameters, meteorological data, load data, and other information are retrieved and combined with a preset multi-objective weighted optimization function to perform calculations. The multi-objective optimization function adopted in this application aims to minimize the annualized total operating cost and the photovoltaic curtailment rate. The specific formula is as follows: Where ω1 is the annualized cost weight and ω2 is the curtailment rate weight; This represents the system's annualized total operating cost. The calculation process strictly adheres to multiple constraints, including power balance, thermal balance, equipment safety, indoor thermal comfort, system operational resilience, and cross-scenario energy transfer. Employing the Pareto multi-objective optimization method, it iterates through annualized cost and curtailment rate data corresponding to different weight combinations. The optimal knee parameter combination with the best overall performance is selected from the Pareto front to solve for the optimal scheduling scheme, generating six categories of control commands: electrochemical energy storage scheduling commands, thermal energy storage and release scheduling commands, heat pump operation commands, terminal heating commands, cross-scenario energy transfer commands, and system protection commands.

[0085] Based on the four identified heating operation modes—weekdays, weekends / holidays, winter and summer vacations, and extreme low temperatures—corresponding scheduling logic and control commands are matched to utilize pre-stored energy, such as electricity and heat, to carry out heating operations in different areas. Different energy allocation rules are implemented under different operation modes. Simultaneously, the power of heating equipment is adjusted in the order of priority for thermal storage and release, supplemented by air source heat pumps, and then as a backup for electric heat tracing. Based on control commands, precise management of energy storage charging and discharging, thermal storage charging and discharging, heat pump start-up and shutdown, and cross-regional power transmission is achieved. This maintains the indoor temperature of public toilets in schools and rural areas within the corresponding comfortable range, consistently adhering to the 6°C antifreeze lower limit at night, enabling refined and intelligent allocation of pre-stored energy under different operating conditions.

[0086] It should be noted that six categories of control commands are generated based on the multi-objective optimization scheduling results: 1) Electrochemical energy storage scheduling commands, used to control the battery charging and discharging power (positive value charging, negative value discharging), precisely adjust the battery charging and discharging rate, maintain the state of charge (SOC) in the optimal range, and realize the time shift of electrical energy; 2) Thermal energy storage and release scheduling commands, used to control the thermal storage charging power and the opening of the thermal storage and release valves, adjust the charging and releasing rates of solid carbon brick thermal storage and phase change thermal storage, and realize the time shift of thermal energy; 3) Heat pump operation commands, including heat pump start / stop signals and operating level / power control, adjusting the output of the air source heat pump. 4) Terminal heating command: By controlling the start / stop signal of electric heat tracing, the operating power and the opening degree of indoor heating valves, the output of the terminal heating equipment is directly controlled to achieve precise adjustment of indoor temperature; 5) Cross-scenario energy transfer command: Used to control the power and direction of bidirectional transmission lines to realize the transmission of surplus power from the school side to the public toilet side, achieving energy mutual assistance between the two scenarios; 6) System protection command: Includes fault isolation signal, equipment start / stop signal and alarm signal. When equipment fault or parameter over-limit is detected, protection action is quickly executed to ensure the overall safety of the system.

[0087] The above steps involve retrieving historical operating data, equipment parameters, meteorological data, and load data. A multi-objective weighted optimization function is constructed and solved, targeting annualized operating costs and photovoltaic curtailment rates. Under multiple constraints, six categories of control commands are generated: electrochemical energy storage scheduling, thermal energy storage and release scheduling, heat pump operation, terminal heating, cross-scenario energy transfer, and system protection. These commands are then combined with four heating operation modes—weekdays, weekends / holidays, summer / winter breaks, and extreme low temperatures—and matched with differentiated energy allocation rules. Pre-stored energy is called upon and coordinated to regulate the operation of various equipment in the order of "priority to thermal storage and release, supplementation by air source heat pumps, and backup by electric heat tracing." This not only achieves time-shifting of electrical and thermal energy, cross-scenario energy mutual assistance, and system safety protection, effectively reducing annualized operating costs and photovoltaic curtailment rates, but also precisely maintains stable indoor temperatures in various areas under different operating conditions and strictly adheres to the minimum anti-freezing temperature. This achieves refined and intelligent allocation of pre-stored energy, comprehensively improving the system's economic efficiency, energy utilization efficiency, heating comfort, and operational resilience under extreme conditions.

[0088] In some embodiments, the multi-source sensing data includes external sensing data and internal operational data; converting the multi-source sensing data into a multi-source fusion feature vector includes:

[0089] Externally perceived data and internally operational data are converted into external feature vectors and internal feature vectors, respectively.

[0090] The external and internal feature vectors are concatenated and linearly mapped to generate a multi-source fusion feature vector.

[0091] External sensing data mainly includes time data labels, environmental and meteorological data, and pedestrian and load data; internal operational data mainly includes equipment status data and system configuration data. In the process of converting multi-source sensing data into multi-source fusion feature vectors, feature transformation processing is first performed on both types of data separately.

[0092] For the feature transformation of externally perceived data: time data labels (including intraday time periods, weekdays, years / months, seasons, etc.) are converted into embedded vectors; environmental and meteorological data (including outdoor temperature, solar irradiance, photovoltaic power generation prediction, etc.) are converted into environmental feature vectors after normalization; and pedestrian and load data (including controlled pedestrian flow in schools / public toilets, historical usage rates, cold water usage rates, etc.) are converted into user behavior feature vectors through the collection results of infrared sensors and water metering devices, after statistical and normalization processing. The above three types of external data are processed uniformly and then concatenated into a complete external feature vector.

[0093] For the feature transformation of internal operating data: For equipment status data (including battery SOC, thermal storage temperature, water pump outlet water temperature, indoor temperature, electric heat tracing power, etc.), it is collected in real time by various equipment sensors, sampled at an update frequency of 1 minute / time, and after outlier removal and unit unification, it is converted into equipment status feature vector; For system configuration data (including photovoltaic capacity, battery capacity / efficiency, thermal storage capacity / efficiency, water pump COP curve, building heat transfer coefficient, etc.), based on design capacity and on-site calibration results, it is used as static configuration parameters to participate in feature construction and converted into system configuration feature vector; After the above two types of internal data are processed in a unified manner, they are spliced ​​into a complete internal feature vector.

[0094] After constructing the external and internal feature vectors, the system concatenates the two types of feature vectors and then performs feature dimension integration and feature space alignment through linear mapping operations. Finally, it generates a multi-source fusion feature vector with unified dimensions and complete information, providing standardized input data for subsequent time series prediction models.

[0095] In the above steps, by dividing multi-source sensing data into external sensing data and internal operational data and constructing feature vectors for each, and then generating a unified multi-source fusion feature vector through vector concatenation and linear mapping, it is possible to completely and systematically integrate multi-dimensional information such as time, weather, population flow, equipment status and system configuration. This enables the standardized and structured expression of different types of data, eliminates the influence of differences in data format, dimensions and distribution on subsequent model inputs, and provides high-quality input data with complete information and uniform format for time series prediction models. This effectively ensures the prediction accuracy and stability of the model, and thus provides reliable data support for subsequent peak energy consumption prediction, energy pre-storage and refined heating scheduling.

[0096] In some embodiments, the external sensing data includes the building heating rate, the initial outdoor temperature, and the initial indoor temperature; based on the multi-source sensing data for each region, the preheating start-up time for each region is calculated, including:

[0097] Obtain the target temperature for each region;

[0098] Based on the building heating rate, indoor initial temperature, outdoor initial temperature, target temperature, and preset self-learning correction coefficient for each region, the preheating start-up time for each region is calculated.

[0099] First, the system obtains the target temperature of each area based on the current scene. For example, under daytime use conditions, the target indoor comfort temperature for school-supported public toilets can be set at 10℃~12℃, and for rural public toilets it can be set at 8℃~10℃;

[0100] Subsequently, the system invokes the adaptive preheating start-up time calculation formula, and calculates the preheating start-up time M for each region by combining the building heating rate, initial indoor temperature, outdoor temperature, target temperature, and preset self-learning correction coefficient. The calculation formula is as follows: ;

[0101] Where M is the preheating start-up time, in minutes; The self-learning correction coefficient for the heating process is calibrated through initial experiments and then updated online during operation. The correction factor for the influence of outdoor temperature is updated online through self-learning after initial experimental calibration. The target indoor comfort temperature is expressed in °C and represents preset parameters for different scenarios. The initial indoor temperature is in °C and is collected in real time by an indoor temperature sensor. The initial outdoor temperature is in °C and is collected in real time by an outdoor weather station. The net average temperature rise rate of the building is expressed in °C / minute and is calibrated through on-site experiments and periodically recalibrated. The base outdoor temperature, in °C, is determined by combining local climate statistics with experimental calibration. The first part of the formula... The base preheating time is obtained by dividing the target temperature difference by the building heating rate, and is corrected by a self-learning correction coefficient. Compensation will be provided; Part Two This is an outdoor temperature correction item. When the outdoor temperature is lower than the reference temperature, this item is positive, and the preheating time will be increased accordingly to compensate for the building's heat loss. The system calculates the preheating start time independently for each area, providing accurate timing data for subsequent advanced heating control.

[0102] The above steps combine the building heating rate, indoor initial temperature, outdoor initial temperature, target temperature, and self-learning correction coefficient, and use an adaptive formula to calculate the preheating start-up time. This fully considers the differences in thermal characteristics and environmental influences in different areas, and dynamically compensates for calculation errors caused by equipment output deviations and outdoor temperature fluctuations through the self-learning correction coefficient. It can accurately and adaptively determine the heating start-up node for each area, avoiding energy waste caused by premature preheating, and ensuring that the indoor temperature is raised to the target range before the peak energy consumption of users. This achieves refined and intelligent advanced preheating control, effectively reducing system energy consumption while ensuring heating comfort, and improving overall operating efficiency and adaptability under different operating conditions.

[0103] In some embodiments, the time-series prediction results include a pedestrian flow prediction curve, a heat load prediction curve, and a photovoltaic power generation prediction curve.

[0104] Among them, the pedestrian flow prediction curve is used to predict future changes in user traffic in each area; the heat load prediction curve, based on pedestrian flow and environmental data, predicts the heating demand trend in each area; and the photovoltaic power generation prediction curve reflects the future output changes of the photovoltaic array. All three types of curves cover a certain scheduling period in the future, providing time-series data for the system's energy scheduling, pre-storage, and heating control.

[0105] The above embodiments, by generating and outputting three types of time-series prediction curves for pedestrian flow, heat load, and photovoltaic power generation, enable the system to predict future changes in user flow, heating demand trends, and photovoltaic output in advance. This provides complete time-series data support for subsequent multi-objective optimization scheduling, energy pre-storage, and advanced heating control, thereby achieving accurate prediction of peak energy consumption, efficient consumption of photovoltaic power, and advance planning of pre-stored energy. This not only reduces ineffective energy consumption and photovoltaic curtailment rate but also ensures stable indoor heating during peak user periods, improving the system's forward-looking capabilities, economy, and heating reliability.

[0106] Figure 4 This is a schematic diagram of the overall process of the heating scheduling method according to an embodiment of this application. First, time stamps and sensor data are input, and the population flow, heat load, and photovoltaic power generation for the next 72 hours are predicted through a time-series prediction model. If a peak event is predicted to occur within the next N hours, the system enters the pre-storage stage, where it is determined whether the battery charge or heat storage temperature is below a threshold. If the condition is met, a hierarchical pre-storage strategy and photovoltaic priority scheduling are activated. Then, the system enters the preheating stage, where the preheating duration M is calculated according to an adaptive formula, and it is determined whether the interval between the current time and the user's arrival time is less than or equal to M. If the condition is met, the heat pump, heat storage release, and indoor temperature closed-loop control are activated. Finally, six types of scheduling instructions are output, forming a control process of prediction, pre-storage, and preheating.

[0107] The following example, using a typical back-to-back scenario of a school and a rural public toilet in a cold northern region, illustrates the implementation steps and effects of this invention in detail. The specific system configuration in this embodiment is as follows: the school side is equipped with a 9.6kWp photovoltaic array, a 10kWh lithium iron phosphate battery, and a 3kW air source heat pump. This air source heat pump has a COP of 2.3 at an ambient temperature of -20℃, and is also equipped with a 36kWh solid carbon brick thermal storage device. The rural public toilet side is equipped with a mirabilite phase change thermal storage device and a 0.75kW electric heat tracing device. The cross-scenario transmission cable connecting the two areas has a rated power of 5kW and a line transmission efficiency of 0.95. The building heat transfer coefficient is differentiated by day and night conditions: 0.8kW / ℃ during the day and 0.6kW / ℃ at night, with a discount rate of 6%. If the traditional timed scheduling scheme is adopted, in order to meet the anti-freezing operation requirements of 16 consecutive hours of extreme no light conditions, the system needs to be equipped with 18kWh batteries and 60kWh thermal storage equipment. The initial investment of the equipment is 40% higher than that of this scheme. In addition, the surplus energy generated by the school during the holidays cannot be transferred to rural public toilets for utilization, and the annual photovoltaic curtailment rate can reach 18%. The fixed-time start-up preheating operation mode will also cause the overall COP of the system to be low.

[0108] The heating dispatching method proposed in this application can meet the antifreeze resilience requirements under extreme conditions without the need for additional energy storage capacity. It can also effectively improve energy utilization and overall system operating efficiency. The verification effect of this embodiment under extreme conditions of 16 consecutive hours without light can be found in [reference needed]. Figure 5 The reliability of the proposed scheduling strategy was comprehensively verified by analyzing the changes in three curves: battery state of charge, heat storage, and indoor temperature. The upper curve shows that the battery state of charge gradually decreased from an initial 75% to approximately 20%, remaining within the resilience reserve range of 5% to 75% throughout, without reaching the emergency threshold, and the battery maintained a certain power supply redundancy. The middle curve shows that the heat storage gradually decreased from an initial approximately 500 kWh to approximately 120 kWh, consistently remaining above the minimum anti-freeze threshold, indicating sufficient heat storage reserves to support heating needs throughout the day. The lower curve shows that the indoor temperatures in the school area and rural public toilets remained stable at 9–10°C and 7–8°C respectively, both consistently above the 6°C anti-freeze baseline. This demonstrates that even under extreme conditions with no photovoltaic output, the system can effectively maintain indoor temperatures above the safe lower limit, significantly improving its anti-freeze operational resilience.

[0109] In this embodiment, the heating scheduling is executed sequentially according to a predetermined process. First, the system collects 18 parameters in 5 major categories from the signal input layer, including 4D time labels, hourly outdoor temperature, solar irradiance, photovoltaic power generation, pedestrian flow in both schools and rural public toilets, hot water usage, battery SOC, thermal storage temperature, indoor temperature, and system configuration parameters. After parameter collection, preprocessing is performed on all multi-source parameters, including outlier removal and data normalization. Simultaneously, the 4D time labels are converted into embedded vectors, which are then combined with meteorological features, pedestrian flow sequences, and equipment status vectors for feature concatenation and linear mapping. The processed feature vectors are then input into a time-series Transformer encoder consisting of three stacked layers. Calculations are performed using a multi-head self-attention mechanism and a position-by-position feedforward network to output hourly updated pedestrian flow prediction curves, heat load prediction curves, and photovoltaic power generation prediction curves for the next 72 hours. The system will send the actual operating measurement values ​​of the equipment back to the prediction model, and continuously update the model weights through an adaptive learning loop to achieve online self-learning of the model. Ultimately, it will ensure that the photovoltaic prediction error is no more than 8%, the traffic flow prediction accuracy is no less than 92%, and the heat load correction error is no more than 5%.

[0110] The system combines forecast results to assess operating conditions, identifying potential peak periods for pedestrian traffic or photovoltaic (PV) power generation within the next N hours. Upon detection of such events, the system enters a pre-storage scheduling phase. If no such events are detected, the current operating strategy is maintained, and continuous monitoring continues. During this process, the system continuously checks whether the battery SOC is below 75% or the thermal storage temperature is below 440°C. If either condition is met, the system activates a photovoltaic priority consumption and tiered pre-storage strategy, allocating PV output power according to set priorities. Priority is given to real-time electrical load and heat pump power needs, followed by using surplus power for battery charging, and the remaining power is supplied to the thermal storage for heating. Only a very small amount of remaining power is counted as curtailment. The system employs tiered range rules to manage battery SOC and thermal storage temperature. Under normal operating conditions, the battery SOC is maintained between 75% and 95%, and the thermal storage temperature is maintained between 110°C and 440°C. Simultaneously, 5% to 25% of the battery SOC and 50°C to 110°C of the thermal storage temperature are reserved as reserves for freeze protection under extreme conditions.

[0111] The system calculates the preheating start-up time using an adaptive M-value formula. Taking the actual operating conditions in this region in January as an example, when the outdoor temperature is -20℃ and the initial indoor temperature is 6℃, the calculated preheating start-up time M is approximately 90 minutes; when the outdoor temperature is -5℃ and the initial indoor temperature is 9℃, the calculated preheating start-up time M is approximately 25 minutes. The system further determines whether the time remaining until the arrival of the first batch of users is less than or equal to the calculated preheating start-up time M. If the condition is met, the preheating scheduling mode is activated. During operation, the combined heat pump and thermal storage equipment provide heating, following the principle of prioritizing thermal storage heat release and supplementing with air source heat pumps. In extreme low-temperature environments, electric heat tracing equipment is simultaneously activated for auxiliary heating. The system monitors the indoor temperature in real time throughout the process, dynamically adjusting the power distribution of the heat pump, thermal storage heat release, and electric heat tracing to stably control the indoor temperature of school public toilets between 10℃ and 12℃, and the indoor temperature of rural public toilets between 8℃ and 10℃, while ensuring that the indoor temperature at night never falls below the frost protection limit of 6℃.

[0112] Based on a pre-defined multi-objective weighted optimization objective function, the system solves for the optimal scheduling strategy under multiple constraints, including power balance, thermal balance, equipment safety, thermal comfort, system resilience, and cross-scenario transmission, and generates six categories of control commands accordingly. Finally, the system sends these scheduling commands to the hardware execution units on both sides of the school and the rural public toilet, completing the entire control process for the current scheduling cycle.

[0113] To meet the application requirements of this scenario, the staff meticulously tuned the relevant parameters one by one. First, considering the requirements for prediction accuracy and data transmission latency, the parameters of the time-series Transformer model were determined. A 3-layer encoder and an 8-head attention mechanism were selected, with the model learning rate set to 1e-4 and the prediction time domain set to 72 hours. Based on the local extreme anti-freezing requirements under conditions of 16 consecutive hours of no light, the parameters for the layered pre-storage strategy were defined: the normal operating SOC range for the battery was 75%~95%, and the resilience reserve SOC range was 5%~75%; the normal heat release temperature range for the thermal storage device was 110℃~440℃, and the anti-freezing reserve temperature range was 50℃~110℃. Combining the building's heat transfer coefficient and historical indoor temperature rise measurement data, the self-learning correction coefficients in the adaptive preheating time calculation formula were tuned, with α=1.2 and β=0.8. Dynamic parameter optimization was then completed using the M-value sensitivity heatmap. Based on the project's design objective of prioritizing economic efficiency while also considering photovoltaic absorption capacity, the weight coefficients of the multi-objective optimization function are determined, with annualized operating cost having a weight of 0.7 and curtailment rate having a weight of 0.3.

[0114] Based on the Pareto multi-objective optimization front, the annualized cost and curtailment rate data corresponding to different weight combinations were traversed. Finally, the optimal parameter combination located at the knee of the Pareto front was selected. This combination corresponds to an annualized operating cost of 45,000 yuan / year and an annual photovoltaic curtailment rate of 5%, achieving the optimal balance between operational economy and photovoltaic absorption rate. Finally, a 16-hour continuous no-sunlight extreme condition simulation verification was conducted. During the operation, the battery SOC decreased from an initial 75% to 22%, and the thermal storage temperature decreased from an initial 420℃ to 310℃. Neither parameter reached the emergency lower limit of equipment operation, and the indoor temperature remained above 6℃ throughout, fully meeting the anti-freeze resilience requirements. This completed the tuning of all system control parameters.

[0115] This embodiment also provides a heating scheduling device for implementing the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the terms "module," "unit," "subunit," etc., can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0116] Figure 6 This is a structural block diagram of a heating dispatching device according to an embodiment of this application, such as... Figure 6 As shown, the device includes:

[0117] The fusion module 61 is used to acquire multi-source sensing data from multiple regions and convert the multi-source sensing data into multi-source fusion feature vectors.

[0118] Prediction module 62 is used to input multi-source fusion feature vectors into a preset time-series prediction model to generate time-series prediction results for each region;

[0119] Pre-storage module 63 is used to store pre-stored energy based on time-series forecast results and a preset photovoltaic priority consumption pre-storage strategy;

[0120] The preheating module 64 is used to calculate the preheating start-up time of each region based on the multi-source sensing data of each region.

[0121] Heating module 65 is used to provide heating to each area based on the preheating start-up time and pre-stored energy of each area.

[0122] It should be noted that the above modules can be functional modules or program modules, and can be implemented by software or hardware. For modules implemented by hardware, the above modules can reside in the same processor; or the above modules can be located in different processors in any combination. Specific examples in this embodiment can be found in the examples described in the above embodiments and optional implementations, and will not be repeated in this embodiment.

[0123] In addition, in conjunction with the heating scheduling methods in the above embodiments, this application embodiment can provide a storage medium for implementation. The storage medium stores a computer program; when the computer program is executed by a processor, it implements any of the heating scheduling methods in the above embodiments.

[0124] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0125] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0126] Those skilled in the art should understand that the technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments have been described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0127] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A heat supply scheduling method, characterized by, include: Acquire multi-source sensing data from multiple regions and convert the multi-source sensing data into a multi-source fusion feature vector; The multi-source fusion feature vector is input into a preset time-series prediction model to generate time-series prediction results for each region; Based on the time-series prediction results and the preset photovoltaic priority consumption and storage strategy, pre-stored energy is stored; Based on the multi-source sensing data of each region, the preheating start-up time of each region is calculated respectively; Heating is provided to each region based on the preheating start-up time and the pre-stored energy.

2. The heating dispatching method according to claim 1, characterized in that, The step of storing pre-stored energy based on the time-series prediction results and the preset photovoltaic priority consumption and storage strategy includes: Based on the time-series prediction results, determine whether there will be peak energy consumption events in each region within a preset time period in the future; If a peak energy consumption event is determined to occur in any region, energy storage status parameters are collected; when the energy storage status parameters are lower than a preset threshold, pre-stored energy is stored according to a preset photovoltaic priority consumption tiered pre-storage strategy.

3. The heating dispatching method according to claim 1, characterized in that, The heating of each region is based on the preheating start-up time and the pre-stored energy, including: Based on the time-series prediction results for each region, the peak energy consumption arrival time for each region is determined. When the remaining time before the peak energy consumption time is less than or equal to the preheating start time, the pre-stored energy is used to heat each area respectively.

4. The heating dispatching method according to claim 3, characterized in that, The method of using the pre-stored energy to heat each area includes: Based on the multi-source sensing data, the heating operation mode is determined; Based on the aforementioned heating operation mode, the pre-stored energy is used to provide heating to each area.

5. The heating dispatching method according to claim 4, characterized in that, The heating operation mode, based on the aforementioned method, utilizes the pre-stored energy to provide heating to each area, including: Based on historical data and a preset multi-objective optimization function, control commands are generated. Based on the heating operation mode and the control command, the pre-stored energy is used to heat each area.

6. The heating dispatching method according to claim 1, characterized in that, The multi-source sensing data includes external sensing data and internal operational data; the conversion of the multi-source sensing data into a multi-source fusion feature vector includes: The external sensing data and the internal operating data are converted into external feature vectors and internal feature vectors, respectively. The external feature vector and the internal feature vector are concatenated and linearly mapped to generate the multi-source fusion feature vector.

7. The heating dispatching method according to claim 6, characterized in that, The external sensing data includes the building heating rate, initial outdoor temperature, and initial indoor temperature; the calculation of the preheating start-up time for each region based on the multi-source sensing data for each region includes: Obtain the target temperature for each region; Based on the building heating rate, indoor initial temperature, outdoor initial temperature, target temperature, and preset self-learning correction coefficient for each region, the preheating start-up time for each region is calculated.

8. The heating dispatching method according to claim 1, characterized in that, The time-series prediction results include the pedestrian flow prediction curve, the heat load prediction curve, and the photovoltaic power generation prediction curve.

9. A heating dispatching device, characterized in that, The device includes: The fusion module is used to acquire multi-source sensing data from multiple regions and convert the multi-source sensing data into a multi-source fusion feature vector. The prediction module is used to input the multi-source fusion feature vector into a preset time-series prediction model to generate time-series prediction results for each region. The pre-storage module is used to store pre-stored energy based on the time-series prediction results and the preset photovoltaic priority consumption pre-storage strategy; The preheating module is used to calculate the preheating start-up time for each region based on the multi-source sensing data of each region. The heating module is used to provide heating to each area based on the preheating start-up time and the pre-stored energy.

10. A storage medium, characterized in that, The storage medium stores a computer program, wherein the computer program is configured to execute the heating scheduling method according to any one of claims 1 to 8 when it runs.

Images