Air energy heat pump control method and device based on narrowband internet of things, and medium

Abstract

This invention discloses a control method, device, and medium for air-source heat pumps based on narrowband Internet of Things (NIoT), relating to the field of heat pump control technology. The method includes: real-time acquisition of operating status data and environmental parameter data of the air-source heat pump unit to obtain a multi-dimensional sensor dataset; edge-side differential encoding compression to generate compressed data packets; transmission of the compressed data packets to a cloud control platform for decompression and reconstruction to obtain a real-time operating profile of the heat pump; load prediction to generate multiple load prediction parameters for control strategy optimization, generating a target control parameter set; and distribution of the target control parameter set to the heat pump controller to execute adaptive adjustment of the air-source heat pump. This invention solves the technical problems of high power consumption in heat pump sensor data transmission, coarse control strategies, and poor operating energy efficiency and stability in existing technologies, achieving low-power data transmission and intelligent cloud-based control of heat pumps, thus improving the operating energy efficiency and stability of the heat pump.

Description

Technical Field

[0001] This invention relates to the field of heat pump control technology, and more specifically to air source heat pump control methods, equipment and media based on narrowband Internet of Things. Background Technology

[0002] Existing air source heat pumps mostly use traditional communication methods to upload operating data, which has problems such as high power consumption, limited bandwidth, and low efficiency of massive sensor data transmission. In addition, they mostly rely on simple local control logic, which makes it difficult to combine environmental changes and unit operating status to achieve accurate load prediction and dynamic strategy optimization. This can easily lead to control lag, low operating energy efficiency, and insufficient operating stability, and cannot meet the actual application requirements of refined intelligent control of heat pumps in narrowband IoT low power consumption scenarios.

[0003] Existing heat pump technologies suffer from high power consumption in data transmission, crude control strategies, and poor operational energy efficiency and stability. Summary of the Invention

[0004] This application provides a control method, device, and medium for air source heat pumps based on narrowband Internet of Things, which is used to address the technical problems of high power consumption of heat pump sensor data transmission, coarse control strategies, and poor operating energy efficiency and stability in the prior art.

[0005] In view of the above problems, this application provides a method, device and medium for controlling air source heat pumps based on narrowband Internet of Things.

[0006] The first aspect of this application provides a control method for an air source heat pump based on narrowband Internet of Things (IoT), the method comprising: Real-time acquisition of operating status data and environmental parameter data of the air source heat pump unit yields a multi-dimensional sensor dataset. Edge-side differential encoding compression is performed on the multi-dimensional sensor dataset to generate a compressed data packet. This compressed data packet is transmitted to a cloud control platform via narrowband IoT for decompression and reconstruction, resulting in a real-time operating profile of the heat pump. Load forecasting is performed based on this real-time operating profile, generating multiple load forecasting parameters for control strategy optimization, resulting in a target control parameter set. This target control parameter set is then distributed to the heat pump controller via narrowband IoT to execute adaptive adjustment of the air source heat pump.

[0007] A second aspect of this application provides an electronic device comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the air source heat pump control method based on narrowband Internet of Things provided in this application.

[0008] In a third aspect of this application, a computer-readable storage medium is provided storing a computer program for executing the air source heat pump control method based on narrowband Internet of Things provided in this application.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages: Real-time acquisition of operating status data and environmental parameter data from air-source heat pump units yields a multi-dimensional sensor dataset. Edge-side differential encoding compression is performed to generate compressed data packets. These compressed data packets are transmitted to a cloud control platform via narrowband IoT for decompression and reconstruction, resulting in a real-time operating profile of the heat pump. Load forecasting is performed, generating multiple load forecast parameters for control strategy optimization, resulting in a target control parameter set. This target control parameter set is then sent to the heat pump controller to execute adaptive adjustments of the air-source heat pump. This achieves low-power data transmission and intelligent cloud-based control of the heat pump, improving its operational efficiency and stability. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 A schematic flowchart of an air source heat pump control method based on narrowband Internet of Things provided in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0012] Explanation of reference numerals in the attached drawings: Processor 21, Memory 22, Input device 23, Output device 24. Detailed Implementation

[0013] This application provides a method, device, and medium for controlling air source heat pumps based on narrowband Internet of Things, which addresses the technical problems of high power consumption in data transmission of heat pump sensors, coarse control strategies, and poor operating energy efficiency and stability in existing technologies.

[0014] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0015] Example 1, as Figure 1 As shown, this application provides an air source heat pump control method based on narrowband Internet of Things, the method comprising: Step S100: Collect real-time operating status data and environmental parameter data of the air source heat pump unit to obtain a multi-dimensional sensor dataset.

[0016] Specifically, the system collects real-time operating status data and environmental parameter data of the air source heat pump unit during operation through sensors. The operating status data includes core operating parameters of the unit such as intake pressure, exhaust pressure, intake temperature, and exhaust temperature, while the environmental parameter data includes external condition information such as outdoor ambient temperature. The above-mentioned multi-type and multi-dimensional collected data are synchronously integrated according to the time of collection to form a multi-dimensional sensor dataset for subsequent edge computing and cloud transmission.

[0017] Step S200: Perform edge-side differential coding compression based on the multidimensional sensor dataset to generate compressed data packets.

[0018] Specifically, differential coding compression is performed on the multidimensional sensor dataset at the edge. First, neighbor point analysis and first-order difference operation are performed on the multidimensional time series data to generate a multidimensional difference sequence. Then, a difference frequency histogram is constructed by calculating the difference frequency distribution and a coding mapping table is generated. At the same time, continuous sampling is performed according to the acquisition cycle to obtain sampled data frames. Inert channels and active channels are distinguished by inter-frame residual and amplitude calculation. Data frame compression coding verification is completed, and finally, compressed data packets that can be used for narrowband IoT transmission are generated.

[0019] Step S300: Transmit the compressed data packet to the cloud control platform via narrowband Internet of Things for decompression and reconstruction to obtain a real-time operation profile of the heat pump.

[0020] Specifically, compressed data packets generated at the edge are transmitted to the cloud control platform via narrowband IoT. The cloud first parses the data packets to obtain the frame header identifier and the encoding mapping version identifier, and then uses this to backtrack and decode to recover the multi-dimensional sensor dataset. Subsequently, the recovered data is time-synchronized and injected into the digital twin model of the air source heat pump. Through simulation calculation, difference comparison and recursive correction, the correction state vector is obtained. Then, multi-dimensional indicators and real-time energy efficiency ratio curves are calculated and structured and aggregated. Finally, a real-time operation profile of the heat pump that can comprehensively reflect the unit's operating conditions is reconstructed.

[0021] Step S400: Based on the real-time operation profile of the heat pump, perform load prediction, generate multiple load prediction parameters, optimize the control strategy, and generate a target control parameter set.

[0022] Specifically, based on the real-time operation profile of the heat pump, a BP neural network machine learning algorithm is used to complete load prediction and control strategy optimization. First, the corrected state vector after digital twin model correction, real-time energy efficiency ratio curve, multi-dimensional operation indicators, and historical time-series sensor data are used as input layer features, and load prediction parameters such as heating load, operating power consumption, and operating condition adaptation coefficient are used as output layer to construct a three-layer BP neural network structure. During training, massive amounts of historical heat pump operation data and corresponding load labels are input into the network. The output error is calculated through forward propagation, and then the gradient descent method is used to update the weights and biases from the input layer to the hidden layer and from the hidden layer to the output layer. The model training is completed iteratively until it converges below the error threshold. In the online inference stage, the real-time operation profile data is input into the trained BP network, which quickly outputs multiple load prediction parameters. Then, with the goal of optimal energy efficiency, optimization calculations are performed in the controllable parameter space such as compressor frequency, fan speed, and electronic expansion valve opening. Finally, a target control parameter set adapted to the current operating condition is generated.

[0023] Step S500: The target control parameter set is sent to the heat pump controller via narrowband Internet of Things to perform adaptive adjustment of the air source heat pump.

[0024] Specifically, the cloud-based control platform generates standard control command frames based on the target control parameter set. These command frames are then sent to the on-site heat pump controller via narrowband IoT for parsing, extracting key parameters such as the target outlet water temperature, the target compressor operating frequency, and the target fan speed. The heat pump controller adjusts the inverter output frequency based on the target compressor operating frequency to determine the target compressor speed, adjusts the electronic expansion valve opening based on the target outlet water temperature to match the target heating capacity, and adjusts the fan drive module based on the target fan speed to output the target air volume. Ultimately, these mechanisms work together to achieve adaptive operation and adjustment of the air source heat pump.

[0025] In one possible implementation, step S200 further includes: Step S210: Traverse the multidimensional sensor dataset to perform multidimensional time series analysis, extract multidimensional time series data for neighbor point analysis, and determine multiple neighboring sampling points.

[0026] Step S220: Perform first-order difference operation based on the multiple adjacent sampling points to generate a multidimensional difference sequence.

[0027] Step S230: Traverse the multidimensional difference sequence to calculate the frequency distribution and construct a difference frequency histogram.

[0028] Step S240: Encode based on the differential frequency histogram to obtain an encoding mapping table.

[0029] Step S250: Compress and encode the data frames according to the encoding mapping table to generate compressed data packets.

[0030] Specifically, the multidimensional sensor dataset is first traversed channel by channel along the time axis. Time alignment and timestamp normalization are performed on the sensor data of each dimension, such as inhalation pressure, exhaust pressure, inhalation temperature, exhaust temperature, and ambient temperature, to complete the multidimensional time series analysis. Then, with a fixed sampling period as the window, the sampling records of adjacent moments are retrieved sequentially for each dimension of continuous sensor data. By using time neighborhood matching and time sequence position association, the sampling values ​​of the previous moment and the current moment corresponding to each sampling point are extracted. This accurately determines multiple sets of adjacent sampling points used for difference calculation, ensuring the continuity of time series and the accuracy of sampling point pairing.

[0031] For each of the determined sets of adjacent sampling points, a first-order difference operation is performed dimension by dimension. The value of the current sampling point under the same sensing channel and the same time series is subtracted from the value of the adjacent sampling point at the previous moment. The single-dimensional difference value reflecting the instantaneous change amplitude of the data is calculated point by point. Then, the difference results of all dimensions such as intake pressure, exhaust pressure, intake temperature, exhaust temperature, and outdoor ambient temperature are aligned and integrated according to the same time axis to form a multi-dimensional difference sequence that can completely characterize the fluctuation characteristics of multi-dimensional sensing data, which is used for subsequent compression encoding processing.

[0032] The generated multidimensional difference sequence is traversed dimension by dimension. First, all difference results are divided into numerical intervals and classified for statistical analysis. Then, the frequency of each specific difference value in the entire sequence is counted to obtain the occurrence frequency of each difference value. The difference value is used as the horizontal axis and the corresponding occurrence frequency is used as the vertical axis. The difference values ​​are arranged in order of numerical size to complete the difference frequency distribution calculation. Finally, a difference frequency histogram is constructed to intuitively reflect the distribution pattern of the occurrence frequency of each difference value, providing a basis for subsequent efficient coding.

[0033] Efficient compression coding is achieved using Huffman coding based on the differential frequency histogram. First, the differential values ​​are sorted from high to low frequency, and the optimal prefix coding tree is constructed by assigning shorter binary codes to higher frequencies and longer binary codes to lower frequencies. Then, a unique short code is generated for each differential value, and each original differential value is bound to its corresponding code one by one, finally forming a coding mapping table for data compression and decoding. This ensures that high-frequency differential data is represented with the minimum bitstream, significantly reducing the amount of data transmitted.

[0034] First, the multidimensional sensor dataset is continuously sampled according to a preset acquisition cycle to form multiple sampling data frames. Then, the original difference values ​​in each data frame are replaced with corresponding short codes according to the encoding mapping table. At the same time, the inert channel and the active channel are distinguished by inter-frame residual calculation and amplitude determination. The transmission data of the inert channel with slow changes is further simplified, while the complete encoding information is retained for the active channel with significant changes. After completing the encoding verification and frame structure encapsulation, a compressed data packet adapted to the low bandwidth transmission of narrowband IoT is generated.

[0035] In one possible implementation, step S250 further includes: Step S251: Set multiple data acquisition cycles, and continuously sample the multidimensional sensor dataset according to the multiple data acquisition cycles to obtain multiple sampled data frames.

[0036] Step S252: Perform inter-frame calculations based on the multiple sampled data frames to obtain the inter-frame residual sequence.

[0037] Step S253: Perform amplitude calculation based on the inter-frame residual sequence to generate inter-frame variation amplitude values.

[0038] Step S254: Compare the inter-frame change amplitude value with the change threshold to determine the sensor channel whose inter-frame change amplitude value is less than the change threshold and mark it to generate an inert channel.

[0039] Step S255: Extract sensor channels whose inter-frame change amplitude values ​​are greater than or equal to the change threshold, identify them, and generate active channels.

[0040] Step S256: Perform data frame compression verification analysis through the lazy channel and the active channel to construct the compressed data packet.

[0041] Specifically, based on the operating characteristics of the air source heat pump and the edge computing resources, multiple data acquisition cycles of different lengths are first set. Then, according to these cycles, multi-dimensional sensor datasets containing parameters such as intake pressure, exhaust pressure, intake temperature, exhaust temperature, and outdoor ambient temperature are synchronously and continuously sampled. The sensor data of each dimension collected at the same time are encapsulated in a fixed format to form a set of sampled data frames arranged in chronological order and with a regular structure, providing standard data units for subsequent inter-frame calculations and compression coding.

[0042] For multiple sampled data frames arranged in chronological order, inter-frame calculation is performed by grouping adjacent frames together and calculating each sensor channel independently: First, the values ​​at the same position of the same sensor channel in the current frame and the previous frame are taken, and the inter-frame difference is obtained by subtracting the corresponding data in the previous frame from the data in the current frame. All difference calculations are completed point by point and channel by channel. Then, the inter-frame differences calculated by all channels at each time point are arranged in order according to the time axis and channel number, and finally an inter-frame residual sequence that can accurately characterize the fluctuation differences between sampled data frames is formed.

[0043] For the inter-frame residual sequence corresponding to each sensor channel, amplitude calculation is performed channel by channel. First, the absolute value of multiple consecutive inter-frame differences under a single channel is extracted. Then, by calculating the root mean square value or the maximum value of the absolute value of the residual of that channel, the overall fluctuation intensity of the channel data is quantified. The calculation result is used as the inter-frame change amplitude value of the corresponding channel, so as to accurately measure the intensity of data change of different sensor channels in the continuous sampling period.

[0044] A fixed change threshold is pre-set based on the normal fluctuation range of each sensor channel during normal operation of the air source heat pump. The inter-frame change amplitude value corresponding to each sensor channel is compared with the change threshold in turn. Sensor channels with inter-frame change amplitude values ​​less than the preset change threshold, smooth data fluctuation and small change amplitude are selected. A special inert channel identifier is added to this type of channel to complete the division and marking of inert channels for subsequent differentiated compression processing.

[0045] The inter-frame variation amplitude values ​​corresponding to each sensor channel are compared one by one with the preset variation threshold. Sensor channels with inter-frame variation amplitude values ​​greater than or equal to the variation threshold, significant data fluctuations, and high real-time requirements are selected. Active channel exclusive identifiers are added to these channels to complete the division and marking of active channels, so that they can be processed using a high-precision complete encoding method in the future.

[0046] Differential data frame compression verification analysis was conducted for the marked inert and active channels. For the inert channel, interval sampling and simplified coding were adopted to omit redundant and repetitive data and compress it according to the coding mapping table. For the active channel, full sampling and complete coding transmission were maintained without deleting key real-time data. Subsequently, the temporal integrity, coding format correctness and numerical validity of the compressed data of the two types of channels were verified, and abnormal and invalid data were removed. The verified inert channel compressed data and active channel coded data were uniformly encapsulated with frame header, timing tag and check bit, and finally a compressed data packet with a standardized structure and adapted to narrowband IoT transmission was constructed.

[0047] In one possible implementation, step S300 further includes: Step S310: Transmit the compressed data packet to the cloud control platform via narrowband IoT for parsing to obtain frame header identification data and encoding mapping version identification data.

[0048] Step S320: Retrieve data from the cloud control platform according to the frame header identifier data and the encoding mapping version identifier data to generate encoding retrieval results.

[0049] Step S330: Decode and recover based on the encoded retrieval results to obtain a multidimensional sensor recovery dataset.

[0050] Step S340: Construct a digital twin model of the air source heat pump, synchronize the multi-dimensional sensor recovery dataset to the digital twin model of the air source heat pump for state estimation, and construct a real-time operation profile of the heat pump.

[0051] Specifically, the compressed data packet is first transmitted from the edge device to the cloud control platform via narrowband IoT. The cloud receiving end performs layered parsing of the compressed data packet according to the preset communication protocol. First, the physical transport layer verification information of the data packet is stripped off. Then, the frame header area at the beginning of the data packet is read to extract the frame header identification data used to identify the data frame type, device number and timing information. Next, the version field at the preset position in the data packet is read and the encoding mapping version identification data corresponding to the Huffman coding rule is matched and extracted to complete the parsing of the identification information of the compressed data packet, providing a basis for subsequent decoding and retrieval.

[0052] The cloud control platform pre-establishes a database index storing different versions of encoding mapping tables. Based on the parsed frame header identifier data, it locates the corresponding air source heat pump equipment and data acquisition channel. Then, combined with the encoding mapping version identifier data, it uses a precise key-value matching retrieval method in the database to retrieve the Huffman encoding mapping table of the current version used by the equipment. After verifying the version consistency and data integrity, it outputs the encoding retrieval results containing complete encoding mapping rules, providing a matching basis for subsequent data decoding.

[0053] Based on the encoding mapping table obtained from the cloud, the encoded bit stream in the compressed data packet is read bit by bit according to the Huffman inverse decoding rule. The compressed encoding is reverse-mapped one by one to restore the corresponding difference value, resulting in a multidimensional difference sequence. Then, the first-order difference inverse operation is performed on the multidimensional difference sequence, and the difference values ​​of adjacent time steps are superimposed to restore the original time-series sampling data of each dimension. Then, abnormal jump data in the transmission process is removed and time-series alignment verification is completed, finally obtaining a complete and ordered multidimensional sensor recovery dataset.

[0054] A digital twin model of an air-source heat pump is constructed, consisting of four layers: a hardware parameter layer, a thermodynamic mechanism layer, a state simulation layer, and an error correction layer. The hardware parameter layer inputs the rated structural parameters of equipment such as the compressor, electronic expansion valve, fan, and heat exchanger. The thermodynamic mechanism layer embeds thermodynamic mathematical models of the heat pump heating cycle, heat transfer loss, and load matching. The state simulation layer is used for operational condition simulation, and the error correction layer integrates the Kalman recursive correction algorithm. During the model training phase, historical sensor data and actual energy efficiency data are used to iteratively optimize key parameters such as the heat transfer coefficient, loss coefficient, and thermal efficiency within the model, ensuring that the simulation results converge with the actual operating conditions. When using the model, the multi-dimensional sensor data is first processed... Unified timestamp alignment and timing calibration are used to generate a multi-dimensional sensor recovery synchronization dataset, which is then injected into the digital twin model of the air source heat pump for real-time simulation calculations. The model prediction sequence is output, and the multi-dimensional sensor dataset is converted into corresponding data sensor sequences. The two sets of sequences are compared point by point to calculate the sequence difference value. Based on the sequence difference value, the model is recursively corrected in real time using Kalman filtering, and a correction state vector is output. Based on the correction state vector, multi-dimensional operating indicators such as temperature, pressure, power consumption, and heating capacity are calculated, and a real-time energy efficiency ratio curve is generated. Finally, the multi-dimensional operating indicators and the real-time energy efficiency ratio curve are structurally integrated to comprehensively construct a real-time operating profile of the heat pump that can map the actual operating status of the equipment.

[0055] In one possible implementation, step S340 further includes: Step S341: Synchronize the data time of the multidimensional sensor recovery dataset to generate a multidimensional sensor recovery synchronization dataset.

[0056] Step S342: Inject the multidimensional sensing recovery synchronization dataset into the air source heat pump digital twin model for simulation calculation to generate a model prediction sequence.

[0057] Step S343: Convert the multidimensional sensing dataset into a data sensing sequence, compare the data sensing sequence with the model prediction sequence, and generate sequence difference values.

[0058] Step S344: Based on the sequence difference values, perform recursive correction on the digital twin model of the air source heat pump to generate a correction state vector.

[0059] Step S345: Calculate multi-dimensional indicators based on the corrected state vector to generate a real-time energy efficiency ratio curve.

[0060] Step S346: Structure the multidimensional indicators and the real-time energy efficiency ratio curve to construct a real-time operation profile of the heat pump.

[0061] Specifically, for the multidimensional sensor recovery dataset obtained by cloud decoding, the original timestamps corresponding to each sensor data are extracted. According to a unified sampling time benchmark, the sensor data of different dimensions such as inhalation pressure, exhaust temperature, and ambient temperature are time-series aligned and deviation calibrated. Data points with abnormal timestamps and delays are removed, and valid data at missing times are filled in, so that the data of each dimension strictly correspond to the same sampling time. Data time synchronization processing is completed, and finally a time-ordered and dimension-aligned multidimensional sensor recovery synchronization dataset is generated.

[0062] The time-aligned multi-dimensional sensor data is restored to a synchronous dataset and input into the digital twin model of the air source heat pump in real time. The model relies on the built-in heat pump thermodynamic cycle mechanism, compressor work model, heat exchanger heat exchange model, and electronic expansion valve flow regulation model. It substitutes real-time parameters such as current pressure, temperature, and environmental conditions to carry out dynamic simulation calculations. It extrapolates state parameters such as compressor discharge volume, heat exchange efficiency, and heat load output at each time step and outputs predicted values ​​of each dimension according to a unified time sequence. Finally, it generates a model prediction sequence that matches the actual sampling period and contains multi-dimensional prediction results.

[0063] First, the time-synchronized multidimensional sensor recovery synchronization dataset is organized into a continuous and ordered data sensing sequence according to a unified time axis. Then, the sequence is compared with the model prediction sequence output by the digital twin model in a one-to-one correspondence at each time step and each sensing dimension. The difference between the actual sensing data and the model prediction data at the same time step is calculated. After the difference is normalized, the sequence difference value that can characterize the model simulation deviation is obtained and used for subsequent model correction.

[0064] Based on the sequence difference values ​​obtained in the previous step, a recursive Kalman filter algorithm is used to correct the digital twin model of the air source heat pump in real time. The specific process is as follows: First, the model correction parameters are initialized, and the sequence difference values ​​are used as the observation error input into the observation equation of the Kalman filter. Combined with the model's own state transition equation, the state estimate and covariance matrix at the current moment are calculated. Then, the weights of the observation error and the model prediction error are weighed using the Kalman gain formula, and key parameters inside the model, such as heat transfer coefficient, compressor loss coefficient, and expansion valve flow coefficient, are recursively corrected time by time to eliminate the deviation between model simulation and actual operation. After the iteration is completed, a corrected state vector containing the corrected model state parameters, error range, and operational stability index is output to ensure that the simulation results of the digital twin model are highly consistent with the actual operating state of the air source heat pump, providing accurate support for the subsequent construction of the operational profile.

[0065] Based on the corrected state vector, core operating parameters such as real-time input power, heating capacity, evaporation temperature, condensation temperature, and ambient temperature of the heat pump are extracted, and multi-dimensional index calculations are carried out sequentially: real-time input electrical power is calculated by the compressor's work parameters, instantaneous heating capacity is calculated by combining the heat exchanger's heat exchange state, and real-time energy efficiency ratio is obtained based on the ratio of heating capacity to input power; at the same time, auxiliary operating indicators such as exhaust superheat, heat exchange efficiency, load rate, and power consumption coefficient are calculated simultaneously. The real-time energy efficiency ratios of continuous sampling times are arranged in chronological order, fitted and connected, and finally a real-time energy efficiency ratio curve that can intuitively reflect the dynamic changes of heat pump energy efficiency is generated.

[0066] First, the multi-dimensional operating indicators such as pressure, temperature, power, heat exchange efficiency, and load rate obtained by the corrected state vector calculation are classified and organized with the real-time energy efficiency ratio curve. A structured data framework is established according to three dimensions: equipment operating status, thermodynamic cycle status, and energy efficiency performance status. The discrete numerical indicators are normalized and labeled, and trend features, fluctuation features, and extreme value features are extracted from the real-time energy efficiency ratio curve. Then, the indicator data, curve feature data, and time series data of each dimension are time-series aligned and associated and bound, and uniformly packaged into a standardized dataset containing numerical values, curves, and feature labels. Finally, the structured aggregation is completed to construct a real-time operating profile of the heat pump that can comprehensively, in real time, and intuitively map the actual operating conditions and performance status of the air source heat pump.

[0067] In one possible implementation, step S342 further includes: Based on the multidimensional sensor recovery synchronization dataset, multiple acquisition time parameter sets are obtained in real time according to the acquisition time. The multiple acquisition time parameter sets include inhalation pressure parameters, exhaust pressure parameters, inhalation temperature parameters, and exhaust temperature parameters.

[0068] Boundary analysis is performed on the digital twin model of the air source heat pump based on the intake pressure parameters and the exhaust pressure parameters to construct the model input boundary conditions.

[0069] Based on the intake temperature parameters and the exhaust temperature parameters, a state analysis is performed on the digital twin model of the air source heat pump to construct model observation variable data.

[0070] The multidimensional sensing recovery synchronization dataset is injected into the air source heat pump digital twin model according to the model input boundary conditions and the model observation variable data to drive the simulation and generate the model prediction sequence.

[0071] Specifically, following a unified acquisition time sequence, the multidimensional sensor recovery synchronization dataset after time synchronization is completed is extracted in real time, and the corresponding inhalation pressure parameters, exhaust pressure parameters, inhalation temperature parameters, and exhaust temperature parameters are sequentially selected. The four core operating parameters extracted at each time are integrated and categorized to form multiple independent and time-aligned acquisition time parameter sets.

[0072] Based on the intake and exhaust pressure parameters collected at each time point, the actual pressure fluctuation range, extreme operating condition range, and normal operating threshold of the high and low pressure sides of the heat pump system are statistically analyzed. Pressure condition boundary analysis is then performed on the digital twin model of the air source heat pump to limit the upper and lower limits of the compressor intake and exhaust pressures and load constraints during model simulation. This is done to construct model input boundary conditions that are adapted to real operating conditions and to avoid invalid simulations that exceed operating conditions.

[0073] Based on the intake and exhaust temperature parameters corresponding to each acquisition time, the compressor inlet and outlet temperature difference, temperature change rate, and superheat characteristic value are calculated. Combined with the heat pump thermodynamic cycle characteristics, state analysis is carried out to determine the actual working state of the compressor and the operating condition of the heat exchange system. After removing abnormal temperature jump data, the filtered and optimized temperature time series data are organized into standardized parameters to construct model observation variable data for model state comparison and error correction.

[0074] The simulation operating range of the heat pump system is limited by the input boundary conditions of the constructed pressure model. The observed variable data of the temperature model is used as the basis for state verification. The time-aligned multi-dimensional sensor recovery synchronization dataset is injected into the digital twin model of the air source heat pump. Real-time thermodynamic cycle simulation is carried out under boundary constraints. The internal operating parameters of the model are corrected in real time by combining the observed variables. The multi-dimensional predicted values ​​of suction pressure, exhaust pressure, suction temperature, and exhaust temperature are deduced and output time by time, and finally a time-continuous and dimension-matched model prediction sequence is generated.

[0075] In one possible implementation, step S345 further includes: Instantaneous identification is performed based on the running time period to determine multiple instantaneous data points collected during the running time period.

[0076] Heating calculations are performed based on the multiple instantaneous data collected to obtain multiple instantaneous heating performance coefficients.

[0077] The multiple instantaneous heating performance coefficients are sorted in descending order according to the outdoor ambient temperature parameters and fitted to generate the energy efficiency ratio characteristic curve.

[0078] Performance degradation analysis is performed based on the energy efficiency ratio characteristic curve to determine the extent of performance deterioration.

[0079] The real-time energy efficiency ratio curve is constructed by backtracking to the correction state vector according to the performance degradation magnitude and calculating multi-dimensional indicators.

[0080] Specifically, the target time period of the actual operation of the air source heat pump is selected, and the time period is sliced ​​based on the preset sampling frequency. An instantaneous timestamp is added to each time slice node to complete the instantaneous identification. The real-time sensor data corresponding to each moment in the operation period is accurately located and extracted through the timestamp. Multiple instantaneous data with continuous time sequence and clear time nodes are selected to achieve accurate acquisition of instantaneous operating condition data during the operation of the heat pump.

[0081] The instantaneous input power, evaporator side temperature, condenser side temperature, and suction and discharge pressure of the heat pump are extracted sequentially from each instantaneous data collection. Combining the thermodynamic formula of the heat pump heating cycle, the instantaneous heating capacity is first calculated based on the pressure and temperature parameters. Then, the instantaneous heating capacity is calculated by ratioing the instantaneous input power to the instantaneous heating performance coefficient at that moment. The above heating calculation is performed for all instantaneous data collections one by one, and finally, multiple instantaneous heating performance coefficients for the corresponding operating time period are obtained.

[0082] The outdoor ambient temperature parameters corresponding to each instantaneous heating performance coefficient are extracted. With the outdoor ambient temperature as the independent variable and the instantaneous heating performance coefficient as the dependent variable, all data points are first sorted in descending order according to the outdoor ambient temperature from high to low. Then, a polynomial fitting method is used to smooth the sorted discrete data points to eliminate data fluctuation interference. Finally, an energy efficiency ratio characteristic curve that can reflect the change law of the heating performance coefficient of the air source heat pump with the outdoor temperature is generated.

[0083] Based on the generated energy efficiency ratio characteristic curve, a random forest regression algorithm is used to analyze the performance degradation of air source heat pumps and calibrate the degree of performance deterioration. This algorithm consists of a decision tree ensemble layer, a feature input layer, an error calculation layer, and a degradation quantification layer. First, a sample dataset is constructed. The input features are outdoor ambient temperature, operating time, intake and exhaust pressure, and intake and exhaust temperature. The output label is the baseline heating performance coefficient under the corresponding operating conditions. The factory-calibrated energy efficiency data of the equipment is used as the baseline sample, and historical normal operating energy efficiency data is used as the training sample, dividing the dataset into training and validation sets. During the training phase, the number of decision trees is set to 120, and a bootstrap sampling method is used to extract samples. This training method uses the minimization of mean squared error as the node splitting criterion, controls the tree depth to avoid overfitting, and completes model parameter tuning through out-of-bag data validation to learn the benchmark mapping law between outdoor temperature and heating performance coefficient. After the model training is completed, the temperature-energy efficiency data corresponding to the current energy efficiency ratio characteristic curve is input into the model to obtain the standard predicted energy efficiency value for each temperature range. The difference between the predicted and actual energy efficiency values ​​is calculated point by point to obtain the energy efficiency deviation. Then, the degradation rate of each temperature range is calculated by the ratio of the deviation to the standard predicted value. The overall performance degradation rate is obtained by combining the running time weights of different temperature ranges, thereby accurately calibrating the extent of heat pump performance degradation.

[0084] Based on the calibrated performance degradation range, a weighted recursive correction algorithm is used to backtrack and update the correction state vector and complete the calculation of multi-dimensional indicators. First, the performance degradation range is decomposed into three types of degradation weights: compressor attenuation weight, heat exchange attenuation weight, and system loss weight. These weights are then substituted into the correction state vector item by item to perform weighted compensation correction on core parameters such as suction pressure, exhaust temperature, input power, and heat transfer coefficient. Subsequently, based on the corrected state vector, the thermodynamic steady-state calculation method is used to solve the problem time-by-time. The instantaneous heating capacity is obtained by subtracting the evaporation enthalpy from the condensation enthalpy and multiplying by the mass flow rate. The product of the real-time voltage and current effective values ​​is then used to obtain the... The instantaneous input power is obtained, and the instantaneous heating capacity is divided by the instantaneous input power to obtain the real-time energy efficiency ratio. The load factor is calculated by the ratio of the instantaneous actual input power to the rated input power of the heat pump. The superheat is solved by subtracting the saturation temperature at the corresponding suction pressure from the actual suction temperature. The heat exchange efficiency is calculated by the ratio of the actual heat exchange of the condenser to the theoretical maximum heat exchange. While solving the instantaneous heating performance coefficient at each moment, the above three auxiliary indicators are simultaneously quantified. Finally, the real-time energy efficiency ratio at continuous moments is linearly smoothed and fitted along the time axis to finally construct an accurate real-time energy efficiency ratio curve that includes performance degradation characteristics.

[0085] In one possible implementation, step S500 further includes: Step S510: Generate a control command frame based on the target control parameter set using the cloud control platform.

[0086] Step S520: Send the control command frame to the heat pump controller via narrowband Internet of Things for parsing, and extract the target outlet water temperature, target compressor operating frequency, and target fan speed.

[0087] Step S530: Based on the heat pump controller adjusting the output frequency of the inverter according to the target compressor operating frequency, determine the compressor target speed data.

[0088] Step S540: Based on the heat pump controller adjusting the opening of the electronic expansion valve according to the target outlet water temperature, determine the target heating capacity data.

[0089] Step S550: Based on the heat pump controller adjusting the fan drive module according to the target fan speed, determine the target air volume data.

[0090] Specifically, after receiving the target control parameter set, the cloud control platform encapsulates, verifies, and splices the frame headers and trailers of each control parameter according to the preset narrowband IoT communication protocol format, generating a control command frame containing complete control information and completing the standardized encoding of the command for easy subsequent distribution and parsing.

[0091] The cloud control platform transmits the encapsulated control command frames to the local heat pump controller of the heat pump equipment through a low-power, long-range wireless communication network such as narrowband Internet of Things. The heat pump controller performs frame verification, protocol decoding, and data parsing operations on the received control command frames. From the valid data segment of the command frame, it extracts three control command parameters: the target outlet water temperature for controlling the hot water output temperature, the target compressor operating frequency for adjusting the compressor operating load, and the target fan speed for regulating the condenser heat exchange air volume.

[0092] The heat pump controller receives the target compressor operating frequency obtained from the analysis, sends a frequency adjustment control signal to the frequency converter, changes the power supply frequency output by the frequency converter to the compressor, and accurately matches and determines the target speed data corresponding to the compressor through the correspondence between the power supply frequency and the compressor speed, thereby realizing quantitative control of the compressor operating speed.

[0093] Based on the target outlet water temperature obtained through analysis, the heat pump controller outputs an opening adjustment signal to control the electronic expansion valve to change the valve opening, thereby adjusting the circulation flow rate of the refrigerant in the heat pump system. Combining the correspondence between refrigerant flow rate and heating capacity, the controller calculates and determines the target heating output data required by the heat pump, ensuring that the outlet water temperature stably reaches the set target.

[0094] Based on the target fan speed obtained from the analysis, the heat pump controller outputs a speed control signal to the fan drive module, adjusts the output power and drive duty cycle of the fan drive module, controls the actual operating speed of the fan, and determines the target air volume data required for heat pump heat exchange based on the corresponding conversion relationship between fan speed and ventilation volume, so as to ensure the stable heat exchange conditions of the heat pump system.

[0095] Example 2, Figure 2 This is a schematic diagram of the electronic device provided by the air source heat pump control method based on narrowband Internet of Things of the present invention, showing an exemplary electronic device suitable for implementing the embodiments of the present invention. Figure 2 The electronic device shown is merely an example and should not be construed as limiting the functionality or scope of the embodiments of the present invention. Figure 2 As shown, the electronic device includes a processor 21, a memory 22, an input device 23, and an output device 24; the number of processors 21 in the electronic device can be one or more. Figure 2 Taking a processor 21 as an example, the processor 21, memory 22, input device 23, and output device 24 in an electronic device can be connected via a bus or other means. Figure 2 Taking the example of a connection between China and Israel via a bus.

[0096] Example 3: Based on the same inventive concept as the air source heat pump control method based on narrowband IoT in the previous examples, this example provides a computer-readable storage medium for storing software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the air source heat pump control method based on narrowband IoT in this application. The processor executes the software programs, instructions, and modules stored in the memory to perform various functional applications and data processing of the computer device, thereby realizing the aforementioned air source heat pump control method based on narrowband IoT.

[0097] It should be noted that the order of the embodiments described above is for descriptive purposes only and does not represent the superiority or inferiority of the embodiments. Specific embodiments of this specification have been described above. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0098] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0099] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A control method for air source heat pumps based on narrowband Internet of Things, characterized in that, The method includes: Real-time acquisition of operating status data and environmental parameter data of air source heat pump units to obtain multi-dimensional sensor datasets; Based on the multidimensional sensor dataset, edge-side differential coding compression is performed to generate compressed data packets; The compressed data packet is transmitted to the cloud control platform via narrowband Internet of Things for decompression and reconstruction to obtain a real-time operating profile of the heat pump. Based on the real-time operation profile of the heat pump, load prediction is performed, multiple load prediction parameters are generated for control strategy optimization, and a target control parameter set is generated. The target control parameter set is sent to the heat pump controller via narrowband Internet of Things to perform adaptive adjustment of the air source heat pump.

2. The air source heat pump control method based on narrowband Internet of Things as described in claim 1, characterized in that, Based on the multidimensional sensor dataset, edge-side differential coding compression is performed to generate compressed data packets. The method includes: Multidimensional time series analysis is performed by traversing the multidimensional sensor dataset, and neighbor point analysis is performed by extracting multidimensional time series data to determine multiple neighboring sampling points. A first-order difference operation is performed based on the multiple adjacent sampling points to generate a multidimensional difference sequence; The frequency distribution of the multidimensional difference sequence is calculated by traversing it, and a difference frequency histogram is constructed. Encoding is performed based on the differential frequency histogram to obtain an encoding mapping table; Data frames are compressed and encoded according to the encoding mapping table to generate compressed data packets.

3. The air source heat pump control method based on narrowband Internet of Things as described in claim 2, characterized in that, The method for compressing and encoding data frames according to the encoding mapping table to generate compressed data packets includes: Multiple data acquisition cycles are set, and the multidimensional sensor dataset is continuously sampled according to the multiple data acquisition cycles to obtain multiple sampled data frames; Inter-frame calculations are performed based on the multiple sampled data frames to obtain an inter-frame residual sequence; Amplitude calculation is performed based on the inter-frame residual sequence to generate inter-frame variation amplitude values; The inter-frame change amplitude value is compared with the change threshold to determine the sensor channel whose inter-frame change amplitude value is less than the change threshold and then identified to generate an inert channel. Sensor channels whose inter-frame change amplitude values ​​are greater than or equal to the change threshold are extracted, identified, and active channels are generated. The compressed data packet is constructed by performing data frame compression verification analysis through the lazy channel and the active channel.

4. The air source heat pump control method based on narrowband Internet of Things as described in claim 1, characterized in that, The compressed data packet is transmitted to a cloud control platform via narrowband Internet of Things (IoT) for decompression and reconstruction to obtain a real-time operational profile of the heat pump. The method includes: The compressed data packet is transmitted to the cloud control platform via narrowband Internet of Things for parsing to obtain frame header identification data and encoding mapping version identification data. The frame header identifier data and the encoding mapping version identifier data are used to trace back to the cloud control platform for retrieval, and encoding retrieval results are generated. Based on the encoded retrieval results, decoding and recovery are performed to obtain a multidimensional sensor recovery dataset; A digital twin model of an air source heat pump is constructed, and the multidimensional sensor recovery dataset is synchronized to the digital twin model of the air source heat pump for state estimation, thereby constructing a real-time operation profile of the heat pump.

5. The air source heat pump control method based on narrowband Internet of Things as described in claim 4, characterized in that, A digital twin model of an air source heat pump is constructed, and the multi-dimensional sensor recovery dataset is synchronized to the digital twin model for state estimation to construct a real-time operational profile of the heat pump. The method includes: The multidimensional sensor recovery dataset is synchronized with the data time to generate a multidimensional sensor recovery synchronization dataset. The multidimensional sensor recovery synchronization dataset is injected into the digital twin model of the air source heat pump for simulation calculation to generate the model prediction sequence; The multidimensional sensing dataset is converted into a data sensing sequence, and the data sensing sequence is compared with the model prediction sequence to generate a sequence difference value. Based on the sequence difference values, the digital twin model of the air source heat pump is recursively corrected to generate a correction state vector. Based on the corrected state vector, multi-dimensional index calculations are performed to generate a real-time energy efficiency ratio curve. The multidimensional indicators are structurally aggregated with the real-time energy efficiency ratio curve to construct a real-time operation profile of the heat pump.

6. The air source heat pump control method based on narrowband Internet of Things as described in claim 5, characterized in that, The method involves injecting the multidimensional sensor recovery synchronization dataset into the digital twin model of an air source heat pump for simulation calculations to generate a model prediction sequence. Based on the multidimensional sensor recovery synchronization dataset, multiple acquisition time parameter sets are obtained in real time according to the acquisition time. The multiple acquisition time parameter sets include inhalation pressure parameters, exhaust pressure parameters, inhalation temperature parameters, and exhaust temperature parameters. Based on the intake pressure parameters and the exhaust pressure parameters, a boundary analysis is performed on the digital twin model of the air source heat pump to construct the model input boundary conditions. Based on the intake temperature parameters and the exhaust temperature parameters, a state analysis is performed on the digital twin model of the air source heat pump to construct model observation variable data; The multidimensional sensing recovery synchronization dataset is injected into the air source heat pump digital twin model according to the model input boundary conditions and the model observation variable data to drive the simulation and generate the model prediction sequence.

7. The air source heat pump control method based on narrowband Internet of Things as described in claim 5, characterized in that, Based on the corrected state vector, multi-dimensional index calculations are performed to generate a real-time energy efficiency ratio curve. The method includes: Instantaneous identification is performed based on the running time period to determine multiple instantaneous data points collected during the running time period; Heating calculations are performed based on the multiple instantaneous data collected to obtain multiple instantaneous heating performance coefficients; The multiple instantaneous heating performance coefficients are sorted and fitted in descending order according to the outdoor ambient temperature parameters to generate the energy efficiency ratio characteristic curve. Based on the energy efficiency ratio characteristic curve, a performance degradation analysis is performed to determine the extent of performance deterioration. The real-time energy efficiency ratio curve is constructed by backtracking to the correction state vector according to the performance degradation magnitude and calculating multi-dimensional indicators.

8. The air source heat pump control method based on narrowband Internet of Things as described in claim 1, characterized in that, The method of transmitting the target control parameter set to the heat pump controller via narrowband Internet of Things to perform adaptive adjustment of the air source heat pump includes: Based on the cloud-based control platform, control command frames are generated according to the target control parameter set; The control command frame is sent to the heat pump controller via narrowband Internet of Things for parsing, and the target outlet water temperature, target compressor operating frequency, and target fan speed are extracted. The heat pump controller adjusts the output frequency of the inverter according to the operating frequency of the target compressor to determine the target compressor speed data; The target heating capacity data is determined by adjusting the opening of the electronic expansion valve based on the target outlet water temperature using the heat pump controller. The heat pump controller adjusts the fan drive module according to the target fan speed to determine the target air volume data.

9. An electronic device, characterized in that, The electronic device includes: processor; Memory used to store the processor's executable instructions; The processor is used to execute the air source heat pump control method based on narrowband Internet of Things as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the air source heat pump control method based on narrowband Internet of Things as described in any one of claims 1-8.

Images