A method for variable frequency coordinated operation control of multi-stage series air source heat pump

Abstract

This invention belongs to the technical field of variable frequency coordinated operation control methods for air source heat pumps, and particularly relates to a multi-stage series air source heat pump variable frequency coordinated operation control method. It uses sensors to collect outdoor ambient temperature, outdoor relative humidity, system return water temperature, target supply water temperature, rated circulation flow rate of a single branch, and the maximum number of parallel branches in the system in real time, transmitting all data to a central controller. A time-series rolling prediction algorithm is used to predict the total system heat load for the next hour. A frost-free performance model and a three-dimensional defrosting model of the air source heat pump are pre-trained and embedded. The frost-free performance model uses inlet water temperature, compressor operating frequency, and outdoor ambient temperature as independent variables, outputting the steady-state heating power and steady-state power consumption of a single unit. The three-dimensional defrosting model uses compressor operating frequency, outdoor ambient temperature, and outdoor relative humidity as independent variables, outputting the defrosting duration, defrosting return water heat absorption power, and defrosting power consumption within one hour.

Description

Technical Field

[0001] This invention belongs to the technical field of variable frequency coordinated operation control method for air source heat pumps, and particularly relates to a multi-stage series air source heat pump variable frequency coordinated operation control method. Background Technology

[0002] Currently, most multi-stage series air source heat pump heating systems employ traditional fixed-frequency or automatic variable-frequency control methods based solely on water temperature feedback, lacking the ability to accurately predict unit frosting and defrosting conditions. The unit's operating frequency is largely adjusted passively based on return and supply water temperatures, without considering outdoor temperature and humidity or the unit's operating status for overall optimization. In low-temperature, high-humidity environments, this leads to frequent defrosting and significant reductions in heating capacity. This not only causes large fluctuations in supply water temperature but also significantly increases overall system energy consumption, making it difficult to fully utilize the unit's heating efficiency.

[0003] Existing control strategies lack dedicated optimization logic for systems with multiple parallel branches and two-stage heat pumps connected in series. They fail to coordinate the matching relationship between the number of branch start-ups and shutdowns, the operating frequency of the two-stage units, and the intermediate water temperature, and generally suffer from simple calculation logic and a single operating scheme. Furthermore, conventional control methods lack data model support and computational pruning mechanisms, making it difficult to maximize system energy efficiency while ensuring heating redundancy and stability. Overall, there is still significant room for improvement in operational economy and adaptability. Summary of the Invention

[0004] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide a multi-stage series air source heat pump variable frequency coordinated operation control method for an air source heat pump heating system with multiple branches in parallel and two condensers connected in series on the water side in each branch. The system operates at a constant flow rate and the heat pump unit supports direct setting of the compressor operating frequency. The method includes the following steps: Step 1: Collect outdoor ambient temperature, outdoor relative humidity, system return water temperature, target supply water temperature, rated circulation flow of a single branch, and maximum number of parallel branches in the system in real time through sensors, and transmit all data to the central controller. Step 2: Use a time series rolling forecasting algorithm to predict the total system heat load for the next hour; Step 3: Pre-train and embed the air source heat pump frost-free performance model and the air source heat pump three-dimensional defrosting model. The frost-free performance model uses the inlet water temperature, compressor operating frequency, and outdoor ambient temperature as independent variables, and outputs the steady-state heating power and steady-state power consumption of a single unit. The three-dimensional defrosting model uses the compressor operating frequency, outdoor ambient temperature, and outdoor relative humidity as independent variables, and outputs the defrosting time, defrosting return water heat absorption power, and defrosting power consumption within 1 hour. Step 4: Calculate the maximum heating capacity of a single branch at the compressor peak frequency using the dual data-driven model. Round up the predicted total heat load to obtain the minimum number of branches to be opened. Combine this with the maximum number of parallel branches in the system to determine the closed interval of the number of selectable branches. Eliminate invalid branches outside the interval to complete the first pruning. Step 5: Enumerate the number of open branches in the selectable range one by one, calculate the heating task and temperature rise required for each branch, check whether the temperature rise of a single branch exceeds the maximum temperature rise capacity of the unit at the peak frequency, remove the number of branches that do not meet the conditions, and complete the second pruning. Step 6: For the number of open branches retained after the second pruning, enumerate the intermediate water temperature of the two-stage heat pump between the return water temperature and the target supply water temperature with a fixed step size, and divide the temperature rise tasks of the first-stage heat pump and the second-stage heat pump according to the intermediate water temperature. Step 7: Divide the defrost operation domain and the frost operation domain according to the real-time outdoor temperature and humidity. Combine the corresponding data-driven model to solve the minimum operating frequency of the two-stage heat pump to meet the temperature rise and heating tasks. In the frost operation domain, first use the three-dimensional defrost model to complete the heating power and power consumption reduction calculation, then solve the operating frequency, eliminate intermediate water temperature schemes whose operating frequency exceeds the unit frequency range, and complete the third pruning. Step 8: Summarize all enumeration and pruning results to obtain multiple feasible alternative schemes, calculate the system heating redundancy rate and system heating performance coefficient (COP) for each scheme; screen out schemes with heating redundancy rates not lower than a preset threshold, and select the scheme with the highest system COP as the optimal operating scheme. Step 9: The central controller issues control commands according to the optimal operating plan to control the start and stop of the corresponding branch, and directly sets the operating frequency of the compressors of the primary and secondary heat pumps in each branch, so that the unit can operate stably at a fixed frequency.

[0005] Preferably, the dual data-driven model described in step three is constructed using the XGBoost algorithm; The training set of the model shall contain no less than 2,000 sets of data, and the test set shall be divided into 20% sets with no less than 500 sets of data. The operating conditions of the training set and the test set shall be consistent. The model evaluation criteria are as follows: the coefficient of determination R² for the frost-free performance model is ≥0.85, the coefficient of determination R² for the three-dimensional defrosting model is ≥0.85, and the mean relative error (MAPE) of both models is no higher than 5%.

[0006] Preferably, the fixed step size for enumerating the intermediate water temperature in step six is ​​0.5℃; The intermediate water temperature between the two-stage heat pumps is the outlet water temperature of the first-stage heat pump and the inlet water temperature of the second-stage heat pump.

[0007] Preferably, in step seven, under the frost-prone operating domain, the formula for calculating the effective heating power of the unit after reduction is as follows: ; The formula for calculating the total power consumption of the unit in 1 hour is: ; in, This is the frost-free steady-state heating power. For defrosting time, The heat absorption capacity of the defrost return water. Power consumption for defrosting.

[0008] Preferably, the operating frequency range of the heat pump unit compressor is 30Hz-120Hz; The preset minimum threshold for heating redundancy rate in step eight is 5%.

[0009] Preferably, the system has a maximum of 10 parallel branches, and the rated circulating flow rate of a single branch is 2.5 m³ / h; The specific heat capacity of water at constant pressure is taken as: The density of water is taken as 1000 kg / m³.

[0010] Preferably, the method uses a 1-hour prediction optimization cycle, performs a complete optimization calculation once per hour, and the execution time of a single optimization calculation does not exceed 10 seconds; After three pruning strategies, the overall average computational cost was reduced by 40%.

[0011] Preferably, the criteria for determining the anti-frost operating domain are: With an outdoor temperature of 7°C and a relative humidity of 60%, the unit has no risk of frost formation within one hour under these operating conditions, and there is no need to reduce its defrosting performance. The criteria for determining the frost-affected operating region are: With an outdoor temperature of -7℃ and a relative humidity of 90%, the unit is at risk of frost formation under these conditions, and a defrosting performance reduction calculation must be performed.

[0012] Preferably, in the defrost suppression mode, the system water supply temperature fluctuation is maintained within 1℃; In frost-prone operation mode, the system water supply temperature fluctuation is less than 2.5℃; The accuracy rate of the three-dimensional defrosting model in determining the frost suppression / frost presence condition is no less than 85%. After adopting this method for control, the system's hourly heating performance coefficient (COP) can be increased by up to 20%, the seasonal performance coefficient (SCOP) can be increased by about 15%, and the total energy consumption of the system in a single heating season can be reduced by about 17%. Preferably, the heating system includes multiple parallel heating branches, circulating water pumps, a main supply pipe, a main return pipe, temperature sensors, flow sensors, outdoor temperature and humidity sensors, and a central controller; each heating branch includes two independent air source heat pumps, with the condensers of the two heat pumps connected in series on the water side, and the return water is heated step by step through a primary heat pump and a secondary heat pump before flowing into the main supply pipe; the water valve corresponding to the shutdown branch is closed, and the total circulating water volume of the system remains constant.

[0013] Based on the above, by constructing a frost-free performance model and a three-dimensional defrosting model, the unit's defrosting status can be accurately predicted, and heating and energy consumption parameters can be quantified. Combined with a multi-round pruning strategy, the computational load is significantly reduced, and the calculation time for a single operation within the optimization cycle does not exceed 10 seconds, ensuring the real-time performance of the control scheme. This control method effectively suppresses water temperature fluctuations, controlling the supply water temperature fluctuation within 1℃ under frost suppression conditions and below 2.5℃ under frost conditions. The accuracy rate of defrosting condition judgment can reach over 85%. The total energy consumption of the system in a single heating season can be reduced by approximately 17%, the hourly heating performance coefficient can be increased by up to 20%, and the seasonal energy efficiency coefficient can be increased by approximately 15%, significantly improving the system's operational energy efficiency.

[0014] This invention achieves global coordinated optimization of parallel branch start-up and shutdown, two-stage heat pump frequency conversion, and intermediate water temperature, strictly ensuring a system heating redundancy rate of no less than 5%, and can operate stably under different outdoor temperature and humidity conditions. Customized and optimized logic is used for the system structure of two-stage series and multiple parallel branches, fully leveraging the unit's heating capacity and solving the problems of poor adaptability, low energy efficiency, and insufficient heating stability of traditional control methods. This significantly improves the overall operational quality and economic value of multi-stage series air source heat pump heating systems. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the water circuit structure of a multi-stage series air source heat pump heating system, which is provided by the multi-stage series air source heat pump variable frequency coordinated operation control method in this embodiment of the invention. Figure 2 This is a flowchart of the dual-data-driven model construction of the multi-stage series air source heat pump variable frequency coordinated operation control method provided in this embodiment of the invention; Figure 3 This is an overall execution flowchart of the control method for the multi-stage series air source heat pump variable frequency coordinated operation control method provided in the embodiments of the present invention. Detailed Implementation

[0016] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1 This is a flowchart illustrating a multi-stage series air source heat pump variable frequency coordinated operation control method provided in one embodiment of the present invention. The multi-stage series air source heat pump variable frequency coordinated operation control method will be described in detail below.

[0017] The heating system suitable for this method is a multi-branch parallel arrangement, with two air source heat pumps installed in each branch and the condenser water side connected in series to achieve step-by-step heating of the return water. The system operates at a constant flow rate, and the heat pump unit supports direct setting of the compressor operating frequency (different from the existing automatic frequency conversion method based on water temperature).

[0018] For specific air source heat pump systems and defrosting strategies, a dual data-driven model is constructed using measured data to achieve early prediction of defrosting conditions and accurate quantification of heat pump heating performance: an air source heat pump frost-free performance model (predicting the heating and power consumption of a single unit under frost-free conditions) and an air source heat pump three-dimensional defrosting model (predicting the defrosting duration, defrosting power consumption, and heat absorption power from the return water within one hour under defrosting conditions).

[0019] Using a 1-hour prediction and optimization cycle, the total heat load for the next 1 hour is obtained through rolling prediction and other methods. First, the range of optional open branch numbers is determined, and then the number of branches and the intermediate water temperature of the two-stage units are enumerated one by one. The operating frequency of the two-stage units that meet the heating task is solved by combining the dual model, and a pruning strategy is added simultaneously to reduce the amount of calculation.

[0020] After generating all feasible solutions, the optimal solution with the highest system heating performance coefficient is selected from the solutions that meet the heating redundancy requirements. This achieves global coordinated optimization of multi-branch start-up and shutdown, two-stage unit frequency conversion, and intermediate water temperature, taking into account heating stability, operational energy efficiency, and real-time performance.

[0021] Model training details, dataset sample size: To ensure the quality of model training, the dataset samples should be sufficiently covered in terms of independent variable dimensions, and the model training set data size should be no less than 2000 sets.

[0022] Training set data coverage: To ensure the generalization ability of the data model, the dual-model training data should fully cover the dimensions involved in the independent variables. For example, the frost-free performance model of an air source heat pump should be fully covered in the three-dimensional space of adjustable inlet water temperature, outdoor temperature, and operating frequency.

[0023] Test set partitioning rules: The test set can be divided into 20% segments, and its operating condition coverage should be consistent with that of the training set, with no less than 500 groups.

[0024] Model validation using quantitative data; Coefficient of determination It should be no less than 0.85 Mean relative error (MAPE): should not exceed 5%; Key performance data; System COP improvement rate: Based on case calculations, hourly COP can be improved by up to 20%, and seasonal energy coefficient SCOP can be improved by about 15%.

[0025] Water supply temperature fluctuation under defrosting conditions: In the defrosting mode, the water supply temperature fluctuation can be maintained within 1℃; in the frost mode, the water supply temperature fluctuation can be lower than 2.5℃, effectively reducing water temperature fluctuations caused by centralized defrosting.

[0026] Total system energy consumption reduction: Based on case calculations, total energy consumption can be reduced by approximately 17% in a single heating season.

[0027] Defrosting prediction accuracy: The 3D defrosting model can achieve an accuracy of over 85% in determining whether there is frost suppression or frost presence.

[0028] Heating redundancy rate target value: The heating redundancy rate is mainly related to the number of units configured in the system. The more units there are, the stronger the heating guarantee capability, and the easier it is to meet the redundancy rate requirements in low temperature environments.

[0029] Implementation Case: Core Configuration of a Single Branch Circuit: Two air source heat pumps with condensers connected in series on the water side. The rated heating capacity of a single branch circuit is 7kW (rated operating conditions: outdoor dry-bulb temperature 7℃, supply water temperature 40℃, compressor rated frequency 70Hz). When the circuit is shut down, the water valve is closed, and the total circulating water volume of the system remains unchanged. The dual data-driven model is constructed based on the XGBoost algorithm. The frost-free performance model has an R² of ≥0.85, the three-dimensional defrosting model has an R² of ≥0.85, and the average relative error is ≤5%.

[0030] The reduction in computational workload after pruning: It has been verified that the average computational workload can be reduced by 40% after pruning, and the computational savings rate is even higher when the outdoor temperature is low.

[0031] Execution time per session: This method performs optimization calculations once per hour, with a maximum execution time of 10 seconds per session.

[0032] System fixed operating parameter table;

[0033] Implementation example of the anti-frost operating domain; Operating conditions; ①Real-time environmental parameters: Outdoor temperature t air =7℃, relative humidity RH=60%, meets the criteria for anti-frost operation zone, there is no risk of frost formation within 1 hour, and no reduction in defrosting performance is required; ②System operating parameters: Return water temperature t return =35℃, target water supply temperature t supply =42℃, the total temperature rise requirement of a single branch circuit is Δt=7℃; ③ Heat load forecast: Based on the ARIMA algorithm rolling forecast, the total heat load of the system in the next hour is Q=120kW.

[0034] 2) Optimize the complete execution process of the algorithm; ① Determine the range of possible branches and the first pruning; Based on the frost-free performance model, the performance of a single branch at a peak frequency of 120Hz and t is calculated. air =7℃、t return Maximum heating capacity at 35℃ =21kW; Calculate the minimum number of open branches N1 to meet the heat load requirement. =6 paths; the closed interval for the number of possible branches is determined to be [6,10], and invalid schemes with N=1~5 are removed in the first pruning.

[0035] Branch count enumeration and second pruning; Enumerate the number of branches N in the interval one by one, calculate the heating task and temperature rise feasibility of a single branch, verify whether the heating capacity exceeds the maximum heating capacity of a single branch at the peak frequency, and complete the second pruning.

[0036]

[0037] ③ Enumeration of intermediate water temperatures, frequency calculation, and third pruning; For each valid N value, within the range of 35℃ to 42℃, enumerate the intermediate water temperature t in steps of 0.5℃. mid There are a total of 15 valid enumeration values; for each t mid Divide the temperature rise task of the two-stage heat pump: Level 1 heat pump temperature rise task: ΔT1=t mid -t return (Return water → Intermediate water temperature); Secondary heat pump temperature rise task: ΔT2=t supply -t mid (Intermediate water temperature → target water supply); Based on the frost-free performance model, the minimum operating frequencies f1 (first stage) and f2 (second stage) of the two-stage heat pump to meet the corresponding temperature rise task are solved. The third pruning directly eliminates invalid schemes with frequencies exceeding the 120Hz peak value.

[0038] Example solution (N=8, tmid=38.5℃): Two-stage temperature rise task: Stage 1 Δt1 = 3.5℃, Stage 2 Δt2 = 3.5℃; Based on the frost-free performance model, the minimum operating frequencies required to meet the single-stage 7.5kW heating task are obtained: f1=70Hz, f2=75Hz. The frequencies are all in the range of 30~120Hz, and the third pruning was performed.

[0039] ④ Feasible solution generation and optimal solution selection; After completing all enumerations and pruning, a total of 75 feasible alternative schemes were generated. The system COP and redundancy rate were calculated for each scheme. Schemes with a heating redundancy rate greater than 5% were first selected, and then the scheme with the highest total system COP was chosen from the results. The final optimal scheme is as follows:

[0040] 3) Implementation of the plan; ① The central controller issues a command to open 8 parallel branches and close the remaining 2 branches; ② Set the operating frequency of the primary heat pump unit of each branch to 70Hz and the secondary unit to 75Hz. After receiving the command, the unit will continue to operate at a fixed frequency. ③After the system is running stably, the actual measured data is fed back: the water supply temperature is 42℃, the absolute value of the water supply temperature fluctuation is less than 0.5℃, the measured value of the total COP of the system is 3.45, and the prediction deviation meets the design and operation requirements.

[0041] (2) Complete implementation of the frost-covered operating domain; 1) Operating conditions; ①Real-time environmental parameters: Outdoor temperature The relative humidity RH=90% does not meet the criteria for the anti-frost operating range and there is a risk of frost formation. Therefore, a calculation to reduce the defrosting performance is required. ②System operating parameters: Return water temperature t return =35℃, target water supply temperature t supply =42℃, the total temperature rise requirement of a single branch circuit is Δt=7℃; ③ Heat load prediction: Based on the ARIMA algorithm rolling prediction, the total heat load of the system in the next hour is Q=140kW.

[0042] 2) Optimize the complete execution process of the algorithm; ① Determine the range of possible branches and the first pruning; Based on dual-model coupled calculation, a single branch at a peak frequency of 120Hz, t air Maximum effective heating capacity after reduction at -7℃ operating conditions =17.5kW; Calculate the minimum number of open branches N1 required to meet the heat load demand. =8 items; The number of possible branches is determined to be within the closed interval [8, 10]. The first pruning directly removes invalid schemes with N=1~7.

[0043] ② Branch count enumeration and second pruning; Enumerate the number of branches N in the interval one by one, calculate the heating task, flow rate, and temperature rise feasibility of a single branch, and complete the second pruning verification:

[0044] ③ Enumeration of intermediate water temperatures, frequency calculation, and third pruning; For each valid N value, within the range of 35℃ to 42℃, enumerate the intermediate water temperature t in steps of 0.5℃. mid There are a total of 15 valid enumeration values; for each t mid The temperature rise task is divided into two levels. The defrosting performance is reduced by combining the dual models. The minimum operating frequencies f1 and f2 that meet the effective heating task are solved. Invalid schemes with frequencies exceeding the 120Hz peak value are removed in the third pruning.

[0045] Solution example (N=9, t) mid =38.5℃): Two-stage temperature rise task: Stage 1 Δt1 = 3.5℃, Stage 2 Δt2 = 3.5℃; Single branch flow rate: 2.78 m³ 3 / h, the heating task of a single branch is 15.56kW, and the heating capacity required for a single stage is 7.78kW; Based on the three-dimensional defrosting model, predict t air Under operating conditions of -7℃, RH=90%, and frequency 95Hz, the defrosting time T for a single unit per hour is... def =16min, defrosting power consumption P de =2.8kW, defrosting return water heat absorption power P d =3.2kW; Based on the frost-free performance model, the predicted values ​​are 95Hz frequency, 35℃ inlet water, and t. air Under operating conditions of -7℃, the frost-free heating power P of a single unit h =8.8kW, frost-free power consumption P e =2.6kW; Effective heating power per unit after reduction ; Total power consumption of a single unit per hour ; The same method was used to complete the reduction calculation for the secondary unit. Finally, the minimum operating frequencies f1=95Hz and f2=105Hz were obtained to meet the heating task of a single branch of 15.56kW. Both are within the range of 30~120Hz. This was achieved through the third pruning.

[0046] ④ Feasible solution generation and optimal solution selection; After completing all enumerations and pruning, a total of 45 feasible alternative schemes were generated. First, schemes with a heating redundancy rate of ≥5% were selected. Then, the scheme with the highest system COP was selected from the selection results. The final optimal scheme is as follows:

[0047] 3) Implementation of the plan ① The central controller issues a command to open 9 parallel branches and close the remaining 1 branch; ② Set the operating frequency of the primary heat pump unit of each branch to 95Hz and the secondary unit to 105Hz. After receiving the command, the unit will continue to operate at a fixed frequency. ③ Actual data after the system is running stably: The actual defrosting time is 15 minutes within 1 hour, the steady-state water supply temperature is 42℃, the lowest water supply temperature during defrosting is 40℃, and the measured COP of the system is 2.50, which meets the requirements for heating stability and energy efficiency under low temperature conditions.

[0048] The core structure of the multi-stage series air source heat pump heating system adapted to this invention is as follows: System structure: The system consists of It consists of a parallel air source heat pump heating branch, a circulating water pump, a main water supply pipe, a main water return pipe, a temperature sensor, a flow sensor, an outdoor temperature and humidity sensor, and a central controller.

[0049] Branch structure: Each heating branch is equipped with two independent air source heat pump units, defined as a primary heat pump and a secondary heat pump respectively; the condensers of the two heat pumps are connected in series on the water side, that is, the return water from the main return water pipe of the system is first distributed to the branch, enters the condenser heat exchanger of the primary heat pump on the branch, completes the primary heating, and then directly enters the condenser heat exchanger of the secondary heat pump to complete the secondary heating. Finally, the water outlet flows into the main supply water pipe, realizing the two-stage progressive heating of the return water.

[0050] Operating mode: The water circulation flow rate of a single branch is constant; all heat pump units support direct setting of the compressor operating frequency, and continue to operate at the set frequency after receiving the frequency command from the central controller.

[0051] Control architecture: The central controller is the core execution unit, responsible for data acquisition, model calculation, optimization algorithm execution, and command issuance. All sensor data are connected to the central controller, and all operating commands for the heat pump units are issued uniformly by the central controller.

[0052] 5.2 Construction of the Core Data-Driven Model This invention targets specific air source heat pump units and defrosting strategies. It obtains effective operating condition datasets through laboratory standard operating condition tests or actual field tests, and uses machine learning and other algorithms to pre-build two core data-driven models. After training and evaluation, these models are embedded in a central controller for real-time operating condition prediction and optimization calculations.

[0053] 5.2.1 Frost-free performance model of air source heat pump; This model is used to accurately predict the heating and energy consumption performance of a single heat pump unit under frost-free conditions and is the core foundation of the optimization algorithm.

[0054] Independent variable of the model: inlet water temperature of a single heat pump Compressor operating frequency Outdoor ambient temperature ; Model dependent variable: Steady-state heating power of the unit Steady-state power consumption ; Model mathematical expression: ; Model training requirements: The dataset must cover the entire operating frequency range, the entire applicable inlet water temperature range, and the entire applicable ambient temperature range of the unit, with the test set accounting for 20% of the data.

[0055] 5.2.2 Three-dimensional defrosting model of air source heat pump; This model is used to predict the defrosting performance of the unit within a 1-hour cycle under different operating conditions, quantify the impact of the defrosting process on the system's heating capacity and energy consumption, and achieve early prediction of the impact of defrosting.

[0056] Independent variable in the model: compressor operating frequency Outdoor ambient temperature Outdoor relative humidity ; Dependent variable in the model: average defrosting time per hour (Unit: seconds) Average heat absorption power from the return water during defrosting Average power consumption for defrosting ; Model mathematical expression: ; Model training requirements: The dataset must cover the entire operating frequency range of the unit, the entire applicable environmental temperature and humidity range, and the entire relative humidity range, with the test set accounting for 20% of the data.

[0057] 5.3 Detailed execution flow of the global collaborative optimization algorithm; This optimization algorithm uses a 1-hour prediction and optimization cycle, updating environmental parameters, predicted heat load values, and optimized operation plans once per cycle. The algorithm execution cycle is synchronized with the unit frequency adjustment cycle. The specific execution steps are as follows: Step 1: Collect and read runtime data; The system sensors read the current outdoor ambient temperature in real time. Outdoor relative humidity System return water temperature System target water supply temperature Rated circulating flow rate of a single branch Maximum number of parallel branches in the system Basic parameters and real-time operating data are synchronized to the central controller.

[0058] Step 2: Predict the heat load for the next hour; Based on historical system heat load data, historical ambient temperature and humidity data, and indoor set temperature data, algorithms such as time series rolling forecasting are used to predict the total system heat load for the next hour. .

[0059] Step 3: Determine the selectable range for the number of branch lines to be opened; Based on the two performance models mentioned above, the maximum stable heating capacity of a single branch circuit under the compressor peak frequency and current environmental conditions is calculated. ; Based on the predicted total heat load Calculate the minimum number of open branches required to meet the heat load demand. The calculation formula is: ,in It is a rounding function; The selectable range for the number of branches to be opened is: ,in The maximum number of branches configured for the system; Step 4: Determine the heating task for a single branch circuit Enumerate intervals one by one The number of integer branches within ; For each enumeration Calculate the heating load that a single branch needs to handle. The calculation formula is: ; Based on the system's constant flow operation mode, and according to the basic formula of heat load... Verification of single-branch temperature rise task, among which The specific heat capacity of water at constant pressure; Implement pruning strategy: If the current The corresponding single-branch temperature rise requirement exceeds the unit's maximum temperature rise capacity at peak frequency, therefore this option is directly excluded. The value terminates all subsequent calculations corresponding to that branch number.

[0060] Step 5: Enumerate the intermediate water temperature of the two-stage units; For each feasible number of open branches With fixed step size (recommend ), at the return water temperature and target water supply temperature Between, list the intermediate water temperatures one by one. ;in This refers to the target water temperature between the primary and secondary heat pumps within a single branch, specifically the target outlet water temperature of the primary heat pump and the inlet water temperature of the secondary heat pump.

[0061] Step 6: Match the operating frequencies of the two-stage heat pump; For each enumeration First, determine the temperature rise task of the two-stage heat pump: the first-stage heat pump is responsible for... The secondary heat pump undertakes the task of raising the temperature. The temperature rise task; and combining the two performance models mentioned above, the minimum operating frequency of the two-stage heat pump to meet the temperature rise and heating tasks is solved respectively. (Level 1 heat pump) (Two-stage heat pump), the specific solution logic is as follows: Operating domain partitioning: based on current ambient temperature relative humidity Pre-divide the frost suppression operation domain and the frost-prone operation domain.

[0062] Determining the frost-suppressing operating frequency: Based on the frost-free performance model and with the constraint of satisfying the step-by-step temperature rise task, the minimum operating frequency of the first-stage heat pump is determined. Minimum operating frequency of a two-stage heat pump This ensures that the steady-state heating capacity of the two-stage units at this frequency can meet the heating requirements.

[0063] Frosty operating domain frequency solution and performance reduction calculation: First, based on the three-dimensional defrosting model, prediction Defrosting time within 1 hour under the corresponding operating conditions Defrosting return water heat absorption power Defrosting power consumption ; The frost-free heating capacity of the unit is reduced, and the effective heating power after reduction is calculated using the following formula: ; The total power consumption of the computer group in one hour is calculated using the following formula: Using the constraint that the reduced effective heating power meets the heating task, the minimum operating frequency required to meet the requirements is recalculated. ; Implement a pruning strategy: If the effective heating power of the unit, after reduction, is still insufficient to meet the corresponding heating task at peak frequency, remove the current unit from the pruning list. The value is then set and subsequent calculations are terminated.

[0064] Step 7: Generate feasible alternative solutions; Complete all and After enumeration, verification, and pruning, all feasible alternative running plans are generated, each plan containing the core parameter: the number of branches opened. Intermediate water temperature Primary heat pump operating frequency Secondary heat pump operating frequency Run the domain label (anti-frost / frost present).

[0065] Step 8: Optimal solution selection and execution; Calculate the total system COP and heating redundancy rate for all feasible alternative schemes; Pre-set the system heating redundancy control threshold (5%~10% recommended) and screen out alternative solutions that meet the redundancy requirements; Among the selected solutions, the one with the highest overall system COP is chosen as the final optimal operating solution. The central controller sends the optimal operating instructions to the corresponding heat pump units, controls the start and stop of the corresponding branches, and directly sets the operating frequency of the two-stage heat pumps to execute the optimal operating scheme.

[0066] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

[0067] The embodiments of this application have been described above with reference to the accompanying drawings. Unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other. This application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method for variable frequency coordinated operation control of a multi-stage series air source heat pump, used in an air source heat pump heating system with multiple branches in parallel and two condensers connected in series on the water side in each branch, wherein the system operates at a constant flow rate and the heat pump unit supports direct setting of the compressor operating frequency, characterized in that, Includes the following steps: Step 1: Collect outdoor ambient temperature, outdoor relative humidity, system return water temperature, target supply water temperature, rated circulation flow of a single branch, and maximum number of parallel branches in the system in real time through sensors, and transmit all data to the central controller. Step 2: Use a time series rolling forecasting algorithm to predict the total system heat load for the next hour; Step 3: Pre-train and embed the air source heat pump frost-free performance model and the air source heat pump three-dimensional defrosting model. The frost-free performance model uses the inlet water temperature, compressor operating frequency, and outdoor ambient temperature as independent variables, and outputs the steady-state heating power and steady-state power consumption of a single unit. The three-dimensional defrosting model uses the compressor operating frequency, outdoor ambient temperature, and outdoor relative humidity as independent variables, and outputs the defrosting time, defrosting return water heat absorption power, and defrosting power consumption within 1 hour. Step 4: Calculate the maximum heating capacity of a single branch at the compressor peak frequency using the dual data-driven model. Round up the predicted total heat load to obtain the minimum number of branches to be opened. Combine this with the maximum number of parallel branches in the system to determine the closed interval of the number of selectable branches. Eliminate invalid branches outside the interval to complete the first pruning. Step 5: Enumerate the number of open branches in the selectable range one by one, calculate the heating task and temperature rise required for each branch, check whether the temperature rise of a single branch exceeds the maximum temperature rise capacity of the unit at the peak frequency, remove the number of branches that do not meet the conditions, and complete the second pruning. Step 6: For the number of open branches retained after the second pruning, enumerate the intermediate water temperature of the two-stage heat pump between the return water temperature and the target supply water temperature with a fixed step size, and divide the temperature rise tasks of the first-stage heat pump and the second-stage heat pump according to the intermediate water temperature. Step 7: Divide the defrost operation domain and the frost operation domain according to the real-time outdoor temperature and humidity. Combine the corresponding data-driven model to solve the minimum operating frequency of the two-stage heat pump to meet the temperature rise and heating tasks. In the frost operation domain, first use the three-dimensional defrost model to complete the heating power and power consumption reduction calculation, then solve the operating frequency, eliminate intermediate water temperature schemes whose operating frequency exceeds the unit frequency range, and complete the third pruning. Step 8: Summarize all enumeration and pruning results to obtain multiple feasible alternative schemes, calculate the system heating redundancy rate and system heating performance coefficient (COP) for each scheme; screen out schemes with heating redundancy rates not lower than a preset threshold, and select the scheme with the highest system COP as the optimal operating scheme. Step 9: The central controller issues control commands according to the optimal operating plan to control the start and stop of the corresponding branch, and directly sets the operating frequency of the compressors of the primary and secondary heat pumps in each branch, so that the unit can operate stably at a fixed frequency.

2. The method for variable frequency coordinated operation control of a multi-stage series air source heat pump according to claim 1, characterized in that: The dual data-driven model described in step three is constructed using the XGBoost algorithm; The training set of the model shall contain no less than 2,000 sets of data, and the test set shall be divided into 20% sets with no less than 500 sets of data. The operating conditions of the training set and the test set shall be consistent. The model evaluation criteria are as follows: the coefficient of determination R² for the frost-free performance model is ≥0.85, the coefficient of determination R² for the three-dimensional defrosting model is ≥0.85, and the mean relative error (MAPE) of both models is no higher than 5%.

3. The method for variable frequency coordinated operation control of a multi-stage series air source heat pump according to claim 1, characterized in that: In step six, the fixed step size for enumerating the intermediate water temperature is 0.5℃; The intermediate water temperature between the two-stage heat pumps is the outlet water temperature of the first-stage heat pump and the inlet water temperature of the second-stage heat pump.

4. The method for variable frequency coordinated operation control of a multi-stage series air source heat pump according to claim 1, characterized in that: In step seven, under the frost-free operating domain, the formula for calculating the effective heating power of the unit after reduction is as follows: ; The formula for calculating the total power consumption of the unit in 1 hour is: ; in, This is the frost-free steady-state heating power. For defrosting time, The heat absorption capacity of the defrost return water. Power consumption for defrosting.

5. The method for variable frequency coordinated operation control of a multi-stage series air source heat pump according to claim 1, characterized in that: The preset minimum threshold for heating redundancy rate in step eight is 5%.

6. The method for variable frequency coordinated operation control of a multi-stage series air source heat pump according to claim 1, characterized in that: The method uses a 1-hour prediction optimization cycle, performs a complete optimization calculation once per hour, and the execution time of a single optimization calculation does not exceed 10 seconds. After three pruning strategies, the overall average computational cost was reduced by 40%.

7. The method for variable frequency coordinated operation control of a multi-stage series air source heat pump according to claim 1, characterized in that: In anti-frost operation mode, the fluctuation value of the system water supply temperature is maintained within 1℃; In frost-prone operation mode, the system water supply temperature fluctuation is less than 2.5℃; The accuracy rate of the three-dimensional defrosting model in determining the frost suppression / frost presence condition is no less than 85%. After adopting this control method, the system's hourly heating performance coefficient (COP) can be increased by up to 20%, the seasonal performance coefficient (SCOP) can be increased by about 15%, and the total energy consumption of the system in a single heating season can be reduced by about 17%.

8. The method for variable frequency coordinated operation control of a multi-stage series air source heat pump according to claim 1, characterized in that: The heating system includes multiple parallel heating branches, circulating water pumps, a main supply pipe, a main return pipe, temperature sensors, flow sensors, outdoor temperature and humidity sensors, and a central controller; each heating branch includes two independent air source heat pumps, with the condensers of the two heat pumps connected in series on the water side, and the return water is heated step by step by a primary heat pump and a secondary heat pump before flowing into the main supply pipe. When the water valve corresponding to the shutdown branch is closed, the total circulating water volume of the system remains constant.

Images