Intelligent control method and system for refrigeration unit based on load dynamic matching

Abstract

This application relates to a method and system for intelligent control of chiller units based on dynamic load matching. The method includes: acquiring multi-dimensional state data during the operation of the chiller unit, and filtering, verifying, and converting the multi-dimensional operating data to obtain standard operating data; analyzing the performance characteristics of the chiller unit under different operating conditions based on the standard operating data to obtain equipment performance characteristic data; inputting the standard operating data into a cooling load prediction model for load change time-series pattern analysis to obtain cooling load prediction data; performing multi-objective collaborative optimization on chilled water temperature setpoint, unit start-up and shutdown status, cooling tower fan frequency, and cooling water pump frequency based on the equipment performance characteristic data, cooling load prediction data, and real-time terminal demand data from the standard operating data to obtain a control optimization scheme; and generating collaborative control commands based on the control optimization scheme. This method can improve the operating energy efficiency and cooling stability of the chiller unit, achieving dynamic load matching.

Description

Technical Field

[0001] This invention belongs to the field of data processing, and in particular relates to a method and system for intelligent control of refrigeration units based on dynamic load matching. Background Technology

[0002] With the development of centralized cooling technology, refrigeration units, as the core cold source equipment in building air conditioning systems, directly affect the energy consumption and economy of the entire cooling system through their operation and control. Currently, the operation and control of refrigeration units mainly rely on manual experience or simple setpoint control methods.

[0003] In traditional technologies, refrigeration units are typically managed by setting fixed chilled water temperature values. The operating parameters of cooling tower fans and cooling water pumps are also mostly controlled by fixed frequencies or simple start-stop mechanisms. There is a lack of coordination and linkage mechanisms between various devices, and control decisions are mainly based on the experience and judgment of maintenance personnel.

[0004] However, traditional methods have the following problems: First, the chilled water temperature setpoint is fixed and cannot be dynamically adjusted according to changes in terminal load, resulting in over-cooling at low loads and under-cooling at high loads, causing the units to deviate from their high-efficiency operating range for a long time; second, there is a lack of scientific start-up and shutdown scheduling strategies among the various chiller units, making it difficult to achieve balanced load distribution and reasonable equipment rotation; third, chiller units, cooling towers, cooling water pumps, and other equipment operate independently, lacking system-level collaborative optimization, resulting in low overall energy efficiency; and fourth, there is a lack of effective prediction of cooling load change trends, and regulation is always in a passive response state, making it difficult to achieve forward-looking optimization regulation. Summary of the Invention

[0005] Therefore, it is necessary to provide a method and system for intelligent control of refrigeration units based on dynamic load matching to address the aforementioned technical problems.

[0006] In a first aspect, this application provides a method for intelligent control of refrigeration units based on dynamic load matching, including:

[0007] S1. Acquire multi-dimensional status data during the operation of the chiller unit, and perform filtering, verification, and unit conversion on the multi-dimensional operating data to obtain standard operating data; among which, the multi-dimensional operating data includes temperature data on the chilled water side, pressure data on the chilled water side, temperature data on the cooling water side, pressure data on the cooling water side, unit electrical parameter data, terminal valve opening data, outdoor meteorological parameter data, and water quality monitoring data;

[0008] S2. Based on standard operating data, analyze the performance characteristics of the refrigeration unit under different operating conditions to obtain equipment performance characteristic data;

[0009] S3. Input the standard operating data into the cooling load prediction model to perform load change time sequence analysis and obtain cooling load prediction data;

[0010] S4. Based on the equipment performance characteristics data, cooling load forecast data and real-time terminal demand data in the standard operation data, multi-objective collaborative optimization is performed on the chilled water temperature setpoint, unit start-up and shutdown status, cooling tower fan frequency and cooling water pump frequency to obtain the control optimization scheme.

[0011] S5. Generate coordinated control instructions based on the control optimization scheme; wherein, the coordinated control instructions are used to control the operating status and operating parameters of each device in the refrigeration unit.

[0012] In one embodiment, S2 includes:

[0013] S21. Based on the chilled water flow rate, chilled water inlet temperature and chilled water outlet temperature in the standard operating data, calculate the cooling capacity of each chiller unit and obtain the cooling capacity data.

[0014] S22. Based on the cooling capacity data and the unit input power in the standard operating data, calculate the performance coefficient of each refrigeration unit under different operating conditions to obtain the performance coefficient data.

[0015] S23. Perform polynomial surface fitting analysis on the relationship between the performance coefficient data and the changes in chilled water outlet temperature, cooling water inlet temperature and partial load rate to obtain the performance coefficient characteristic surface data.

[0016] S24. Based on the performance coefficient characteristic surface data, determine the range of chilled water outlet temperature, cooling water inlet temperature, and partial load rate corresponding to the performance coefficient of each refrigeration unit being higher than the preset high efficiency threshold, and obtain the equipment performance characteristic data.

[0017] In one embodiment, S3 includes:

[0018] S31. Extract the characteristic variables of historical periods from the standard operating data, normalize the characteristic variables, and obtain the normalized characteristic sequence; among them, the characteristic variables include the start-stop status of the air conditioning unit, the temperature of the water distributor, the temperature of the water collector, the outdoor dry bulb temperature, the outdoor relative humidity, the historical cooling load and time characteristics.

[0019] S32. Input the normalized feature sequence into the forget gate of the preset cold load prediction model to selectively forget the historical time series information and obtain the forget gate output data.

[0020] S33. Input the normalized feature sequence and the output data of the forget gate into the input gate of the cold load prediction model, selectively store the new information at the current moment, and obtain the updated unit state data.

[0021] S34. Based on the updated unit state data, the time-series features are filtered and output through the output gate of the cold load prediction model, and nonlinear mapping is performed through a fully connected layer to output the cold load prediction value for future periods, thus obtaining the cold load prediction data.

[0022] In one embodiment, S4 includes:

[0023] S41. Based on the cooling load forecast data and equipment performance characteristic data, perform a matching analysis between the cooling load demand and the high-efficiency operating range of each refrigeration unit, determine the initial set value of chilled water temperature to make the refrigeration unit operate in the high-efficiency range, and obtain the initial chilled water temperature data.

[0024] S42. Based on the terminal valve opening data in the standard operating data, iteratively correct the initial chilled water temperature data to obtain the chilled water temperature setpoint.

[0025] S43. Based on the cooling load forecast data, chilled water temperature setpoint and equipment performance characteristic data, optimize the start-up and shutdown timing and number of operating units of each chiller unit to obtain the unit start-up and shutdown plan.

[0026] S44. Based on the cooling water temperature data in the standard operating data, the frequency of the cooling tower fan and the frequency of the cooling water pump are coordinated and optimized to obtain auxiliary equipment control data.

[0027] S45. Integrate the chilled water temperature setpoint, unit start-up and shutdown scheme, and auxiliary equipment control data to obtain an optimized control scheme.

[0028] In one embodiment, S41 includes:

[0029] S411. Based on the equipment performance characteristic data, determine the range of cooling capacity corresponding to the performance coefficient of each refrigeration unit being higher than the preset high efficiency threshold under the current cooling water inlet temperature, and obtain the high efficiency cooling capacity range data.

[0030] S412. Based on the cooling load demand and high-efficiency cooling capacity range data in the cooling load forecast data, determine whether the cooling load demand falls within the high-efficiency cooling capacity range of a single refrigeration unit, and obtain the load matching judgment result.

[0031] S413. When the load matching judgment result is yes, the optimization objective is to maximize the performance coefficient of the chiller unit. The solution is performed under the conditions of satisfying the cooling capacity constraint and the upper and lower limits of the chilled water temperature to obtain the optimal set value of the chilled water temperature when the unit is running.

[0032] S414. When the load matching judgment result is negative, determine the number of chiller units that need to be operated based on the cooling load demand, and perform load distribution calculation with the goal of balancing the load rate of each operating chiller unit to obtain the initial set value of chilled water temperature when multiple units are running.

[0033] S415. The optimal setpoint for chilled water temperature during single-unit operation or the initial setpoint for chilled water temperature during multi-unit operation shall be combined to form the initial chilled water temperature data.

[0034] In one embodiment, S42 includes:

[0035] S421. Calculate the weighted average valve opening of each terminal air conditioning unit in the standard operating data to obtain the weighted average valve opening data.

[0036] S422. The weighted average valve opening data and the preset valve opening target range are deviated to obtain valve opening deviation data;

[0037] S423. When the weighted average valve opening of the valve opening deviation data is continuously higher than the upper limit of the preset valve opening target range and the duration exceeds the preset time threshold, the chilled water temperature setting value in the initial chilled water temperature data is reduced according to the preset temperature step size to obtain the reduced chilled water temperature data.

[0038] S424. When the weighted average valve opening of the valve opening deviation data is continuously lower than the lower limit of the preset valve opening target range and the duration exceeds the preset time threshold, the chilled water temperature setting value in the initial chilled water temperature data is increased according to the preset temperature step size to obtain the increased chilled water temperature data.

[0039] S425. Perform upper and lower limit constraint verification on the chilled water temperature data for lowering or raising the chilled water temperature data to obtain the chilled water temperature setpoint.

[0040] In one embodiment, S43 includes:

[0041] S431. Based on the cooling load forecast data and the chilled water temperature setpoint, determine whether the currently operating unit meets the preset start-up or shutdown conditions, and obtain the start-up / shutdown trigger signal; wherein, the start-up conditions include the chilled water supply temperature being higher than the setpoint for more than a preset time and the existing operating unit having reached full load, and the shutdown conditions include the operating current of the operating unit being lower than the preset lower limit.

[0042] S432. When the start / stop trigger signal is a start signal, calculate the comprehensive priority score of each chiller unit to be started based on its predicted performance coefficient and cumulative operating time under the current operating conditions, and obtain the unit start-up order data; wherein, the expression for the comprehensive priority score is:

[0043]

[0044] In the formula, For the first The overall priority score of the refrigeration units to be started. For the first The predicted coefficient of performance of the chiller unit to be started under current operating conditions. The maximum predicted coefficient of performance among all chiller units to be started. For the first The cumulative operating time of the refrigeration units awaiting startup This represents the maximum cumulative running time among all chiller units awaiting startup. and These are the weighting coefficients, and ;

[0045] S433. When the start / stop trigger signal is a stop signal, calculate the comprehensive stop priority score for each operating chiller unit based on its load rate and cumulative operating time, and obtain the unit stop ranking data; wherein, the expression for the comprehensive stop priority score is:

[0046]

[0047] In the formula, For the first Overall shutdown priority score for operating refrigeration units in Taiwan. For the first Load rate of the refrigeration units in operation This represents the minimum load rate among all operating refrigeration units. For the first The cumulative operating time of the refrigeration units in operation. This represents the maximum cumulative operating time among all operating refrigeration units. and These are the weighting coefficients, and ;

[0048] S434. Based on the unit start-up sequence data or unit shutdown sequence data, and combined with the anti-vibration constraints of minimum running time and minimum downtime, determine the start-up and shutdown status of each refrigeration unit to obtain the unit start-up and shutdown scheme.

[0049] In one embodiment, S44 includes:

[0050] S441. Perform a deviation analysis between the cooling water outlet temperature in the standard operating data and the preset target range of cooling water temperature to obtain cooling water temperature deviation data.

[0051] S442. Based on the cooling water temperature deviation data, the number of operating units and the operating frequency of the cooling tower fan are adjusted in a stepwise manner to obtain the cooling tower fan control data.

[0052] S443. Perform a deviation analysis between the total inlet and outlet temperature difference of cooling water in the standard operating data and the preset reasonable range of cooling water temperature difference to obtain cooling water temperature difference deviation data.

[0053] S444. Based on the cooling water temperature difference deviation data, the number of operating cooling water pumps and their operating frequency are adjusted in a stepwise manner to obtain cooling water pump control data.

[0054] S445. Integrate the cooling tower fan control data and cooling water pump control data to obtain auxiliary equipment control data.

[0055] Secondly, this application also provides an intelligent control system for refrigeration units based on dynamic load matching, comprising:

[0056] The data acquisition and preprocessing module is used to acquire multi-dimensional status data during the operation of the chiller unit, and to filter, verify and convert the multi-dimensional operating data to obtain standard operating data. The multi-dimensional operating data includes temperature data on the chilled water side, pressure data on the chilled water side, temperature data on the cooling water side, pressure data on the cooling water side, electrical parameter data of the unit, opening data of terminal valves, outdoor meteorological parameter data and water quality monitoring data.

[0057] The equipment performance characteristic analysis module is used to analyze the performance characteristics of the refrigeration unit under different operating conditions based on standard operating data, and obtain equipment performance characteristic data.

[0058] The cooling load time-series prediction module is used to input standard operating data into the cooling load prediction model to analyze the time-series patterns of load changes and obtain cooling load prediction data.

[0059] The multi-objective collaborative optimization module is used to perform multi-objective collaborative optimization on chilled water temperature setpoint, unit start-up and shutdown status, cooling tower fan frequency, and cooling water pump frequency based on equipment performance characteristic data, cooling load prediction data, and real-time terminal demand data in standard operation data, to obtain a control optimization scheme.

[0060] The coordinated control instruction generation module is used to generate coordinated control instructions based on the control optimization scheme; among them, the coordinated control instructions are used to control the operating status and operating parameters of each device in the refrigeration unit.

[0061] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect.

[0062] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.

[0063] The aforementioned intelligent control method and system for chiller units based on dynamic load matching acquires and standardizes multi-dimensional operating data of the chiller units, combines equipment performance characteristic analysis with cooling load prediction, and achieves dynamic adjustment of the chilled water temperature setpoint. This avoids the problems of excessive cooling at low loads and insufficient cooling at high loads, ensuring that the units always operate within their high-efficiency range. Simultaneously, multi-objective collaborative optimization enables coordinated operation between the chiller units, cooling towers, cooling water pumps, and other equipment, addressing the issue of low energy efficiency due to independent operation of individual devices. Cooling load prediction transforms control from passive response to proactive prediction, and unit start-up and shutdown optimization strategies achieve balanced load distribution and rational equipment rotation, effectively reducing system energy consumption, improving cooling stability, extending equipment lifespan, and enhancing the operational economy and intelligence level of the chiller unit system. Attached Figure Description

[0064] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0065] Figure 1 This is a flowchart illustrating an intelligent control method for a refrigeration unit based on dynamic load matching in one embodiment.

[0066] Figure 2 This is a schematic diagram of the structure of a refrigeration unit intelligent control system based on dynamic load matching in one embodiment. Detailed Implementation

[0067] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0068] In one embodiment, reference Figure 1 The document presents a flowchart illustrating the intelligent control method for chiller units based on dynamic load matching provided in this application. This embodiment uses the application of this method to an intelligent control terminal (hereinafter referred to as the terminal) as an example. It is understood that this method can also be applied to a server, and further to a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0069] S1. Acquire multi-dimensional status data during the operation of the refrigeration unit, and perform filtering, verification, and unit conversion on the multi-dimensional operating data to obtain standard operating data.

[0070] Optionally, the multidimensional operating data includes temperature data on the chilled water side, pressure data on the chilled water side, temperature data on the cooling water side, pressure data on the cooling water side, unit electrical parameter data, terminal valve opening data, outdoor meteorological parameter data, and water quality monitoring data;

[0071] For example, the intelligent control terminal collects multi-dimensional status data of the chiller unit in real time through a distributed sensor network. Specifically, chilled water temperature and pressure data can be obtained from thermocouples and pressure transmitters located in the manifold and collector pipes; cooling water parameters can be monitored by similar sensors in the cooling tower circulating water pipeline; unit electrical parameters can be recorded by the unit's built-in electrical parameter acquisition module; terminal valve opening data can be transmitted to the terminal via MODBUS protocol from feedback signals from the electric actuators in each air-conditioned room; outdoor meteorological parameters can be obtained in real time through a group of temperature and humidity sensors installed on the building facade; and water quality monitoring data can be obtained from the conductivity sensor and turbidity detection unit integrated in the cooling water circulation system. The collected data undergoes two stages of preprocessing. First, a sliding window mid-range filtering algorithm is used to eliminate short-term fluctuation noise. Then, all data units are standardized, and an outlier removal mechanism is introduced. When the fluctuation of a certain type of data exceeds a threshold, a manual review procedure is triggered to generate standard operating data.

[0072] S2. Based on standard operating data, analyze the performance characteristics of the refrigeration unit under different operating conditions to obtain equipment performance characteristic data.

[0073] For example, a performance database for the chiller unit is constructed based on preprocessed standard operating data. The formula for calculating cooling capacity is derived using the law of conservation of mass and the energy balance equation. The Coefficient of Performance (COP) is calculated using input power data, establishing a four-dimensional parameter matrix consisting of COP, chilled water outlet temperature, cooling water inlet temperature, and partial load rate. A cubic polynomial surface fitting technique can be used to nonlinearly model the parameter matrix, and the fitting parameters are optimized using the least squares method to obtain a three-dimensional surface model characterizing the equipment's operating efficiency. This model uses partial derivative analysis to determine the boundary conditions of the high-efficiency operating range: when the COP gradient change rate is less than a preset threshold, it is determined to be the high-efficiency operating point. This generates performance characteristic data containing temperature and load rate ranges; the performance characteristic data can support real-time evaluation of the equipment's operating status under dynamic conditions.

[0074] S3. Input the standard operating data into the cooling load prediction model to analyze the time sequence of load changes and obtain the cooling load prediction data.

[0075] For example, the intelligent control terminal can input standard operating data into the cooling load prediction model for analysis of the temporal patterns of load changes, thereby obtaining cooling load prediction data. The cooling load prediction model can employ a Long Short-Term Memory (LSTM) network architecture to achieve deep mining of time-series data. The model input layer can receive normalized feature sequences, which may include air conditioner start / stop status, distributor temperature, collector temperature, outdoor dry-bulb temperature, relative humidity, historical cooling load, and timestamps. The forget gate can use the tanh activation function and sigmoid gating mechanism to filter the importance of historical information and selectively retain key temporal features. The input gate can update the cell state based on new information at the current moment. After nonlinear mapping through a fully connected layer, the output layer can output the predicted cooling load value for future periods using the softmax function. Feature normalization can use the Z-score standardization method to eliminate the influence of different dimensions on model convergence. Model training can use the cross-entropy loss function to optimize the backpropagation process, and an early stopping strategy can be used to prevent overfitting.

[0076] S4. Based on the equipment performance characteristics data, cooling load forecast data, and real-time terminal demand data in the standard operation data, multi-objective collaborative optimization is performed on the chilled water temperature setpoint, unit start-up and shutdown status, cooling tower fan frequency, and cooling water pump frequency to obtain the control optimization scheme.

[0077] For example, the intelligent control terminal can construct a multi-objective optimization framework based on fused data. First, it determines the initial setpoint of chilled water temperature through a load-capacity matching algorithm. When the maximum cooling capacity of a single unit meets the predicted cooling load, it solves for the optimal chilled water outlet temperature with the objective of maximizing COP. When the demand exceeds the capacity of a single unit, a K-means clustering algorithm can be used to distribute the cooling load evenly among multiple units, and the variance of the load rate of each unit is set to not exceed a preset threshold. Subsequently, a terminal valve opening feedback mechanism can be introduced to calculate the weighted average of the electric valve openings in each area. When the valve opening continuously deviates from the target opening range... When the set time window is exceeded, the chilled water temperature setpoint is dynamically adjusted according to the preset step size; the unit start-up and shutdown decision can adopt a two-layer priority evaluation system, which comprehensively evaluates the predicted COP value and cumulative running time of the standby units when starting up; when shutting down, the units are sorted according to the load rate and running time of the operating units; the cooling tower fan frequency is adjusted in segments according to the deviation of the cooling water outlet temperature, and when the temperature exceeds the preset range, the frequency is gradually adjusted according to the frequency increment corresponding to each temperature step size; the number and frequency of the cooling water pumps are controlled according to the deviation between the total cooling water temperature difference and the preset range, and one pump is added or removed and the frequency is adjusted for each temperature difference exceeding the limit.

[0078] S5. Generate coordinated control instructions based on the control optimization scheme.

[0079] Optionally, the coordinated control command is used to control the operating status and operating parameters of each device in the refrigeration unit.

[0080] For example, the intelligent control terminal can convert optimization results into a structured instruction set. The chilled water temperature setpoint is written to each unit controller via the Modbus RTU protocol. Unit start-up and shutdown commands can be transmitted to the PLC control system via industrial Ethernet. Cooling tower fan frequency regulation can be achieved through closed-loop control via a variable frequency drive. Cooling water pump variable frequency speed regulation can be matched to the actual flow demand via a PID controller. The instruction execution process includes a dual verification mechanism: first, it verifies the compatibility between the setpoint and the equipment's allowable range; second, it checks whether the response delay of each device is lower than the system time constant. When execution conflicts occur, priority is given to ensuring the stable operation of critical equipment, while secondary equipment enters standby mode. The entire control process forms a closed-loop feedback chain, and can also achieve data interaction with the building management system via the OPC UA protocol, supporting remote monitoring and fault diagnosis functions.

[0081] The aforementioned intelligent control method for chiller units based on dynamic load matching provides reliable data support for control by filtering, verifying, and standardizing multi-dimensional operating data. Combined with equipment performance characteristic analysis, it enables dynamic adjustment of chilled water temperature setpoints, avoiding over-cooling at low loads and under-cooling at high loads. It achieves proactive control through cooling load prediction, overcoming the limitations of traditional passive response. Through multi-objective collaborative optimization, it realizes coordinated linkage between chiller units, cooling towers, and cooling water pumps, while optimizing unit start-up and shutdown scheduling and load distribution to achieve reasonable equipment rotation and load balance, improving overall system energy efficiency. Finally, it ensures the implementation of the solution by generating collaborative control commands, improving the operating economy and cooling stability of the refrigeration system, reducing equipment losses, and solving problems such as unreasonable control and low energy efficiency caused by traditional manual experience or setpoint control.

[0082] In an optional embodiment, S2 includes:

[0083] S21. Based on the chilled water flow rate, chilled water inlet temperature and chilled water outlet temperature in the standard operating data, calculate the cooling capacity of each chiller unit to obtain the cooling capacity data.

[0084] Optionally, the intelligent control terminal can collect chilled water flow rate, inlet water temperature and outlet water temperature in real time through a distributed sensor network, and derive the cooling capacity using the mass-energy conservation law; the intelligent control terminal has a built-in data verification module, which eliminates instantaneous fluctuation interference through sign consistency check and sliding window filtering algorithm to ensure the stability of the calculation results, and supports parallel calculation of multiple units in the process.

[0085] S22. Based on the cooling capacity data and the unit input power in the standard operating data, calculate the performance coefficient of each refrigeration unit under different operating conditions to obtain the performance coefficient data.

[0086] Optionally, after synchronously acquiring the unit's input power data, the intelligent control terminal can generate the coefficient of performance (COP) by combining the cooling capacity calculation results. By standardizing to eliminate dimensional differences, a correlation between COP and operating parameters is established. At the same time, an outlier detection mechanism can be introduced, triggering a manual review process when a power surge exceeds a set threshold, thus ensuring data reliability.

[0087] S23. Perform polynomial surface fitting analysis on the relationship between performance coefficient data and chilled water outlet temperature, cooling water inlet temperature and partial load rate to obtain performance coefficient characteristic surface data.

[0088] Optionally, the intelligent control terminal can use a nonlinear regression algorithm to model the relationship between the performance coefficient and the chilled water outlet temperature, cooling water inlet temperature, and partial load rate; a high-order mapping model can be constructed using cubic polynomial surface fitting technology, and cross-validation can be used to optimize the parameter weights; the model output includes a gradient sensitivity index; wherein, the gradient sensitivity index can be used to identify the sensitive direction of efficiency changes with parameters.

[0089] S24. Based on the performance coefficient characteristic surface data, determine the range of chilled water outlet temperature, cooling water inlet temperature, and partial load rate corresponding to the performance coefficient of each refrigeration unit being higher than the preset high efficiency threshold, and obtain the equipment performance characteristic data.

[0090] Optionally, the intelligent control terminal generates a three-dimensional decision boundary based on a surface model, defining the operating area where the COP exceeds a preset threshold under various operating conditions. The intelligent control terminal can determine the boundary conditions of the high-efficiency range through gradient analysis and can generate performance characteristic data including temperature range and load rate range; the performance characteristic data can support a dynamic update mechanism, retraining the model with periodic measured data to adapt to performance drift caused by equipment aging or changes in operating conditions.

[0091] In an optional embodiment, S3 includes:

[0092] S31. Extract the feature variables of historical periods from the standard operating data, normalize the feature variables, and obtain the normalized feature sequence.

[0093] Optionally, the characteristic variables include the start / stop status of the air conditioning unit, the temperature of the manifold, the temperature of the collector, the outdoor dry-bulb temperature, the outdoor relative humidity, the historical cooling load, and time characteristics.

[0094] Optionally, the intelligent control terminal can extract characteristic variables such as the start-stop status of the air conditioning unit, the temperature of the water distributor, the temperature of the water collector, the outdoor dry-bulb temperature, the relative humidity, the historical cooling load, and the timestamp from the standard operating data. To eliminate dimensional differences and enhance model convergence, the intelligent control terminal can use normalization processing to scale temperature data to the [0,1] interval by global minimum-maximum scaling, and cooling load and humidity data can be standardized by Z-score. The normalized feature sequence serves as the input basis for the LSTM model, ensuring that different physical quantities are comparable in time series analysis.

[0095] S32. Input the normalized feature sequence into the forget gate of the preset cold load prediction model to selectively forget the historical time series information and obtain the forget gate output data.

[0096] Optionally, the intelligent control terminal inputs the normalized feature sequence into the forget gate module of the LSTM. The forget gate module can generate a forget weight matrix through the sigmoid activation function to selectively retain historical time series information: when the historical data has a low correlation with the current time, the weight approaches zero, realizing "forgetting"; while recent valid data is given a higher weight; the output data of the forget gate can be used as a reference for updating the hidden layer state and eliminating redundant historical interference terms.

[0097] S33. Input the normalized feature sequence and the output data of the forget gate into the input gate of the cold load prediction model, selectively store the new information at the current moment, and obtain the updated unit state data.

[0098] Optionally, the intelligent control terminal can input the normalized feature sequence and the forget gate output data into the input gate module of the LSTM. The input gate can adopt a dual-channel processing mechanism: first, the sigmoid function generates the input weight matrix to filter the new information that needs attention at the current moment; second, the tanh function performs nonlinear encoding on the candidate state vector. After the two are combined, the cell state is updated through element-wise multiplication, so that the model can both inherit historical trends and absorb real-time changes, generating dynamic unit state data containing temporal evolution information.

[0099] S34. Based on the updated unit state data, the time-series features are filtered and output through the output gate of the cold load prediction model, and nonlinear mapping is performed through a fully connected layer to output the cold load prediction value for future periods, thus obtaining the cold load prediction data.

[0100] Optionally, the intelligent control terminal drives the output gate module of the LSTM based on the updated cell state data. The output gate can use the sigmoid function to filter the temporal features of the hidden layer state, retaining only the temporal patterns strongly correlated with the target cooling load. Subsequently, the fully connected layer performs a nonlinear mapping on the filtering results, converting the high-dimensional hidden features into predicted cooling load values. The prediction results are output in time series form, supporting the dynamic decision-making of subsequent control strategies of the terminal. The nonlinear mapping can employ the ReLU activation function.

[0101] In an optional embodiment, S4 includes:

[0102] S41. Based on the cooling load forecast data and equipment performance characteristic data, perform a matching analysis between the cooling load demand and the high-efficiency operating range of each refrigeration unit, determine the initial set value of the chilled water temperature that enables the refrigeration unit to operate in the high-efficiency range, and obtain the initial chilled water temperature data.

[0103] Optionally, the intelligent control terminal can first analyze the cooling load forecast data and equipment performance characteristic data to establish a mapping relationship between cooling load demand and the efficient operating range of the chiller units; it can identify the efficient cooling capacity range of each unit at the current cooling water inlet temperature through gradient analysis; when the maximum cooling capacity of a single unit can cover the forecast cooling load, the optimal chilled water temperature setpoint is solved with the goal of maximizing the coefficient of performance (COP); when the demand exceeds the limit, a clustering algorithm is used to distribute the cooling load evenly to multiple units, and a load rate variance constraint is set, introducing a constraint optimization algorithm to ensure that the temperature setpoint always falls within the boundary of the efficient range.

[0104] S42. Based on the terminal valve opening data in the standard operating data, iteratively correct the initial chilled water temperature data to obtain the chilled water temperature setpoint.

[0105] Optionally, the intelligent control terminal can continuously collect the opening data of the electric valves of the terminal air conditioning units and calculate the overall opening level of the area through a weighted average algorithm. When the weighted opening deviates from the target range for more than a preset time window, a temperature correction mechanism is triggered: if the opening is too high, the chilled water temperature setpoint is reduced step by step according to a preset step size; if the opening is too low, the temperature is adjusted in the opposite direction. The correction process can be embedded with anti-oscillation logic to avoid system fluctuations caused by frequent fine-tuning.

[0106] S43. Based on the cooling load forecast data, chilled water temperature setpoint and equipment performance characteristic data, optimize the start-up and shutdown timing and number of operating units for each refrigeration unit to obtain the unit start-up and shutdown plan.

[0107] Optionally, the intelligent control terminal can integrate cooling load forecasts, current temperature settings, and equipment performance data to construct a two-layer priority evaluation system. During the startup phase, standby units are ranked according to a composite weight of "predicted COP potential × remaining lifespan," prioritizing the activation of high-efficiency and low-wear-risk equipment. During the shutdown phase, operating units are scored in reverse according to "current load rate × cumulative running time," prioritizing the elimination of high-load or overworked units. The anti-vibration mechanism avoids frequent start-stops by setting a minimum switching interval, while reserving redundant capacity to cope with sudden load fluctuations.

[0108] S44. Based on the cooling water temperature data in the standard operating data, the frequency of the cooling tower fan and the frequency of the cooling water pump are coordinated and optimized to obtain auxiliary equipment control data.

[0109] Optionally, the intelligent control terminal can implement graded control based on cooling water temperature data; the cooling tower fan frequency is dynamically adjusted according to the deviation between the outlet water temperature and the target range, adding or removing one fan and matching the corresponding frequency increment when the preset temperature step exceeds the limit; the cooling water pump controls the number and frequency of the control based on the deviation between the total cooling water temperature difference and the reasonable range, adding or removing one pump and adjusting the frequency when the temperature difference exceeds the preset temperature step; the two systems are linked through a closed-loop feedback chain to ensure precise matching between cooling capacity and refrigeration demand.

[0110] S45. Integrate the chilled water temperature setpoint, unit start-up and shutdown scheme, and auxiliary equipment control data to obtain an optimized control scheme.

[0111] Optionally, the intelligent control terminal integrates chilled water temperature setpoints, unit start-up and shutdown plans, and cooling equipment control parameters into a structured instruction set. This set is then distributed to each execution unit via an industrial communication protocol, and a dual verification mechanism is enabled to verify the compatibility of the temperature setpoints with the equipment's allowable range and the matching of the cooling water pump frequency with the pipeline pressure. In the event of an execution conflict, priority is given to ensuring the stable operation of the core chiller unit, while secondary equipment enters standby mode. Simultaneously, an execution log is generated for subsequent fault diagnosis and energy efficiency analysis. The industrial communication protocol can be Modbus / TCP, etc.; the secondary equipment can be a backup cooling tower.

[0112] In an optional embodiment, S41 includes:

[0113] S411. Based on the equipment performance characteristic data, determine the range of cooling capacity corresponding to the coefficient of performance of each refrigeration unit being higher than the preset high efficiency threshold under the current cooling water inlet temperature, and obtain the high efficiency cooling capacity range data.

[0114] Optionally, the intelligent control terminal can first analyze the three-dimensional surface model in the equipment performance characteristic data to extract the cooling capacity range where the COP is higher than the preset threshold under the current cooling water inlet temperature condition. This process can identify the efficiency-sensitive direction through gradient analysis. When moving along a specific parameter direction, such as when moving with an increase in temperature, if the COP decrease rate exceeds the threshold, the region is determined to be an inefficient range. This generates the high-efficiency cooling capacity range data for each unit under the current operating conditions.

[0115] S412. Based on the cooling load demand and high-efficiency cooling capacity range data in the cooling load forecast data, determine whether the cooling load demand falls within the high-efficiency cooling capacity range of a single refrigeration unit, and obtain the load matching judgment result.

[0116] Optionally, the intelligent control terminal can compare and analyze the real-time cooling load forecast value with the high-efficiency cooling capacity range; if the forecast cooling load falls entirely within the high-efficiency cooling capacity range of a single unit, it is determined to be "load matching"; if it exceeds the range, the multi-unit collaborative mode needs to be activated; at this stage, a fuzzy logic judgment mechanism can be introduced to allow for a margin of error to cope with the forecast.

[0117] S413. When the load matching judgment result is yes, the optimization objective is to maximize the performance coefficient of the chiller unit. The solution is performed under the conditions of satisfying the cooling capacity constraint and the upper and lower limits of the chilled water temperature to obtain the optimal set value of the chilled water temperature when the unit is running.

[0118] Optionally, when the intelligent control terminal determines that the load matching result is load matching, it establishes a constrained nonlinear programming model with the optimization objective of maximizing COP, and solves the optimal setpoint of chilled water temperature that satisfies the constraints through an iterative algorithm to ensure that the unit always operates within the boundary of the high-efficiency range.

[0119] S414. When the load matching judgment result is negative, determine the number of chiller units that need to be operated based on the cooling load demand, and perform load distribution calculation with the goal of balancing the load rate of each operating chiller unit to obtain the initial set value of chilled water temperature when multiple units are running.

[0120] Optionally, when the intelligent control terminal determines that the load matching judgment result is a load matching failure, it dynamically calculates the required number of units based on the cooling load demand. It can use a hierarchical clustering algorithm to evenly distribute the total cooling load to multiple units. During the allocation process, a load rate balancing factor is introduced, and a weighted polling mechanism is used to avoid long-term overload operation of a single unit. For each unit participating in operation, the terminal further optimizes its initial set value of chilled water temperature, so that each unit can achieve optimal global energy efficiency while meeting its own high-efficiency range.

[0121] S415. The optimal setpoint for chilled water temperature during single-unit operation or the initial setpoint for chilled water temperature during multi-unit operation shall be combined to form the initial chilled water temperature data.

[0122] Optionally, the intelligent control terminal performs a logical OR operation between the optimal temperature setpoint in single-unit mode and the initial temperature values ​​of each unit in multi-unit mode to generate initial chilled water temperature data in a unified format; the initial chilled water temperature data may include the temperature range and the corresponding operating condition constraints.

[0123] In an optional embodiment, S42 includes:

[0124] S421. Perform a weighted average calculation on the opening degree of the electric regulating valve of each terminal air conditioning unit in the standard operating data to obtain the weighted average valve opening degree data.

[0125] Optionally, the intelligent control terminal can collect the real-time opening signal of the electric regulating valve of each terminal air conditioning unit. The real-time opening signal can be transmitted to the terminal through the Modbus protocol, and the weighted average valve opening can be calculated by combining the area weight.

[0126] S422. The weighted average valve opening data and the preset valve opening target range are deviated to obtain valve opening deviation data.

[0127] Optionally, the intelligent control terminal compares the weighted average valve opening with a preset target range. A sliding window statistical method can be used to calculate the deviation trend over a continuous period, generating quantitative deviation data. The deviation direction is categorized as positive or negative, and the deviation amplitude is standardized to eliminate dimensional differences.

[0128] S423. When the weighted average valve opening of the valve opening deviation data is continuously higher than the upper limit of the preset valve opening target range and the duration exceeds the preset time threshold, the chilled water temperature setting value in the initial chilled water temperature data is reduced according to the preset temperature step size to obtain the reduced chilled water temperature data.

[0129] Optionally, when the intelligent control terminal detects that the weighted average valve opening is continuously higher than the target upper limit and the duration exceeds the preset time window, the terminal triggers a cooling command; the cooling range gradually reduces the initial value of the chilled water temperature according to the preset step size, and an observation period is set after each adjustment to verify the effect; if the opening still cannot be reduced to the target range after multiple rounds of adjustment, redundant control logic is activated to link the fan frequency to fine-tune the cooling.

[0130] S424. When the weighted average valve opening of the valve opening deviation data is continuously lower than the lower limit of the preset valve opening target range and the duration exceeds the preset time threshold, the chilled water temperature setting value in the initial chilled water temperature data is increased according to the preset temperature step size to obtain the increased chilled water temperature data.

[0131] Optionally, when the intelligent control terminal detects that the valve opening is continuously lower than the target lower limit and timeout, the terminal performs a heating operation: gradually increasing the chilled water temperature set value according to the preset step size, and synchronously monitoring the unit current change after each adjustment to prevent overload; if the heating causes a certain unit to operate close to full load, the remaining load is preferentially allocated to other low-load units to avoid the risk of local overheating.

[0132] S425. Perform upper and lower limit constraint verification on the chilled water temperature data for lowering or raising the chilled water temperature data to obtain the chilled water temperature setpoint.

[0133] Optionally, the intelligent control terminal can perform dual constraint verification on the adjusted chilled water temperature setpoint. On the one hand, it ensures that the temperature always falls within the allowable range of the equipment; on the other hand, it verifies the matching with the current cooling water temperature difference and water pump frequency. If a conflict occurs, such as the temperature setpoint exceeding the boundary of the high-efficiency range, it will fall back to the previous level of optimization result and trigger alarm log recording. The final output temperature setpoint is synchronously uploaded to the building management system, supporting remote monitoring and manual intervention.

[0134] In an optional embodiment, S43 includes:

[0135] S431. Based on the predicted cooling load data and the chilled water temperature setpoint, determine whether the currently operating unit meets the preset start-up or shutdown conditions, and obtain the start-up / shutdown trigger signal.

[0136] Optionally, the start-up conditions may include the chilled water supply temperature continuously exceeding the set value for a preset time and the existing operating units reaching full load, while the shutdown conditions may include the operating current of the operating units continuously falling below a preset lower limit.

[0137] Optionally, the intelligent control terminal can trigger start-stop commands by monitoring the chilled water supply temperature and the load status of the operating units in real time; when the chilled water temperature continuously deviates from the set value and the existing units reach the full load threshold, a start signal is generated; if the operating current of a certain unit is continuously lower than the preset lower limit for more than a set time, a shutdown signal is triggered; this process can use the sliding window statistical method to verify the persistence of conditions and introduce redundant judgment logic, so that global adjustment is only performed when both the main control unit and the standby unit meet the conditions, avoiding misoperation caused by local fluctuations.

[0138] S432. When the start-stop trigger signal is a start signal, calculate the comprehensive priority score of each chiller unit to be started based on the predicted performance coefficient and cumulative running time of each chiller unit under the current operating conditions, and obtain the unit start-up order data.

[0139] Alternatively, the expression for the comprehensive priority score can be:

[0140]

[0141] In the above expression, For the first The overall priority score of the refrigeration units to be started. For the first The predicted coefficient of performance of the chiller unit to be started under current operating conditions. The maximum predicted coefficient of performance among all chiller units to be started. For the first The cumulative operating time of the refrigeration units awaiting startup This represents the maximum cumulative running time among all chiller units awaiting startup. and These are the weighting coefficients, and ; The weighting of cumulative operating time is used to prioritize the startup of units with shorter operating times, thereby achieving reasonable unit rotation and extending equipment lifespan. The comprehensive priority score of each unit to be started is calculated, and the units are sorted from high to low to obtain the unit startup order data.

[0142] For example, when the start / stop trigger signal is a start signal, the intelligent control terminal can calculate the comprehensive priority score of each chiller unit to be started based on its predicted performance coefficient and cumulative operating time under the current operating conditions, thus obtaining the unit start-up order data; in the above expression The weighting of the predicted performance coefficient is used to prioritize the startup of units with high energy efficiency. For example, newly started units receive higher priority because their cumulative operating time is close to zero, while long-running units need to compensate for their disadvantages through the predicted performance coefficient.

[0143] S433. When the start / stop trigger signal is a stop signal, calculate the comprehensive stop priority score of each operating chiller unit based on the load rate and cumulative running time of each operating chiller unit to obtain the unit stop ranking data.

[0144] Alternatively, the expression for the comprehensive shutdown priority score can be:

[0145]

[0146] In the above expression, For the first Overall shutdown priority score for operating refrigeration units in Taiwan. For the first Load rate of the refrigeration units in operation This represents the minimum load rate among all operating refrigeration units. For the first The cumulative operating time of the refrigeration units in operation. This represents the maximum cumulative operating time among all operating refrigeration units. and These are the weighting coefficients, and ; The weight used to reflect the load rate is to prioritize the shutdown of the unit with the lowest load rate to avoid energy waste caused by low-load operation; The weighting of cumulative operating time is used to prioritize units with longer downtime, enabling unit rotation. The comprehensive downtime priority score of each operating unit is calculated, and the downtime ranking data is obtained by sorting the scores from high to low.

[0147] Optionally, for units that need to be shut down, the intelligent control terminal constructs a two-dimensional load-life assessment model; the terminal can normalize the load rate. Combined with runtime The model calculates a composite score and prioritizes eliminating those with low workloads. near And it is close to the design life cycle ( near The equipment is configured to retain some units with shorter operating times as a buffer for adjustment. For example, when multiple units have similar load rates, the unit with the shortest cumulative operating time is selected for shutdown first, in order to balance the equipment rotation frequency.

[0148] S434. Based on the unit start-up sequence data or unit shutdown sequence data, and combined with the anti-vibration constraints of minimum running time and minimum downtime, determine the start-up and shutdown status of each refrigeration unit to obtain the unit start-up and shutdown scheme.

[0149] Optionally, the terminal implements dual constraints on the sorting results: a minimum running time is set during startup to avoid repeated start-ups and shutdowns; a minimum stable duration is set during shutdown to prevent system oscillations; when multiple units have the same score, a secondary sorting rule is introduced to ensure the uniqueness of the decision; the final generated start-up and shutdown scheme is sent to the PLC control system through the industrial communication protocol, synchronously updates the unit status database, and triggers the energy consumption early warning module to pre-respond to potential anomalies.

[0150] In an optional embodiment, S44 includes:

[0151] S441. Perform a deviation analysis between the cooling water outlet temperature in the standard operating data and the preset target range of cooling water temperature to obtain cooling water temperature deviation data.

[0152] Optionally, the intelligent control terminal collects the real-time deviation between the cooling tower outlet water temperature and the preset target range. This data can be obtained through a distributed temperature sensor network and filtered by a sliding window to eliminate short-term fluctuation interference. The deviation direction is divided into positive deviation (temperature too high) and negative deviation (temperature too low), and the amplitude is quantified into a standardized deviation value through normalization processing. At this stage, a trend prediction module is introduced, which can determine whether the deviation continues to worsen through a sliding window of historical data.

[0153] S442. Based on the cooling water temperature deviation data, the number of operating units and the operating frequency of the cooling tower fans are adjusted in a stepwise manner to obtain the cooling tower fan control data.

[0154] Optionally, the terminal implements a stepped adjustment strategy based on the cooling water temperature deviation: when the cooling tower outlet water temperature remains above the upper limit and exceeds the timeout period, the number of operating fans is gradually increased by a preset step size, and the corresponding frequency is matched; if the temperature still exceeds the limit, redundant fans are activated and an observation period is set to verify the effect; conversely, when the temperature is below the lower limit, the fan configuration is decreased by a reverse step size. The adjustment process incorporates anti-oscillation logic, and a stable period monitoring system response is set after each adjustment to avoid energy waste caused by frequent switching.

[0155] S443. Perform a deviation analysis on the total inlet and outlet temperature difference of cooling water in the standard operating data and the preset reasonable range of cooling water temperature difference to obtain cooling water temperature difference deviation data.

[0156] Optionally, the intelligent control terminal analyzes the deviation relationship between the total inlet and outlet temperature difference of the cooling water and the preset reasonable range; the temperature difference data can be calculated jointly by the electromagnetic flow meter and the temperature sensor group, and the Kalman filter algorithm can be used to correct the measurement error; the deviation analysis focuses on the direction and magnitude of the deviation of the total inlet and outlet temperature difference of the cooling water from the center value. In the process, a load correlation factor can be introduced to couple the temperature difference deviation with the current cooling load demand for analysis, thereby improving the adjustment accuracy.

[0157] S444. Based on the cooling water temperature difference deviation data, the number of operating cooling water pumps and their operating frequency are adjusted in a stepwise manner to obtain cooling water pump control data.

[0158] Optionally, based on the temperature difference deviation, the intelligent control terminal implements dual-dimensional adjustment of the number and frequency of cooling water pumps. When the cooling water temperature difference deviation data significantly deviates from the reasonable range, the number of pumps is increased or decreased in a step-by-step manner, and the frequency of each pump is dynamically adjusted through the frequency converter. If the abnormal temperature difference is accompanied by a sudden increase in cooling load, the number of pumps is increased first to meet the flow demand; otherwise, the number of operating pumps is reduced and a low-frequency mode is matched. The pipeline pressure is checked simultaneously during the adjustment process to prevent system oscillation caused by over-adjustment.

[0159] S445. Integrate the cooling tower fan control data and cooling water pump control data to obtain auxiliary equipment control data.

[0160] Optionally, the intelligent control terminal can integrate the control commands of cooling tower fans and cooling water pumps into structured data. The structured data can include key information such as equipment number, operating status, and parameter settings. It can be sent to the execution unit through industrial communication protocols and a dual verification mechanism can be enabled to verify the matching of fan frequency and temperature difference correction requirements, and to confirm the consistency of water pump flow rate and cooling load requirements. When execution conflicts occur, a priority scheduling strategy is activated to prioritize the cooling needs of the core refrigeration unit, while secondary equipment enters standby mode and triggers alarm log recording. The final auxiliary equipment control scheme supports remote monitoring and manual intervention, forming a closed-loop feedback chain.

[0161] The aforementioned intelligent control method for refrigeration units based on dynamic load matching acquires and processes multi-dimensional operating data of the refrigeration units through an intelligent control terminal. This analyzes equipment performance characteristics, predicts changes in cooling load, and then performs multi-objective collaborative optimization of chilled water temperature, unit start-up and shutdown, and auxiliary equipment parameters. This effectively solves the problems of fixed chilled water temperature, lack of unit coordination, insufficient load prediction, and passive control in traditional control methods. By dynamically matching the cooling load demand at the terminal, the refrigeration units always operate in the high-efficiency range, achieving coordinated linkage among various devices, reducing situations of over- or under-cooling, balancing unit load distribution, and rationally rotating equipment, thereby improving the overall energy efficiency and cooling stability of the system, reducing equipment losses, achieving forward-looking optimized control, and ensuring the efficient, stable, and economical operation of the refrigeration system.

[0162] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0163] Based on the same inventive concept, this application also provides an intelligent control system for refrigeration units based on dynamic load matching, used to implement the aforementioned intelligent control method for refrigeration units based on dynamic load matching. The solution provided by this system is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the intelligent control system for refrigeration units based on dynamic load matching provided below can be found in the limitations of the intelligent control method for refrigeration units based on dynamic load matching described above, and will not be repeated here.

[0164] In one exemplary embodiment, such as Figure 2 As shown, a schematic diagram of the structure of a refrigeration unit intelligent control system 10 based on dynamic load matching is provided, including:

[0165] The data acquisition and preprocessing module 11 is used to acquire multi-dimensional status data during the operation of the chiller unit, and to filter, verify and convert the multi-dimensional operating data to obtain standard operating data. The multi-dimensional operating data includes temperature data on the chilled water side, pressure data on the chilled water side, temperature data on the cooling water side, pressure data on the cooling water side, electrical parameter data of the unit, opening data of terminal valves, outdoor meteorological parameter data and water quality monitoring data.

[0166] The equipment performance characteristic analysis module 12 is used to analyze the performance characteristics of the refrigeration unit under different operating conditions based on standard operating data, and obtain equipment performance characteristic data.

[0167] The cooling load time-series prediction module 13 is used to input standard operating data into the cooling load prediction model to perform load change time-series pattern analysis and obtain cooling load prediction data.

[0168] The multi-objective collaborative optimization module 14 is used to perform multi-objective collaborative optimization on chilled water temperature setpoint, unit start-up and shutdown status, cooling tower fan frequency and cooling water pump frequency based on equipment performance characteristic data, cooling load prediction data and real-time demand data of the terminal in standard operation data, to obtain a control optimization scheme.

[0169] The coordinated control instruction generation module 15 is used to generate coordinated control instructions based on the control optimization scheme; wherein, the coordinated control instructions are used to control the operating status and operating parameters of each device in the refrigeration unit.

[0170] Furthermore, the equipment performance characteristic analysis module 12 can also be used for:

[0171] S21. Based on the chilled water flow rate, chilled water inlet temperature and chilled water outlet temperature in the standard operating data, calculate the cooling capacity of each chiller unit and obtain the cooling capacity data.

[0172] S22. Based on the cooling capacity data and the unit input power in the standard operating data, calculate the performance coefficient of each refrigeration unit under different operating conditions to obtain the performance coefficient data.

[0173] S23. Perform polynomial surface fitting analysis on the relationship between the performance coefficient data and the changes in chilled water outlet temperature, cooling water inlet temperature and partial load rate to obtain the performance coefficient characteristic surface data.

[0174] S24. Based on the performance coefficient characteristic surface data, determine the range of chilled water outlet temperature, cooling water inlet temperature, and partial load rate corresponding to the performance coefficient of each refrigeration unit being higher than the preset high efficiency threshold, and obtain the equipment performance characteristic data.

[0175] Furthermore, the cooling load time-series prediction module 13 can also be used for:

[0176] S31. Extract the characteristic variables of historical periods from the standard operating data, normalize the characteristic variables, and obtain the normalized characteristic sequence; among them, the characteristic variables include the start-stop status of the air conditioning unit, the temperature of the water distributor, the temperature of the water collector, the outdoor dry bulb temperature, the outdoor relative humidity, the historical cooling load and time characteristics.

[0177] S32. Input the normalized feature sequence into the forget gate of the preset cold load prediction model to selectively forget the historical time series information and obtain the forget gate output data.

[0178] S33. Input the normalized feature sequence and the output data of the forget gate into the input gate of the cold load prediction model, selectively store the new information at the current moment, and obtain the updated unit state data.

[0179] S34. Based on the updated unit state data, the time-series features are filtered and output through the output gate of the cold load prediction model, and nonlinear mapping is performed through a fully connected layer to output the cold load prediction value for future periods, thus obtaining the cold load prediction data.

[0180] Furthermore, the multi-objective collaborative optimization module 14 can also be used for:

[0181] S41. Based on the cooling load forecast data and equipment performance characteristic data, perform a matching analysis between the cooling load demand and the high-efficiency operating range of each refrigeration unit, determine the initial set value of chilled water temperature to make the refrigeration unit operate in the high-efficiency range, and obtain the initial chilled water temperature data.

[0182] S42. Based on the terminal valve opening data in the standard operating data, iteratively correct the initial chilled water temperature data to obtain the chilled water temperature setpoint.

[0183] S43. Based on the cooling load forecast data, chilled water temperature setpoint and equipment performance characteristic data, optimize the start-up and shutdown timing and number of operating units of each chiller unit to obtain the unit start-up and shutdown plan.

[0184] S44. Based on the cooling water temperature data in the standard operating data, the frequency of the cooling tower fan and the frequency of the cooling water pump are coordinated and optimized to obtain auxiliary equipment control data.

[0185] S45. Integrate the chilled water temperature setpoint, unit start-up and shutdown scheme, and auxiliary equipment control data to obtain an optimized control scheme.

[0186] Furthermore, the multi-objective collaborative optimization module 14 can also be used for:

[0187] S411. Based on the equipment performance characteristic data, determine the range of cooling capacity corresponding to the performance coefficient of each refrigeration unit being higher than the preset high efficiency threshold under the current cooling water inlet temperature, and obtain the high efficiency cooling capacity range data.

[0188] S412. Based on the cooling load demand and high-efficiency cooling capacity range data in the cooling load forecast data, determine whether the cooling load demand falls within the high-efficiency cooling capacity range of a single refrigeration unit, and obtain the load matching judgment result.

[0189] S413. When the load matching judgment result is yes, the optimization objective is to maximize the performance coefficient of the chiller unit. The solution is performed under the conditions of satisfying the cooling capacity constraint and the upper and lower limits of the chilled water temperature to obtain the optimal set value of the chilled water temperature when the unit is running.

[0190] S414. When the load matching judgment result is negative, determine the number of chiller units that need to be operated based on the cooling load demand, and perform load distribution calculation with the goal of balancing the load rate of each operating chiller unit to obtain the initial set value of chilled water temperature when multiple units are running.

[0191] S415. The optimal setpoint for chilled water temperature during single-unit operation or the initial setpoint for chilled water temperature during multi-unit operation shall be combined to form the initial chilled water temperature data.

[0192] Furthermore, the multi-objective collaborative optimization module 14 can also be used for:

[0193] S421. Calculate the weighted average valve opening of each terminal air conditioning unit in the standard operating data to obtain the weighted average valve opening data.

[0194] S422. The weighted average valve opening data and the preset valve opening target range are deviated to obtain valve opening deviation data;

[0195] S423. When the weighted average valve opening of the valve opening deviation data is continuously higher than the upper limit of the preset valve opening target range and the duration exceeds the preset time threshold, the chilled water temperature setting value in the initial chilled water temperature data is reduced according to the preset temperature step size to obtain the reduced chilled water temperature data.

[0196] S424. When the weighted average valve opening of the valve opening deviation data is continuously lower than the lower limit of the preset valve opening target range and the duration exceeds the preset time threshold, the chilled water temperature setting value in the initial chilled water temperature data is increased according to the preset temperature step size to obtain the increased chilled water temperature data.

[0197] S425. Perform upper and lower limit constraint verification on the chilled water temperature data for lowering or raising the chilled water temperature data to obtain the chilled water temperature setpoint.

[0198] Furthermore, the multi-objective collaborative optimization module 14 can also be used for:

[0199] S431. Based on the cooling load forecast data and the chilled water temperature setpoint, determine whether the currently operating unit meets the preset start-up or shutdown conditions, and obtain the start-up / shutdown trigger signal; wherein, the start-up conditions include the chilled water supply temperature being higher than the setpoint for more than a preset time and the existing operating unit having reached full load, and the shutdown conditions include the operating current of the operating unit being lower than the preset lower limit.

[0200] S432. When the start-stop trigger signal is a start signal, calculate the comprehensive priority score of each chiller unit to be started based on the predicted performance coefficient and cumulative running time of each chiller unit under the current operating conditions, and obtain the unit start-up order data.

[0201] S433. When the start-stop trigger signal is a shutdown signal, calculate the comprehensive shutdown priority score of each operating chiller unit based on the load rate and cumulative running time of each operating chiller unit to obtain the unit shutdown ranking data.

[0202] S434. Based on the unit start-up sequence data or unit shutdown sequence data, and combined with the anti-vibration constraints of minimum running time and minimum downtime, determine the start-up and shutdown status of each refrigeration unit to obtain the unit start-up and shutdown scheme.

[0203] Furthermore, the multi-objective collaborative optimization module 14 can also be used for:

[0204] S441. Perform a deviation analysis between the cooling water outlet temperature in the standard operating data and the preset target range of cooling water temperature to obtain cooling water temperature deviation data.

[0205] S442. Based on the cooling water temperature deviation data, the number of operating units and the operating frequency of the cooling tower fan are adjusted in a stepwise manner to obtain the cooling tower fan control data.

[0206] S443. Perform a deviation analysis between the total inlet and outlet temperature difference of cooling water in the standard operating data and the preset reasonable range of cooling water temperature difference to obtain cooling water temperature difference deviation data.

[0207] S444. Based on the cooling water temperature difference deviation data, the number of operating cooling water pumps and their operating frequency are adjusted in a stepwise manner to obtain cooling water pump control data.

[0208] S445. Integrate the cooling tower fan control data and cooling water pump control data to obtain auxiliary equipment control data.

[0209] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the intelligent control method for refrigeration units based on dynamic load matching as described above.

[0210] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0211] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0212] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A method for intelligent control of refrigeration units based on dynamic load matching, characterized in that, The method includes: S1. Acquire multi-dimensional status data during the operation of the chiller unit, and filter, verify, and convert the multi-dimensional operating data to obtain standard operating data; wherein, the multi-dimensional operating data includes temperature data on the chilled water side, pressure data on the chilled water side, temperature data on the cooling water side, pressure data on the cooling water side, electrical parameter data of the unit, opening degree data of terminal valves, outdoor meteorological parameter data, and water quality monitoring data; S2. Based on the standard operating data, analyze the performance characteristics of the refrigeration unit under different operating conditions to obtain equipment performance characteristic data; S3. Input the standard operating data into the cooling load prediction model to perform load change time sequence analysis and obtain cooling load prediction data; S4. Based on the equipment performance characteristic data, the cooling load prediction data, and the terminal real-time demand data in the standard operation data, perform multi-objective collaborative optimization on the chilled water temperature setpoint, unit start-up and shutdown status, cooling tower fan frequency, and cooling water pump frequency to obtain a control optimization scheme. S5. Generate coordinated control instructions based on the control optimization scheme; wherein, the coordinated control instructions are used to control the operating status and operating parameters of each device in the refrigeration unit.

2. The method according to claim 1, characterized in that, S2 includes: S21. Based on the chilled water flow rate, chilled water inlet temperature and chilled water outlet temperature in the standard operating data, calculate the cooling capacity of each chiller unit to obtain cooling capacity data; S22. Based on the cooling capacity data and the unit input power in the standard operating data, calculate the performance coefficient of each refrigeration unit under different operating conditions to obtain performance coefficient data. S23. Perform polynomial surface fitting analysis on the relationship between the performance coefficient data and the changes in chilled water outlet temperature, cooling water inlet temperature and partial load rate to obtain performance coefficient characteristic surface data. S24. Based on the performance coefficient characteristic surface data, determine the range of chilled water outlet temperature, cooling water inlet temperature, and partial load rate corresponding to the performance coefficient of each refrigeration unit being higher than the preset high efficiency threshold, and obtain the equipment performance characteristic data.

3. The method according to claim 1, characterized in that, S3 includes: S31. Extract historical period feature variables from the standard operating data, normalize the feature variables to obtain a normalized feature sequence; wherein, the feature variables include the start-stop status of the air conditioning unit, the temperature of the water distributor, the temperature of the water collector, the outdoor dry bulb temperature, the outdoor relative humidity, the historical cooling load and time characteristics. S32. Input the normalized feature sequence into the forget gate of the preset cold load prediction model to selectively forget the historical time series information and obtain the forget gate output data. S33. Input the normalized feature sequence and the forget gate output data into the input gate of the cold load prediction model, selectively store the new information at the current moment, and obtain the update unit state data. S34. Based on the status data of the update unit, the time-series features are filtered and output through the output gate of the cold load prediction model, and nonlinear mapping is performed through a fully connected layer to output the cold load prediction value for future periods, thereby obtaining the cold load prediction data.

4. The method according to claim 1, characterized in that, S4 includes: S41. Based on the predicted cooling load data and the equipment performance characteristic data, perform a matching analysis between the cooling load demand and the high-efficiency operating range of each refrigeration unit, determine the initial set value of the chilled water temperature that enables the refrigeration unit to operate in the high-efficiency range, and obtain the initial chilled water temperature data. S42. Based on the terminal valve opening data in the standard operating data, the initial chilled water temperature data is iteratively corrected to obtain the chilled water temperature setpoint. S43. Based on the cooling load forecast data, the chilled water temperature setpoint and the equipment performance characteristic data, optimize the start-up and shutdown timing and number of operating units of each refrigeration unit to obtain the unit start-up and shutdown scheme. S44. Based on the cooling water temperature data in the standard operating data, the frequency of the cooling tower fan and the frequency of the cooling water pump are coordinated and optimized to obtain auxiliary equipment control data. S45. Integrate the chilled water temperature setpoint, the unit start-up and shutdown scheme, and the auxiliary equipment control data to obtain the control optimization scheme.

5. The method according to claim 4, characterized in that, S41 includes: S411. Based on the equipment performance characteristic data, determine the range of cooling capacity corresponding to the performance coefficient of each refrigeration unit being higher than the preset high efficiency threshold under the current cooling water inlet temperature, and obtain the high efficiency cooling capacity range data. S412. Based on the cooling load demand and the high-efficiency cooling capacity range data in the cooling load forecast data, determine whether the cooling load demand falls within the high-efficiency cooling capacity range of a single refrigeration unit, and obtain the load matching judgment result. S413. When the load matching judgment result is yes, the optimal set value of chilled water temperature during single-unit operation is obtained by solving the problem under the conditions of satisfying the cooling capacity constraint and the upper and lower limit constraints of chilled water temperature, with the maximization of the performance coefficient of the chiller unit as the optimization objective. S414. When the load matching judgment result is negative, determine the number of chiller units that need to be operated based on the cooling load demand, and perform load distribution calculation with the goal of balancing the load rate of each operating chiller unit to obtain the initial set value of chilled water temperature when multiple units are running. S415. The optimal setting value of chilled water temperature during single-unit operation or the initial setting value of chilled water temperature during multi-unit operation shall be used to form the initial chilled water temperature data.

6. The method according to claim 4, characterized in that, S42 includes: S421. Calculate the weighted average valve opening of each terminal air conditioning unit in the standard operating data to obtain the weighted average valve opening data. S422. The weighted average valve opening data and the preset valve opening target range are deviated to obtain valve opening deviation data; S423. When the weighted average valve opening of the valve opening deviation data is continuously higher than the upper limit of the preset valve opening target range and the duration exceeds the preset time threshold, the chilled water temperature setting value in the initial chilled water temperature data is reduced according to the preset temperature step size to obtain the reduced chilled water temperature data. S424. When the weighted average valve opening of the valve opening deviation data is continuously lower than the lower limit of the preset valve opening target range and the duration exceeds the preset time threshold, the chilled water temperature setting value in the initial chilled water temperature data is increased according to the preset temperature step size to obtain the increased chilled water temperature data. S425. Perform chilled water temperature upper and lower limit constraint verification on the lowered chilled water temperature data or the uppered chilled water temperature data to obtain the chilled water temperature set value.

7. The method according to claim 4, characterized in that, S43 includes: S431. Based on the predicted cooling load data and the set value of chilled water temperature, determine whether the currently operating unit meets the preset start-up or shutdown conditions, and obtain a start-up / shutdown trigger signal; wherein, the start-up conditions include the chilled water supply temperature being continuously higher than the set value for more than a preset time and the existing operating unit having reached full load, and the shutdown conditions include the operating current of the operating unit being continuously lower than a preset lower limit value. S432. When the start / stop trigger signal is a start signal, the comprehensive priority score of each chiller unit to be started is calculated based on the predicted performance coefficient and cumulative operating time of each chiller unit under the current operating conditions, thus obtaining the unit start-up order data; wherein, the expression for the comprehensive priority score is: ； In the formula, For the first The overall priority score of the refrigeration units to be started. For the first The predicted coefficient of performance of the chiller unit to be started under current operating conditions. The maximum predicted coefficient of performance among all chiller units to be started. For the first The cumulative operating time of the refrigeration units awaiting startup This represents the maximum cumulative running time among all chiller units awaiting startup. and These are the weighting coefficients, and ; S433. When the start / stop trigger signal is a shutdown signal, the comprehensive shutdown priority score of each operating chiller unit is calculated based on the load rate and cumulative operating time of each operating chiller unit to obtain the unit shutdown ranking data; wherein, the expression for the comprehensive shutdown priority score is: ； In the formula, For the first Overall shutdown priority score for operating refrigeration units in Taiwan. For the first Load rate of the refrigeration units in operation This represents the minimum load rate among all operating refrigeration units. For the first The cumulative operating time of the refrigeration units in operation. This represents the maximum cumulative operating time among all operating refrigeration units. and These are the weighting coefficients, and ; S434. Based on the unit start-up sequence data or the unit shutdown sequence data, and combined with the anti-vibration constraints of minimum running time and minimum downtime, determine the start-up and shutdown status of each refrigeration unit to obtain the unit start-up and shutdown scheme.

8. The method according to claim 4, characterized in that, S44 includes: S441. Perform a deviation analysis between the cooling water outlet temperature in the standard operating data and the preset cooling water temperature target range to obtain cooling water temperature deviation data. S442. Based on the cooling water temperature deviation data, the number of operating units and the operating frequency of the cooling tower fan are adjusted in a stepwise manner to obtain cooling tower fan control data. S443. Perform a deviation analysis on the total inlet and outlet temperature difference of cooling water in the standard operating data and the preset reasonable range of cooling water temperature difference to obtain cooling water temperature difference deviation data. S444. Based on the cooling water temperature difference deviation data, the number of operating cooling water pumps and their operating frequency are adjusted in a stepwise manner to obtain cooling water pump control data. S445. Integrate the cooling tower fan control data and the cooling water pump control data to obtain the auxiliary equipment control data.

9. A smart control system for refrigeration units based on dynamic load matching, characterized in that, The system includes: The data acquisition and preprocessing module is used to acquire multi-dimensional status data during the operation of the chiller unit, and to filter, verify, and convert the multi-dimensional operating data to obtain standard operating data. The multi-dimensional operating data includes temperature data on the chilled water side, pressure data on the chilled water side, temperature data on the cooling water side, pressure data on the cooling water side, electrical parameter data of the unit, opening data of terminal valves, outdoor meteorological parameter data, and water quality monitoring data. The equipment performance characteristic analysis module is used to analyze the performance characteristics of the refrigeration unit under different operating conditions based on the standard operating data, and obtain equipment performance characteristic data. The cooling load time-series prediction module is used to input the standard operating data into the cooling load prediction model to perform load change time-series pattern analysis and obtain cooling load prediction data. The multi-objective collaborative optimization module is used to perform multi-objective collaborative optimization on chilled water temperature setpoint, unit start-up and shutdown status, cooling tower fan frequency, and cooling water pump frequency based on the equipment performance characteristic data, the cooling load prediction data, and the terminal real-time demand data in the standard operation data, to obtain a control optimization scheme. The coordinated control instruction generation module is used to generate coordinated control instructions based on the control optimization scheme; wherein, the coordinated control instructions are used to control the operating status and operating parameters of each device in the refrigeration unit.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 8.

Images