Central air conditioner intelligent identification control system and method

Abstract

The present application belongs to the technical field of building HVAC intelligent control, and discloses a central air conditioning intelligent identification control system and method. The method collects operation data, generates and executes a perturbation identification instruction in a stable operation state; deducts the baseline from the operation data before and after the perturbation to obtain a net temperature response, a net pressure difference response and a net energy consumption response; extracts a thermal response fingerprint according to the net response quantity, establishes an actual association matrix between the execution object and the controlled area; compares the actual association matrix with a drawing association table or an installation registration table, identifies the end mismatch, the weak response area and the abnormal coupling area; corrects the area load demand according to the identification result, generates a candidate control quantity, and performs constraint projection under the constraints of equipment safety, comfort and actual association matrix to output the final control instruction. The present application is suitable for the intelligent identification, operation regulation and control and energy saving control of the central air conditioning system.

Description

Technical Field

[0001] This invention belongs to the field of building heating, ventilation, air conditioning and intelligent control technology, specifically relating to a central air conditioning intelligent identification control system and method based on end-point perturbation identification, thermal response fingerprint and actual correlation matrix. Background Technology

[0002] Central air conditioning systems typically include chilled water pumps, cooling water pumps, fans, air valves, terminal valves, sensors, and multiple controlled areas. Their operation is affected by factors such as building load, occupant activity, outdoor weather, terminal installation status, and hydraulic-wind coupling. With increasing demands for energy-efficient operation in public buildings, central air conditioning systems are gradually evolving from manual settings, timed start-stop systems, and single-parameter feedback control to intelligent optimization control based on multi-source operational data.

[0003] For example, prior art document CN101424436B discloses a central air conditioning intelligent optimization control system and method. This system acquires operating parameters of each system through sensors, uses an intelligent controller to dynamically analyze and predict user cooling demand, and controls system equipment operation based on the load prediction results to achieve optimized control of the central air conditioning system. The document also points out that existing central air conditioning energy-saving technologies, such as variable frequency water pumps and fuzzy control of chilled water systems, are mostly concentrated on individual equipment or local systems, lacking global considerations and easily leading to localized energy savings while limiting overall energy efficiency.

[0004] The aforementioned technologies can improve the overall operation strategy of central air conditioning systems through parameter acquisition and load forecasting. However, their focus remains on adjusting equipment operating parameters based on predicted loads or empirical strategies, typically assuming that the "equipment-zone" correspondence recorded in drawings, installation registers, or control logic is accurate. In actual engineering projects, after long-term operation, renovation, maintenance, or terminal replacement, deviations may occur in the correspondence between valves, dampers, fan coil units, or zone sensors. Examples include incorrect terminal wiring, inconsistencies between valve control zones and registered zones, weakened response of a particular terminal, and abnormal coupling between multiple zones. If the control system still relies on the original drawing relationships for load allocation and command output, even with relatively accurate load forecasting, problems such as mismatched controlled objects, localized overcooling or overheating, frequent actuator adjustments, and increased energy consumption may arise.

[0005] Therefore, it is necessary to provide an intelligent control method that can proactively identify actual control relationships during the operation of central air conditioning systems. This method enables the system to not only sense operational data such as temperature, pressure difference, and energy consumption, but also to identify the true response relationship between the execution object and the controlled area through small-amplitude controllable perturbations, baseline subtraction, and response fingerprint analysis. Based on this, load correction and control command constraint generation can be performed to improve the control accuracy and operational stability of central air conditioning systems in complex retrofitting, weak response, and abnormal coupling scenarios. Summary of the Invention

[0006] The technical objective of this invention is to provide a central air conditioning intelligent identification control system and method, which establishes the actual correlation between the execution object and the controlled area through perturbation identification and thermal response fingerprinting, and performs load correction and constraint control based on the actual correlation.

[0007] To achieve the above-mentioned technical objectives, the present invention provides the following technical solution.

[0008] In a first aspect, the present invention discloses a central air conditioning intelligent identification and control method, comprising the following steps:

[0009] S1. Collect operating data of the cold source, water pump, fan, terminal valve, air valve and each controlled area in the central air conditioning system. The operating data includes temperature and humidity, personnel status, energy consumption, supply and return water temperature difference, pipeline pressure difference and terminal opening.

[0010] S2. When the central air conditioning system is in a stable operating state, apply a perturbation identification command to at least one terminal valve or air valve in a preset sequence, and record the temperature change rate, pressure difference change and energy consumption change of each controlled area in the perturbation window.

[0011] S3. Obtain the temperature change trend, pressure difference change trend and energy consumption change trend within the stability window before the perturbation. Perform baseline subtraction on the response data within the perturbation window to obtain the net temperature response, net pressure difference response and net energy consumption response caused by the perturbation identification command. Generate a thermal response fingerprint based on the net temperature response, net pressure difference response, net energy consumption response and response start lag time, and establish the actual correlation matrix between the execution object and the controlled area.

[0012] S4. Compare the actual correlation matrix with the pre-stored drawing correlation table or installation registration table to identify end-point mismatch, weak response area and abnormal coupling area, and use the identification results as the basis for correcting the area load demand and subsequent control command allocation.

[0013] S5. Generate candidate control quantities based on the corrected regional load demand and the actual correlation matrix, and project the candidate control quantities onto the feasible constraint set under the constraints of equipment safety, comfort and actual correlation matrix to obtain the final control commands for cold source output, water pump frequency, fan frequency, terminal valve opening and air valve opening.

[0014] Specifically, in step S1, the personnel status includes one or more of the following: personnel number, personnel density, personnel stay time, and area occupancy status.

[0015] The personnel status is obtained by at least one of the following: camera, infrared sensor, access control record, wireless terminal access record, or carbon dioxide concentration change data;

[0016] The terminal opening degree includes the terminal water valve opening degree and / or air valve opening degree, and the pipeline pressure difference includes one or more of the terminal branch pressure difference, the supply and return water main pressure difference, or the pressure difference before and after the terminal.

[0017] Specifically, in step S2, the perturbation identification instruction includes applying an amplitude change to the opening degree of the end water valve, the opening degree of the air valve, the frequency of the fan, or the frequency of the water pump for a duration not exceeding one control cycle.

[0018] The amplitude change is 5% to 15% of the current adjustable range of the corresponding execution object, and does not exceed the preset safety range, so that the comfort deviation of the controlled area does not exceed the preset allowable range;

[0019] The stable operating state is defined as follows: within the stable window before the perturbation identification command is applied, the rate of change of temperature in the controlled area, the change of pipeline differential pressure, and the change of terminal opening are all within the corresponding stable threshold range.

[0020] Specifically, in step S3, the thermal response fingerprint includes at least the response start lag time, peak temperature change rate, temperature recovery time, pipeline pressure difference change, end opening change, and regional energy consumption change.

[0021] The thermal response fingerprint is written into the thermal response fingerprint database according to the execution object number and the controlled area number;

[0022] When the same execution object generates responses in multiple controlled regions, the response start lag time and net temperature response intensity of each controlled region are recorded to distinguish between the main controlled region, the weakly controlled region, and the coupled controlled region.

[0023] Specifically, in step S3, the actual correlation matrix is ​​A, and the elements of the actual correlation matrix A are... This represents the strength of the influence of the i-th executing object on the j-th controlled region. satisfy:

[0024] ;

[0025] In the formula, For the first The execution object for the first The intensity of the impact on each controlled area; This represents the normalized net temperature response intensity. The normalized net pressure difference response intensity; This represents the normalized net energy consumption response intensity. The normalized response start lag time; , , , These are the weighting coefficients.

[0026] when When the value is greater than the first association threshold, the i-th execution object is determined to be the main associated object of the j-th controlled region; when When the value falls between the first and second association thresholds, the i-th execution object is determined to be a weakly associated object of the j-th controlled region; when... When the value is less than the second association threshold, the i-th execution object is determined to be a non-associative object of the j-th controlled region.

[0027] Specifically, in step S4, identifying end-mismatch, weak response regions, and abnormal coupling regions includes:

[0028] When the drawing association table or installation registration table records that the i-th execution object corresponds to the j-th controlled area, but in the actual association matrix... Less than the second association threshold, and there exists another controlled region k such that When the value exceeds the first association threshold, it is determined that the i-th execution object has an end-point mismatch;

[0029] When all execution objects corresponding to the controlled region j If all values ​​are below the first correlation threshold, and the temperature deviation of the controlled area continues to exceed the comfort threshold within a preset time, the controlled area is determined to be a weak response area.

[0030] When the influence intensity of at least two executing objects on the same controlled area is greater than the first association threshold, and the difference in their response start lag time is less than the preset lag threshold, it is determined that there is abnormal coupling in the controlled area.

[0031] Specifically, in step S5, the set of feasible constraints includes equipment safety constraints, comfort constraints, and actual correlation matrix constraints;

[0032] The equipment safety constraints include upper and lower limits for cold source output, upper and lower limits for water pump frequency, upper and lower limits for fan frequency, upper and lower limits for terminal valve opening, and upper and lower limits for air valve opening.

[0033] The comfort constraints include the permissible deviation range of temperature in the controlled area, the permissible deviation range of humidity, and the upper limit of the rate of temperature change.

[0034] The actual correlation matrix constraints include: prioritizing the execution objects that are primarily associated with the target controlled area as adjustment objects, restricting non-associated objects from participating in the load adjustment of the target controlled area, and setting a collaborative adjustment ratio for multiple execution objects in abnormally coupled areas;

[0035] When projecting candidate control variables onto the set of feasible constraints, the final control command is obtained according to the objective of minimizing the correction magnitude of candidate control variables, regional temperature deviation, and system energy consumption increment.

[0036] Furthermore, it also includes step S6: after executing step S5, continue to collect response data of each controlled area, and write back the deviation between the actual response and the predicted response to the thermal response fingerprint database.

[0037] When the response deviation between the same execution object and the same controlled area exceeds a preset number of consecutive times, steps S2 to S3 are retried to update the hot response fingerprint and the actual correlation matrix.

[0038] When the difference between the updated actual association matrix and the unupdated actual association matrix exceeds a preset matrix difference threshold, step S4 is re-executed to update the identification results of end mismatch, weak response region and abnormal coupling region.

[0039] Secondly, the present invention also discloses a central air conditioning intelligent identification and control system for implementing the central air conditioning intelligent identification and control method described in the first aspect, the system comprising:

[0040] The data acquisition module is used to collect operating data from the cold source, water pump, fan, terminal valve, air valve, and each controlled area in the central air conditioning system.

[0041] The perturbation identification module is used to apply perturbation identification commands to at least one terminal valve or air valve in a preset sequence when the central air conditioning system is in a stable operating state.

[0042] The response fingerprint generation module is used to perform baseline subtraction on the response data within the perturbation window, generate a thermal response fingerprint, and establish an actual correlation matrix between the execution object and the controlled area.

[0043] The anomaly identification module is used to compare the actual correlation matrix with the pre-stored drawing correlation table or installation registration table to identify end-point mismatch, weak response areas and abnormal coupling areas.

[0044] The control optimization module is used to generate candidate control quantities based on the corrected regional load demand and the actual correlation matrix, and project the candidate control quantities onto the feasible constraint set under equipment safety constraints, comfort constraints and actual correlation matrix constraints to obtain the final control command.

[0045] Specifically, the response fingerprint generation module includes a baseline subtraction unit, a response feature extraction unit, and an association matrix generation unit;

[0046] The baseline subtraction unit is used to subtract the temperature change trend, pressure difference change trend, and energy consumption change trend within the stability window before the perturbation.

[0047] The response feature extraction unit is used to extract net temperature response, net pressure difference response, net energy consumption response, response start lag time, peak temperature change rate, and temperature recovery time.

[0048] The correlation matrix generation unit is used to calculate the influence strength between the execution object and the controlled area based on the response characteristics, and generate the actual correlation matrix.

[0049] The control optimization module includes a candidate control quantity generation unit and a constraint projection correction unit;

[0050] The candidate control quantity generation unit is used to generate initial control commands based on the regional load demand.

[0051] The constraint projection correction unit is used to correct the initial control command into a final control command that satisfies the equipment safety constraints, comfort constraints, and actual correlation matrix constraints.

[0052] The system also includes a feedback update module, which is used to write back the deviation between the actual response and the predicted response to the thermal response fingerprint database, and to trigger the update of the thermal response fingerprint and the actual correlation matrix when the response deviation exceeds a preset number of consecutive times.

[0053] This invention applies small-amplitude perturbation identification commands to actuators such as valves, air valves, water pumps, fans, or cold sources under stable operation of a central air conditioning system. Baseline subtraction is performed on the temperature, pressure difference, and energy consumption data before and after the perturbation to obtain the net response after eliminating the influence of natural load fluctuations. By extracting features such as response initiation lag time, peak temperature change rate, pressure difference change amplitude, energy consumption change amplitude, and recovery time, a thermal response fingerprint is formed, and an actual correlation matrix between the actuator and the controlled area is established. By comparing the actual correlation matrix with the drawing correlation table or installation registration table, terminal mismatch, weak response areas, and abnormal coupling areas can be identified, avoiding the continued output of control commands based on erroneous or distorted engineering registration relationships. During the control phase, this invention corrects the regional load demand based on the identification results and projects candidate control quantities into a set of feasible constraints jointly defined by equipment safety constraints, comfort constraints, and actual correlation matrix constraints. This ensures that the final control command matches the actual response relationship on site, reducing the risks of control mismatch, over-adjustment, and local comfort fluctuations, and improving the control accuracy, operational stability, and energy-saving operation capability of the central air conditioning system. Attached Figure Description

[0054] Figure 1 This is a schematic diagram of the intelligent identification and control method for central air conditioning according to the present invention;

[0055] Figure 2 This is a schematic diagram of the perturbation identification, actual correlation matrix, and constraint control closed loop of the present invention;

[0056] Figure 3 This is a schematic diagram of the central air conditioning intelligent identification control system of the present invention. Detailed Implementation

[0057] The embodiments of the present invention will be further described below with reference to the accompanying drawings and examples. It should be understood that the following embodiments are used to illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention. For those skilled in the art, equivalent substitutions made to the sensor type, communication method, control cycle, threshold value, number of execution objects, building zoning method, and equipment model without departing from the technical solutions of the present invention can all be used to implement the present invention.

[0058] The central air conditioning system described in this embodiment can be applied to building scenarios such as office buildings, commercial complexes, hospitals, schools, data center auxiliary areas, and industrial clean areas. The central air conditioning system includes chilled water equipment, water pumps, cooling towers or heat exchange equipment, air handling units, fan coil units, variable air volume (VAV) terminals, terminal water valves, air valves, zone temperature and humidity sensors, pipeline differential pressure sensors, energy consumption metering units, and a central controller. For water systems, the actuators can be chilled water valves, water pumps, fan coil unit fans, or air handling unit water valves; for air systems, the actuators can be VAV terminal air valves, supply fan frequencies, or reheat devices; for water-air hybrid systems, the actuators can simultaneously include water-side actuators and air-side actuators.

[0059] I. Definitions and Explanations of Terms

[0060] A controlled area refers to a space served by one or more central air conditioning terminals, such as an office area, meeting room, ward, corridor, business area, or equipment room.

[0061] The execution object refers to the equipment or component that can receive control commands and change the distribution of cooling capacity, air volume or water volume, such as terminal valves, air valves, water pumps, fans or cold source output regulating units.

[0062] The stability window refers to the time window used to determine whether the system is in a stable operating state before the perturbation identification command is applied.

[0063] The perturbation window refers to the time window used to collect response data after the executing object receives the perturbation identification command.

[0064] Thermal response fingerprint refers to a data combination consisting of features such as net temperature response, net pressure difference response, net energy consumption response, response start lag time, peak temperature change rate, and temperature recovery time of the controlled area after a small-amplitude, short-duration perturbation identification command is applied to the execution object.

[0065] The actual correlation matrix is ​​a matrix composed of the actual influence strength of each executing object on each controlled area, used to characterize the real service relationship on site.

[0066] This implementation method employs rule-based computation, matrix association, and constraint optimization, eliminating the need for pre-trained neural network models and dependence on specific training datasets. If a machine learning prediction module is introduced into the engineering implementation, it can serve as an auxiliary module for temperature prediction, load prediction, or anomaly identification, without affecting the basic implementation of perturbation identification, thermal response fingerprint generation, actual correlation matrix establishment, and constraint control in this embodiment.

[0067] II. Specific Implementation Steps of the Method of the Invention

[0068] like Figure 1 As shown, a central air conditioning intelligent identification and control method includes the following steps.

[0069] 2.1 Operational Data Acquisition

[0070] S1. Collect operating data of the cold source, water pump, fan, terminal valve, air valve and each controlled area in the central air conditioning system. The operating data includes temperature and humidity, personnel status, energy consumption, supply and return water temperature difference, pipeline pressure difference and terminal opening.

[0071] In this embodiment, the central controller collects various operational data through a building automation network. The building automation network can use BACnet, Modbus, LonWorks, KNX, OPC UA, or Ethernet communication methods. The collected data includes at least the temperature of each controlled area. relative humidity Personnel status carbon dioxide concentration End valve opening Air valve opening Fan frequency chilled water supply temperature Return water temperature Supply and return water pressure difference chilled water flow rate Cooling source power Water pump power Fan power and energy consumption by item .

[0072] In the formula, Indicates the sequence number of the object to be executed; Indicates the controlled area number; Indicates the sampling time.

[0073] If multiple temperature sensors are installed within the controlled area, the weighted average temperature can be used as the temperature of that area. The weights can be determined based on sensor location, historical stability, or effective coverage area. If personnel status is obtained from records accessed by cameras, infrared sensors, access control systems, wireless terminals, or calculated from changes in carbon dioxide concentration, the central controller can uniformly convert personnel status into occupancy status, number of personnel, personnel density, or personnel dwell time.

[0074] The terminal opening degree includes the terminal water valve opening degree and / or air valve opening degree, and the pipeline pressure difference includes one or more of the following: terminal branch pressure difference, supply and return main water pressure difference, pressure difference before and after the terminal, or duct static pressure. For systems that do not have a separate pressure difference sensor installed in each controlled area, the terminal branch pressure difference, main water pressure difference, and terminal opening degree changes can be combined to estimate the pressure difference response of the corresponding controlled area or terminal branch.

[0075] During data acquisition, the central controller performs basic preprocessing on the acquired data, including timestamp alignment, missing value imputation, glitch removal, unit standardization, and validity verification. For continuous data, sliding median filtering or amplitude limiting filtering can be used; for state data, state preservation methods can be used to compensate for short-term communication interruptions. For temperature, pressure difference, and energy consumption data, the sampling period can be standardized before entering the subsequent identification process, for example, to a sampling interval of 30s, 60s, or 120s.

[0076] 2.2 Perturbation Identification Command Generation and Secure Execution

[0077] S2. When the central air conditioning system is in a stable operating state, apply perturbation identification commands to at least one terminal valve or air valve in a preset sequence, and record the temperature change rate, pressure difference change and energy consumption change of each controlled area in the perturbation window.

[0078] Before performing perturbation identification, the central controller first determines whether the central air conditioning system is in a stable operating state. The stable operating state refers to the stable window before the perturbation identification command is applied. Within this range, the controlled area's temperature change rate, pipeline pressure difference change, terminal opening change, and system power fluctuation all remain within their respective stability threshold ranges. For example, the stability window... The timeframe can be 5–15 minutes, preferably 10 minutes; the temperature change rate in the controlled area should not exceed 0.05℃ / min, the pipeline pressure difference change should not exceed 5% of the average pressure difference in the stable window, the change in the opening degree of the terminal valve or damper should not exceed 3%, and the cold source, water pump, and fan should not be in a start / stop switching state. If the above conditions are not met, perturbation identification should be paused, and conventional control should continue until the system returns to a stable operating state.

[0079] This step is used to identify the actual impact range of a given controllable area on each controlled area without significantly affecting personnel comfort and equipment safety. Because engineering sites often have issues such as incorrect end-point numbering, discrepancies between drawings and the actual site, reversed valve installation, duct crosstalk, hydraulic imbalance, or sensor installation misalignment, relying solely on installation records for control can easily lead to a mismatch between the controlled area and the actual controllable area. This step obtains the true on-site response relationship by making short-term, small adjustments to the controllable area and observing the response in each area.

[0080] For the For each execution object, the central controller applies an amplitude based on its current opening degree or frequency. The perturbation recognition command. For water valves or air valves, It can be taken as 5% to 15% of the current adjustable range; for the frequency of fans or water pumps, The frequency can be 3% to 8% of the rated frequency. The perturbation direction can be either increasing or decreasing. To reduce the impact of environmental disturbances, a positive perturbation can be performed first on the same target, followed by a recovery or reverse perturbation. The perturbation duration can be set to 3 to 10 minutes, preferably 5 to 6 minutes; the post-perturbation recovery observation window can be set to 5 to 10 minutes.

[0081] The perturbation amplitude can be determined by the following formula:

[0082]

[0083] In the formula, For the first The perturbation amplitude of each execution object; This is the perturbation scaling factor, with a value ranging from 0.05 to 0.15; For the first The maximum allowable opening or frequency for each execution object; For the first The minimum allowed opening or minimum frequency for each execution object; The maximum permissible variation to meet comfort constraints and equipment safety constraints.

[0084] During perturbation execution, the central controller monitors the temperature deviation of each controlled area in real time. When the temperature of any area deviates from the set value by more than the preset safety boundary, or when the pipeline pressure difference, duct static pressure, or motor current exceeds the allowable range, the perturbation identification command is immediately terminated, and the executed object is restored to the state before the perturbation. Preferably, the comfort safety boundary is set to a temperature deviation of no more than 1.0℃, a humidity deviation of no more than 8%, and a pipeline pressure difference change of no more than 15% of the normal operating pressure difference.

[0085] When multiple execution objects need to be identified, the central controller applies perturbation identification commands one by one in a preset order. The preset order can be determined according to floor, air conditioning zone, terminal number, branch topology, or historical anomaly records. To avoid overlapping responses caused by simultaneous actions of adjacent terminals, it is preferable to perform perturbation identification on only one execution object at a time, and an interval of no less than one recovery observation window should be set between terminals in adjacent areas.

[0086] 2.3 Baseline Subtraction and Net Response Calculation

[0087] like Figure 2 As shown, the core control closed loop of this invention includes three parts: perturbation identification and net response calculation, actual correlation matrix establishment and comparison, and constraint control command generation. The central controller first collects response data within the stability window and perturbation window, and performs baseline subtraction to obtain the net temperature response, net pressure difference response, and net energy consumption response. Then, it establishes an actual correlation matrix based on the thermal response fingerprint and compares it with the drawing correlation table or installation registration table to identify end-point mismatch, weak response regions, and abnormal coupling regions. Finally, it generates candidate control quantities based on the corrected regional load demand and projects them onto the feasible constraint set under equipment safety constraints, comfort constraints, and actual correlation matrix constraints to obtain the final control command. Closed-loop correction is achieved through feedback updates.

[0088] S3. Obtain the temperature change trend, pressure difference change trend, and energy consumption change trend within the pre-perturbation stabilization window. Perform baseline subtraction on the response data within the perturbation window to obtain the net temperature response, net pressure difference response, and net energy consumption response caused by the perturbation identification command. Generate a thermal response fingerprint based on the net temperature response, net pressure difference response, net energy consumption response, and response start lag time, and establish the actual correlation matrix between the execution object and the controlled area.

[0089] 2.3.1 Net response quantity caused by perturbation identification command

[0090] After the perturbation command is executed, baseline subtraction processing is performed on the response data within the perturbation window to eliminate interference caused by changes in outdoor temperature, personnel load, solar radiation, and other actions of the executed object. Baseline subtraction processing includes: based on the pre-perturbation stabilization window... The trend within the data is used to establish baselines for predicting temperature, pressure difference, and energy consumption under perturbation-free conditions; then the perturbation window is adjusted. The measured data within the data set are compared with the corresponding baseline to obtain the net response caused by the perturbation identification command.

[0091] Stability window before perturbation Inside, the central controller obtains the first The temperature change trend of each controlled area was analyzed, and baseline temperature forecasts were established. When the temperature changes approximately linearly within the baseline window, linear extrapolation can be used; when temperature fluctuations are large, a weighted moving average can be used. Linear extrapolation method:

[0092]

[0093] In the formula, To achieve the first in the absence of perturbations Each controlled area at any time Baseline temperature prediction; The start time of the perturbation The regional temperature; Stability window before perturbation The slope of the temperature change within; The sampling time is within the perturbation window.

[0094] No. The execution object for the first The net temperature response generated by each controlled area is calculated using the following formula:

[0095]

[0096] In the formula, For the first The execution object for the first Normalized net temperature response of a controlled region; For the first The perturbation amplitude of each execution object; For perturbation response window; For the first Each controlled area at any time The measured temperature; This represents the baseline temperature prediction value for the corresponding time.

[0097] For differential pressure data, the central controller uses the pre-perturbation stabilization window. Establish baseline differential pressure forecasts based on the trend of differential pressure changes within the system. . No. The execution object for the first The net differential pressure response intensity of a controlled area or its corresponding branch can be calculated using the following formula:

[0098]

[0099] In the formula, For the first The execution object for the first The net differential pressure response intensity of a controlled area or its corresponding branch; For the first Each controlled area corresponds to a branch at time... The measured pressure difference; This represents the predicted baseline pressure difference under perturbation-free conditions. For the first The perturbation amplitude of each execution object; This is a perturbation window.

[0100] For energy consumption data, the central controller uses the pre-perturbation stability window. Establish baseline energy consumption forecasts based on the trend of energy or power changes within the region. . No. The execution object for the first The net energy consumption response intensity of a controlled area or its associated equipment can be calculated using the following formula:

[0101]

[0102] In the formula, For the first The execution object for the first Net energy consumption response intensity of a controlled area or its associated equipment; For the first A controlled area or its associated device at a time The measured energy consumption or power; This represents the baseline energy consumption or power prediction under perturbation-free conditions. For the first The perturbation amplitude of each execution object; This is a perturbation window.

[0103] No. The execution object for the first The initial response lag time of each controlled region can be determined by the following formula:

[0104]

[0105] In the formula, For the first The execution object for the first Response start lag time for each controlled area; This is the start time of the perturbation; For the first Each controlled area at any time The measured temperature; This is the baseline temperature prediction value; The effective temperature response threshold can be set between 0.05℃ and 0.15℃.

[0106] If in the perturbation window If the effective response threshold is not reached, then... This is recorded as an invalid response value, and the influence between the executing object and the controlled region is reduced in the subsequent actual correlation matrix calculation.

[0107] Peak temperature change rate It can be calculated using the following formula:

[0108]

[0109] In the formula, For the first The controlled area in the first Peak temperature change rate during the perturbation of each execution object; For the first Each controlled area at any time Temperature; The sampling interval; This is a perturbation window.

[0110] Temperature recovery time This refers to the period after the perturbation is withdrawn. The time required for the temperature deviation of a controlled area to fall back below the effective response threshold. If the area temperature does not fall back within the preset recovery window, the fingerprint is marked as a delayed response or an abnormal response.

[0111] 2.3.2 Thermal Response Fingerprint Generation

[0112] For each execution object and each controlled area The central controller builds a thermal response fingerprint. The thermal response fingerprint includes at least the net temperature response intensity, net pressure difference response intensity, net energy consumption response intensity, response onset lag time, peak temperature change rate, temperature recovery time, end-point opening change, and response stability marker. The thermal response fingerprint is written into the thermal response fingerprint database according to the execution object number and the controlled area number, and is used for subsequent establishment of the actual correlation matrix, identification of end-point mismatch, and updating of the control strategy.

[0113] Thermal response fingerprints can be represented as:

[0114]

[0115] In the formula, For the first The execution object and the first Thermal response fingerprints between controlled regions; This is the normalized net temperature response. For response lag time; This represents the peak value of the rate of temperature change. Temperature recovery time; This is the pressure difference response quantity; Energy consumption response quantity; For response stability tagging.

[0116] Among them, response stability tag The determination can be based on the consistency of response characteristics obtained from multiple perturbation identifications of the same execution object. For example, when the net temperature response intensity deviation obtained from two or three consecutive perturbation identifications of the same execution object for the same controlled area does not exceed 20%, and the deviation of the response start lag time does not exceed 2 minutes, then... Record it as a stable response; otherwise, record it as a response pending verification.

[0117] To avoid direct superposition of data with different dimensions, the central controller normalizes each response characteristic. Taking the net temperature response intensity as an example, its normalized value can be calculated using the following formula:

[0118]

[0119] In the formula, This represents the normalized net temperature response intensity. This represents the original net temperature response intensity. and These represent the minimum and maximum values ​​of the net temperature response intensity in the same round of identification; To prevent tiny constants with a denominator of zero.

[0120] A shorter response start lag time generally indicates a stronger association between the executing object and the controlled area; therefore, reverse normalization can be used.

[0121]

[0122] In the formula, The normalized response start lag time; This is the initial lag time of the original response; and These are the minimum and maximum values ​​of the effective response start lag time, respectively. To prevent tiny constants with a denominator of zero.

[0123] 2.3.3 Establishment of the Actual Correlation Matrix

[0124] The central controller generates the actual correlation matrix based on thermal response fingerprints. .matrix elements in Indicates the first The execution object for the first The intensity of the impact of each controlled area. It can be calculated using the following formula:

[0125]

[0126] In the formula, For the first The execution object for the first The intensity of the impact on each controlled area; This represents the normalized net temperature response intensity. The normalized net pressure difference response intensity; This represents the normalized net energy consumption response intensity. The normalized response start lag time; , , , These are the weighting coefficients.

[0127] In a preferred embodiment Take 0.45, Take 0.20, Take 0.15, Set the value to 0.20. For scenarios with densely packed temperature sensors, this can improve... For scenarios with a relatively complete deployment of differential pressure sensors in water systems, it can improve... For scenarios with relatively complete sub-metering equipment, it can improve... .

[0128] Central controller according to The relationship with the threshold determines the type of relationship between the execution object and the controlled region. When When it is greater than the first association threshold, the first... The execution object is determined to be the first The main associated object of a controlled region; when When the value is between the first association threshold and the second association threshold, the first... The execution object is determined to be the first Weakly related objects in a controlled region; when When it is less than the second association threshold, the first The execution object is determined to be the first The unassociated objects in the controlled area. Preferably, the first association threshold can be 0.60 to 0.75, and the second association threshold can be 0.25 to 0.40, or it can be set according to the building type, sensor accuracy and control cycle.

[0129] 2.4 Identification of End-Mismatch, Weak Response Regions, and Abnormal Coupling Regions

[0130] S4. Compare the actual correlation matrix with the pre-stored drawing correlation table or installation registration table to identify end-point mismatch, weak response areas and abnormal coupling areas, and use the identification results as the basis for correcting regional load demand and subsequent control command allocation.

[0131] 2.4.1 Identify end-mismatch, weak response regions, and anomalous coupling regions.

[0132] The central controller will actually associate the matrix Association table with pre-stored drawings or installation registration form Compare the drawings with the relevant tables or installation registration forms. It can be generated from design drawings, construction registration forms, building automation point tables, or equipment commissioning records. At that time, it indicates that the drawing or registration form considers the first The execution object service number A controlled area; when When the two are unrelated in design, it means that they are not related.

[0133] For each execution object The central controller searches the actual association matrix. Largest controlled area :

[0134]

[0135] In the formula, For the first The main response area identified on-site for each execution object; For the first The execution object for the first The intensity of the impact on each controlled area; Indicates to make Take the index of the controlled region with the maximum value.

[0136] When the drawing association table or installation registration table records the first The execution object corresponds to the first There are controlled regions, but in the actual correlation matrix Less than the second correlation threshold, and there is another controlled region. make When the value is greater than the first association threshold, determine the first... An execution object has an end-point mismatch, and the controlled region... This serves as the actual primary associated region for the executed object. The end-point mismatch marker includes the executed object number, drawing registration area, actual primary response region, impact intensity, identification time, and confidence level. In subsequent control, the central controller prioritizes assigning control commands according to the actual primary response region.

[0137] When the controlled area All execution objects corresponding If all values ​​are below the first correlation threshold, and the temperature deviation of the controlled area continuously exceeds the comfort threshold within a preset time, the controlled area is determined to be a weak response area. Weak response areas may be caused by valve jamming, filter clogging, hydraulic imbalance, insufficient airflow, or improper sensor placement. For weak response areas, the central controller increases the adjustment priority of execution objects that have a primary or weak correlation with that area and generates maintenance prompts.

[0138] When the influence intensity of at least two executing objects on the same controlled area is greater than the first association threshold, and the difference in their response start lag time is less than the preset lag threshold, the controlled area is determined to have abnormal coupling. Abnormal coupling may be caused by duct crosstalk, waterway branch intersections, the combined effect of multiple terminals, or unclear control point binding relationships. For areas with abnormal coupling, the central controller does not use a single executing object to make large adjustments, but instead uses a coordinated adjustment method of multiple executing objects to reduce the risk of local over-adjustment.

[0139] 2.4.2 Regional Load Demand Generation and Adjustment

[0140] After completing the on-site real-world correlation identification, the central controller calculates the initial regional load demand for each controlled area. The initial regional load demand can be determined based on the current temperature deviation, humidity deviation, personnel status, outdoor weather conditions, and historical response characteristics.

[0141] Initial regional load demand It can be calculated using the following formula:

[0142]

[0143] In the formula, For the first Initial load demand indicators for each controlled area; For the first Current temperature of each controlled area; For the first Target temperature for each controlled area; The current relative humidity; Target relative humidity; For personnel status indicators; Outdoor temperature; For the first Historical load correction items for each controlled area; , , , , These are the weighting coefficients.

[0144] After generating the initial regional load demand, the central controller corrects the regional load demand based on the identification results of terminal mismatch, weak response areas, and abnormal coupling areas, thus obtaining the corrected regional load demand. For mismatched objects at the end, the control allocation relationship of the original registered region is transferred to the actual main associated region; for weak response regions, the adjustment priority of execution objects that have a main or weak association with this region is increased; for abnormally coupled regions, the large adjustment of a single execution object is broken down into the coordinated adjustment of multiple execution objects to reduce the risk of over-adjustment and crosstalk.

[0145] In one implementation, the modified regional load demand can be expressed as:

[0146]

[0147] In the formula, For the first The adjusted regional load demand for each controlled area; For the first Initial load demand for each controlled area; This is a term for end-mismatch correction. For weak response regions, correction terms are provided. This is a correction term for abnormal coupling regions.

[0148] 2.5. Generation of control commands based on the actual correlation matrix

[0149] S5. Generate candidate control quantities based on the corrected regional load demand and the actual correlation matrix, and project the candidate control quantities onto the feasible constraint set under the constraints of equipment safety, comfort and actual correlation matrix to obtain the final control commands for cold source output, water pump frequency, fan frequency, terminal valve opening and air valve opening.

[0150] The central controller first determines the regional load demand based on the revised data. and actual correlation matrix The regional demand is allocated to the relevant execution objects, generating candidate control quantities. The candidate adjustment amount for the execution object can be calculated using the following formula:

[0151]

[0152] In the formula, For the first Candidate control variables for each execution object; For the first The current control quantity of each execution object; For the first Adjustment ratio coefficient for each execution object; Total number of controlled areas; For the first The execution object for the first The intensity of the impact on each controlled area; For the revised first Load demand in controlled areas.

[0153] After obtaining candidate control variables, the central controller projects these variables onto the feasible constraint set to obtain the final control command. The feasible constraint set includes equipment safety constraints, comfort constraints, and actual correlation matrix constraints. Equipment safety constraints include upper and lower limits for cold source output, water pump frequency, fan frequency, terminal valve opening, and damper opening; comfort constraints include the allowable temperature deviation range, humidity deviation range, and upper limit for the rate of temperature change in the controlled area; actual correlation matrix constraints include prioritizing execution objects primarily associated with the target controlled area as adjustment objects, restricting non-associated objects from participating in load regulation of the target controlled area, and setting coordinated adjustment ratios for multiple execution objects in abnormally coupled areas.

[0154] The projection correction can be solved by optimizing the following objective:

[0155]

[0156] In the formula, The final set of control variables to be solved; For the first The final control quantity of an execution object; For the first Candidate control variables for each execution object; For the first The adjusted weights of each execution object; For the first Comfort weights for each controlled area; For the first Predicted temperature of each controlled area; For the first The set temperature of each controlled area; Energy consumption weighting; To predict system power; Number of objects to be executed; The number of controlled areas.

[0157] The corresponding constraints include:

[0158]

[0159] In the formula, For the first The minimum amount of control allowed for each execution object; For the first The maximum amount of control allowed per execution object; For the first The final control of an execution object.

[0160]

[0161] In the formula, For the first Predicted temperature of each controlled area; For the first The set temperature of each controlled area; To allow for temperature deviations.

[0162]

[0163] In the formula, To predict the pressure difference between supply and return water or the static pressure of the duct; To the minimum allowable pressure difference; To the maximum allowable pressure difference.

[0164]

[0165] In the formula, To predict system power; The maximum allowable power for a building or station.

[0166] For execution objects identified as having end-mismatch, the control system no longer allocates control quantities according to the original drawing area, but instead allocates them according to the main associated area corresponding to the actual association matrix; for weak response areas, the control system prioritizes adjusting multiple execution objects that have a main or weak association with it, and restricts excessive adjustment of a single execution object; for abnormally coupled areas, the control system reduces the proportion of rapid and large-scale adjustments and adopts a multi-cycle gradual adjustment method.

[0167] The final control commands are output to the cold source, water pump, fan, and terminal actuators. After each control command is executed, the system continues to collect response data and compares the actual response with the predicted response. If the deviation exceeds the threshold for multiple consecutive control cycles, the perturbation identification process is restarted to update the thermal response fingerprint and the actual correlation matrix.

[0168] 2.6 Feedback Update

[0169] After executing step S5, the response data of each controlled area continues to be collected, and the deviation between the actual response and the predicted response is written back to the thermal response fingerprint database. When the response deviation between the same execution object and the same controlled area exceeds a preset number of times, steps S2 to S3 are triggered again to update the thermal response fingerprint and the actual correlation matrix. When the difference between the updated actual correlation matrix and the original actual correlation matrix exceeds a preset matrix difference threshold, step S4 is executed again to update the identification results of end mismatch, weak response area and abnormal coupling area.

[0170] In this embodiment, the predicted response can be calculated based on the thermal response fingerprint and the final control command, while the actual response is collected by the controlled area sensor after the control command is executed. The response deviation can be expressed as:

[0171]

[0172] In the formula, For the first After the adjustment of the first execution object Each controlled area at any time Response deviation; For the first Each controlled area at any time The actual temperature; For the first Each controlled area at any time The predicted temperature.

[0173] when If the temperature continuously exceeds a preset deviation threshold, for example, if it exceeds 0.5°C for 3 to 5 consecutive control cycles, it indicates that the response relationship between the actuator and the controlled area may have changed, and the central controller triggers local perturbation identification. Local perturbation identification is performed only on the relevant actuator and its adjacent controlled areas, without having to rescan the entire end effector, thus reducing identification time and impact on comfort.

[0174] III. System Structure and Implementation

[0175] like Figure 3 As shown, the present invention also provides a central air conditioning intelligent identification and control system for implementing the above-mentioned central air conditioning intelligent identification and control method. The system includes a data acquisition module, a disturbance identification module, a response fingerprint generation module, an anomaly identification module, a control optimization module, an instruction output module, a data storage module, and a feedback update module.

[0176] The data acquisition module communicates with temperature and humidity sensors, differential pressure sensors, energy consumption metering units, valve actuators, damper actuators, water pump frequency converters, fan frequency converters, and cold source controllers to collect operational data.

[0177] The disturbance identification module is used to determine the disturbance object, amplitude, direction, and duration when the central air conditioning system is in a stable operating state, and to apply disturbance identification commands to terminal valves, air dampers, fans, or water pumps. The disturbance identification module includes a safety judgment unit that immediately cancels the disturbance command when temperature deviation, pressure difference, motor current, or equipment status exceeds the allowable range.

[0178] The response fingerprint generation module includes a baseline subtraction unit, a response feature extraction unit, and an association matrix generation unit. The baseline subtraction unit is used to subtract the temperature change trend, pressure difference change trend, and energy consumption change trend within the stabilization window before perturbation; the response feature extraction unit is used to extract the net temperature response, net pressure difference response, net energy consumption response, response onset lag time, peak temperature change rate, and temperature recovery time; the association matrix generation unit is used to calculate the influence strength between the executed object and the controlled area based on the response features and generate the actual association matrix.

[0179] The anomaly identification module compares the actual correlation matrix with pre-stored drawing correlation tables or installation registration tables to identify end-point mismatches, weak response areas, and abnormal coupling areas, and generates corresponding markers. The control optimization module includes a candidate control quantity generation unit and a constraint projection correction unit. The candidate control quantity generation unit generates initial control commands based on the corrected regional load requirements, while the constraint projection correction unit corrects the initial control commands to final control commands that satisfy equipment safety constraints, comfort constraints, and actual correlation matrix constraints.

[0180] The feedback update module continuously collects the actual response of the controlled area after the final control command is executed, compares the actual response with the predicted response obtained from the thermal response fingerprint, and writes the deviation back to the thermal response fingerprint database. When the response deviation between the same execution object and the same controlled area exceeds a preset number of consecutive times, the feedback update module triggers the perturbation identification module to re-execute the local perturbation identification, and the response fingerprint generation module updates the thermal response fingerprint and the actual correlation matrix.

[0181] The above modules can be integrated into the building automation server or deployed on edge controllers. For large buildings, multiple edge controllers can be set up according to floors or air conditioning zones. Each edge controller completes the perturbation identification and terminal control in its area, while the building-level server is responsible for coordinating the cooling source, water pumps, and total energy consumption.

[0182] IV. Examples

[0183] 4.1 Example 1: Identification and Control of Central Air Conditioning System in Office Building

[0184] A 12-story office building with a floor area of ​​approximately 28,000 m² was selected for the test. The central air conditioning system includes two screw chillers, three chilled water pumps, three cooling water pumps, four air handling units, and 48 terminal fan coil units. The test area consisted of 24 controlled zones on floors 6 through 8, with one temperature and humidity sensor and one carbon dioxide sensor installed in each zone. The system sampling cycle was 60 seconds, and the control cycle was 5 minutes. The comfort setting temperature was 24℃, with an allowable deviation of ±1℃.

[0185] In this embodiment, a 10-minute window is selected before perturbation identification. Within the stable window, the temperature change rate of each controlled area does not exceed 0.05℃ / min, the supply and return water pressure difference fluctuation does not exceed 5% of the average pressure difference, and the end valve opening change does not exceed 3%. Therefore, the system is determined to be in a stable operating state.

[0186] During the perturbation identification phase, perturbation identification commands of 8% opening degree were sequentially applied to 48 end valves. Each perturbation lasted for 6 minutes, with a baseline window of 10 minutes before the perturbation and an observation window of 8 minutes after the perturbation. To avoid interference caused by simultaneous operation of adjacent end valves, perturbation was performed on only one end valve at a time, and a 15-minute interval was set between adjacent end valves.

[0187] Temperature, pressure difference, and energy consumption data collected within the perturbation window are all deducted from the baseline trend obtained by extrapolation from the stability window to obtain the net temperature response, net pressure difference response, and net energy consumption response, respectively, and the actual correlation matrix is ​​calculated accordingly.

[0188] The identification results show that the system detected 3 mismatched objects at the ends, 2 weak response regions, and 1 set of abnormally coupled regions. Some of the identification results are shown in Table 1.

[0189] Table 1. Air Conditioning System Area Identification Results for Example 1

[0190]

[0191] As shown in Table 1, taking V-0612 as an example, in the drawing registration table, this valve corresponds to the East Conference Room on the 6th floor. However, after perturbation, the net temperature response intensity of the East Conference Room on the 6th floor is low, with a response start lag time exceeding 10 minutes; while the net temperature response intensity of the East Open Office Area on the 6th floor is high, with a response start lag time of 3 minutes. In the actual correlation matrix... The value is 0.82. Therefore, the system marks V-0612 as a service area mismatch and corrects its primary associated area to the 6th floor East Open Office Area in subsequent controls.

[0192] After identification was completed, a comparative test was conducted for 14 consecutive days. For the first 7 days, the original building automation system was used for control according to the drawing area and a fixed PID strategy. For the following 7 days, the actual correlation matrix constraint control of this implementation method was used. During the test, the average outdoor temperature was 31.5℃, the average daily number of personnel was approximately 850, and the working hours were from 8:00 to 19:00.

[0193] The test results are shown in Table 2.

[0194] Table 2 Test results of different control methods in Example 1

[0195]

[0196] As shown in Table 2, after identifying and correcting the end-service relationships, the system can avoid assigning adjustment commands to incorrect ends, thereby reducing regional temperature deviations and over-limit durations. Simultaneously, because weak response areas utilize multi-executor collaborative adjustment and abnormally coupled areas employ gradual adjustment, pump and fan power fluctuations are reduced, resulting in an overall decrease in energy consumption.

[0197] 4.2 Example 2: Control Verification of a Conference Room with Variable Load

[0198] Six meeting rooms and four open office areas are located on the 10th floor of the same office building, where the personnel density varies significantly. This implementation method is used to identify the 10 controlled areas. Personnel status is determined jointly by access control records and changes in carbon dioxide concentration. 15 minutes before the start of meeting room use, the system generates pre-adjustment demand based on reservation information; when the difference between the actual number of people entering and the number of people reserving exceeds 30%, the system recalculates the area load demand based on the real-time personnel status.

[0199] In the 10-layer test, eight meetings within a day were selected as samples to compare the temperature stability during the meetings before and after using the actual correlation matrix control. The test results are shown in Table 3.

[0200] Table 3 Test results of different control methods in Example 2

[0201]

[0202] Table 3 shows that for areas with rapid changes in personnel, this implementation method allocates the regional load demand to the real and effective terminals through the actual correlation matrix, which can quickly offset the temperature rise caused by personnel load and reduce ineffective cooling after the meeting.

[0203] 4.3 Example 3: Verification of Weak Response Region Maintenance Prompt

[0204] In the experiment of Example 1, the north office on the 6th floor was marked as a weak response area. On-site inspection revealed that the fan coil unit filters in this area had a lot of dust accumulation, and the electric water valves exhibited opening feedback lag. After maintenance personnel cleaned the filters and calibrated the water valves, the system re-performed perturbation identification. Before maintenance, the influence strength of the main associated execution object in this area was 0.34; after maintenance, the influence strength increased to 0.71, the response start lag time was shortened from 9 minutes to 4 minutes, and the time to reach the set temperature was shortened from 42 minutes to 23 minutes.

[0205] This embodiment illustrates that this implementation method can not only be used for operation control, but also assist maintenance personnel in locating problems of insufficient terminal cooling or air supply capacity by marking weak response areas.

[0206] 4.4 Example 4: Verification of Actual Association Matrix Update

[0207] To verify the adaptive update capability of the actual correlation matrix, the control point binding relationship of the two end valves on the west side of the 7th floor was adjusted in the later stage of the experiment. During subsequent operation, the system found that the deviation between the predicted temperature response and the actual response exceeded the threshold for four consecutive control cycles, automatically triggering local perturbation identification. Identification was performed only on the relevant 6 execution objects and 4 controlled areas, rather than rescanning all end valves. Local identification took approximately 95 minutes, during which the local actual correlation matrix was regenerated and the binding relationship was corrected. After correction, the average temperature deviation in the relevant areas decreased from 1.05℃ to 0.46℃.

[0208] In summary, this embodiment, through steady-state judgment, perturbation identification, baseline subtraction, thermal response fingerprint generation, actual correlation matrix establishment, terminal mismatch and weak response identification, and constraint projection control, enables the central air conditioning system to perform zonal control according to the true response relationship even when there are deviations in on-site configuration or changes in long-term operating status. Those skilled in the art can implement the technical solution described in this invention by following the above-described data acquisition methods, perturbation parameter ranges, response calculation methods, matrix judgment rules, and constraint control procedures.

[0209] The foregoing description of embodiments of the present invention, through which those skilled in the art are able to implement or use the present invention, will be readily apparent to those skilled in the art. Various modifications to these embodiments will be readily apparent to those skilled in the art. The general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novelty disclosed herein.

Claims

1. A method for intelligent identification and control of central air conditioning, characterized in that, include: S1. Collect operating data of the cold source, water pump, fan, terminal valve, air valve and each controlled area in the central air conditioning system. The operating data includes temperature and humidity, personnel status, energy consumption, supply and return water temperature difference, pipeline pressure difference and terminal opening. S2. When the central air conditioning system is in a stable operating state, apply a perturbation identification command to at least one terminal valve or air valve in a preset sequence, and record the temperature change rate, pressure difference change and energy consumption change of each controlled area in the perturbation window. S3. Obtain the temperature change trend, pressure difference change trend and energy consumption change trend within the stability window before the perturbation. Perform baseline subtraction on the response data within the perturbation window to obtain the net temperature response, net pressure difference response and net energy consumption response caused by the perturbation identification command. Generate a thermal response fingerprint based on the net temperature response, net pressure difference response, net energy consumption response and response start lag time, and establish the actual correlation matrix between the execution object and the controlled area. S4. Compare the actual correlation matrix with the pre-stored drawing correlation table or installation registration table to identify end-point mismatch, weak response area and abnormal coupling area, and use the identification results as the basis for correcting the area load demand and subsequent control command allocation. S5. Generate candidate control quantities based on the corrected regional load demand and the actual correlation matrix, and project the candidate control quantities onto the feasible constraint set under the constraints of equipment safety, comfort and actual correlation matrix to obtain the final control commands for cold source output, water pump frequency, fan frequency, terminal valve opening and air valve opening.

2. The intelligent identification and control method for central air conditioning according to claim 1, characterized in that, In step S1, the personnel status includes one or more of the following: personnel quantity, personnel density, personnel stay time, and area occupancy status. The personnel status is obtained by at least one of the following: camera, infrared sensor, access control record, wireless terminal access record, or carbon dioxide concentration change data; The terminal opening degree includes the terminal water valve opening degree and / or air valve opening degree, and the pipeline pressure difference includes one or more of the terminal branch pressure difference, the supply and return water main pressure difference, or the pressure difference before and after the terminal.

3. The intelligent identification and control method for central air conditioning according to claim 1, characterized in that, In step S2, the perturbation identification instruction includes applying an amplitude change to the opening degree of the end water valve, the opening degree of the air valve, the frequency of the fan, or the frequency of the water pump for a duration not exceeding one control cycle. The amplitude change is 5% to 15% of the current adjustable range of the corresponding execution object, and does not exceed the preset safety range, so that the comfort deviation of the controlled area does not exceed the preset allowable range; The stable operating state is defined as follows: within the stable window before the perturbation identification command is applied, the rate of change of temperature in the controlled area, the change of pipeline differential pressure, and the change of terminal opening are all within the corresponding stable threshold range.

4. The intelligent identification and control method for central air conditioning according to claim 1, characterized in that, In step S3, the thermal response fingerprint includes at least the response start lag time, peak temperature change rate, temperature recovery time, pipeline pressure difference change, end opening change, and regional energy consumption change. The thermal response fingerprint is written into the thermal response fingerprint database according to the execution object number and the controlled area number; When the same execution object generates responses in multiple controlled regions, the response start lag time and net temperature response intensity of each controlled region are recorded to distinguish between the main controlled region, the weakly controlled region, and the coupled controlled region.

5. The intelligent identification and control method for central air conditioning according to claim 1, characterized in that, In step S3, the actual association matrix is ​​A, and the elements of the actual association matrix A are... This represents the strength of the influence of the i-th executing object on the j-th controlled region. satisfy: ； In the formula, For the first The execution object for the first The intensity of the impact on each controlled area; This represents the normalized net temperature response intensity. The normalized net pressure difference response intensity; This represents the normalized net energy consumption response intensity. The normalized response start lag time; , , , These are the weighting coefficients; when When the value is greater than the first association threshold, the i-th execution object is determined to be the main associated object of the j-th controlled region; when When the value falls between the first and second association thresholds, the i-th execution object is determined to be a weakly associated object of the j-th controlled region; when... When the value is less than the second association threshold, the i-th execution object is determined to be a non-associative object of the j-th controlled region.

6. The intelligent identification and control method for central air conditioning according to claim 1, characterized in that, In step S4, identifying end-mismatch, weak response regions, and anomalous coupling regions includes: When the drawing association table or installation registration table records that the i-th execution object corresponds to the j-th controlled area, but in the actual association matrix... Less than the second association threshold, and there exists another controlled region k such that When the value exceeds the first association threshold, it is determined that the i-th execution object has an end-point mismatch; When all execution objects corresponding to the controlled region j If all values ​​are below the first correlation threshold, and the temperature deviation of the controlled area continues to exceed the comfort threshold within a preset time, the controlled area is determined to be a weak response area. When the influence intensity of at least two executing objects on the same controlled area is greater than the first association threshold, and the difference in their response start lag time is less than the preset lag threshold, it is determined that there is abnormal coupling in the controlled area.

7. The intelligent identification and control method for central air conditioning according to claim 1, characterized in that, In step S5, the feasible constraint set includes equipment safety constraints, comfort constraints, and actual correlation matrix constraints; The equipment safety constraints include upper and lower limits for cold source output, upper and lower limits for water pump frequency, upper and lower limits for fan frequency, upper and lower limits for terminal valve opening, and upper and lower limits for air valve opening. The comfort constraints include the permissible deviation range of temperature in the controlled area, the permissible deviation range of humidity, and the upper limit of the rate of temperature change. The actual correlation matrix constraints include: prioritizing the execution objects that are primarily associated with the target controlled area as adjustment objects, restricting non-associated objects from participating in the load adjustment of the target controlled area, and setting a collaborative adjustment ratio for multiple execution objects in abnormally coupled areas; When projecting candidate control variables onto the set of feasible constraints, the final control command is obtained according to the objective of minimizing the correction magnitude of candidate control variables, regional temperature deviation, and system energy consumption increment.

8. The intelligent identification and control method for central air conditioning according to any one of claims 1 to 7, characterized in that, It also includes step S6: After executing step S5, continue to collect response data for each controlled area and write back the deviation between the actual response and the predicted response to the thermal response fingerprint database. When the response deviation between the same execution object and the same controlled area exceeds a preset number of consecutive times, steps S2 to S3 are retried to update the hot response fingerprint and the actual correlation matrix. When the difference between the updated actual association matrix and the unupdated actual association matrix exceeds a preset matrix difference threshold, step S4 is re-executed to update the identification results of end mismatch, weak response region and abnormal coupling region.

9. A central air conditioning intelligent identification control system, characterized in that, The system is used to implement the intelligent identification and control method for central air conditioning according to any one of claims 1 to 7, the system comprising: The data acquisition module is used to collect operating data from the cold source, water pump, fan, terminal valve, air valve, and each controlled area in the central air conditioning system. The perturbation identification module is used to apply perturbation identification commands to at least one terminal valve or air valve in a preset sequence when the central air conditioning system is in a stable operating state. The response fingerprint generation module is used to perform baseline subtraction on the response data within the perturbation window, generate a thermal response fingerprint, and establish an actual correlation matrix between the execution object and the controlled area. The anomaly identification module is used to compare the actual correlation matrix with the pre-stored drawing correlation table or installation registration table to identify end-point mismatch, weak response areas and abnormal coupling areas. The control optimization module is used to generate candidate control quantities based on the corrected regional load demand and the actual correlation matrix, and project the candidate control quantities onto the feasible constraint set under equipment safety constraints, comfort constraints and actual correlation matrix constraints to obtain the final control command.

10. The intelligent identification control system for central air conditioning according to claim 9, characterized in that, The response fingerprint generation module includes a baseline subtraction unit, a response feature extraction unit, and an association matrix generation unit; The baseline subtraction unit is used to subtract the temperature change trend, pressure difference change trend, and energy consumption change trend within the stability window before the perturbation. The response feature extraction unit is used to extract net temperature response, net pressure difference response, net energy consumption response, response start lag time, peak temperature change rate, and temperature recovery time. The correlation matrix generation unit is used to calculate the influence strength between the execution object and the controlled area based on the response characteristics, and generate the actual correlation matrix. The control optimization module includes a candidate control quantity generation unit and a constraint projection correction unit; The candidate control quantity generation unit is used to generate initial control commands based on the regional load demand. The constraint projection correction unit is used to correct the initial control command into a final control command that satisfies the equipment safety constraints, comfort constraints, and actual correlation matrix constraints. The system also includes a feedback update module, which is used to write back the deviation between the actual response and the predicted response to the thermal response fingerprint database, and to trigger the update of the thermal response fingerprint and the actual correlation matrix when the response deviation exceeds a preset number of consecutive times.

Images