Energy efficiency strategy intelligent execution control system adapted to water-cooled air conditioner

Abstract

The present application relates to water-cooled air conditioner control technology field, especially to the energy efficiency strategy intelligent execution control system suitable for water-cooled air conditioner, including facility joint control center, modeling and collaborative management module, perception and strategy evaluation module, local anomaly correction module, global multi-objective optimization module, collaborative arbitration module and execution response module, the present application is to establish the unified abstract model and collaborative control matrix of equipment, before the instruction output, the response rate normalization compensation and energy-quality balance interlocking check are carried out, eliminate the adjustment conflict and thermal balance deviation of multiple equipment, while improving the energy efficiency of water-cooled air conditioning system, taking into account the equipment health degree and operation stability, with the ability of anomaly adaptive correction, multi-device collaborative arbitration and system self-evolution, suitable for intelligent energy-saving control of various large buildings and industrial water-cooled air conditioning system.

Description

Technical Field

[0001] This invention relates to the field of water-cooled air conditioning control technology, and in particular to an intelligent execution and control system for energy efficiency strategies adapted to water-cooled air conditioners. Background Technology

[0002] Water-cooled air conditioning systems are major energy-consuming equipment in large public buildings (such as shopping malls, office buildings, hospitals, and data centers) and industrial plants. With the advancement of the "dual carbon" target and the adjustment of electricity price policies, improving the operating energy efficiency of water-cooled air conditioning systems and reducing carbon emissions has become an important issue in the field of building energy conservation.

[0003] Traditional water-cooled air conditioning control strategies mainly employ methods such as rule-based start-stop control. This relies heavily on the experience of maintenance personnel, controlling equipment start-up and shutdown through fixed schedules or simple thresholds (such as the upper limit of chilled water outlet temperature). While this method is simple and easy to implement, it cannot adapt to dynamic factors such as load changes, environmental fluctuations, and equipment aging, resulting in low energy efficiency. In other words, conventional water-cooled air conditioning control technology has the following main shortcomings:

[0004] 1. Traditional methods often use energy efficiency ratio as the sole optimization target, ignoring the impact of equipment health (such as compressor wear and heat exchanger scaling) on ​​long-term reliability. In pursuit of short-term energy efficiency, the load is often concentrated on a few high-efficiency devices, leading to a rapid decline in their health. At the same time, there is a lack of effective constraints on large fluctuations in chilled water temperature, affecting end-point comfort. Furthermore, when the system experiences energy efficiency decline, load imbalance, or command conflicts, existing methods only trigger alarms and cannot automatically identify the type and severity of abnormal indicators, nor can they proactively match correction strategies. Maintenance personnel need to manually analyze and intervene, resulting in slow response, reliance on experience, and difficulty in ensuring the long-term efficient operation of the system.

[0005] 2. Chillers have high thermal inertia, while pumps, valves and other equipment respond quickly. Traditional methods do not align the timing, resulting in fast-response equipment repeatedly adjusting while slow-response equipment is not yet in place. This can easily lead to command conflicts, overshoot oscillations, and even frequent equipment start-ups and shutdowns. Furthermore, there is a dynamic coupling between the cooling tower's heat dissipation capacity and the chiller's cooling demand. However, existing methods lack real-time detection of thermal balance deviations on both sides. When heat dissipation is excessive or insufficient, automatic reconfiguration is not possible, leading to long-term hidden operational conflicts that affect energy efficiency and equipment safety.

[0006] To address the aforementioned technical shortcomings, a solution is proposed. Summary of the Invention

[0007] The purpose of this invention is to provide an intelligent execution and control system for energy efficiency strategies adapted to water-cooled air conditioners, in order to solve the aforementioned technical defects.

[0008] The objective of this invention can be achieved through the following technical solution: an intelligent execution and control system for energy efficiency strategies adapted to water-cooled air conditioners, including a facility joint control center, a modeling and collaborative management module, a perception and strategy evaluation module, a local anomaly correction module, a global multi-objective optimization module, a collaborative arbitration module, and an execution response module;

[0009] The modeling and collaborative management module is used to perform unified abstract modeling of all controllable devices in water-cooled air conditioners, and to build a dynamically updated device collaborative control matrix and a dictionary of coupling relationships between devices.

[0010] The perception and strategy evaluation module is used to normalize the collected multi-dimensional index data, calculate the current strategy comprehensive score S using a weighted scoring method, and perform discrimination processing. Based on the discrimination results, it outputs a qualified signal, a trigger signal for entering local correction, or a unqualified signal.

[0011] The local anomaly correction module is used to filter out abnormal indicators, calculate the abnormal change rate of abnormal indicators and sort them from high to low. It retains the abnormal indicators corresponding to the first and second highest values ​​as abnormal indicators to be corrected, matches them with the preset indicator-correction decision library to output correction decisions, calculates the predicted load balance and the predicted health of each device at the future Tp time after correction, triggers early adjustment, and then performs load transfer and distribution. If it is still unqualified after correction, it outputs a correction unqualified signal.

[0012] When generating or correcting unqualified signals, the global multi-objective optimization module is used to construct a multi-objective function that prioritizes energy efficiency while also considering health and stability. The module uses a preset hybrid model to solve the objective function and outputs the target instruction set of the global energy efficiency optimization strategy.

[0013] The arbitration module is used to perform response rate normalization compensation and energy-mass balance interlock verification in conjunction with the device collaborative control matrix and coupling relationship dictionary before outputting the target instruction set, and outputs the modified execution instruction set to each sub-device.

[0014] Preferably, the analysis process of the unified abstract modeling is as follows:

[0015] Register the following key attributes for all controllable devices in the water-cooled air conditioner: Device identifier: unique ID, type, response characteristic parameters: thermal inertia / response time constant τj, upper and lower limits of adjustment rate [Vjmin, Vjmax] and initial health value, and mark the device as j (j>0).

[0016] Obtain the dynamic health decay rate δj and predict the health Hj(t+Tp): Calculate the health after Tp (Tp>0) time based on the current load rate and decay rate;

[0017] Hj(t+Tp)=Hj(t)-δj×(lj(t) / ljmax)×Tp to update the current health status, where ljmax is the rated maximum load rate of device j;

[0018] Construct a dynamically updated device cooperative control matrix C∈R N×M Where N is the total number of devices, M is the dimension of control attributes, and the coupling relationships between devices are predefined, and a coupling relationship dictionary G is established.

[0019] Preferably, the analysis process for the current strategy comprehensive score S is as follows:

[0020] S1: Collect multi-dimensional index data of the energy efficiency strategy currently being implemented by the water-cooled air conditioner within a set period. The multi-dimensional index data includes energy efficiency index E, stability index W, equipment load balance degree J, and command conflict frequency C.

[0021] S2: Normalize the multidimensional indicator data, and calculate the current strategy's comprehensive score S based on the weighted scoring method after normalization;

[0022] S3: Compare the current strategy comprehensive score S with the preset current strategy comprehensive score range [Smin, Smax] and output a qualified signal, a local correction, or a unqualified signal.

[0023] Preferably, the analysis process of the local anomaly correction module is as follows:

[0024] Each indicator is compared with its corresponding preset threshold to filter out abnormal indicators, obtain the abnormal change rate of the abnormal indicators, sort the abnormal change rates from high to low, retain the abnormal indicators corresponding to the first and second highest abnormal change rates, and set them as abnormal indicators to be corrected.

[0025] Retrieve preset indicators and correction decisions, match the abnormal indicators to be corrected with the preset indicators and correction decisions, and output the correction decisions.

[0026] Preferably, after correcting the abnormal indicators, the predicted load balance and the predicted health of each device are calculated at the future Tp time.

[0027] If the predicted load balance is greater than the preset load balance threshold, or if the predicted health is less than the preset health threshold, or if a high-load device with high attenuation exceeds the preset duration, then the load distribution will be adjusted in advance.

[0028] Preferably, the pre-adjustment load allocation analysis process is as follows:

[0029] Construct a source device set and a target device set, obtain the remaining capacity of each target device, then calculate the total remaining capacity of all target devices, and calculate the utilization ratio of each target device. Utilization ratio = remaining capacity / total remaining capacity;

[0030] The total load to be transferred is allocated according to the proportion of the load to be borne. The load capacity of the target equipment is obtained, and the load capacity is converted into a load rate increment. Adjustment is performed based on the load rate increment. After adjustment, the set values ​​of each equipment are updated, and the health prediction column is refreshed synchronously through the equipment collaborative control matrix. If the predicted load balance and predicted health meet the requirements after adjustment, steps S1-S3 are not executed and the current strategy is maintained directly. Otherwise, steps S1-S3 are executed. If the result is still unqualified after correction, a correction unqualified signal is generated.

[0031] Preferably, the analysis process of the target instruction set is as follows: Construct a multi-objective function that prioritizes energy efficiency while also considering health and stability; combine the constraints; use a preset hybrid model to solve the objective function; and output the target instruction set of the global energy efficiency optimization strategy.

[0032] Preferably, the analysis process of the collaborative arbitration module is as follows:

[0033] By combining the equipment collaborative control matrix and the coupling relationship dictionary, the response rate normalization compensation and energy-mass balance interlock verification are performed, and the corrected execution instruction set is output to each sub-device.

[0034] The normalization compensation process for the equipment response rate is as follows: Read the thermal inertia time constant τ of the chiller unit, apply dynamic damping slope constraint to the adjustment command issued by the fast response equipment: Adjustment amount (t) = original expected adjustment amount (t) × min(1, thermal inertia time constant τ / (fast response equipment's own response time constant + preset zero offset number)).

[0035] Preferably, the quantity-mass balance interlock verification process is as follows:

[0036] Calculate the heat dissipation on the cooling side and the cooling demand on the freezing side. If |heat dissipation on the cooling side - cooling demand on the freezing side| > ε×max(heat dissipation on the cooling side, cooling demand on the freezing side), where ε represents the relative tolerance coefficient allowed by the preset thermal balance, then it is determined that there is a hidden risk of operational conflict.

[0037] Pre-intercept conflicting instruction sets and trigger forced refactoring until the balance condition is met;

[0038] The forced reconstruction analysis process is as follows: obtain the difference ΔQ between the heat dissipation on the cooling side and the cooling requirement on the freezing side, perform discrimination processing on ΔQ, and output whether the cooling side is excessive or insufficient.

[0039] Select the equipment that can be adjusted and the current adjustable margin ΔQT. When it is insufficient or excessive: ΔQT = heat dissipation on the cooling side - cooling required on the freezing side + ε × max (heat dissipation on the cooling side, cooling required on the freezing side).

[0040] Obtain the preset thermodynamic influence coefficient K for each device, combine it with the device health H, calculate the adjustment weight w for each device, and allocate the single-step adjustment amount for each device: Single-step adjustment amount = ΔQT / set number of steps × w × (1 / preset thermodynamic influence coefficient K).

[0041] After an adjustment is performed, the new heat dissipation on the cooling side and the cooling demand on the refrigeration side are recalculated, and the energy-mass balance interlock check is performed again. If no hidden operational conflict risk is triggered, the refactoring is exited and the adjusted execution instruction set is output. If a hidden operational conflict risk is triggered, an alarm signal is triggered.

[0042] The beneficial effects of this invention are as follows:

[0043] (1) This invention establishes a local correction mechanism of indicator collection-scoring-anomaly screening-correction decision matching-load transfer, automatically identifies abnormal indicators and sorts them by change rate, prioritizes correcting the two with the greatest impact, matches preset decisions and triggers load transfer, realizes rapid adaptive adjustment under abnormal working conditions, reduces reliance on human experience, and constructs a multi-objective function with the goals of maximizing total energy efficiency ratio, minimizing health decay and minimizing outlet water temperature fluctuation. At the same time, it applies high load penalty to low health equipment, avoids the problem of excessively damaging a few high-efficiency equipment in pursuit of short-term energy efficiency, and extends the overall service life of the system.

[0044] (2) In addition, the present invention also addresses the rate difference between chiller units (slow response) and water pumps and valves (fast response) by designing dynamic damping slope constraints to compress the commands of fast-response equipment, force timing alignment, avoid repeated adjustments, overshoot oscillations and command offsetting, improve operational stability, and at the same time detect the thermal balance deviation between the cooling side and the freezing side in real time. When the deviation exceeds the limit, it automatically triggers forced reconstruction, selects adjustable equipment according to the deviation direction, and adjusts the weights in combination with health calculation to gradually correct until the balance meets the requirements, thereby reducing surge risk and energy efficiency loss. Attached Figure Description

[0045] The invention will now be further described with reference to the accompanying drawings;

[0046] Figure 1 This is a flowchart of the system of the present invention;

[0047] Figure 2 This is a schematic diagram for analysis of the perception and strategy evaluation module;

[0048] Figure 3 This is a schematic diagram illustrating the construction and analysis of the objective function. Detailed Implementation

[0049] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0050] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments;

[0051] Example 1: Please refer to Figures 1 to 3 As shown, this invention is an intelligent execution and control system for energy efficiency strategies adapted to water-cooled air conditioners, including a facility control center, a modeling and collaborative management module, a sensing and strategy evaluation module, a local anomaly correction module, a global multi-objective optimization module, a collaborative arbitration module, and an execution response module. The facility control center and the modeling and collaborative management module have a bidirectional communication connection, the facility control center and the sensing and strategy evaluation module have a unidirectional communication connection, the sensing and strategy evaluation module and the local anomaly correction module and the execution response module have both unidirectional communication connections, the local anomaly correction module and the global multi-objective optimization module and the execution response module have both unidirectional communication connections, the global multi-objective optimization module and the collaborative arbitration module have both unidirectional communication connections, and the collaborative arbitration module and the execution response module have both unidirectional communication connections.

[0052] The modeling and collaborative management module is used to perform unified abstract modeling of all controllable devices in water-cooled air conditioners, and to build a dynamically updated device collaborative control matrix and a dictionary of coupling relationships between devices, specifically including:

[0053] All controllable devices in the water-cooled air conditioner are modeled in a unified abstract way. Each device is registered with the following key attributes: Device identifier: unique ID, type (chiller / variable frequency pump / electric valve / cooling tower fan, etc.), response characteristic parameters: thermal inertia / response time constant τj, upper and lower limits of adjustment rate [Vjmin, Vjmax], initial health value: initially set to 1.0, and the device is marked as j (j>0).

[0054] Dynamic health decay rate δj: The health decrease value under rated load rate (100%) per unit time (hour). The initial value is preset according to the equipment type (e.g., 0.0005 for chiller, 0.0003 for variable frequency pump, 0.0001 for electric valve, and 0.0002 for cooling tower fan). It supports online identification and updating.

[0055] Predicted health Hj(t+Tp): Calculates the health status after a future time Tp (Tp>0) based on the current load rate and decay rate, for use in forward scheduling;

[0056] Health status dynamic update equation: In each control cycle, the current health status is updated according to Hj(t+Tp)=Hj(t)-δj×(lj(t) / ljmax)×Tp, with a lower limit of 0 and an upper limit of 1, where ljmax is the rated maximum load rate of equipment j;

[0057] Current operating status: Online / Offline, Current output value, Fault indicator;

[0058] Construct a dynamically updated device cooperative control matrix C∈R N×M Where N is the total number of devices and M is the control attribute dimension, including response time constant, current health status, etc.

[0059] The health-related columns are refreshed before each optimization solution is obtained.

[0060] To support command conflict detection, the system predefines the coupling relationships between devices, including:

[0061] Chiller-pump coupling: The combination of chilled / cooling water pumps corresponding to a certain chiller unit;

[0062] Valve-terminal coupling: the correspondence between electric valves and terminal air conditioning units;

[0063] Balance constraint pair: Thermodynamic coupling between heat dissipation on the cooling side and cooling demand on the refrigeration side;

[0064] The coupling relationship dictionary G = {(device A, device B): coupling type, coupling coefficient} is stored in the facility control center.

[0065] The perception and strategy evaluation module is used to normalize the collected multi-dimensional indicator data, calculate the current strategy comprehensive score S using a weighted scoring method, and perform discrimination processing. Based on the discrimination result, it outputs a qualified signal, a trigger signal for local correction, or a qualified signal, specifically including:

[0066] S1: Collect multi-dimensional index data of the energy efficiency strategy currently being implemented by the water-cooled air conditioner within a set period. The multi-dimensional index data includes energy efficiency index E, stability index W, equipment load balance degree J, and command conflict frequency C.

[0067] Among them, the energy efficiency index represents the ratio of actual cooling capacity to total input power;

[0068] The stability index represents the variance of the chilled water outlet temperature fluctuation;

[0069] Equipment load balance is represented by the maximum value of current health status × load rate among all devices, divided by the minimum value of current health status × load rate among all devices.

[0070] S2: Normalize the multidimensional indicator data, and calculate the current strategy comprehensive score S based on the weighted scoring method after normalization. S = E × preset weight 1 + W × preset weight 2 + J × preset weight 3 + (1-C) × preset weight 4.

[0071] S3: Compare the current strategy's overall score S with the preset range of current strategy's overall score [Smin, Smax]:

[0072] If the current strategy's overall score S > Smax, it is deemed qualified, and a qualified signal is generated.

[0073] If the current strategy's overall score S∈[Smin, Smax], then proceed to the local anomaly correction module for local correction;

[0074] If the current strategy's overall score S < Smin, it is deemed unqualified, and an unqualified signal is generated.

[0075] The execution response module is used to respond to qualified or unqualified signals. It immediately displays the preset text prompts corresponding to the qualified or unqualified signals so as to intuitively understand whether the current energy efficiency strategy is qualified and to take appropriate action based on the feedback information.

[0076] Example 2: The local anomaly correction module is used to filter anomaly indicators, calculate the abnormal change rate of the anomaly indicators and sort them from high to low. The first and second-highest anomaly indicators are retained as the anomaly indicators to be corrected. A correction decision is output by matching the pre-set indicator-correction decision library. After correction, the predicted load balance and predicted health of each device at future time Tp are calculated. If early adjustment is triggered, load transfer and allocation are executed. If the correction is still unsatisfactory, a correction failure signal is output, specifically including:

[0077] Each indicator is compared with its corresponding preset threshold, and abnormal indicators are screened out, such as energy efficiency indicator E being lower than the preset energy efficiency indicator threshold, stability indicator W being lower than the preset stability indicator threshold, equipment load balance J being lower than the preset equipment load balance threshold, and command conflict frequency C exceeding the preset command conflict frequency threshold.

[0078] Obtain the abnormal change rate of abnormal indicators, sort the abnormal change rates from high to low, retain the abnormal indicators corresponding to the first and second highest abnormal change rates, and set them as abnormal indicators to be corrected.

[0079] Retrieve preset indicators and correction decisions, match the abnormal indicators to be corrected with the preset indicators and correction decisions, and output the correction decisions;

[0080] For example: Abnormal indicators to be corrected: Energy efficiency index is too low - slightly increase the chilled water set temperature (e.g., 0.5°C) and decrease the cooling water set temperature;

[0081] After the abnormal indicators are corrected, the predicted load balance and the predicted health of each device at the future time Tp are calculated.

[0082] If the predicted load balance is greater than the preset load balance threshold, or if the predicted health is less than the preset health threshold, or if there are high-degradation devices (health degradation rate > preset health degradation rate threshold) with high load (load rate > preset load rate) for more than the preset duration, then the load allocation will be adjusted in advance.

[0083] The process of adjusting load allocation in advance is as follows:

[0084] The load rate of each device is redistributed using a load transfer heuristic method (or a lightweight optimization model) to construct a source device set. The source device set represents the set of devices that meet the trigger condition that the predicted health value is less than the preset minimum health value, or that there are high-decaying devices (health decay rate > preset health decay rate threshold) with high load (load rate > preset load rate) for a preset duration.

[0085] Construct a target device set, where the target devices represent the three devices with the lowest health decay rate, and whose current load rate is less than the preset load rate and whose predicted health rate is greater than the preset maximum health rate;

[0086] Obtain the remaining capacity of each target device, then calculate the total remaining capacity of all target devices, and calculate the utilization ratio of each target device. Utilization ratio = remaining capacity / total remaining capacity;

[0087] The total load to be transferred is allocated according to the proportion of the load to be borne. The load to be borne by the target equipment is calculated as: Load to be borne = Proportion of the load to be borne × Total load transferred from the source equipment. The load to be borne is converted into a load factor increment, which is calculated as: Load factor increment = Load to be borne / Rated capacity. Adjustments are then made based on the load factor increment.

[0088] After adjustment, update the settings of each device (chilled water outlet temperature, pump frequency, etc.) and refresh the health prediction column synchronously through the device collaborative control matrix. If the predicted load balance and predicted health meet the requirements after adjustment, the S1-S3 steps are not executed and the current strategy is maintained directly. Otherwise, the S1-S3 steps are executed. If the correction is still not qualified, a correction failure signal is generated.

[0089] The execution response module is used to respond to and correct non-compliance signals. It immediately displays the preset text prompt corresponding to the non-compliance signal so as to intuitively understand whether the current abnormal indicator has been corrected and can be dealt with accordingly.

[0090] Example 3: When generating or correcting a non-compliant signal, the global multi-objective optimization module constructs a multi-objective function that prioritizes energy efficiency while also considering health and stability. It solves the objective function using a pre-defined hybrid model and outputs a set of target instructions for the global energy efficiency optimization strategy, specifically including:

[0091] Collect current operating data of water-cooled air conditioners and perform preprocessing (such as cleaning, normalization, etc.). The operating data includes load side (such as chilled water supply and return water temperature difference, terminal pressure difference), equipment side (such as current output, health, and online status of each piece of equipment) and environmental side (such as outdoor wet-bulb temperature and cooling water return temperature).

[0092] Construct a multi-objective function that prioritizes energy efficiency while also considering health and stability:

[0093] The decision variable X includes the chilled water outlet temperature setpoint, the cooling water outlet / return temperature setpoint, and the cooling water pump frequency, etc.

[0094] Objective function: ;

[0095] Where N is the total number of controllable devices, Hj is the current health status of device j, lj is the load rate undertaken by device j, a1 and a2 are preset weighting coefficients, both a1 and a2 are greater than zero, and FC represents the standard deviation of chilled water outlet temperature fluctuation.

[0096] Constraints: Equipment output upper and lower limits: Ujmin ≤ Uj ≤ Ujmax;

[0097] Adjustment rate limit: |Uj(t)-Uj(t-1)|≤ΔUjmax, where t represents time, t>0;

[0098] Thermal balance constraint: Heat absorbed by the evaporator side (cooling required on the freezing side) < Heat dissipated by the condenser side (heat dissipated on the cooling side).

[0099] Health constraint: Equipment with Hj < health threshold has its output set to zero;

[0100] The objective function is solved using a pre-defined hybrid model, and the target instruction set of the global energy efficiency optimization strategy is output, including the outlet water temperature setpoint of each chiller unit, the frequency setpoint of each water pump, the opening setpoint of each electric valve, and the start / stop or speed setpoint of each cooling tower fan.

[0101] For example: Hybrid optimization strategy:

[0102] Coarse optimization: A neural network (NN) or rule base trained on historical data quickly provides an initial feasible solution;

[0103] Fine optimization: Particle swarm optimization (PSO) or Bayesian optimization is used to iteratively find the best solution within the feasible region, with a maximum of 50 iterations.

[0104] The collaborative arbitration module, before outputting the target instruction set, combines the device collaborative control matrix and coupling relationship dictionary to perform response rate normalization compensation and energy-mass balance interlock verification, and outputs the corrected execution instruction set to each sub-device. Specifically, this includes:

[0105] By combining the equipment collaborative control matrix and the coupling relationship dictionary, the response rate normalization compensation and energy-mass balance interlock verification are performed, and the corrected execution instruction set is output to each sub-device.

[0106] The process of performing device response rate normalization compensation is as follows:

[0107] Read the thermal inertia time constant τ of the chiller unit, and apply dynamic damping slope constraints to the adjustment commands issued by fast-response equipment such as variable frequency pumps and electric valves:

[0108] Adjustment amount (t) = Original expected adjustment amount (t) × min(1, thermal inertia time constant τ / (fast response device's own response time constant + preset zero offset number)), forcibly realizes the timing alignment of the adjustment process of fast and slow devices, and avoids overshoot and adjustment offset caused by the difference in response rate;

[0109] The energy-mass balance interlock verification process is as follows:

[0110] Calculate the heat dissipation on the cooling side and the cooling demand on the freezing side. If |heat dissipation on the cooling side - cooling demand on the freezing side| > ε×max(heat dissipation on the cooling side, cooling demand on the freezing side), where ε represents the relative tolerance coefficient allowed by the preset thermal balance, then it is determined that there is a hidden risk of operational conflict.

[0111] Pre-intercept conflicting instruction sets and trigger forced refactoring until the balance condition is met;

[0112] The forced reconstruction analysis process is as follows:

[0113] Obtain the difference ΔQ between the heat dissipation on the cooling side and the cooling requirement on the refrigeration side. Perform a judgment process on ΔQ. If ΔQ > 0, the cooling side is in excess, and the heat dissipation on the cooling side is reduced first. If ΔQ < 0, the cooling side is insufficient, and the heat dissipation on the cooling side is increased first.

[0114] Select the equipment that can be adjusted (controllable equipment in the water-cooled air conditioning system that has adjustment capabilities and participates in resolving thermal balance conflicts under the current operating state, such as cooling water pumps and cooling tower fans) and the current adjustable margin ΔQT:

[0115] When there is insufficient or excessive heat: ΔQT = Heat dissipation on the cooling side - Cooling requirement on the freezing side + ε × max (Heat dissipation on the cooling side, Cooling requirement on the freezing side);

[0116] Obtain the preset thermodynamic influence coefficient K for each device, and calculate the adjustment weight for each device in combination with the device health H: w = (K × H) / weighted sum of the products of preset thermodynamic influence coefficient and health in the set of adjustable devices, and device health H = Hj(t) - preset look-ahead coefficient × δj × (lj / ljmax).

[0117] Allocation of single-step adjustment amount for each device: Single-step adjustment amount = ΔQT / set number of steps × w × (1 / preset thermodynamic influence coefficient K);

[0118] After an adjustment is performed, the new heat dissipation on the cooling side and the cooling demand on the refrigeration side are recalculated, and the energy-mass balance interlock verification is performed again. If no hidden operational conflict risk is triggered, the refactoring is exited and the adjusted execution instruction set is output. If a hidden operational conflict risk is triggered, an alarm signal is triggered. The execution response module is used to respond to the alarm signal and immediately make the preset early warning operation corresponding to the alarm signal so that timely control can be carried out through manual intervention.

[0119] The adjusted set of execution instructions is sent to each sub-device (chiller unit, water pump, valve, cooling tower fan, etc.). Each device returns an execution confirmation instruction. If no execution confirmation instruction is returned, or the return time (indicating the time of sending) exceeds the preset return time, or no confirmation is received within the return time, an instruction warning signal is triggered.

[0120] Real-time acquisition of actual equipment response values ​​(such as actual outlet water temperature and actual frequency).

[0121] Calculate the deviation between the commanded value and the actual value: Deviation rate = |Actual value - Commanded value| / Commanded value

[0122] If the deviation rate is greater than the preset deviation threshold, then the execution deviation signal is triggered;

[0123] The execution response module is used to respond to instruction warning signals or execution deviation signals, and display the preset warning text corresponding to the instruction warning signal or execution deviation signal to intuitively understand the execution status of the instruction set;

[0124] In summary, by establishing a local correction mechanism of indicator collection, scoring, anomaly screening, correction decision matching, and load transfer, abnormal indicators are automatically identified and sorted by rate of change. The two indicators with the greatest impact are corrected first, and the preset decision is matched and load transfer is triggered. This enables rapid adaptive adjustment under abnormal operating conditions, reduces reliance on human experience, and constructs a multi-objective function with the goals of maximizing the overall energy efficiency ratio, minimizing the decline in health status, and minimizing the fluctuation of the outlet water temperature. At the same time, high load penalties are imposed on equipment with low health status to avoid the problem of excessively damaging a few high-efficiency equipment in pursuit of short-term energy efficiency, thus extending the overall service life of the system.

[0125] In addition, considering the rate difference between chillers (slow response) and pumps and valves (fast response), a dynamic damping slope constraint is designed to compress commands from fast-response equipment, force timing alignment, avoid repeated adjustments, overshoot oscillations, and command offsetting, and improve operational stability. At the same time, the thermal balance deviation between the cooling side and the refrigeration side is detected in real time. When the deviation exceeds the limit, a forced reconfiguration is automatically triggered: adjustable equipment is selected according to the direction of the deviation, and the weight is adjusted in combination with the health calculation to gradually correct the balance to meet the requirements, thereby reducing the risk of surge and energy efficiency loss.

[0126] The threshold is set for comparative analysis of results to determine whether they are good or bad. The value of the threshold is determined by a combination of large-scale model analysis of sample data and human experience. It can also be adjusted appropriately based on seasonal or common-sense influencing factors.

[0127] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An intelligent execution and control system for energy efficiency strategies adapted to water-cooled air conditioners, characterized in that: It includes a facility control center, a modeling and collaborative management module, a perception and strategy evaluation module, a local anomaly correction module, a global multi-objective optimization module, a collaborative arbitration module, and an execution response module; The modeling and collaborative management module is used to perform unified abstract modeling of all controllable devices in water-cooled air conditioners, and to build a dynamically updated device collaborative control matrix and a dictionary of coupling relationships between devices. The perception and strategy evaluation module is used to normalize the collected multi-dimensional index data, calculate the current strategy comprehensive score S using a weighted scoring method, and perform discrimination processing. Based on the discrimination results, it outputs a qualified signal, a trigger signal for entering local correction, or a unqualified signal. The local anomaly correction module is used to filter out abnormal indicators, calculate the abnormal change rate of abnormal indicators and sort them from high to low. It retains the abnormal indicators corresponding to the first and second first as abnormal indicators to be corrected, matches them with the preset indicator-correction decision library to output correction decisions, calculates the predicted load balance and the predicted health of each device at the future Tp time after correction, triggers early adjustment, and then performs load transfer and distribution. If it is still unqualified after correction, it outputs a correction unqualified signal. When generating or correcting unqualified signals, the global multi-objective optimization module is used to construct a multi-objective function that prioritizes energy efficiency while also considering health and stability. The module uses a preset hybrid model to solve the objective function and outputs the target instruction set of the global energy efficiency optimization strategy. The arbitration module is used to perform response rate normalization compensation and energy-mass balance interlock verification in conjunction with the device collaborative control matrix and coupling relationship dictionary before outputting the target instruction set, and outputs the modified execution instruction set to each sub-device.

2. The intelligent execution and control system for energy efficiency strategies adapted to water-cooled air conditioners according to claim 1, characterized in that, The analysis process of the unified abstract modeling is as follows: Register the following key attributes for all controllable devices in the water-cooled air conditioner: Device identifier: unique ID, type, response characteristic parameters: thermal inertia / response time constant τj, upper and lower limits of adjustment rate [Vjmin, Vjmax] and initial health value, and mark the device as j (j>0). Obtain the dynamic health decay rate δj and predict the health Hj(t+Tp): Calculate the health after Tp (Tp>0) time based on the current load rate and decay rate; Hj(t+Tp)=Hj(t)-δj×(lj(t) / ljmax)×Tp to update the current health status, where ljmax is the rated maximum load rate of device j; Construct a dynamically updated device cooperative control matrix C∈R N×M Where N is the total number of devices, M is the dimension of control attributes, and the coupling relationships between devices are predefined, and a coupling relationship dictionary G is established.

3. The intelligent execution and control system for energy efficiency strategies adapted to water-cooled air conditioners according to claim 1, characterized in that, The analysis process for the current strategy's overall score S is as follows: S1: Collect multi-dimensional index data of the energy efficiency strategy currently being implemented by the water-cooled air conditioner within a set period. The multi-dimensional index data includes energy efficiency index E, stability index W, equipment load balance degree J, and command conflict frequency C. S2: Normalize the multidimensional indicator data, and calculate the current strategy's comprehensive score S based on the weighted scoring method after normalization; S3: Compare the current strategy comprehensive score S with the preset current strategy comprehensive score range [Smin, Smax] and output a qualified signal, a local correction, or a unqualified signal.

4. The intelligent execution and control system for energy efficiency strategies adapted to water-cooled air conditioners according to claim 1, characterized in that, The analysis process of the local anomaly correction module is as follows: Each indicator is compared with its corresponding preset threshold to filter out abnormal indicators, obtain the abnormal change rate of the abnormal indicators, sort the abnormal change rates from high to low, retain the abnormal indicators corresponding to the first and second highest abnormal change rates, and set them as abnormal indicators to be corrected. Retrieve preset indicators and correction decisions, match the abnormal indicators to be corrected with the preset indicators and correction decisions, and output the correction decisions.

5. The intelligent execution and control system for energy efficiency strategies adapted to water-cooled air conditioners according to claim 4, characterized in that, After the abnormal indicators are corrected, the predicted load balance and the predicted health of each device at the future time Tp are calculated. If the predicted load balance is greater than the preset load balance threshold, or if the predicted health is less than the preset health threshold, or if a high-load device with high attenuation exceeds the preset duration, then the load distribution will be adjusted in advance.

6. The intelligent execution and control system for energy efficiency strategies adapted to water-cooled air conditioners according to claim 5, characterized in that, The process of pre-adjusting load allocation analysis is as follows: Construct a source device set and a target device set, obtain the remaining capacity of each target device, then calculate the total remaining capacity of all target devices, and calculate the utilization ratio of each target device. Utilization ratio = remaining capacity / total remaining capacity; The total load to be transferred is allocated according to the proportion of the load to be borne. The load capacity of the target equipment is obtained, and the load capacity is converted into a load rate increment. Adjustment is performed based on the load rate increment. After adjustment, the set values ​​of each equipment are updated, and the health prediction column is refreshed synchronously through the equipment collaborative control matrix. If the predicted load balance and predicted health meet the requirements after adjustment, steps S1-S3 are not executed and the current strategy is maintained directly. Otherwise, steps S1-S3 are executed. If the result is still unqualified after correction, a correction unqualified signal is generated.

7. The intelligent execution and control system for energy efficiency strategies adapted to water-cooled air conditioners according to claim 1, characterized in that, The analysis process of the target instruction set is as follows: Construct a multi-objective function that prioritizes energy efficiency while also considering health and stability. Combined with constraints, use a preset hybrid model to solve the objective function and output the target instruction set of the global energy efficiency optimization strategy.

8. The intelligent execution and control system for energy efficiency strategies adapted to water-cooled air conditioners according to claim 1, characterized in that, The analysis process of the collaborative arbitration module is as follows: By combining the equipment collaborative control matrix and the coupling relationship dictionary, the response rate normalization compensation and energy-mass balance interlock verification are performed, and the corrected execution instruction set is output to each sub-device. The normalization compensation process for the equipment response rate is as follows: Read the thermal inertia time constant τ of the chiller unit, apply dynamic damping slope constraint to the adjustment command issued by the fast response equipment: Adjustment amount (t) = original expected adjustment amount (t) × min(1, thermal inertia time constant τ / (fast response equipment's own response time constant + preset zero bias number)).

9. The intelligent execution and control system for energy efficiency strategies adapted to water-cooled air conditioners according to claim 8, characterized in that, The quantity-mass balance interlock verification process is as follows: Calculate the heat dissipation on the cooling side and the cooling demand on the freezing side. If |heat dissipation on the cooling side - cooling demand on the freezing side| > ε×max(heat dissipation on the cooling side, cooling demand on the freezing side), where ε represents the relative tolerance coefficient allowed by the preset thermal balance, then it is determined that there is a hidden risk of operational conflict. Pre-intercept conflicting instruction sets and trigger forced refactoring until the balance condition is met; The forced reconstruction analysis process is as follows: obtain the difference ΔQ between the heat dissipation on the cooling side and the cooling requirement on the freezing side, perform discrimination processing on ΔQ, and output whether the cooling side is excessive or insufficient. Select the equipment that can be adjusted and the current adjustable margin ΔQT. When it is insufficient or excessive: ΔQT = heat dissipation on the cooling side - cooling required on the freezing side + ε × max (heat dissipation on the cooling side, cooling required on the freezing side). Obtain the preset thermodynamic influence coefficient K for each device, combine it with the device health H, calculate the adjustment weight w for each device, and allocate the single-step adjustment amount for each device: Single-step adjustment amount = ΔQT / set number of steps × w × (1 / preset thermodynamic influence coefficient K). After an adjustment is performed, the new heat dissipation on the cooling side and the cooling demand on the refrigeration side are recalculated, and the energy-mass balance interlock check is performed again. If no hidden operational conflict risk is triggered, the refactoring is exited and the adjusted execution instruction set is output. If a hidden operational conflict risk is triggered, an alarm signal is triggered.

Images