Energy-saving and consumption-reducing method for water system of central air conditioner

Abstract

The invention relates to the technical field of central air conditioners, and discloses a central air conditioner water system energy-saving and consumption-reducing method which comprises the following steps: S1, acquiring operation data of a central air conditioner water system, and preprocessing the operation data; and S2, a cold load prediction model is established based on the operation data, and the model combines wavelet threshold denoising, a convolutional neural network and a long and short term memory network to accurately predict the future cold load demand. According to the energy-saving and consumption-reducing method for the water system of the central air conditioner, through the synergistic effect of a cooling load prediction model and a parameter optimization strategy, the system can dynamically match real-time load requirements, the model is combined with a convolutional neural network to extract local features and capture time correlation of a long-short-term memory network, and high-precision prediction of cooling loads in the next two hours is achieved; and compared with a traditional fixed parameter operation mode, invalid energy consumption of equipment such as a water pump and a fan is reduced, and energy conservation is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of central air conditioners, and particularly to an energy-saving and consumption-reducing method for the water system of a central air conditioner. Background Art

[0002] The central air conditioning system is one of the main energy-consuming equipment in modern buildings, accounting for about 25%-40% of the total building energy consumption.

[0003] Since the traditional control system is designed according to the full-load state, but in actual applications, it operates under partial load in most cases, resulting in serious energy waste. In recent years, with the increasingly severe global environmental problems, various countries have successively formulated policies to promote energy conservation and emission reduction. Therefore, how to effectively reduce the energy consumption of the central air conditioning water system and reduce carbon emissions has become an urgent problem to be solved, and thus an energy-saving and consumption-reducing method for the water system of a central air conditioner is proposed. Summary of the Invention

[0004] (I) Technical Problems to be Solved

[0005] In view of the deficiencies of the prior art, the present invention provides an energy-saving and consumption-reducing method for the water system of a central air conditioner. Through accurate cooling load prediction and advanced intelligent optimization algorithms, comprehensive optimization control of the central air conditioning water system is achieved, thereby achieving the purpose of energy conservation and carbon reduction, and solving the problem of serious energy waste caused by the fact that the traditional control system is designed according to the full-load state, but in actual applications, it operates under partial load in most cases.

[0006] (II) Technical Solutions

[0007] To achieve the above-mentioned comprehensive optimization control of the central air conditioning water system through accurate cooling load prediction and advanced intelligent optimization algorithms, thereby achieving the purpose of energy conservation and carbon reduction, the present invention provides the following technical solutions: An energy-saving and consumption-reducing method for the water system of a central air conditioner, characterized by comprising the following steps:

[0008] S1. Collect the operation data of the central air conditioning water system and perform preprocessing;

[0009] S2: Based on the operation data, establish a cooling load prediction model, which combines wavelet threshold denoising, convolutional neural network and long short-term memory network to accurately predict the future cooling load demand;

[0010] S3: According to the cooling load prediction result, optimize the operation parameters of the chilled water and cooling water circulation systems, specifically including adjusting the flow rates of the chilled water pump and the cooling water pump, the chilled water supply temperature, and the frequency of the cooling tower fan;

[0011] S4: Use the multi-strategy fusion improved sparrow search algorithm (MISSA) to globally optimize the operating parameters of the chiller, chilled water pump, cooling water pump, air-conditioning terminal fan, and cooling tower, with the minimum total energy consumption as the objective function.

[0012] Preferably, in step S2, the specific construction steps of the cooling load prediction model are as follows:

[0013] S2.1: Use wavelet threshold denoising technology to denoise the original operating data;

[0014] S2.2: Input the denoised data into a convolutional neural network (CNN) to extract local features in the time series;

[0015] S2.3: Use the features extracted by the CNN as input and input them into a long short-term memory network (LSTM) to further extract time correlation features and output the cooling load prediction results;

[0016] S2.4: Identify and optimize the unknown parameters in the cooling load prediction model through the Levenberg–Marquardt parameter identification algorithm.

[0017] Preferably, in step S3, the operating parameter optimization method for the chilled water and cooling water circulation systems includes:

[0018] The chilled water flow rate range is 30 m 3 / h to 550 m 3 / h;

[0019] The cooling water flow rate range is 390 m 3 / h to 650 m 3 / h;

[0020] The chilled water inlet temperature range is 7°C to 12°C;

[0021] The chilled water outlet temperature range is 8°C to 15°C;

[0022] The cooling water inlet temperature range is 15°C to 35°C;

[0023] The operating frequency range of the cooling tower fan is 30 Hz to 50 Hz.

[0024] Preferably, in step S4, the specific implementation steps of the MISSA algorithm are as follows:

[0025] Initialize the population and use Circle chaotic mapping to generate the initial sparrow positions;

[0026] The position update formula for the discoverers is:

[0027]

[0028] Among them, X is the position of the sparrow, t represents the current iteration number, j = 1, 2, …, n, α is a random number, R 2 and ST are the alarm value and the safety threshold respectively, Q is a random number obeying the normal distribution, and L is the identity matrix;

[0029] The position update formula for the joiner is:

[0030]

[0031] Among them, is the worst position, is the current position of the discoverer, is the best position;

[0032] Randomly generate some sparrows as scouts. When the fitness value is poor, according to the formula:

[0033]

[0034] perform position update, where β is a random number of Gaussian distribution, K is a random coefficient of direction and step size, f i is the fitness value of the current sparrow, f worst is the worst fitness value.

[0035] Preferably, in the step S4, the objective function is:

[0036]

[0037] Among them, E is the total energy consumption, and the unit is kilowatt (kW); N is the number of devices; P chiller,i is the energy consumption of the i-th chiller, and the unit is kilowatt (kW); P cpump,i is the energy consumption of the i-th chilled water pump, and the unit is kilowatt (kW); P fpump,i is the energy consumption of the i-th cooling water pump, and the unit is kilowatt (kW); P ctower,i is the energy consumption of the i-th cooling tower, and the unit is kilowatt (kW).

[0038] Preferably, in the step S4, the constraint conditions include:

[0039] The number of chillers N chiller ranges from 1 to 4;

[0040] The number of chilled water pumps N cpump ranges from 1 to 4;

[0041] The number of cooling water pumps N fpump ranges from 1 to 4;

[0042] The number of cooling towers N ctowerRange from 1 to 16;

[0043] The number of operating devices needs to meet the actual load demand and be dynamically adjusted within the above range.

[0044] Preferably, in the step S4, a carbon emission objective function is introduced, and the calculation formula is:

[0045]

[0046] Where C is the total carbon emission, C chiller,i is the carbon emission of the i-th chiller, and the unit is kilograms of carbon dioxide per hour (kgCO 2 / h); C cpump,i is the carbon emission of the i-th chilled water pump, and the unit is kilograms of carbon dioxide per hour (kgCO 2 / h); C fpump,i is the carbon emission of the i-th cooling water pump, and the unit is kilograms of carbon dioxide per hour (kgCO 2 / h); C ctower,i is the carbon emission of the i-th cooling tower, and the unit is kilograms of carbon dioxide per hour (kgCO 2 / h).

[0047] Preferably, in the step S4, the grey wolf optimization algorithm (GWO) is used to optimize the controllable variables of the cooling system, and the specific steps are as follows:

[0048] Initialize the grey wolf population, and define the leader α, followers β and δ;

[0049] Update the position of each grey wolf individual according to the formula:

[0050] D = |C·X p (t) - X(t)|

[0051] X(t + 1) = X p (t) - A·D

[0052] Where D is the distance vector, X p (t) is the position vector of the prey, X(t) is the position vector of the grey wolf individual, and A and C are control parameters;

[0053] Through multiple iterations, gradually approach the optimal solution, and finally determine the best operating parameter combination of each device.

[0054] (III) Beneficial effects

[0055] Compared with the prior art, the present invention provides a method for energy saving and consumption reduction of the water system of a central air conditioner, and has the following beneficial effects:

[0056] 1. The energy-saving and consumption-reducing method for the water system of the central air conditioner enables the system to dynamically match the real-time load demand through the synergistic effect of the cooling load prediction model and the parameter optimization strategy. Specifically, the model filters out data noise through wavelet threshold denoising technology, extracts local features by combining a convolutional neural network, and captures temporal correlations using a long short-term memory network to achieve high-precision prediction of the cooling load for the next 2 hours. Based on the prediction results, the system adjusts the chilled water flow rate, supply water temperature, and cooling tower fan frequency to accurately match the equipment operating parameters with the actual load, reducing the ineffective energy consumption of equipment such as pumps and fans compared to the traditional fixed-parameter operation mode and achieving energy savings.

[0057] 2. The energy-saving and consumption-reducing method for the water system of the central air conditioner further improves the system energy efficiency by using a multi-strategy fusion improved sparrow search algorithm to globally optimize the operating parameters of the chiller, pump, and cooling tower. The algorithm initializes the population using Circle chaotic mapping and combines the dynamic position update strategies of discoverers, joiners, and scouts to search for the optimal solution under the dual-objective constraints of energy consumption and carbon emissions. For example, after optimizing parameters such as the number of chillers and the chilled water pump frequency using MISSA, the total system energy consumption is reduced by 12% - 18%. At the same time, the grey wolf optimization algorithm is used to assist in adjusting the cooling system variables, and through the iterative approximation of leaders, followers, and scouts, the cooling efficiency is further improved and carbon emissions are reduced. This method breaks through the limitations of traditional single-device independent control, realizes the collaborative optimization of multiple devices, and significantly reduces the overall operating cost and environmental load of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a schematic structural diagram of the system of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0059] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the embodiments and drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0060] Please refer to Figure 1 , an energy-saving and consumption-reducing method for the water system of a central air conditioner, comprising the following steps:

[0061] S1. Collect the operation data of the central air conditioner water system and perform preprocessing;

[0062] S2: Establish a cooling load prediction model based on the operation data. This model combines wavelet threshold denoising, convolutional neural network, and long short-term memory network to accurately predict the future cooling load demand;

[0063] S3: According to the predicted chilled load results, optimize the operating parameters of the chilled water and cooling water circulation systems, specifically including adjusting the flow rates of the chilled water pumps and cooling water pumps, the chilled water supply temperature, and the frequency of the cooling tower fans;

[0064] S4: Use the multi-strategy fusion improved sparrow search algorithm (MISSA) to globally optimize the operating parameters of the chiller, chilled water pumps, cooling water pumps, air-conditioning terminal fans, and cooling towers, with the objective function of minimizing the total energy consumption.

[0065] Example 1:

[0066] The method for constructing and real-time applying a chilled load prediction model based on deep learning provided in this example has the following technical key points: The model adopts a hybrid convolutional neural network (CNN) and long short-term memory network (LSTM) architecture and is deployed on edge computing devices to achieve low-latency prediction.

[0067] In specific implementation, the data acquisition layer uses devices such as PT100 temperature sensors and EMF-100 flow meters to collect parameters such as chilled water temperature, cooling water flow rate, and air-conditioning terminal temperature and humidity in real time. The data acquisition frequency is 1 time per minute. After the original data is denoised by wavelet threshold (using db4 wavelet basis for 3-layer decomposition), it is input into the CNN module for feature extraction.

[0068] The CNN contains 3 convolutional layers (kernel size = 3×3, stride = 1, ReLU activation function), and the number of output channels is 32, 64, and 128 respectively. Each layer is followed by a max pooling layer (pool size = 2×2). Finally, the size of the output feature map is 8×64. The feature map is input into a bidirectional LSTM network with 128 hidden layer nodes, and the time step is set to 24 hours (corresponding to 96 sampling points) to output the predicted chilled load value for the next 2 hours.

[0069] The model parameters are optimized online by the Levenberg-Marquardt algorithm, and the objective function is to minimize the sum of squared prediction errors, as shown in the following formula:

[0070]

[0071] where the unit of Loss is kilowatt squared (kW 2 ) and are in the unit of kilowatt (kW).

[0072] The iteration termination condition is set as the error ≤ 0.5% or the number of iterations reaches 500 times. The model training dataset contains 3 years of historical operation data (about 100,000 sets of samples), and the accuracy of the validation set reaches 98.2%. After deployment, the online parameter update module is automatically called every 24 hours to fine-tune the model weights according to real-time data, ensuring that the prediction accuracy dynamically adapts to seasonal changes.

[0073] Example 2:

[0074] This example describes a dynamic optimization method for the chilled water and cooling water circulation systems, and its core innovation points include a parameter adjustment strategy that combines PID control and fuzzy control.

[0075] The chilled water circulation system realizes flow control through a variable-frequency water pump (model CP-500, frequency range 30 - 50Hz) and an electric control valve (model V2000). According to the predicted value of the cooling load (Q load ), the PID controller dynamically adjusts the water pump frequency, and the formula is:

[0076]

[0077] where Q cold target is in cubic meters per hour (m 3 / h); Q load is in kilowatts (kW), C p is in kilojoules per kilogram per degree Celsius (kJ / (kg·℃)); T outcold and T incold are in degrees Celsius (℃);

[0078] C p = 4.18 kJ / kg·℃, T incold is set to 7℃ to 12℃, T outcold is PID-tuned to incold + 1℃, and the PID parameters are set as: proportional coefficient K p = 0.8, integral time T i = 10 minutes, derivative time T d = 2 minutes. When the chilled water flow rate fluctuates by more than ±5%, the stable flow rate is adjusted by the opening of the electric control valve.

[0079] The cooling water system adopts a cooling tower group control strategy and dynamically adjusts the fan operation mode according to the cooling water outlet temperature (T outcool ): when T outcool > 32℃, the cooling tower fans switch to the high-speed mode (frequency 50Hz), and all cooling towers (up to 16 units) are turned on; when 28℃ < T outcool ≤ 32℃, it switches to the medium-speed mode (frequency 40Hz), and some cooling towers are turned off; when T outcoolWhen the temperature is ≤ 28°C, switch to the low-speed mode (frequency 30 Hz), and only keep 1 - 2 units running.

[0080] The frequency of the cooling water pump is adjusted by a fuzzy control algorithm. The input variable is the cooling water temperature difference (ΔT cool = T incool - T outcool ), and the output variable is the pump frequency. Example of fuzzy rules: If ΔT cool > 5°C → increase the frequency by 2 Hz; if ΔT cool < 2°C → decrease the frequency by 2 Hz.

[0081] Example 3:

[0082] The improved sparrow search algorithm (MISSA) proposed in this example is innovative in combining Circle chaotic mapping initialization and dynamic step size adjustment strategy.

[0083] The initial population size of the algorithm is 50, and the dimensions include parameters such as the number of operating chiller units (1 - 4), the frequency of the chilled water pump (30 - 50 Hz), and the cooling tower fan gear (low / medium / high). The initial population is generated through Circle chaotic mapping:

[0084]

[0085] where u i is a uniformly distributed random number. During the iterative process of the algorithm, the update formulas for the discoverer, joiner, and scout are respectively:

[0086]

[0087] where α is a Gaussian parameter with variance decreasing with the number of iterations. The objective function simultaneously minimizes energy consumption (E) and carbon emissions (C):

[0088]

[0089] where EF is the carbon emission coefficient. The constraint conditions include the number of operating chiller units ≤ the physical number of units, the chilled water temperature range (7°C - 12°C), and the upper limit of the cooling water temperature (32°C). After the algorithm converges, the optimal parameter combination is output and sent to the PLC controller in real time to adjust the operating parameters of the equipment. F

[0090] Example 4:

[0091] This example describes the cooperative optimization method of the GWO algorithm and MISSA, which is innovative in combining cooling system parameter optimization and multi-objective fusion strategy.

[0092] The population size of the GWO algorithm is initialized to 30, and the dimensions include the cooling water pump frequency (35 - 60 Hz), the opening of the cooling tower fan (0 - 100%), and the set cooling water temperature (25°C ± 1°C). The positions of the leader (Alpha), followers (Beta), and scouts (Delta) are updated through the following formula:

[0093] X(t + 1) = X p (t) - A·D

[0094] During control execution, the cooling water pump frequency is adjusted by a frequency converter (model VFD - 700), and the response time ≤ 2 seconds; the opening of the cooling tower fan is controlled by a PWM signal with a step size of 5%.

[0095] Example Five:

[0096] This example describes the system integration solution and real - time monitoring method, and its innovations include the OPC UA communication protocol and fault diagnosis logic.

[0097] The central controller connects to the sensor network, optimization controller, and PLC through the OPC UA protocol, and the communication cycle is 10 seconds.

[0098] The cold load prediction model transmits the results to the controller through the MQTT protocol, and the device control instructions are issued through ModbusTCP / IP.

[0099] The real - time monitoring interface displays the chilled water temperature (T incold ), the cooling water flow rate (Q cool ), the total system energy consumption (kW), and the carbon emissions (kgCO 2 / h).

[0100] The fault handling logic includes:

[0101] When the chilled water pump frequency is lower than 30 Hz for 5 minutes continuously, an alarm is triggered and the standby water pump is switched.

[0102] When the cooling tower fan stops, the corresponding cooling water pump is automatically shut down and the standby cooling tower is started.

[0103] The vibration sensor (model VB - 5) monitors the state of the water pump bearing in real - time. When the vibration value exceeds 5 mm / s, it automatically shuts down and records the fault code (such as bearing wear).

[0104] The data is stored in the MySQL database, supporting historical data query and trend analysis. An energy - saving report is automatically generated every month to compare the energy consumption and carbon emission data before and after optimization; the storage, analysis of energy consumption data and the automatic generation of energy - saving reports are realized through the following core data tables, supporting historical data query, trend analysis, and comparison of optimization effects:

[0105] Table (1) Equipment Basic Information Table

[0106] Field Name Data Type Remarks device_id INTPRIMARYKEY Unique Device Identifier device_type VARCHAR(50) Device Type (e.g., Chiller, Pump) location VARCHAR(100) Device Installation Location

[0107] This data table stores the static information of the equipment and provides an associated benchmark for dynamic data.

[0108] Table (2) Real-time Operation Data Table

[0109] Field Name Data Type Remarks data_id BIGINTPRIMARYKEY Auto - increment Primary Key device_id INT Foreign Key, associated with device_info timestamp DATETIME Data Record Time power_consumption FLOAT Real - time Energy Consumption (unit: kW) carbon_emission FLOAT <![CDATA[Real-time carbon emissions (unit: kgCO 2 )]]>

[0110] This data table records the real-time operation parameters of the equipment and supports historical backtracking and high-frequency queries.

[0111] Table (3) Historical Data Summary Table

[0112] Field Name Data Type Remarks summary_id INTPRIMARYKEY Primary Key time_period DATETIME Start Time of Time Period (e.g., 2023 - 10 - 01) period_type VARCHAR(10) Time Granularity (e.g., "hour", "month") total_energy FLOAT Total Energy Consumption within Specified Time Period total_carbon FLOAT Total Carbon Emission within Specified Time Period

[0113] This data table aggregates data according to the time granularity and supports trend analysis.

[0114] Table (4) Optimization Parameter Table

[0115]

[0116]

[0117] This data table stores the optimized parameters and effects and supports the comparison of algorithm effects.

[0118] Table (5) Energy Saving Report Table

[0119] Field Name Data Type Remarks report_id INTPRIMARYKEY Primary Key report_month DATE Report Month (e.g., 2023 - 10) pre_optimization_energy FLOAT Total Energy Consumption before Optimization post_optimization_energy FLOAT Total Energy Consumption after Optimization energy_saving_percent FLOAT Energy Saving Percentage (e.g., 15.2%) pre_optimization_carbon FLOAT Total Carbon Emission before Optimization carbon_reduction_percent FLOAT Carbon Emission Reduction Percentage (e.g., 12.8%)

[0120] This data table stores the energy saving report data automatically generated monthly and compares the energy consumption and carbon emissions before and after optimization.

[0121] Example Six:

[0122] This example verifies the energy saving effect under the synergistic action of the cooling load prediction model and the parameter optimization strategy, and analyzes the influence of the adjustment of chilled water flow rate, supply water temperature and cooling tower fan frequency on the comprehensive energy consumption.

[0123] Experimental Preparation:

[0124] (1) Experimental Environment Setup

[0125] Select the central air-conditioning system of a commercial building, and configure chillers (4 sets), chilled water pumps (3 sets), cooling water pumps (2 sets), and cooling towers (8 groups);

[0126] Deploy a sensor network: temperature sensors (PT100), flow meters (EMF-100), power meters (WT3000);

[0127] Build edge computing devices and deploy a cold load prediction model (CNN-LSTM) and an optimization controller.

[0128] (2) Data collection and preprocessing

[0129] Collect historical operation data (for 3 years) for model training, and the accuracy of the validation set should be ≥98%;

[0130] Collect parameters such as chilled water temperature, cooling water flow rate, and cold load demand in real time, with a frequency of 1 time per minute.

[0131] (3) Benchmark parameter setting (traditional mode)

[0132] Table (6) Benchmark parameter setting table

[0133] Parameter Name Value Description Chilled Water Flow Rate <![CDATA[120m 3 / h]]> Fixed Flow Rate Operation Chilled Water Supply Temperature 7℃ Lower Limit of Design Temperature Cooling Tower Fan Frequency 50Hz Full - load Operation Number of Operating Chillers 4 units All Turned On

[0134] The experimental results are as follows:

[0135] Table (7) Energy-saving report form

[0136]

[0137] By dynamically adjusting the parameters, the comprehensive energy consumption is reduced by 29.3%. Among them, the adjustment of the cooling tower fan frequency contributes the most (energy-saving rate of 42%), followed by the chilled water flow rate (26.7%) and the optimization of the supply water temperature (12.4%); Taking a commercial building as an example, the average annual electricity bill is reduced from 1.2 million yuan to 849,000 yuan, saving 351,000 yuan. At the same time, the annual carbon emissions are reduced by about 42 tons of carbon dioxide (based on an emission factor of 0.5 kg of carbon dioxide / kWh), significantly improving energy efficiency and contributing to the carbon neutrality goal.

[0138] Example 7:

[0139] In this example, through the collaborative optimization of the MISSA algorithm and the GWO algorithm, the parameters of the chiller, water pump, and cooling tower in the central air-conditioning water system are dynamically adjusted, verifying the effectiveness of the algorithm in reducing system energy consumption and carbon emissions and breaking through the limitations of the traditional fixed control mode.

[0140] Implementation steps: (1) Set the number of operating chillers to 5 units (full-load configuration), the frequency of the chilled water pump to 50 Hz (rated frequency), and the rotational speed of the cooling tower fan to 60% of the rated power; (2) Initialize the population through the Circle chaotic mapping, combine the dynamic position update strategies of the discoverer, joiner, and scout, input real-time environmental temperature, cold load demand, and equipment energy consumption data, and search for the optimal solution under the dual-objective constraints of energy consumption and carbon emissions; (3) Iteratively optimize the cooling system variables (such as the rotational speed of the cooling tower fan and the frequency of the cooling water pump) to further improve the cooling efficiency.

[0141] The comparison results of the optimization effect and parameters are as follows:

[0142] Table (8) Comparison Table of Optimization Effect and Parameters

[0143]

[0144]

[0145] Through multi-device collaborative optimization, redundant operation of the chiller is avoided, and the dynamic adjustment of the pump frequency and fan speed significantly reduces mechanical losses. Finally, the total system energy consumption is reduced by 15% (within the expected range), and carbon emissions are reduced by 15%.

[0146] In summary, for the energy-saving and consumption-reducing method of the water system of this central air conditioner, through the synergistic effect of the cooling load prediction model and the parameter optimization strategy, the system can dynamically match the real-time load demand. Specifically, the model filters data noise through wavelet threshold denoising technology, extracts local features by combining convolutional neural networks, and captures time correlations by long short-term memory networks to achieve high-precision prediction of the cooling load for the next 2 hours. Based on the prediction results, the system adjusts the chilled water flow rate, supply water temperature, and cooling tower fan frequency to accurately match the equipment operation parameters with the actual load. Compared with the traditional fixed-parameter operation mode, it reduces the ineffective energy consumption of equipment such as pumps and fans, achieving energy conservation.

[0147] Moreover, for the energy-saving and consumption-reducing method of the water system of this central air conditioner, the sparrow search algorithm improved by multi-strategy fusion globally optimizes the operation parameters of the chiller, pump, and cooling tower, further improving the system energy efficiency. The algorithm initializes the population using Circle chaotic mapping and combines the dynamic position update strategies of discoverers, joiners, and scouts to search for the optimal solution under the dual-objective constraints of energy consumption and carbon emissions. For example, after optimizing parameters such as the number of chillers and the chilled water pump frequency through MISSA, the total system energy consumption is reduced by 12% - 18%. At the same time, the grey wolf optimization algorithm assists in adjusting the cooling system variables, and further improves the cooling efficiency and reduces carbon emissions through the iterative approximation of leaders, followers, and scouts. This method breaks through the limitations of traditional single-device independent control, realizes multi-device collaborative optimization, significantly reduces the overall system operation cost and environmental load, and solves the problem of serious energy waste caused by the fact that traditional control systems are designed according to the full-load state but operate at partial load in most actual applications.

[0148] All relevant modules involved in this system are hardware system modules or functional modules formed by combining computer software programs or protocols in the prior art with hardware. The computer software programs or protocols themselves involved in such functional modules are well-known technologies to those skilled in the art and are not the improvements of this system. The improvement of this system lies in the interaction relationships or connection relationships among the modules, that is, the improvement of the overall structure of the system to solve the corresponding technical problems to be solved by this system.

[0149] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for energy saving and consumption reduction of a water system of a central air conditioner, characterized in that: The following steps are involved: S1. Collect the operation data of the central air-conditioning water system and perform pre-processing; S2: establishing a cooling load prediction model based on the operation data, the model combining wavelet threshold denoising, convolutional neural network and long short-term memory network to accurately predict future cooling load demand; S3: Optimizing the operating parameters of the chilled water and cooling water circulation systems according to the cooling load prediction results, specifically including adjusting the flow rates of the chilled water pump and the cooling water pump, the chilled water supply temperature, and the cooling tower fan frequency; S4: The multi-strategy fusion improved sparrow search algorithm (MISSA) is used to globally optimize the operating parameters of the chiller, chilled water pump, cooling water pump, air conditioning terminal fan and cooling tower, with the objective function of minimizing the total energy consumption.

2. A method for energy saving and consumption reduction of a water system of a central air conditioner according to claim 1, characterized in that: In step S2, the specific steps of constructing the cooling load prediction model are as follows: S2.1: De-noising the original running data using wavelet threshold denoising technology; S2.2: Input the denoised data into a convolutional neural network (CNN) to extract local features in the time series; S2.3: The features extracted by CNN are used as input to the long short-term memory network (LSTM), which further extracts the time-correlation features and outputs the cooling load prediction results; S2.4: Identify and optimize unknown parameters in the cooling load prediction model using the Levenberg–Marquardt parameter identification algorithm.

3. A method for energy saving and consumption reduction of a water system of a central air conditioner according to claim 1, characterized in that: In step S3, the method for optimizing the operating parameters of the chilled water and cooling water circulation system includes: Chilled water flow range is 30m 3 / h to 550m 3 / h; Cooling water flow range is 390m 3 / h to 650m 3 / h; The chilled water inlet temperature ranges from 7°C to 12°C; The chilled water outlet temperature ranges from 8°C to 15°C; The cooling water inlet temperature ranges from 15°C to 35°C; Cooling tower fans operate in the frequency range of 30Hz to 50Hz.

4. A method for energy saving and consumption reduction of a water system of a central air conditioner according to claim 1, characterized in that: In step S4, the specific implementation steps of the MISSA algorithm are as follows: Initialize the population and use Circle chaos mapping to generate the initial sparrow positions; The finder's position update formula is: Where X is the sparrow position, t is the current iteration number, j = 1, 2, ..., n, α is a random number, R2 and ST are the alarm value and safety threshold respectively, Q is a random number that obeys the normal distribution, and L is the unit matrix; The position update formula of the joiner is: in, The worst position, is the current location of the finder, For the best position; Randomly generate some sparrows as scouts. When the fitness value is poor, follow the formula: Perform position update, where β is a Gaussian distribution random number, K is the random coefficient of direction and step length, and f i is the fitness value of the current sparrow, f worst is the worst fitness value.

5. A method for energy saving and consumption reduction of a water system of a central air conditioner according to claim 1, characterized in that: In step S4, the objective function is: Where, E is the total energy consumption in kilowatts (kW); N is the number of devices; P chiller,i is the energy consumption of the i-th chiller, in kilowatts (kW); P cpump,i is the energy consumption of the i-th chilled water pump, in kilowatts (kW); P fpump,i is the energy consumption of the i-th cooling water pump, in kilowatts (kW); P ctower,i is the energy consumption of the i-th cooling tower, in kilowatts (kW).

6. A method for energy saving and consumption reduction of a water system of a central air conditioner according to claim 1, characterized in that: In step S4, the constraints include: Number of chillers N chiller The range is 1 to 4; Number of chilled water pumps N cpump The range is 1 to 4; Number of cooling water pumps N fpump The range is 1 to 4; Number of cooling towers N ctower The range is 1 to 16; The number of operating equipment must meet the actual load demand and be dynamically adjusted within the above range.

7. A method for energy saving and consumption reduction of a water system of a central air conditioner according to claim 1, characterized in that: In step S4, the carbon emission objective function is introduced, and the calculation formula is: Among them, C is the total carbon emissions, C chiller,i is the carbon emission of the i-th chiller, in kilograms of carbon dioxide per hour (kgCO2 / h); C cpump,i is the carbon emission of the i-th chilled water pump, in kilograms of carbon dioxide per hour (kgCO2 / h); C fpump,i is the carbon emission of the i-th cooling water pump, in kilograms of carbon dioxide per hour (kgCO2 / h); C ctower,i is the carbon emission of the i-th cooling tower, in kilograms of carbon dioxide per hour (kgCO2 / h).

8. A method for energy saving and consumption reduction of a water system of a central air conditioner according to claim 1, characterized in that: In step S4, the controllable variables of the cooling system are optimized using the Grey Wolf Optimization Algorithm (GWO), and the specific steps are as follows: Initialize the gray wolf group and define the leader α, followers β and δ; Update the position of each individual gray wolf according to the formula: D=|C·X p (t)-X(t)| X(t+1)=X p (t)-A·D Where D is the distance vector, X p (t) is the position vector of the prey, X(t) is the position vector of the individual gray wolf, and A and C are control parameters; Through multiple iterations, the optimal solution is gradually approached, and finally the best operating parameter combination of each device is determined.

Images