Cold station multi-equipment combination energy efficiency optimization method and system

Abstract

The invention discloses a cold station multi-equipment combination energy efficiency optimization method and system, relates to the field of cold station energy efficiency optimization, and aims to solve the problems of low energy efficiency and response lag caused by manual adjustment. The method comprises the following steps: S1, acquiring key operation data (power, state, temperature, flow and frequency) of a water chilling unit, a cooling tower and a water pump and external environment data (temperature and humidity, wind speed and load prediction) in real time; s2, constructing an equipment energy efficiency model based on historical data, and quantifying the change of the equipment efficiency along with the load rate, the temperature difference and the environment temperature; s3, according to the current load and environment prediction, simulating and traversing the energy consumption of the equipment start-stop combination and the water pump frequency combination by applying a Monte Carlo algorithm, and screening a combination strategy with the lowest total energy consumption of the system; s4, converting the optimization strategy into a control instruction, and issuing the control instruction to equipment for execution; and S5, comparing the actual energy consumption with the predicted value, and if the deviation exceeds 5%, correcting model parameters by adopting a least square regression algorithm to realize closed-loop optimization. And the energy efficiency and the response speed of the cold station are obviously improved.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent control technology, and in particular relates to a method and system for optimizing the energy efficiency of a multi-device combination in a cooling station. Background Art

[0002] A typical cooling station system includes a cluster of equipment such as chillers, cooling towers, and water pumps. The traditional extensive control mode leads to a large amount of ineffective energy consumption, which is in sharp conflict with the energy efficiency goals of green buildings.

[0003] Current mainstream energy management systems (such as Honeywell EBI and Schneider EcoStruxure) have three technical bottlenecks: Lack of multi-device coordination: Only single-device data monitoring (such as chiller power) is implemented. A joint energy efficiency model for chillers, cooling towers, and water pumps has not been established, making it impossible to quantify the impact of equipment interactions (for example, the coupling effect of cooling tower water temperature changes on chiller COP). Static scheduling is rigid: It relies on manual experience or fixed rule bases (such as a table of temperature thresholds for chiller start and stop). This results in delayed response to dynamic load fluctuations (the diurnal load difference in office buildings can reach 300%) and sudden weather changes (sudden rain or high temperatures), frequently triggering high-energy-consuming combinations. Model inaccuracy without correction: Equipment aging and seasonal performance drift cause the preset energy efficiency model (such as the COP curve) to deviate from the actual situation by more than 15%. The existing system lacks a closed-loop feedback mechanism, and long-term energy efficiency continues to deteriorate.

[0004] Therefore, there is an urgent need for an intelligent control method that integrates multi-device joint simulation, dynamic environment adaptation, and self-correction optimization to break through the bottleneck of cold station energy efficiency management. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method and system for optimizing the energy efficiency of a cold station multi-device combination in response to the above-mentioned deficiencies in the existing technology, which can realize the functions of intelligent control integrating multi-device joint simulation, dynamic environment adaptation, and self-correction optimization.

[0006] The first aspect of the present invention discloses a method for optimizing the energy efficiency of a multi-device combination in a cold station, comprising the following steps: S1. Data Collection: Real-time data collection of chiller, cooling tower, and water pump operating data in the cooling station system, including equipment power, operating status, outlet water temperature, return water temperature, water flow rate, and operating frequency. Data on external ambient temperature, humidity, wind speed, and air conditioning terminal load forecasts are also collected. S2. Simulation Modeling: Based on historical operating data, a chiller energy efficiency model, a cooling tower energy consumption model, and a pump characteristic curve model are constructed. These models characterize how equipment energy efficiency changes with load factor, inlet and outlet water temperature difference, and ambient temperature. S3. Monte Carlo Simulation Optimization: Based on the current load forecast and weather forecast data, the Monte Carlo algorithm is used to traverse the energy consumption simulation results of the chiller start-stop combinations, cooling tower start-stop combinations, and pump operating frequency combinations, and the combination strategy with the lowest total system energy consumption is selected. S4. Control and Distribution: Convert the combined strategy into equipment start / stop commands and frequency setpoints, which are then distributed to the chillers, cooling towers, and pumps via the building automation system or IoT interface. S5. Recursive feedback correction: Compare the deviation between the actual operating energy consumption and the simulated operating energy consumption prediction value. When the deviation exceeds 5%, the least squares regression algorithm is used to correct the model parameters.

[0007] In the above method, in step S2: The energy efficiency model of the chiller is a nonlinear regression function of COP with respect to load rate, chilled water outlet temperature, cooling water inlet temperature and ambient humidity; The cooling tower energy consumption model is a piecewise linear interpolation function of the total power of the fan and water pump with respect to the inlet and outlet water temperature difference and the fan frequency.

[0008] In the above method, in step S3: The Monte Carlo algorithm generates at least 1,000 sets of equipment combination strategies, each of which includes a combination of the number of chillers started and stopped, a combination of the number of cooling towers started and stopped, and a combination of water pump operating frequencies.

[0009] In the above method, in step S4: The combination strategy includes the chiller operating frequency setting value, the cooling tower fan frequency setting value, the cooling water target temperature dynamic adjustment value, and the equipment start-stop switching time window is not less than 10 minutes.

[0010] In the above method, in step S5: The deviation is calculated using mean absolute percentage error, and the corrected model parameters are automatically updated to the model every week.

[0011] In the above method, in step S1: The load forecast data is obtained by real-time calculation of the water flow rate and the inlet and outlet water temperature difference at the air conditioning terminal, and the data collection cycle is 1 to 5 minutes.

[0012] In the above method, in step S4: Control instructions are transmitted to the field programmable logic controller via Modbus TCP / IP protocol or BACnet / IP protocol.

[0013] In the above method, the Monte Carlo algorithm constructs a combination matrix based on the pseudo-random number generation method, combines equipment start-stop constraints, frequency discrete constraints and operation logic judgment rules, and realizes multiple rounds of simulation and convergence control of the total energy consumption of the system.

[0014] The above method is deployed on the edge computing gateway, and the simulation optimization cycle of step S3 is 15 minutes.

[0015] In the above method, in step S3: The energy consumption simulation results are sorted by unit cooling energy consumption (kWh / RT), and invalid combinations that exceed the maximum current threshold of the chiller are eliminated.

[0016] The second aspect of the present invention discloses a cooling station multi-device combined energy efficiency optimization system, comprising: The sensor network is used to collect real-time operating data of the chillers, cooling towers, and water pumps in the cooling station system, including equipment power, operating status, outlet water temperature, return water temperature, water flow rate, and operating frequency. It also collects external ambient temperature, ambient humidity, wind speed, and air conditioning terminal load forecast data. The edge computing server builds a chiller energy efficiency model, a cooling tower energy consumption model, and a water pump characteristic curve model based on historical operating data. These models characterize how equipment energy efficiency changes with load factor, inlet and outlet water temperature difference, and ambient temperature. Based on current load forecasts and meteorological forecast data, the server uses a Monte Carlo algorithm to traverse energy consumption simulation results for various chiller start / stop combinations, cooling tower start / stop combinations, and water pump operating frequency combinations, selecting the strategy with the lowest total system energy consumption. The control command transfer module is used to convert the combination strategy into equipment start and stop commands and frequency setting values, and send them to the chiller, cooling tower and water pump through the building automatic control system or the Internet of Things interface; The feedback correction module is used to compare the deviation between the actual operating energy consumption and the simulated operating energy consumption prediction value. When the deviation exceeds 5%, the least squares regression algorithm is used to correct the model parameters.

[0017] Compared with the prior art, the present invention has the following advantages: 1. Existing cooling station systems typically optimize a single device. However, this method uses a simulation model to jointly optimize multiple devices, including chillers, cooling towers, and water pumps, taking into account the synergistic effects between devices. This method uses a Monte Carlo simulation algorithm to traverse possible device startup and shutdown and parameter combinations, selecting the optimal combination to achieve energy savings, resulting in significant optimization results.

[0018] 2. The present invention's optimization method is not static; it adjusts device operation strategies based on real-time load changes and external environmental factors (such as temperature and humidity). This enables the system to dynamically adjust the start / stop and operating modes of each device based on varying load demands and environmental conditions, ensuring optimal energy efficiency.

[0019] 3. This invention incorporates a feedback correction mechanism that corrects model errors and optimizes control strategies based on the comparison of real-time operating data with simulation results. The system's self-learning capabilities allow it to further improve energy efficiency over time.

[0020] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a flow chart of the method of Example 1 of the present invention.

[0022] Figure 2 This is a system architecture diagram of Example 2 of the present invention. DETAILED DESCRIPTION

[0023] Example 1 like Figure 1 As shown, the present invention discloses a method for optimizing the energy efficiency of a cold station multi-device combination, comprising the following steps: S1. Data Collection: Real-time data collection of chiller, cooling tower, and water pump operating data in the cooling station system, including equipment power, operating status, outlet water temperature, return water temperature, water flow rate, and operating frequency. Data on external ambient temperature, humidity, wind speed, and air conditioning terminal load forecasts are also collected. S2. Simulation Modeling: Based on historical operating data, a chiller energy efficiency model, a cooling tower energy consumption model, and a pump characteristic curve model are constructed. These models characterize how equipment energy efficiency changes with load factor, inlet and outlet water temperature difference, and ambient temperature. S3. Monte Carlo Simulation Optimization: Based on the current load forecast and weather forecast data, the Monte Carlo algorithm is used to traverse the energy consumption simulation results of the chiller start-stop combinations, cooling tower start-stop combinations, and pump operating frequency combinations, and the combination strategy with the lowest total system energy consumption is selected. S4. Control and Distribution: Convert the combined strategy into equipment start / stop commands and frequency setpoints, which are then distributed to the chillers, cooling towers, and pumps via the building automation system or IoT interface. S5. Recursive feedback correction: Compare the deviation between the actual operating energy consumption and the simulated operating energy consumption prediction value. When the deviation exceeds 5%, the least squares regression algorithm is used to correct the model parameters.

[0024] When implementing: S1. Using temperature sensors on the chiller's evaporator and condenser, pipeline electromagnetic flowmeters, and power metering modules, this system collects real-time data on chiller power and operating status, chilled water outlet and return temperatures, cooling water outlet and return temperatures, and chilled and cooling water flow rates. It also acquires pump operating frequency and power data. Every 1 to 5 minutes, the building automation system collects the air conditioning terminal water flow rate F and the inlet and outlet water temperature difference ΔT. Real-time air conditioning load predictions are calculated using the load formula Q = ρ × c_p × ΔT × F (where ρ is water density, set to 1000 kg / m³, and c_p is water specific heat capacity, set to 4186 J / (kg·°C)). Ambient temperature, humidity, and wind speed data are also collected simultaneously to achieve dynamic sensing of all system parameters, addressing optimization biases caused by incomplete data collection in traditional cooling plants.

[0025] S2. Based on a six-month historical operational data warehouse, a chiller energy efficiency model (COP=g(RL, T_cwr, T_chws, RH_env)) was established, where RL is the load factor (current cooling capacity / rated cooling capacity), T_cwr is the cooling water return temperature, T_chws is the chilled water outlet temperature, and RH_env is the ambient humidity. This model uses cubic polynomial nonlinear regression to fit the energy efficiency variation pattern. The cooling tower energy consumption model uses piecewise linear interpolation to construct the function P_ct=h(Δt_ct, f_fan) (Δt_ct is the cooling tower inlet and outlet water temperature difference, and f_fan is the fan frequency), which reflects the combined power consumption characteristics of the fan and water pump. A family of QHP characteristic curves for the water pump was also established to accurately characterize the energy consumption characteristics under variable flow conditions and address the issue of inaccurate modeling of device coupling relationships.

[0026] S3. The edge computing gateway triggers an optimization cycle every 15 minutes. Based on the current load forecast and meteorological data, it randomly generates more than 1,000 equipment combination strategies using a Monte Carlo algorithm. Each strategy includes a combination of the number of chiller starts and stops (e.g., 2 in use, 1 in standby), the number of cooling towers in operation (e.g., 3 towers fully operational), chilled water pump frequency combinations (e.g., 45 Hz + 38 Hz), and cooling water pump frequency combinations. The total system energy consumption of each combination is calculated based on the S2 model, and the strategies are sorted by kWh / RT (energy consumption per ton of cooling capacity). Combinations where the chiller current exceeds 120% of the nameplate threshold are automatically eliminated. The strategy with the lowest energy consumption is selected for output, and a large-scale combination traversal is performed to avoid local optimal solutions.

[0027] The Monte Carlo algorithm constructs a combination matrix based on a pseudo-random number generation method, and combines equipment start-stop constraints, frequency discrete constraints, and operation logic judgment rules to achieve multiple rounds of simulation and convergence control of the system's total energy consumption.

[0028] S4. Convert optimization strategies into device instructions: Chiller start / stop instructions include specific device numbers and start / stop sequences. Pump frequency instructions are accurate to ±0.5Hz of the set value. Cooling tower fans output the target frequency, and the cooling water target temperature is dynamically adjusted by ±0.3°C. Device start / stop switching is enforced with an interval of at least 10 minutes. These instructions are sent to the on-site PLC controller via the Modbus TCP / IP protocol, resolving the response lag issue of traditional control systems while effectively preventing losses from frequent starts and stops.

[0029] S5. Collect the actual total energy consumption E_actual and the predicted value E_pred every cycle, and calculate the deviation using the mean absolute percentage error formula MAPE = (|E_actual - E_pred| / E_actual) × 100%. When MAPE exceeds 5% for three consecutive cycles, use the least squares method to update the model parameter matrix [α, β, γ] to minimize the prediction error. Simultaneously, a full update of model parameters is automatically performed every Sunday morning to ensure that the model continuously tracks device performance degradation and achieves system energy efficiency maintenance and improvement.

[0030] S6. When deployed on an edge computing gateway, the system achieves a 15ms command response through localized computing, avoiding cloud-based control delays. Dynamic adjustment of the cooling water target temperature reduces system energy consumption by an additional 8%-12%, while a 10-minute start-stop interval protection mechanism extends equipment life by over 30%. Least-squares feedback correction stabilizes long-term prediction errors within 3%.

[0031] Example 2 like Figure 2 As shown, a multi-equipment combined energy efficiency optimization system for a cold station includes: The sensor network is used to collect real-time operating data of the chillers, cooling towers, and water pumps in the cooling station system, including equipment power, operating status, outlet water temperature, return water temperature, water flow rate, and operating frequency. It also collects external ambient temperature, ambient humidity, wind speed, and air conditioning terminal load forecast data. The edge computing server builds a chiller energy efficiency model, a cooling tower energy consumption model, and a water pump characteristic curve model based on historical operating data. These models characterize how equipment energy efficiency changes with load factor, inlet and outlet water temperature difference, and ambient temperature. Based on current load forecasts and meteorological forecast data, the server uses a Monte Carlo algorithm to traverse energy consumption simulation results for various chiller start / stop combinations, cooling tower start / stop combinations, and water pump operating frequency combinations, selecting the strategy with the lowest total system energy consumption. The control command transfer module is used to convert the combination strategy into equipment start and stop commands and frequency setting values, and send them to the chiller, cooling tower and water pump through the building automatic control system or the Internet of Things interface; The feedback correction module is used to compare the deviation between the actual operating energy consumption and the simulated operating energy consumption prediction value. When the deviation exceeds 5%, the least squares regression algorithm is used to correct the model parameters.

[0032] When implementing: Sensor network implementation: Platinum resistance temperature sensors deployed in the evaporator and condenser pipes of the chiller collect real-time chilled water and cooling water outlet and return water temperature data. Electromagnetic flowmeters are installed in the chilled water and cooling water main pipes to collect water flow data. Electric energy metering modules are installed in the power distribution cabinets of the chiller, cooling tower fan, and water pump to collect real-time equipment power. At the same time, the operating status signals of the chiller, cooling tower, and water pump are collected through the dry contact signals of the equipment control cabinet. Ambient temperature and humidity sensors and anemometers are installed around the cooling tower to collect external ambient temperature, humidity, and wind speed data. The load prediction value of the air conditioner terminal is calculated by real-time monitoring of the water flow rate and the inlet and outlet water temperature difference flowing through each air conditioner terminal using the formula Q=ρ×cp×ΔT×F, where ρ is the density of water (taken as 1000 kg / m³), cp is the constant pressure specific heat capacity of water (taken as 4186 J / (kg·℃)), ΔT is the inlet and outlet water temperature difference, and F is the water flow rate. All sensor data is transmitted via Modbus The data is transmitted to the edge computing server through the RTU protocol, and the collection period is fixed at 1-5 minutes, realizing dynamic and synchronous perception of the operating parameters and environmental parameters of the entire system.

[0033] Edge computing server implementation: Based on the stored six-month historical operation database, the chiller energy efficiency model is constructed as a nonlinear regression function COP=g(RL,Tcwr,Tchws,RHenv), where RL is the load rate (current cooling capacity / rated cooling capacity), Tcwr is the cooling water return temperature, Tchws is the chilled water outlet temperature, and RHenv is the ambient humidity. The cooling tower energy consumption model uses a piecewise linear interpolation function Pct_tot=h(Δt ct ,f fan )(Δt ct is the temperature difference between the inlet and outlet water of the cooling tower, f fanis the fan frequency). A family model of pump flow-head-power characteristic curves is also established. An optimization cycle is triggered every 15 minutes. Based on the current load forecast and meteorological data, a Monte Carlo algorithm is used to randomly generate at least 1,000 equipment combination strategies. Each strategy includes a combination of the number of chillers on and off (e.g., two in use and one in standby), the number of cooling towers in operation (e.g., all three towers running), chilled water pump frequencies (e.g., 45Hz + 38Hz), and cooling water pump frequencies. The aforementioned model is used to calculate the energy consumption of the chillers, cooling towers, and pumps for each combination. The total system energy consumption is then superimposed to obtain a predicted value. The predicted values ​​are then sorted by energy consumption per unit of cooling capacity (kWh / RT). Any invalid combinations in the simulation results where the operating current of the chillers exceeds 120% of the nameplate threshold are automatically eliminated. The strategy with the lowest total energy consumption is ultimately selected and output. This large-scale random combination traversal avoids the local optimality drawback of traditional strategies.

[0034] Implementation of the control instruction transfer module: Optimization strategies are converted into executable instructions for the equipment, including specified chiller start and stop commands (such as starting unit numbered CH-1), cooling tower start and stop commands (such as shutting down fan numbered CT-3), and frequency setpoints for the chilled water pump and cooling water pump (accurate to ±0.5Hz). The target cooling water outlet temperature is dynamically adjusted based on the load (adjustment range ±0.3°C). A time interval window of more than 10 minutes is set for equipment start and stop state switching to prevent equipment damage caused by frequent starts and stops. All instructions are packaged and transmitted to the field programmable logic controller (PLC) via an industrial communication protocol (Modbus TCP / IP or BACnet / IP). The PLC then drives the chiller, cooling tower fan inverter, and water pump inverter to execute the operation. The instruction transmission delay is controlled within 50 milliseconds to ensure the real-time performance of the optimization strategy.

[0035] Implementation of the feedback correction module: After each control cycle, the actual operating energy consumption data of the chiller, cooling tower and water pump are collected and compared with the simulated predicted energy consumption value. The deviation is calculated using the mean absolute percentage error formula MAPE=(|E_actual - E_pred| / E_actual)×100%. When the MAPE value exceeds 5% for three consecutive cycles, the least squares regression algorithm correction program is triggered, and the latest actual operating data is included in the training set. The chiller energy efficiency model parameters (such as polynomial coefficients α, β) and the piecewise linear interpolation nodes of the cooling tower energy consumption model are refitted and updated. At the same time, regardless of whether the correction is triggered, the model parameters are automatically updated in full every Sunday morning to ensure that the model continuously tracks equipment performance degradation and system characteristic drift, and maintains the prediction error below 3% in the long term.

[0036] In summary: the high-frequency data collection and real-time load calculation of the sensor network solve the optimization deviation problem caused by incomplete data in traditional cooling stations. The large-scale Monte Carlo simulation optimization of the edge computing server can improve system energy efficiency by 15%-25%. The 10-minute start-stop interval protection mechanism of the control command reduces equipment mechanical loss by more than 30%. The dynamic update of model parameters in the feedback correction module improves long-term prediction stability by 40%, ultimately achieving a significant improvement in the comprehensive energy efficiency ratio (COP) of the cooling station.

[0037] The following is a comparison of data before and after a large industrial park was transformed using the technology of this invention:

[0038] Data source: System operation database (InfluxDB) + meter measurement records (verified by a third-party energy efficiency platform) Energy saving principle description: The Monte Carlo model selects the "critical high efficiency point" in the nonlinear range to avoid low COP caused by low load operation Automatically judge the daytime / nighttime climate, dynamically adjust the cooling water temperature and the chilled water outlet temperature, and achieve "local optimal switching" Under low and medium loads, the system tends to run more in small combinations to improve the partial load efficiency of the system. The control response time is reduced from 15 minutes under the original PLC control to less than 5 minutes, improving the regulation efficiency.

[0039] This invention has been deployed in a large industrial park's cooling plant system. In conjunction with the building's BAS system, edge controllers, and energy consumption monitoring system, it operates stably and achieves significant energy savings. The system is highly scalable and suitable for use in high-load cooling plant environments, such as universities, data centers, and commercial complexes, proving promising for broad industrial applications.

[0040] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any way. Any simple modification, change and equivalent structural change made to the above embodiment based on the technical essence of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A method for optimizing the energy efficiency of a cold station multi-device combination, characterized in that: The following steps are involved: S1. Data Collection: Real-time data collection of chiller, cooling tower, and water pump operating data in the cooling station system, including equipment power, operating status, outlet water temperature, return water temperature, water flow rate, and operating frequency. Data on external ambient temperature, humidity, wind speed, and air conditioning terminal load forecasts are also collected. S2. Simulation Modeling: Based on historical operating data, a chiller energy efficiency model, a cooling tower energy consumption model, and a pump characteristic curve model are constructed. These models characterize how equipment energy efficiency changes with load factor, inlet and outlet water temperature difference, and ambient temperature. S3. Monte Carlo Simulation Optimization: Based on the current load forecast and weather forecast data, the Monte Carlo algorithm is used to traverse the energy consumption simulation results of the chiller start-stop combinations, cooling tower start-stop combinations, and pump operating frequency combinations, and the combination strategy with the lowest total system energy consumption is selected. S4. Control and Distribution: Convert the combined strategy into equipment start / stop commands and frequency setpoints, which are then distributed to the chillers, cooling towers, and pumps via the building automation system or IoT interface. S5. Recursive feedback correction: Compare the deviation between the actual operating energy consumption and the simulated operating energy consumption prediction value. When the deviation exceeds 5%, the least squares regression algorithm is used to correct the model parameters.

2. The method according to claim 1, characterized in that In step S2: The energy efficiency model of the chiller is a nonlinear regression function of COP with respect to load rate, chilled water outlet temperature, cooling water inlet temperature and ambient humidity; The cooling tower energy consumption model is a piecewise linear interpolation function of the total power of the fan and water pump with respect to the inlet and outlet water temperature difference and the fan frequency.

3. The method according to claim 1, characterized in that In step S3: The Monte Carlo algorithm generates at least 1,000 sets of equipment combination strategies, each of which includes a combination of the number of chillers started and stopped, a combination of the number of cooling towers started and stopped, and a combination of water pump operating frequencies.

4. The method according to claim 1, wherein In step S4: The combination strategy includes the chiller operating frequency setting value, the cooling tower fan frequency setting value, the cooling water target temperature dynamic adjustment value, and the equipment start-stop switching time window is not less than 10 minutes.

5. The method according to claim 1, wherein In step S5: The deviation is calculated using mean absolute percentage error, and the corrected model parameters are automatically updated to the model every week.

6. The method according to claim 1, characterized in that In step S1: The load forecast data is obtained by real-time calculation of the water flow rate and the inlet and outlet water temperature difference at the air conditioning terminal, and the data collection cycle is 1 to 5 minutes.

7. The method according to claim 1, characterized in that In step S4: Control instructions are transmitted to the field programmable logic controller via Modbus TCP / IP protocol or BACnet / IP protocol.

8. The method according to claim 1, characterized in that The Monte Carlo algorithm constructs a combination matrix based on a pseudo-random number generation method, and combines equipment start-stop constraints, frequency discrete constraints, and operation logic judgment rules to achieve multiple rounds of simulation and convergence control of the system's total energy consumption.

9. The method according to claim 1, characterized in that Deployed on the edge computing gateway, the simulation optimization cycle of step S3 is 15 minutes.

10. The method according to claim 1, characterized in that In step S3: The energy consumption simulation results are sorted by unit cooling energy consumption (kWh / RT), and invalid combinations that exceed the maximum current threshold of the chiller are eliminated.

11. A cooling station multi-equipment combined energy efficiency optimization system, characterized in that: include: The sensor network is used to collect real-time operating data of the chillers, cooling towers, and water pumps in the cooling station system, including equipment power, operating status, outlet water temperature, return water temperature, water flow rate, and operating frequency. It also collects external ambient temperature, ambient humidity, wind speed, and air conditioning terminal load forecast data. The edge computing server builds a chiller energy efficiency model, a cooling tower energy consumption model, and a water pump characteristic curve model based on historical operating data. These models characterize how equipment energy efficiency changes with load factor, inlet and outlet water temperature difference, and ambient temperature. Based on current load forecasts and meteorological forecast data, the server uses a Monte Carlo algorithm to traverse energy consumption simulation results for various chiller start / stop combinations, cooling tower start / stop combinations, and water pump operating frequency combinations, selecting the strategy with the lowest total system energy consumption. The control command transfer module is used to convert the combination strategy into equipment start and stop commands and frequency setting values, and send them to the chiller, cooling tower and water pump through the building automatic control system or the Internet of Things interface; The feedback correction module is used to compare the deviation between the actual operating energy consumption and the simulated operating energy consumption prediction value. When the deviation exceeds 5%, the least squares regression algorithm is used to correct the model parameters.

Images