An energy-saving optimization control method for an industrial refrigeration system

Abstract

The application discloses an energy-saving optimization control method for an industrial refrigeration system, adopts a cloud global optimization-edge collaborative execution control framework, dynamically allocates operation parameters of a refrigeration host unit, a carrier refrigerant circulating unit, a cooling water circulating unit and a cooling tower unit with the lowest total energy consumption of the system as a target, realizes deep collaboration of the units, and through actual verification, compared with a traditional unit independent control logic, the actual measurement system overall power saving rate can reach 23.5%. Since a multi-stage control strategy is adopted in the anti-icing control module, the system can completely replace carrier refrigerants such as ethylene glycol with softened water with low viscosity and high specific heat capacity, directly utilizes the low viscosity characteristics of water to reduce the pumping power consumption by 12%, and due to the excellent heat exchange characteristics of water and the combination of the high-efficiency heat exchange of the falling-film evaporator, the COP of the host is improved by about 20%, and is particularly suitable for near-ice-point working conditions with the outlet temperature of the carrier refrigerant being 0-2 DEG C.

Description

Technical Field

[0001] This invention relates to the field of industrial energy conservation and automatic control technology, and in particular to an energy-saving optimization control method for industrial refrigeration systems. Background Technology

[0002] In industrial production, refrigeration systems are crucial facilities providing low-temperature cold sources for core processes. These systems include a refrigeration unit, a refrigerant circulation unit, a cooling water circulation unit, a cooling tower unit, and a control unit. The refrigeration unit, as the core of the system's cooling capacity production, includes one or more refrigeration units equipped with an anti-icing control module. The refrigerant circulation unit forms a closed loop with the evaporator circuit of the refrigeration unit and includes a variable frequency chilled pump and a refrigerant temperature / pressure sensor to deliver the cooling capacity generated by the refrigeration unit to the process end. The cooling water circulation unit forms a closed loop with the condenser circuit of the refrigeration unit and includes a variable frequency cooling pump and a throttling component to adjust pipeline characteristics to match new flow requirements. The cooling tower unit includes at least one variable frequency cooling tower fan, and the cooling tower is equipped with a water distributor suitable for low-flow conditions, as well as a wet-bulb temperature sensor and a water temperature sensor for calculating approximation. The unit control unit uses an industrial PC to control each unit.

[0003] Existing refrigeration systems generally adopt static control mode and independent control logic for each unit. Due to the following problems, their energy consumption accounts for 40 to 60% of the factory's total energy consumption.

[0004] 1. Static control mode: Due to the use of static control modes such as fixed cooling water return temperature, fixed approximation degree or fixed load ratio, especially for the refrigeration host unit, which only uses single-unit independent adjustment such as starting and stopping according to fixed load threshold, it cannot respond to the dynamic changes of host load and ambient wet-bulb temperature, resulting in the system deviating from the optimal operating condition for a long time and significant energy efficiency loss.

[0005] 2. Independent Control Logic for Each Unit: Because each unit uses independent control logic, such as cooling water pumps adjusting according to a fixed temperature difference and cooling towers starting and stopping according to a fixed number of units, conflicts arise between local energy saving and overall energy efficiency. For example, while reducing the frequency of cooling tower fans can lower their own energy consumption, insufficient airflow leads to an increase in the condensing temperature of the chiller, which in turn increases the compression ratio of the chiller, significantly increasing the energy consumption per unit of cooling capacity. This can even cause the chiller to shut down due to high-pressure protection. Blindly starting and stopping the chiller further exacerbates the waste of total system energy.

[0006] Furthermore, although existing refrigeration units are equipped with anti-refrigeration refrigerant icing control modules, their control methods cannot guarantee safe and efficient operation of pure water as a refrigerant under near-freezing point conditions (referring to chilled water temperatures between 0 and 2°C). Therefore, the industry commonly uses ethylene glycol, alcohol-water, etc., as refrigerants, which not only leads to high power consumption during transport but also poses a risk of equipment corrosion. For example, because ethylene glycol solutions have high viscosity, if the refrigeration pump flow rate is reduced simultaneously when the refrigeration unit is running at low load, the Reynolds number of the fluid in the pipes can easily drop below the laminar flow critical value (Re < 2300), causing a sharp decline in heat exchange efficiency in the heat exchange tubes. This results in an abnormal drop in the evaporation temperature of the main unit, weakening the unit's operating efficiency and increasing the risk of icing. Therefore, the industry's conventional solution requires maintaining a high flow rate to avoid laminar flow problems, thus resulting in high power consumption during transport. Summary of the Invention

[0007] The present invention aims to solve the aforementioned technical problems existing in the prior art by providing an energy-saving optimization control method for industrial refrigeration systems.

[0008] The technical solution of this invention is: an energy-saving optimization control method for an industrial refrigeration system, using a cloud computing device and an edge control device as control units. The cloud computing device generates and distributes a globally optimal set of operating parameters based on operational status data and embeds an AI algorithm model. The operational status data includes the refrigeration unit load rate, ambient wet-bulb temperature, refrigerant inlet and outlet temperatures and flow rates, and cooling water inlet and outlet temperatures and flow rates. The globally optimal set of operating parameters includes the chilled water pump frequency, cooling water pump frequency, target cooling water temperature difference, cooling tower fan frequency, refrigerant outlet water temperature setpoint, and the number of operating refrigeration units. The AI ​​algorithm model is constructed by calculating the globally optimal set of operating parameters and iteratively training it to minimize total energy consumption. The edge control device executes the globally optimal set of operating parameters distributed by the cloud computing device to perform unified and coordinated control of the operating status of the refrigeration unit, refrigerant circulation unit, cooling water circulation unit, and cooling tower unit.

[0009] The evaporator of the refrigeration unit is a falling film evaporator. The anti-icing control module of the refrigeration unit performs the following multi-level early warning and linkage control based on the refrigerant temperature, evaporation pressure, and refrigerant flow rate at the evaporator outlet: Level 1 early warning threshold: refrigerant temperature ≤ 0℃, or evaporation temperature ≤ -0.5℃, triggering the unit to slowly reduce load; Level 2 early warning threshold: refrigerant temperature ≤ -0.1℃, or evaporation temperature ≤ -1.0℃, or refrigerant flow rate < 40% of rated flow rate, triggering the unit to rapidly reduce load and trigger an alarm; Emergency protection threshold: refrigerant temperature ≤ -0.3℃, or evaporation temperature ≤ -1.5℃, triggering the unit to shut down. Preferably, when the edge control device performs unified and coordinated control of the cooling tower unit's operating status, it prioritizes increasing the number of operating cooling towers and ensuring that the fan frequency of each operating cooling tower is not lower than 30% of the rated frequency.

[0010] When the edge control device performs unified and coordinated control of the operating status of the refrigeration unit, if the refrigeration unit is a high-temperature stage unit and a low-temperature stage unit connected in series, the load of the high-temperature stage unit and the low-temperature stage unit is independently optimized and allocated.

[0011] When the edge control device preferably performs unified and coordinated control of the operating status of the cooling water circulation unit, it adjusts the valve opening of the throttling component based on the target temperature difference of the cooling water.

[0012] Preferably, the cloud computing device operates according to a set cycle and triggers instant optimization when the change in the load rate of the cooling host or the ambient wet-bulb temperature exceeds a preset threshold.

[0013] Compared with the prior art, the present invention has the following advantages:

[0014] 1. By adopting a cloud-based global optimization and edge-based collaborative execution control architecture, the operating parameters of the refrigeration unit, refrigerant circulation unit, cooling water circulation unit, and cooling tower are dynamically allocated with the goal of minimizing total system energy consumption. This achieves deep collaboration among the units, fundamentally solving the energy-saving strategy failure phenomenon of "local energy saving, global energy consumption" caused by independent control of each unit. Actual verification shows that compared to traditional independent unit control logic, the measured overall system energy saving rate can reach 23.5%.

[0015] 2. Because the anti-icing control module employs a multi-level control strategy, it overcomes the technical bottleneck of requiring antifreeze in near-freezing conditions. This allows the system to completely replace refrigerants such as ethylene glycol with low-viscosity, high-specific-heat-capacity softened water. Directly utilizing water's low viscosity reduces pumping power consumption by 12%. Furthermore, the excellent heat exchange characteristics of water, combined with the efficient heat exchange of the falling film evaporator, increase the main unit's COP by approximately 20%. It completely eliminates the corrosion and denaturation risks associated with ethylene glycol solutions, achieving a balance between safety, energy efficiency, and reliability. This makes it particularly suitable for near-freezing conditions where the refrigerant outlet temperature is between 0 and 2°C. Detailed Implementation

[0016] The energy-saving optimization control method for an industrial refrigeration system of the present invention was implemented in a large-scale industrial refrigeration system (hereinafter referred to as a refrigeration station). The following description only uses a beer production enterprise as an example, but its application is not limited to this. For example, the system and method of the present invention can be used to achieve energy-saving optimization in chip cooling refrigeration stations of electronic enterprises, vaccine cold chain refrigeration stations of pharmaceutical enterprises, and process cooling refrigeration stations of chemical enterprises. The specific hardware selection and parameter settings can be adjusted according to the characteristics of different industries.

[0017] Industrial refrigeration system construction:

[0018] (1) Refrigeration Unit: A high-temperature stage refrigeration unit and a low-temperature stage refrigeration unit are selected and connected in series (for example, the high-temperature stage unit has a single cooling capacity of 2740kW, and the low-temperature stage unit has a single cooling capacity of 1880kW, totaling 4620kW), to meet the process requirement of reducing the return water temperature from 23.8℃ to the effluent temperature of 0.5℃. The evaporator outlet of the high-temperature stage unit is connected to the evaporator inlet of the low-temperature stage unit, and the condenser of the high-temperature stage unit is connected in parallel with the condenser of the low-temperature stage unit on the cooling water side. Both units use falling film evaporators and have built-in anti-icing control modules.

[0019] (2) Refrigerant circulation unit: The original ethylene glycol solution is replaced with softened water (softened to ensure that the calcium and magnesium hardness ion content is ≤0.03mmol / L), and two variable frequency chilled pumps (flow rate 182m³ / h, head 28m) are configured to form a closed loop with the evaporator of the refrigeration unit; the variable frequency chilled pumps are configured to safely reduce the refrigerant flow rate to about 50% of the rated flow rate when the system load rate is about 50%.

[0020] (3) Cooling water circulation unit: All cooling water pumps are variable frequency pumps, and a manual regulating valve is installed on the outlet pipeline as a throttling component. The throttling component adjusts the valve opening synchronously when the variable frequency cooling pump adjusts the frequency according to the global optimal operating parameter set, so as to accurately match the pipeline characteristics and ensure that the pump operating point is in the high efficiency zone.

[0021] (4) Cooling Tower Unit: Multiple cooling towers (e.g., 2) are configured, and all fans are controlled by frequency converters. The selection of the cooling towers must ensure that the rated cooling capacity of a single tower is greater than or equal to the total cooling demand of the system ÷ the number of towers × 1.05, which is a safety factor. At the same time, the water distributors inside the towers are replaced with models designed specifically for low-flow conditions (e.g., Φ6mm orifice diameter, 50mm orifice spacing) to ensure uniform water distribution in multi-tower low-speed mode.

[0022] (5) System Coordination and Control Unit: Deploy a cloud computing device (cloud server) with an AI algorithm model trained based on historical data, aiming to minimize the total energy consumption of the system. Deploy an edge control device (such as an industrial PC) in the local computer room, which is connected to all unit controllers, frequency converters and sensors through an RS485 communication network, with a control response time of ≤1 second.

[0023] The AI ​​algorithm model running on the cloud server is a genetic algorithm-neural network fusion model used to establish and optimize the nonlinear relationship between system operating parameters and total energy consumption. The key network and training parameters of this model are as follows: the input layer has 8 neurons, corresponding to 8 types of operating state data (including chiller load rate, ambient wet-bulb temperature, refrigerant inlet and outlet temperatures and flow rates, and cooling water inlet and outlet temperatures and flow rates); the hidden layer has 2 layers, each with 32 neurons, using the ReLU activation function; the output layer has 6 neurons, corresponding to chiller pump frequency, cooling water pump frequency, target cooling water temperature difference, cooling tower fan frequency, refrigerant outlet water temperature setpoint, and the number of chillers operating. The optimization parameters for the genetic algorithm are set as follows: crossover probability 0.8, mutation probability 0.05, population size 100, number of iterations 100, and termination condition is convergence of the fitness function (minimum total system energy consumption). The AI ​​algorithm model possesses periodic self-learning and updating capabilities. It iterates parameters monthly based on recent operational data (such as energy consumption, temperature, and flow rate) to adaptively compensate for long-term performance drift such as equipment scaling and efficiency degradation, thereby ensuring the long-term stability of the optimization effect. Actual testing shows that after self-learning and updating, the model's prediction error is stably controlled within ≤5%, and the calculation time for a single optimization is ≤30 seconds.

[0024] The AI ​​algorithm model performs self-learning and parameter updates monthly based on recent operating data to compensate for performance changes caused by equipment scaling, efficiency degradation, etc., and to ensure the long-term effectiveness of the optimization.

[0025] The present invention provides an energy-saving optimization control method for an industrial refrigeration system, which is as follows:

[0026] 1. Global Dynamic Optimization Steps: The cloud computing device collects real-time operational status data, including chiller load rate, ambient wet-bulb temperature, and inlet and outlet temperatures and flow rates of refrigerant and cooling water. The AI ​​algorithm model runs every 15 minutes and triggers immediate optimization when the chiller load rate changes by more than ±5% or the ambient wet-bulb temperature fluctuates by more than ±1℃. Under this series configuration, the AI ​​algorithm model dynamically outputs a set of parameters including chilled water pump frequency (e.g., 30-50Hz), cooling water pump frequency (e.g., 30-50Hz), cooling tower fan frequency (e.g., 30-50Hz), refrigerant outlet water temperature setpoint (0.5℃ for low-temperature stage outlet water), and the number of chillers in operation.

[0027] 2. Equipment Collaborative Execution Steps: The edge control device strictly executes the parameter set issued by the cloud computing device to ensure that the cooling water pump adjusts the flow rate of the throttling component according to the target temperature difference of the cooling water calculated by the cloud (for example, in this brewery embodiment, the temperature difference is approximately 6°C at full load, verified by system simulation and actual measurement). It controls the cooling tower unit to operate in high-efficiency mode. When the cooling tower unit contains multiple cooling towers, it controls them to operate in a "multi-tower low-speed" state, that is, prioritizing increasing the number of operating cooling towers and ensuring that the fan frequency of each operating cooling tower is not lower than 30% of the rated frequency, to ensure the heat exchange efficiency of a single tower and collaboratively achieve optimal total system energy consumption. Simultaneously, the edge control device coordinates the start-up and shutdown of the refrigeration units within the station (e.g., sorting according to cumulative operating time to achieve balanced wear) and load distribution. In particular, it needs to coordinate the increase / decrease of loads for the high- and low-temperature stage units to ensure load matching and water temperature stability between the two stages. For example, in this embodiment, the outlet water temperature of the high-temperature stage refrigerant is 9°C, and the outlet water temperature of the low-temperature stage refrigerant is 0.5°C to avoid control conflicts.

[0028] The evaporator of the refrigeration unit described in this embodiment of the invention is a falling film evaporator. The anti-icing control module of the refrigeration unit performs the following multi-level early warning and linkage control based on the refrigerant temperature, evaporation pressure, and refrigerant flow rate at the evaporator outlet: The first-level early warning threshold is when the refrigerant temperature is ≤0℃ or the evaporation temperature is ≤-0.5℃, triggering the unit to slowly reduce load; the second-level early warning threshold is when the refrigerant temperature is ≤-0.1℃, or the evaporation temperature is ≤-1.0℃, or the refrigerant flow rate is <40% of the rated flow rate, triggering the unit to rapidly reduce load and trigger an alarm; the emergency protection threshold is when the refrigerant temperature is ≤-0.3℃ or the evaporation temperature is ≤-1.5℃, triggering the unit to shut down. The difference between slow and rapid load reduction lies in the different reduction rates.

[0029] The thresholds at each level can be set according to actual hardware performance and operating conditions. The design concept is not a simple passive alarm, but an active safety redundancy design based on the excellent performance of the falling film evaporator. For example, the alarm threshold for the refrigerant temperature at the outlet of the low-temperature stage main unit is set slightly below 0°C (e.g., -0.1°C). This is not the normal operating point of the system, but a dynamic buffer reserved to cope with extreme operating conditions such as sudden changes in process load. It makes full use of the hardware performance boundaries, ensuring that the system can operate continuously and efficiently under most disturbances, and only triggers the final protection when approaching the physical limit, thus achieving the optimal balance between safety and energy efficiency.

[0030] In this embodiment, the anti-icing control module is specifically designed to ensure near-freezing point operation with a refrigerant outlet temperature ≤0.5℃. Actual testing shows that during long-term operation, the anti-icing control module accurately triggers emergency shutdown protection when the refrigerant temperature drops to -0.3℃, thus fundamentally eliminating the possibility of equipment damage due to icing when the softened water operates below 0.5℃. Simultaneously, due to the combination of the excellent heat exchange characteristics of the softened water and the high-efficiency heat exchange of the falling film evaporator, under the operating conditions described in this embodiment, the COP of the main system is improved by approximately 20% compared to the traditional ethylene glycol refrigerant solution, fully verifying the dual advantages of this invention in terms of safety and energy efficiency.

[0031] Implementation results:

[0032] After a full year of operation, the refrigeration plant achieved the significant energy-saving effects shown in Table 1 after implementing this invention:

[0033] Table 1

[0034]

[0035] The aforementioned effects are achieved for two main reasons. First, the adoption of a cloud-based global optimization and edge-coordinated execution control architecture results in an overall energy saving rate of up to 23.5% compared to traditional unit-based independent control logic. Second, the use of softened water instead of traditional ethylene glycol solution as the refrigerant demonstrates multi-dimensional and profound energy-saving potential, specifically reflected in the following three synergistic advantages:

[0036] 1. Direct reduction in pump power consumption: Under near-freezing conditions (0°C), the viscosity of softened water is only about 1 mPa·s, which is significantly lower than that of ethylene glycol solution at the same temperature (viscosity about 3 mPa·s). Combined with the positive correlation between fluid flow resistance and viscosity, the variable frequency refrigeration pump directly reduces power consumption by about 12% under the same flow rate and head conditions.

[0037] 2. Improved compatibility with low flow rate operation: Due to its low viscosity, softened water can maintain turbulent flow (Re>4000) even when the flow rate of the chilled water pump is significantly reduced (e.g., to 40-50% of the rated flow rate). This ensures stable heat exchange performance of the evaporator and provides key support for flow rate optimization under low load conditions. It further amplifies the energy-saving effect that the pump power consumption is proportional to the cube of the flow rate (e.g., when the flow rate drops to 50%, the pump power consumption is only 12.5% ​​of the rated value).

[0038] 3. Enhanced Main Unit Energy Efficiency: Compared to the commonly used stainless steel plate evaporators in the industry, the falling film evaporator configured in this system improves heat exchange efficiency by approximately 40%. Combined with the enhanced heat exchange effect brought by the low viscosity of softened water, the main unit evaporation temperature is approximately 5°C higher than the ethylene glycol solution. Based on the energy efficiency characteristics of the refrigeration unit (for every 1°C increase in evaporation temperature, COP increases by approximately 3%-4%), the final unit COP is increased by approximately 20%.

Claims

1. An energy saving optimization control method for an industrial refrigeration system, characterized in that: The cloud computing device and the edge control device are used as the control unit. The cloud computing device generates a global optimal operation parameter set based on the running state data and through embedding an AI algorithm model, and the running state data is the refrigeration host load rate, the environmental wet bulb temperature, the inlet and outlet temperature and flow of the secondary refrigerant, the inlet and outlet temperature and flow of the cooling water, and the global optimal operation parameter set is the refrigeration pump frequency, the cooling water pump frequency, the cooling water target temperature difference, the cooling tower fan frequency, the secondary refrigerant outlet water temperature set value and the refrigeration host operation number. The AI algorithm model is constructed by calculating the global optimal operation parameter set and iterative training with the lowest total energy consumption as the target. The edge control device executes the global optimal operation parameter set issued by the cloud computing device, and uniformly and cooperatively controls the running state of the refrigeration host unit, the secondary refrigerant circulation unit, the cooling water circulation unit and the cooling tower unit. The evaporator of the refrigeration host unit is a falling film evaporator, and the anti-icing control module of the refrigeration host unit is based on the outlet temperature of the secondary refrigerant, the evaporation pressure and the flow of the secondary refrigerant, and performs the following multi-level early warning and linkage control: the first level early warning threshold is that the secondary refrigerant temperature is less than or equal to 0℃, or the evaporation temperature is less than or equal to -0.5℃, triggering the host slow load reduction; the second level early warning threshold is that the secondary refrigerant temperature is less than or equal to -0.1℃, or the evaporation temperature is less than or equal to -1.0℃, or the flow of the secondary refrigerant is less than 40% of the rated flow, triggering the host rapid load reduction and alarm; the emergency protection threshold is that the secondary refrigerant temperature is less than or equal to -0.3℃, or the evaporation temperature is less than or equal to -1.5℃, triggering the host shutdown.

2. The energy optimization control method of an industrial refrigeration system according to claim 1, characterized in that: When the edge control device uniformly and cooperatively controls the running state of the cooling tower unit, the running number of the cooling tower is preferentially increased, and the fan frequency of each running cooling tower is not less than 30% of the rated frequency.

3. The energy optimization control method of an industrial refrigeration system according to claim 2, characterized in that: When the edge control device uniformly and cooperatively controls the running state of the refrigeration host unit, if the refrigeration host is a high-temperature unit and a low-temperature unit connected in series, the load of the high-temperature unit and the low-temperature unit is independently optimized and distributed.

4. The energy optimization control method of an industrial refrigeration system according to claim 3, characterized in that: When the edge control device uniformly and cooperatively controls the running state of the cooling water circulation unit, the valve opening of the throttling component is adjusted based on the cooling water target temperature difference.

5. The energy optimization control method of an industrial refrigeration system as claimed in claim 1, 2, 3 or 4, characterized in that: The cloud computing device operates according to a set period, and triggers immediate optimization when the refrigeration host load rate or the environmental wet bulb temperature changes by more than a preset threshold.

Images