Intelligent cold station control system based on machine learning

Abstract

The invention provides an intelligent cold station control system based on machine learning, which relates to the technical field of energy and power engineering and is divided into an infrastructure part and an intelligent auxiliary part, the infrastructure part comprises a data acquisition subsystem, a data analysis and preprocessing subsystem, a machine learning model training subsystem, an intelligent control strategy generation subsystem, a real-time monitoring subsystem, a fault diagnosis and early warning subsystem, a man-machine interaction subsystem, an energy management subsystem and a communication subsystem. The intelligent auxiliary part comprises a self-adaptive optimization subsystem, a load prediction subsystem, an equipment health assessment subsystem, an intelligent cooperation subsystem, a virtual simulation subsystem, a green energy access subsystem and a safety protection subsystem, and the energy management subsystem generates output of the subsystems according to an intelligent control strategy and performs optimization management on energy consumption of the cold station. Through precise load prediction, intelligent control strategy generation and energy management optimization, the energy consumption of the cold station is reduced, and energy-saving operation is realized.

Description

Technical Field

[0001] The present invention belongs to the field of energy and power engineering technology, and more specifically, particularly relates to an intelligent refrigeration station control system based on machine learning. Background Art

[0002] With the development of society and the advancement of science and technology, efficient use and intelligent management of energy have become crucial in all fields. In the field of air conditioning and refrigeration, as the core facility of centralized cooling, the operation efficiency and control strategy of the cold station have a significant impact on the performance, energy consumption and operating cost of the entire air conditioning system.

[0003] Traditional cooling plant control systems usually rely on fixed rules and experience for control, which makes it difficult to cope with complex and changing operating environments and load demands. Such systems have limitations in data collection and analysis, and cannot fully tap the potential information in the data, resulting in inaccurate and inflexible control strategies.

[0004] At the same time, traditional cold station control systems lack optimization mechanisms in energy management, often resulting in energy waste. In terms of equipment maintenance, they mainly rely on regular inspections and repairs after failures, making it difficult to detect potential problems in advance, increasing the risk of equipment failure and maintenance costs.

[0005] In addition, with the development and application of green energy, traditional cooling station control systems are insufficient in integrating renewable energy, unable to fully utilize the advantages of green energy and unable to meet the requirements of sustainable development.

[0006] In the era of intelligence, the control of cooling stations needs to be more accurate, efficient, flexible and energy-saving. Some existing intelligent control technologies have problems such as poor model adaptability, weak coordination, and insufficient security in cooling station applications. Therefore, developing an intelligent cooling station control system based on machine learning to achieve accurate control of cooling stations, efficient and energy-saving operation, equipment health management, and green energy access optimization has become an important issue that needs to be urgently addressed in the current air-conditioning and refrigeration field. Summary of the invention

[0007] In order to solve the above technical problems, the present invention provides an intelligent refrigeration station control system based on machine learning to solve the above problems.

[0008] An intelligent cold station control system based on machine learning is divided into an infrastructure part and an intelligent auxiliary part. The infrastructure part includes a data acquisition subsystem, a data analysis and preprocessing subsystem, a machine learning model training subsystem, an intelligent control strategy generation subsystem, a real-time monitoring subsystem, a fault diagnosis and early warning subsystem, a human-computer interaction subsystem, an energy management subsystem, and a communication subsystem. The intelligent auxiliary part includes an adaptive optimization subsystem, a load forecasting subsystem, an equipment health assessment subsystem, an intelligent collaboration subsystem, a virtual simulation subsystem, a green energy access subsystem, and a safety protection subsystem. The energy management subsystem optimizes the energy consumption of the cold station according to the output of the intelligent control strategy generation subsystem.

[0009] Preferably, the cold station operation data collected by the data acquisition subsystem is transmitted to the data analysis and preprocessing subsystem, and the processed data is provided to the machine learning model training subsystem for model training and optimization. The trained machine learning model plays a role in the intelligent control strategy generation subsystem to generate a control strategy.

[0010] Preferably, the real-time monitoring subsystem feeds back the monitored real-time data to the intelligent control strategy generation subsystem so as to adjust the control strategy in real time. At the same time, the real-time monitoring subsystem also transmits the data to the fault diagnosis and early warning subsystem for fault diagnosis and early warning.

[0011] Preferably, the results of the fault diagnosis and warning subsystem are transmitted to the human-machine interaction subsystem through the communication subsystem to inform the operator. The human-machine interaction subsystem can also receive instructions from the operator to set parameters and intervene in other subsystems.

[0012] Preferably, the adaptive optimization subsystem automatically adjusts the parameters of the machine learning model training subsystem and the strategies of the intelligent control strategy generation subsystem according to the actual operation of the cold station and changes in the external environment.

[0013] Preferably, the equipment health status information evaluated by the equipment health evaluation subsystem is transmitted to the intelligent control strategy generation subsystem to affect the control strategy, and also prompts the operator with equipment maintenance requirements through the human-computer interaction subsystem.

[0014] Preferably, the intelligent collaboration subsystem exchanges information with external related systems and feeds back collaboration requirements to the intelligent control strategy generation subsystem.

[0015] Preferably, the virtual simulation subsystem performs simulation verification before the strategy is implemented, and the results are fed back to the intelligent control strategy generation subsystem for optimization.

[0016] Preferably, the green energy access subsystem transmits the information of renewable energy to the energy management subsystem to achieve comprehensive and optimal utilization of energy.

[0017] Preferably, the security protection subsystem ensures the data security and operational security of the entire system, and maintains real-time security monitoring and protection linkage with other subsystems.

[0018] Compared with the prior art, the present invention has the following beneficial effects:

[0019] 1. Significantly improve energy utilization efficiency: Through accurate load forecasting, intelligent control strategy generation and energy management optimization, it can avoid energy waste, reduce the energy consumption of the cold station, and achieve energy-saving operation.

[0020] 2. Enhance system stability and reliability: The real-time monitoring, fault diagnosis and early warning subsystems can detect and handle equipment failures in a timely manner, and the equipment health assessment subsystem helps to carry out maintenance in advance, thereby reducing the risk of unexpected system shutdown and ensuring stable operation of the cold station.

[0021] 3. Improve control accuracy and response speed: The machine learning-based model can adapt to changes in the operating conditions and load of the cooling station in real time, quickly and accurately adjust the control strategy, and ensure that the cooling station is always in the best operating state.

[0022] 4. Reduce operating costs: Energy-saving operation and effective maintenance management of equipment reduce energy consumption and maintenance costs, while adaptive optimization reduces manual intervention costs.

[0023] 5. Realize the effective use of green energy: The green energy access subsystem promotes the application of renewable energy in cold stations, reduces dependence on traditional energy, reduces carbon emissions, and meets the requirements of sustainable development.

[0024] 6. Improve the adaptability and flexibility of the system: The adaptive optimization subsystem and virtual simulation subsystem enable the system to cope with different external environmental changes and internal demand adjustments, and have stronger adaptability and flexibility.

[0025] 7. Promote the collaborative work of the system: The information interaction between the intelligent collaborative subsystem and external related systems realizes the optimized integration of the energy system of the entire building or park, and improves the overall energy utilization efficiency.

[0026] 8. Ensure data and operation security: The security protection subsystem ensures the data security and stable operation of the system, prevents malicious attacks and data leakage, and protects the normal operation of the system and user privacy.

[0027] 9. Improve user experience: Stable, efficient and accurate cold station control provides users with a more comfortable indoor environment and meets their needs for temperature and humidity.

[0028] 10. Promote technological progress in the industry: Introduced advanced machine learning technology and intelligent management concepts into the field of cold station control, providing innovative solutions and demonstrations for the development of the industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0030] The following embodiments of the present invention are described in further detail in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0031] See also Figure 1 The present invention provides an intelligent cold station control system based on machine learning. Before the system is first run, a series of initialization and configuration work needs to be performed. First, through the human-computer interaction subsystem, the operator inputs basic parameters according to the scale, equipment type, operation requirements, etc. of the cold station. These parameters include but are not limited to the design capacity of the cold station, the technical specifications of the equipment, the energy cost, the environmental requirements, etc. The sensors and monitoring equipment of the data acquisition subsystem are calibrated and debugged to ensure that various data such as temperature, humidity, pressure, flow, equipment operation status, etc. can be accurately collected during the operation of the cold station. At the same time, in the data analysis and preprocessing subsystem, data cleaning and conversion rules are set to remove outliers and noise, and the data is converted into a format suitable for machine learning model processing. The machine learning model training subsystem selects a suitable machine learning algorithm, such as a neural network, a decision tree, a support vector machine, etc., and pre-trains according to historical data. The intelligent control strategy generation subsystem sets the initial control strategy parameters as the default control strategy when the system is started. In the energy management subsystem, the target and restriction conditions of energy consumption, as well as the calculation method of energy cost are set. The communication subsystem configures network parameters to ensure that data can be transmitted stably and efficiently between each subsystem.

[0032] The data acquisition subsystem collects the operation data and environmental parameters of the cold station equipment in real time through various sensors distributed in the cold station, such as temperature sensors, humidity sensors, pressure sensors, flow sensors, etc. These data include the inlet and outlet temperatures of the refrigerator, the pressure of the condenser and evaporator, the flow and speed of the water pump, the inlet and outlet water temperature of the cooling tower, the indoor and outdoor temperature and humidity, etc. The collected data is transmitted to the data analysis and preprocessing subsystem at a certain frequency (such as every second, every minute, etc.). In this subsystem, the data is first cleaned to remove obvious outliers and erroneous data. Then, the data is smoothed to reduce the impact of noise. Next, data conversion and normalization are performed to convert data of different units and ranges into a unified format and map them to the range of [0,1] or [-1,1] for machine learning model processing. The preprocessed data is divided into training set, validation set and test set. The training set is used to update and optimize the parameters of the machine learning model, the validation set is used to evaluate the performance of the model during the training process to select the optimal model structure and parameters, and the test set is used to evaluate the generalization ability of the final trained model.

[0033] The machine learning model training subsystem receives the preprocessed data and trains it using the selected machine learning algorithm. Taking the prediction of cooling load as an example, a time series prediction model such as a long short-term memory network (LSTM) or a gated recurrent unit (GRU) can be used. During the training process, the model learns the patterns and regularities in the data, such as the impact of different times, weather conditions, and indoor personnel activities on the cooling load. By continuously adjusting the parameters of the model, the error between the model's predicted output and the actual cooling load data is minimized. At the same time, the validation set is used to evaluate the model in training, and the optimal model structure and parameters are selected based on the evaluation indicators (such as mean square error, mean absolute error, etc.). When the performance of the model on the validation set no longer improves, the training is stopped and the optimal model is saved.

[0034] The intelligent control strategy generation subsystem generates specific control strategies based on the output of the trained machine learning model, combined with the preset control objectives and constraints. For example, if the machine learning model predicts that the cooling load will increase in the future, the intelligent control strategy generation subsystem may decide to start additional chillers in advance or improve the operating efficiency of existing chillers, while adjusting the speed of the water pump and the fan speed of the cooling tower to meet the expected cooling load demand while ensuring the energy efficiency and stability of the system. The control strategy is sent to each actuator in the form of a digital signal, such as a chiller, water pump, cooling tower, etc.

[0035] The real-time monitoring subsystem continuously monitors the operating status and environmental parameters of the cold station equipment. These monitoring data include not only the measured values ​​of physical quantities, but also the operating mode of the equipment (such as on, off, standby), fault status, etc. On the one hand, the monitored real-time data is fed back to the intelligent control strategy generation subsystem for real-time adjustment of the current control strategy. For example, if the actual cooling load is lower than the predicted value, the control strategy may immediately adjust the operating status of the refrigerator and water pump to avoid over-cooling and energy waste. On the other hand, the real-time monitoring subsystem passes the data to the fault diagnosis and early warning subsystem. The fault diagnosis and early warning subsystem uses rule-based methods or machine learning algorithms to analyze real-time data to detect whether the equipment has abnormal operating modes or potential faults. For example, by monitoring the current, voltage, oil temperature and other parameters of the refrigerator, it is determined whether there are problems such as compressor failure and refrigerant leakage. When a fault or abnormality is detected, the fault diagnosis and early warning subsystem sends an alarm message to the human-computer interaction subsystem through the communication subsystem, including the fault type, location and severity. The operator can understand the fault situation in time through the human-computer interaction interface and take corresponding measures, such as emergency shutdown and maintenance scheduling.

[0036] The human-computer interaction subsystem provides an intuitive and easy-to-use interface through which operators can view the real-time operating status, historical data, control strategies, fault alarms and other information of the cold station. Operators can manually adjust the control parameters of the system according to actual needs, such as modifying the set temperature of the refrigerator, adjusting the speed range of the water pump, etc. The human-computer interaction subsystem transmits the operator's instructions to other subsystems to achieve manual intervention in the operation of the cold station. In addition, the human-computer interaction subsystem can also be used to set the system's operating mode, such as automatic mode, manual mode or semi-automatic mode. In automatic mode, the system mainly relies on the control strategy generated by the machine learning model to run automatically; in manual mode, the operator fully controls the operation of the equipment; in semi-automatic mode, the operator can make certain adjustments and interventions based on the control strategy automatically generated by the system.

[0037] The energy management subsystem generates subsystem outputs based on the intelligent control strategy, and monitors and analyzes the energy consumption of the cold station in real time. By comparing the actual energy consumption with the preset energy consumption target, the energy management subsystem identifies energy-inefficient links and equipment. For example, if the energy consumption of a chiller is too high and its cooling effect does not meet expectations, the energy management subsystem may recommend maintenance or replacement of the chiller. At the same time, the energy management subsystem optimizes the operating time and power of the equipment in combination with price fluctuations in the energy market to reduce energy costs. For example, during periods of low energy prices, the operating power of the refrigeration equipment is appropriately increased, and the cooling capacity is stored in advance to reduce operation during peak price periods, thereby reducing overall energy costs.

[0038] The adaptive optimization subsystem continuously monitors the actual operation of the cold station and changes in the external environment, such as seasonal changes, sudden weather changes, and changes in building usage patterns. When it is found that there is a large deviation between the actual operation and the model prediction, the adaptive optimization subsystem automatically adjusts the parameters of the machine learning model training subsystem and retrains the model to improve the accuracy and adaptability of the model. For example, if the cooling effect of the system cannot always meet the demand during the high temperature period in summer, the adaptive optimization subsystem may increase the model's weight for high temperature weather factors, or introduce new feature variables, such as outdoor solar radiation intensity, to improve the model's predictive ability. At the same time, the adaptive optimization subsystem will also adjust the rules and parameters of the intelligent control strategy generation subsystem based on the actual operation data to make the control strategy more in line with the actual situation.

[0039] The load forecasting subsystem uses historical data, real-time data, and external factors (such as weather forecasts, holiday information, etc.) to perform short-term and long-term cooling load forecasts through machine learning algorithms. Short-term forecasts (such as the next few hours to a day) can help the intelligent control strategy generation subsystem to adjust the equipment operating status in time to cope with upcoming load changes. Long-term forecasts (such as the next week or month) can provide references for equipment maintenance plans and energy procurement strategies. For example, if the load forecasting subsystem predicts that there will be continued high temperatures in the next week, resulting in a significant increase in cooling load, the system can arrange maintenance and overhaul of refrigeration equipment in advance to ensure that the equipment can operate stably during peak loads; at the same time, the energy management subsystem can increase energy reserves or sign more energy supply contracts based on the forecast.

[0040] The equipment health assessment subsystem evaluates the health of the equipment by monitoring the equipment's operating parameters, vibration, noise and other data, combined with information such as the equipment's service life and maintenance history. For key equipment, such as refrigerators and water pumps, model-based methods or expert systems are used for fault diagnosis and remaining life prediction. When the equipment health assessment subsystem determines that the equipment has potential faults or is approaching its service life, maintenance prompts and suggestions are sent to the operator through the human-computer interaction subsystem. At the same time, the results of the equipment health assessment subsystem will also affect the decision-making of the intelligent control strategy generation subsystem. For example, if a refrigerator is assessed to be in poor health, the control strategy may reduce its operating time to avoid failures.

[0041] The intelligent collaborative subsystem exchanges information with external related systems, such as the building automation system (BAS), energy management system (EMS), smart grid, etc. Through integration with BAS, information such as personnel activities, lighting and equipment usage inside the building can be obtained to further optimize the control strategy of the cold station. Interaction with EMS can realize the comprehensive management and allocation of energy and improve the energy utilization efficiency of the entire building or park. Connection with the smart grid can enable the cold station to adjust the operation mode according to the load situation and electricity price signal of the power grid, realize demand response and optimize the use of energy. For example, during the peak load period of the power grid, the power consumption of the cold station can be appropriately reduced to provide support for the power grid and obtain corresponding economic compensation.

[0042] Before the implementation of the new control strategy, the virtual simulation subsystem uses the established mathematical model of the cold station and the simulation environment to perform simulation operation. By simulating different operation scenarios and control strategies, the impact on the performance, energy consumption and equipment operation of the cold station is evaluated. The results of the virtual simulation can provide a reference for the intelligent control strategy generation subsystem, help optimize the control strategy, and reduce risks and errors in actual operation. For example, when considering upgrading or renovating the cold station equipment, first simulate different solutions in the virtual simulation subsystem, compare their effects and costs, and select the best solution before actual implementation.

[0043] The green energy access subsystem is responsible for monitoring and managing the input of renewable energy, such as solar photovoltaic power generation, wind power generation, etc. When the electricity generated by renewable energy is sufficient, green energy is used first to power the cold station equipment to reduce dependence on the traditional power grid. At the same time, combined with energy storage devices (such as battery packs), green energy storage and deployment are achieved to cope with fluctuations in renewable energy output. The green energy access subsystem works in conjunction with the energy management subsystem to optimize the comprehensive utilization of energy and reduce the carbon footprint of the cold station. For example, during periods of abundant sunshine, solar power is used to drive refrigeration equipment, and excess electricity is stored for use at night or on cloudy days.

[0044] The security protection subsystem uses a variety of technical means, such as firewalls, encryption technology, access control, etc., to ensure the data security and operation security of the entire cold station control system. Sensitive data in the system, such as control strategy parameters and equipment operation data, are encrypted for storage and transmission to prevent data leakage and tampering. At the same time, strict permission management is implemented for access to the system, and only authorized personnel can operate and view relevant data. The security protection subsystem also monitors the operating status of the system in real time to prevent network attacks and malware intrusions. When security threats are detected, isolation, alarm and other measures are taken in a timely manner to ensure the stable operation of the system.

[0045] Through the collaborative work of the above subsystems, the intelligent cold station control system based on machine learning of the present invention can achieve efficient, accurate, energy-saving and safe control and management of the cold station, provide users with a comfortable indoor environment, and reduce energy consumption and operating costs.

[0046] Through accurate load forecasting, intelligent control strategy generation and energy management optimization, energy waste can be avoided, energy consumption of cold stations can be reduced, and energy-saving operation can be achieved. The real-time monitoring, fault diagnosis and early warning subsystems can detect and handle equipment failures in a timely manner. The equipment health assessment subsystem helps to carry out maintenance in advance, thereby reducing the risk of unexpected system shutdowns and ensuring the stable operation of cold stations. The machine learning-based model can adapt to changes in cold station operating conditions and loads in real time, quickly and accurately adjust the control strategy, and ensure that the cold station is always in the best operating state. Energy-saving operation and effective maintenance management of equipment reduce energy consumption and maintenance costs. At the same time, adaptive optimization reduces the cost of manual intervention. The green energy access subsystem promotes the application of renewable energy in cold stations, reduces dependence on traditional energy, and reduces Carbon emissions, in line with the requirements of sustainable development, the adaptive optimization subsystem and the virtual simulation subsystem enable the system to cope with different external environmental changes and internal demand adjustments, with stronger adaptability and flexibility, the intelligent collaborative subsystem interacts with external related systems to achieve the optimized integration of the energy system of the entire building or park, and improves the overall energy utilization efficiency, the security protection subsystem ensures the data security and stable operation of the system, prevents malicious attacks and data leakage, protects the normal operation of the system and user privacy, and the stable, efficient and accurate cold station control provides users with a more comfortable indoor environment, meets users' needs for temperature and humidity, and introduces advanced machine learning technology and intelligent management concepts to the field of cold station control, providing innovative solutions and demonstrations for the development of the industry.

[0047] The embodiments of the present invention are given for the purpose of illustration and description, and are not intended to be exhaustive or to limit the invention to the disclosed forms. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are selected and described in order to better illustrate the principles and practical applications of the present invention and to enable those of ordinary skill in the art to understand the present invention and thereby design various embodiments with various modifications suitable for specific uses.

Claims

1. An intelligent cold station control system based on machine learning, characterized in that: The intelligent cold station control system based on machine learning is divided into an infrastructure part and an intelligent auxiliary part; The infrastructure includes data acquisition subsystem, data analysis and preprocessing subsystem, machine learning model training subsystem, intelligent control strategy generation subsystem, real-time monitoring subsystem, fault diagnosis and early warning subsystem, human-computer interaction subsystem, energy management subsystem, and communication subsystem. The intelligent auxiliary part includes adaptive optimization subsystem, load forecasting subsystem, equipment health assessment subsystem, intelligent coordination subsystem, virtual simulation subsystem, green energy access subsystem, and safety protection subsystem; The energy management subsystem generates subsystem outputs based on intelligent control strategies and optimizes the energy consumption of the cooling station.

2. The intelligent cold station control system based on machine learning as claimed in claim 1, characterized in that: The cold station operation data collected by the data acquisition subsystem is transmitted to the data analysis and preprocessing subsystem. The processed data is provided to the machine learning model training subsystem for model training and optimization. The trained machine learning model plays a role in the intelligent control strategy generation subsystem to generate control strategies.

3. The intelligent cold station control system based on machine learning as claimed in claim 2, characterized in that: The real-time monitoring subsystem will feed back the real-time data to the intelligent control strategy generation subsystem so as to adjust the control strategy in real time. At the same time, the real-time monitoring subsystem will also pass the data to the fault diagnosis and early warning subsystem for fault diagnosis and early warning.

4. The intelligent cold station control system based on machine learning as claimed in claim 3, characterized in that: The results of the fault diagnosis and early warning subsystem are transmitted to the human-computer interaction subsystem through the communication subsystem to inform the operator. The human-computer interaction subsystem can also receive instructions from the operator to set parameters and intervene in other subsystems.

5. The intelligent refrigeration plant control system based on machine learning as claimed in claim 4, characterized in that: The adaptive optimization subsystem automatically adjusts the parameters of the machine learning model training subsystem and the strategies of the intelligent control strategy generation subsystem according to the actual operation of the cold station and changes in the external environment.

6. The intelligent refrigeration plant control system based on machine learning as claimed in claim 5, characterized in that: The equipment health status information evaluated by the equipment health assessment subsystem is transmitted to the intelligent control strategy generation subsystem to affect the control strategy. At the same time, the equipment maintenance requirements are also prompted to the operator through the human-computer interaction subsystem.

7. The intelligent refrigeration plant control system based on machine learning as claimed in claim 6, characterized in that: The intelligent collaborative subsystem exchanges information with external related systems and feeds back the collaborative requirements to the intelligent control strategy generation subsystem.

8. The intelligent refrigeration plant control system based on machine learning as claimed in claim 7, characterized in that: The virtual simulation subsystem performs simulation verification before the strategy is implemented, and the results are fed back to the intelligent control strategy generation subsystem for optimization.

9. The intelligent refrigeration plant control system based on machine learning as claimed in claim 8, characterized in that: The green energy access subsystem transmits the information of renewable energy to the energy management subsystem to achieve comprehensive and optimal utilization of energy.

10. The intelligent refrigeration plant control system based on machine learning as claimed in claim 9, characterized in that: The security protection subsystem ensures the data security and operation security of the entire system, and maintains real-time security monitoring and protection linkage with other subsystems.

Images