Automatic exhaust control method and device for water pump of cooling-water machine

Abstract

The invention relates to the technical field of measurement and testing, in particular to an exhaust pressure fault detection method. According to the technical scheme, the automatic exhaust control method for the water pump of the cooling-water machine comprises the following steps that S1, historical data of the working state of the water pump are collected, and a feature set is constructed after data preprocessing; s2, dividing the collected historical data into a training set and a test set, and training a model by using data of the training set; s3, relevant data are obtained in real time in the running process of the water pump, the data monitored in real time are input into the trained model, and whether exhausting is needed or not is judged; and S4, executing according to the judgment condition, and monitoring real-time data. The problems that in the prior art, the exhaust opportunity cannot be judged according to reality, alarm measures are lacked, and self iteration cannot be achieved are solved, the automatic exhaust control method and device for the water pump of the cooling-water machine are provided, and the purposes of high efficiency, high intelligent level, high safety, energy saving and high reliability are achieved.

Description

Technical Field

[0001] The invention relates to the field of measurement and testing technology, and in particular to an exhaust pressure fault detection method. Background Art

[0002] The working surface area of ​​a traditional thermoelectric cooler is generally fixed after the design is finalized and cannot be adjusted a second time. Once the position of the heat source changes, the original thermoelectric cooler will no longer be applicable and needs to be redesigned; secondly, the working area of ​​a traditional thermoelectric cooler is often larger than the heat source area, so when working, it is inevitable that some working area will be wasted, resulting in a portion of electrical energy loss. For example, a Chinese patent with publication number CN110420565A discloses an exhaust method and an exhaust device for a water purifier with a movable water tank, and provides the following technical solutions. The present invention is related to the water treatment industry, and specifically to drinking water filtration. The present invention discloses an exhaust method and an exhaust device for a water purifier with a movable water tank. A plurality of filter galls are connected in series to form a filter channel, and the movable water tank plug-in water port is sealed and plugged with the filter channel water interface, and an exhaust device is arranged in the pre-filter channel between the water interface and the fine filter gall. It includes a front-end interface and an interface sealing switching structure placed in the pre-filter channel, and also includes a rear-end interface and a discharge pipeline with a water interface; the interface sealing switching structure is connected to the pre-filter channel located behind the booster pump through the front-end interface, and is connected to the active water tank through the rear-end interface and the discharge pipeline with a water interface, and is switched between the front-end interface and the rear-end interface for closing or connecting: the interface sealing switching structure is in a closed position or a connecting position, respectively, and each corresponding discharge pipeline is isolated from the filter channel, or the originally closed discharge pipeline is connected to the pre-filter channel. However, the above-mentioned method for exhausting a water purifier with a movable water tank and its exhaust device cannot judge whether exhaust is needed according to actual needs, and cannot take different alarms or corresponding measures for different situations, and lacks corresponding learning iteration functions. Summary of the invention

[0003] The present invention solves the problems in the prior art of being unable to judge the exhaust timing according to actual conditions, lacking alarm measures and being unable to self-iterate, and proposes an automatic exhaust control method and device for a chiller water pump, achieving the goals of high efficiency, high intelligence level, high safety, energy saving and high reliability.

[0004] To achieve the above object, the present invention adopts the following technical solution: A method for automatically exhausting a water pump of a chiller, comprising the following steps: S1: Collect historical data of the working status of the water pump, and construct a feature set after data preprocessing; S2: Divide the collected historical data into training set and test set, and use the training set data to train the model; S3: Obtain relevant data in real time during the operation of the water pump, input the real-time monitoring data into the trained model, and determine whether exhaust is needed; S4: Execute based on judgment and monitor real-time data.

[0005] An automatic exhaust device for a chiller water pump comprises an exhaust device and a control system. The control system comprises a control module and a data detection module. The data detection module is installed in the exhaust device. The data detection module is connected to an alarm module via the control module. The control module is connected to a human-computer interaction module. The exhaust device comprises a water pump and a first solenoid valve and a second solenoid valve located at both ends of the water pump.

[0006] The advantage of this design is that by collecting historical data and preprocessing, relevant features can be extracted, providing a good foundation for subsequent model training. By dividing the data set and training the model, the generalization ability of the model can be effectively verified, thereby improving the accuracy of the prediction.

[0007] Preferably, step S2 comprises the following steps: S2.1: Use the dataset partitioning function to divide the feature set and label; S2.2: Use part of the data as the training set and the remaining data as the test set, and set the random seed; S2.3: Import the random forest classifier, create a random forest classifier object, and set parameters; S2.4: Input the training set features and labels into the model for training, and input the test set features into the model to obtain the prediction results; S2.5: By comparing the prediction results with the actual labels, calculate the various indicators of the model and evaluate the model effect.

[0008] Preferably, the historical data includes flow rate, pressure, temperature and pump operation time; the feature set includes flow rate change rate, pressure change, temperature change and drainage times; and the indicators of the model include model accuracy, precision and recall rate.

[0009] The advantage of this design is that it takes into account a variety of influencing factors, can comprehensively evaluate the model performance, and ensure that the system can accurately judge the exhaust demand under different working conditions.

[0010] Preferably, step S3 comprises the following steps: S3.1: Each detection unit detects relevant data in real time and calculates features; S3.2: Encapsulate the calculated features into an array or list and convert them into the input format of the model; S3.3: Use the prediction method to input the real-time features into the model and obtain the prediction results; S3.4: Determine whether exhaust is required based on the prediction results.

[0011] The advantage of this design is that it enables real-time monitoring and processing of data, ensuring that the system can respond quickly and make correct judgments, avoiding unnecessary energy consumption or damage.

[0012] Preferably, step S4 comprises the following steps: S4.1: Determine whether to perform exhaust according to the situation. If yes, proceed to step S4.2; if not, return to step S3.1; S4.2: Start the solenoid valve C1 to exhaust, open the solenoid valve C2 to replenish water, and close the solenoid valves C1 and C2 according to the set time; S4.3: Each detection unit monitors relevant data in real time and determines whether there is any abnormality; S4.4: Record the results and compare with the actual state, adjust the model parameters or training data based on the feedback and return to step S3.1.

[0013] The advantage of this design is that it provides a closed-loop control system that continuously optimizes exhaust operations through a feedback mechanism, thereby improving system stability and reliability.

[0014] Preferably, in step S4.3, if the flow rate is detected to be 0 for a long time after exhaust is performed, it is determined that the water pump has not been started, and an alarm for water pump not being started is issued; if the flow sensor detects that the flow rate exceeds the set maximum value, an overcurrent alarm is issued and the water pump is controlled to be shut down.

[0015] The advantage of this design is that it ensures system security through anomaly detection mechanisms and prevents hardware damage and resource waste.

[0016] Preferably, in step S4.3, if exhaust is performed multiple times in a short period of time but the water pump flow rate is not restored, a continuous exhaust alarm is issued; if exhaust is performed multiple times in a short period of time and the flow rate is not detected to increase after exhaust, it is determined that the water pump is abnormal and a water pump abnormality alarm is issued.

[0017] The advantage of this design is that by monitoring and analyzing abnormal exhaust behavior, potential faults can be identified and handled in a timely manner to avoid causing greater losses to the system.

[0018] Preferably, in step S2.2, 70% to 80% of the data is used as a training set, and the remaining 20% ​​to 30% is used as a test set.

[0019] Preferably, the front end of the water pump is connected to the second solenoid valve, and the rear end is connected to the first solenoid valve. The front end of the second solenoid valve is connected to a water tank. The data detection module is installed between the water pump and the first solenoid valve. The rear end of the first solenoid valve is connected to an exhaust pipe.

[0020] Compared with the prior art, the invention has the following beneficial effects.

[0021] 1. The automatic exhaust control method for chiller water pumps of the present invention realizes intelligent exhaust decision-making by real-time monitoring and analysis of the working status of the water pump. This method can timely judge the accumulation of gas, thereby avoiding the reduction of the efficiency of the water pump caused by gas and keeping it in the best working state. By optimizing the exhaust process, the overall efficiency is improved, ensuring that the water pump can operate stably and efficiently, thereby improving the overall effect.

[0022] 2. The present invention sets up a variety of alarm mechanisms, including water pump failure alarm and overcurrent alarm, to ensure that the system can promptly detect and handle potential abnormal situations during operation. This real-time monitoring and feedback mechanism greatly enhances the safety of the system, and can effectively reduce the safety risks of equipment and personnel. By responding to abnormalities in a timely manner, it ensures the safe operation of the chiller and reduces the possibility of failures.

[0023] 3. Through the analysis of historical data and model optimization, the present invention enables the system to self-adjust, reducing the need for manual intervention and the maintenance frequency of equipment. At the same time, the alarm and monitoring functions of the system can detect problems in a timely manner and avoid greater losses caused by equipment failure, thereby significantly saving maintenance and repair costs. This efficient maintenance method makes the operation of the chiller water pump more economical and improves the overall economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 The present invention is a method flow chart of an embodiment of a chiller water pump automatic exhaust control method and device of the present invention.

[0025] Figure 2 The present invention is a schematic diagram of an exhaust device of an embodiment of a method and device for automatically exhausting a water pump of a chiller.

[0026] Illustration Description: 1 water tank, 2.1 first solenoid valve, 2.2 second solenoid valve, 3 water pump, 4 data detection module. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical scheme and advantages of the present disclosure clearer, the following will further describe the embodiments of the present disclosure in detail with reference to the accompanying drawings. The proportions of the components are not drawn according to the actual proportions, and the proportions and sizes shown in the accompanying drawings should not limit the substantial technical scheme of the present invention. These embodiments do not describe all the details in detail, nor do they limit the invention to the specific embodiments described.

[0028] See also Figure 1 and Figure 2As shown, a chiller water pump automatic exhaust control method comprises the following steps: S1: Collect historical data of the working status of the water pump, and construct a feature set after data preprocessing; S2: Divide the collected historical data into training set and test set, and use the training set data to train the model; S3: Obtain relevant data in real time during the operation of the water pump, input the real-time monitoring data into the trained model, and determine whether exhaust is needed; S4: Execute based on judgment and monitor real-time data.

[0029] An automatic exhaust device for a chiller water pump comprises an exhaust device and a control system. The control system comprises a control module and a data detection module. The data detection module is installed in the exhaust device. The data detection module is connected to an alarm module via the control module. The control module is connected to a human-computer interaction module. The exhaust device comprises a water pump and a first solenoid valve and a second solenoid valve located at both ends of the water pump.

[0030] The automatic exhaust control method and device for a chiller water pump designed by the present invention has significant advantages. First, it can automatically detect whether there is gas inside the water pump and exhaust it when necessary without manual intervention, thereby reducing the workload of the operator and improving the intelligence level of the system. Secondly, the system monitors the working status of the water pump in real time and can automatically determine whether the water pump is operating normally or starting, which ensures that timely measures are taken when an abnormality occurs, reduces the probability of failure, and improves the reliability of the equipment. Finally, the equipped alarm system can timely detect and feedback abnormal conditions of the water pump, including overcurrent, equipment failure, and continuous exhaust, etc., to help operators quickly identify potential risks and reduce damage to the water pump through timely intervention, thereby extending the service life of the equipment. In summary, the system not only improves the operating efficiency of the water pump, but also reduces maintenance costs and failure risks, ensuring the long-term stable operation of the chiller system.

[0031] like Figure 1 In one embodiment shown, Figure 1 The present invention is a method flow chart of an embodiment of a chiller water pump automatic exhaust control method and device. The present invention is designed to improve the working efficiency and safety of the water pump. The specific steps are as follows: First, the system collects historical data on the working status of the water pump, including flow, pressure, temperature, and running time of the water pump. After preprocessing, the collected data is used to construct a feature set for subsequent model training and prediction. The preprocessing process includes data cleaning, denoising, and standardization to ensure the accuracy and reliability of the data.

[0032] Then, the feature set and label are divided using the dataset partitioning function to ensure that the label corresponding to each feature can be effectively used for training. Next, the collected historical data is divided into a training set and a test set. Usually, 70% to 80% of the data is used as a training set, and the remaining 20% ​​to 30% is used as a test set. A random seed is set to ensure the repeatability of the experiment.

[0033] In S2.3, import the random forest classifier, create a classifier object, and set relevant parameters. Random forest is an ensemble learning method that can effectively process high-dimensional data and nonlinear features and has strong classification capabilities. Subsequently, in S2.4, the features and labels of the training set are input into the model for training, and then the features of the test set are input into the model to obtain the prediction results. Finally, in S2.5, by comparing the model's prediction results with the actual labels, the model's accuracy, precision, recall and other indicators are calculated to evaluate the model's effect and performance.

[0034] After completing the model training, enter step S3. In S3.1, each detection unit detects relevant data in real time and calculates real-time features, such as flow rate change and pressure change. Then in S3.2, the calculated features are encapsulated into an array or list and converted into the input format of the model. Then, the real-time features are input into the model using the prediction method to obtain the prediction results. In S3.4, it is determined whether exhaust is needed based on the prediction results.

[0035] In step S4, firstly, it is decided whether to perform exhaust according to the judgment situation. In S4.1, if exhaust is required, step S4.2 is entered; if not, step S3.1 is returned to continue monitoring data. In S4.2, while starting the solenoid valve C1 for exhaust, the solenoid valve C2 is opened for water replenishment, and the solenoid valves C1 and C2 are closed according to the set time. This step ensures that the normal operation of the water pump will not be affected by lack of water during the exhaust process. Next, in S4.3, each detection unit monitors the relevant data in real time to determine whether there is an abnormal situation. If the flow rate is detected to be 0 for a long time after the exhaust is performed, it is determined that the water pump has not started, and the water pump has not started alarm; if the flow sensor detects that the flow rate exceeds the set maximum value, an over-current alarm is issued, and the water pump is controlled to shut down to protect the safety of the equipment. In addition, if exhaust is performed repeatedly in a short period of time, but the water pump flow rate is not restored, a continuous exhaust alarm is issued; if exhaust is performed multiple times in a short period of time and the flow rate is still not increased after exhaust, it is determined that the water pump is abnormal, and the water pump abnormal alarm is issued. In S4.4, the system will record the operation results and compare them with the actual status. Based on the feedback information, the system can adjust the model parameters or training data and return to step S3.1 to perform new data monitoring and prediction, thereby achieving self-learning and optimization of the system.

[0036] The automatic exhaust control method for chiller water pump designed by the present invention can effectively identify the key factors affecting the operating performance of the water pump by comprehensively collecting and analyzing the historical data of the working status of the water pump, thereby providing a solid foundation for subsequent model training. The feature set constructed based on historical data can reflect the performance of the water pump under different working conditions and environments, making the model training more effective. Secondly, the model training is performed using a random forest classifier, so that the model has a higher accuracy and robustness. As an integrated learning method, random forest can effectively reduce the risk of overfitting and improve the generalization ability of new data through the voting mechanism of multiple decision trees. This feature enables the model to better adapt to the changes in the working status of the water pump and make timely responses and adjustments in practical applications. In addition, the combination of real-time data monitoring and model prediction enables the system to automatically determine whether exhaust operation is required during the operation of the water pump. By acquiring sensor data such as flow and pressure in real time, the system can quickly respond to changes under different working conditions to ensure that the water pump is always in the best operating state. This intelligent control method reduces the need for manual intervention, reduces the workload of operators, and also improves the automation level of the system. In terms of safety, the various alarm mechanisms set up can promptly detect and handle abnormal conditions in the operation of the water pump. For example, if it is detected that the water pump has not started or the flow rate exceeds the maximum value, the system can quickly issue an alarm signal and take corresponding protective measures. This not only effectively prevents equipment damage, but also ensures the safety of operators.

[0037] Finally, the design of the system feedback mechanism enables the model to be continuously optimized and adjusted according to the actual operating conditions. This adaptive capability ensures the stability and reliability of the system in long-term operation, can adapt to changing working conditions, and improves overall economic benefits. By recording operation results and making feedback adjustments, the system can not only correct possible errors in a timely manner, but also continuously improve decision-making capabilities in future operations, thereby achieving more efficient pump management.

[0038] In one embodiment, the specific method steps are as follows: 1. Data Collection and Processing Collect historical data on the working status of the water pump, including flow, pressure, temperature, and running time of the water pump, and then perform data preprocessing, including removing outliers (such as flow cannot be negative), filling missing values ​​(such as using mean or median), and normalizing (scaling the data to between 0 and 1):

[0039] 2. Feature Engineering A feature set is constructed based on the collected data, where the features include: the flow rate change rate of the water pump, pressure change, temperature change, operating time, drainage times, and environmental factors such as the temperature and pressure of the water source.

[0040] 3. Model selection Use a suitable machine learning algorithm. In this embodiment, a random forest algorithm model is used. Specifically: When the system starts, load the pre-trained model. If the system is used for the first time, perform preliminary data collection and model training. Divide the feature set into a training set and a test set. Use the train_test_split function from the scikit-learn library to divide the feature set (X) and the label (y) to train the model and evaluate the performance of the model. In this embodiment, 70% of the data is used for training and 30% of the data is used for testing. Then set the random seed. In this embodiment, random_state = 42 to ensure that the data division is consistent each time it is run, so as to facilitate the reproduction of the results.

[0041] Use the training set data to train the random forest model so that the model can learn the relationship between features and prediction results. Import the RandomForestClassifier class, create a random forest classifier object, set the n_estimators parameter (that is, the number of trees built, set between 100 and 200) and the random seed; use the fit method to input the training set features (X_train) and labels (y_train) into the model for training. The relevant code is: from sklearn.ensemble import RandomForestClassifier from sklearn.model_selection import train_test_split #Assume that X is the feature set and y is the label (need exhaust or not) X_train,X_test,y_train,y_test=train_test_split(X,y,test_size=0.3,random_state=42) #Create a random forest classifier model=RandomForestClassifier(n_estimators=100,random_state=42) #Train model model.fit(X_train,y_train)

[0042] Evaluate the performance of the model on the test set to understand its predictive ability. Use the predict method to input the test set features (X_test) into the model to obtain the predicted results. By comparing the predicted results with the actual labels (y_test), calculate the model's accuracy, precision, recall and other indicators to evaluate the model's effectiveness.

[0043] 4. Real-time monitoring When the water pump is turned on, the data of flow sensor, pressure sensor and temperature sensor are continuously monitored and passed to the model for analysis. The sensor monitoring program is set to ensure that the flow, pressure, temperature and other data can be read in real time. The calculation features mainly include flow rate change rate, pressure change, temperature change, running time and past exhaust times. These features will be passed to the model as input. The specific code is: flow_rate_change=current_flow_rate-previous_flow_rate pressure_change=current_pressure-previous_pressure temperature_change=current_temperature-previous_temperature running_time=get_running_time()#Get the running time of the pump exhaust_count = get_exhaust_count() #Number of past exhaust times

[0044] 5. Model prediction Encapsulate the calculated features into an array or list, format it into the input format required by the model, use the predict method to input the real-time features into the model, obtain the prediction results, and decide the next control strategy based on the prediction results. The model will output a binary classification result: Venting required: if the model predicts the presence of gas; No venting required: if the model predicts the pump is OK.

[0045] The specific code is: #Get real-time data and calculate features current_features=[flow_rate_change,pressure_change,temperature_change,running_time,exhaust_count] #Call the model for prediction prediction=model.predict([current_features]) #Take corresponding control measures based on the prediction results if prediction[0]==1:#Assume that 1 means exhaust is required #Start solenoid valve C1 to exhaust and solenoid valve C2 to replenish water activate_valve_C1() activate_valve_C2() time.sleep(2)#Exhaust lasts for 2 seconds deactivate_valve_C1() deactivate_valve_C2() else: #Continue to monitor the pump status continue_monitoring()

[0046] 6. Execute control logic based on the results If the model predicts that venting is needed: The solenoid valve C1 is started to exhaust, and the solenoid valve C2 is opened to replenish water, and the solenoid valves C1 and C2 are closed according to the set exhaust time (such as 2 seconds in this embodiment).

[0047] If the model predicts that no venting is needed: Continue to monitor the pump status.

[0048] 7. Alarm system over-current alarm: When the flow sensor detects that the flow exceeds the set maximum value, the system will issue an over-current alarm; Treatment measures: Stop the water pump immediately and remind the user to check whether the pipeline is blocked or leaking.

[0049] Equipment failure alarm: When the system detects a failure in the water pump or related equipment (such as motor overload, sensor failure, etc.), an equipment failure alarm is issued; Treatment measures: Stop the operation of all related equipment and prompt the user to perform maintenance inspection.

[0050] Continuous exhaust alarm: If exhaust is predicted to be required for multiple times within a certain period of time (5 minutes in this embodiment), but the water pump flow rate has not been restored, a continuous exhaust alarm is issued; Action: Prompt the user to check the system status to ensure there are no other problems.

[0051] Low pressure alarm: When the water outlet pressure of the water pump is detected to be lower than the set safety threshold, a low pressure alarm will be issued; Treatment measures: Stop the water pump and remind the user to check whether the water inlet valve is open or blocked.

[0052] Frequent exhaust alarm: if the exhaust is automatically exhausted multiple times within a short period of time (within 30 minutes in this embodiment), the system will issue a frequent exhaust alarm; Treatment measures: Prompt the user to check the system settings or operating status to avoid unnecessary exhaust.

[0053] 8. Feedback and Learning After each execution of the control logic, the results are recorded and compared with the actual state, and the model parameters or training data are adjusted based on the feedback to continuously improve the accuracy of the model.

[0054] like Figure 2 In one embodiment shown, Figure 2 The present invention is a schematic diagram of an exhaust device of an embodiment of a chiller water pump automatic exhaust control method and device. The present invention relates to an automatic exhaust device for a chiller water pump, which is composed of an exhaust device and a control system. The control system mainly includes a control module and a data detection module, wherein the data detection module is installed inside the exhaust device and is connected to the alarm module through the control module. In addition, the control module is also connected to a human-computer interaction module for easy operation and monitoring. In the exhaust device, the water pump is a core component, and its front end is connected to the second solenoid valve, and the rear end is connected to the first solenoid valve. The front end of the second solenoid valve is connected to a water tank to ensure that the water pump can smoothly pump water. At the same time, the data detection module is installed between the water pump and the first solenoid valve to facilitate real-time monitoring of the working status of the water pump. Finally, the rear end of the first solenoid valve is connected to an exhaust pipe for discharging excess gas in the water pump to ensure efficient operation of the water pump.

[0055] The benefit of this design is that it takes into account the efficiency and safety of the system. By placing the data detection module between the water pump and the first solenoid valve, the working status of the water pump can be monitored in real time and fed back to the control module in a timely manner. In this way, if the water pump flow is abnormal or not started, the system can quickly send out an alarm signal to protect the equipment and reduce safety risks. At the same time, the connection between the control module and the human-computer interaction module enhances the user's operating convenience, allowing the operator to intuitively monitor the status of the water pump and make corresponding adjustments. In addition, the rationally configured solenoid valve and water pump structure design ensure efficient water flow and gas discharge, which helps to improve the overall performance of the chiller. Through the automated control and monitoring system, the device can achieve more intelligent water pump management and improve the operating efficiency and economic benefits of the chiller.

[0056] In summary, the present invention provides a method and device for automatically exhausting a water pump of a chiller, aiming to improve the working efficiency and safety of the water pump. The system includes an exhaust device and a control system, wherein the control system is composed of a control module and a data detection module, the latter of which is installed inside the exhaust device and connected to the alarm module through the control module, and is connected to the human-computer interaction module for operation and monitoring. The control method mainly includes the following steps: First, the historical data of the water pump (such as flow, pressure, temperature and running time) is collected, and the data is preprocessed to construct a feature set. Then, the historical data is divided into a training set and a test set, and the model is trained using a random forest classifier. After the model training is completed, the system can obtain the relevant data of the water pump in real time, and input these data into the trained model to determine whether exhaust is required. When performing the exhaust operation, the system starts the solenoid valve for exhaust and water replenishment according to the prediction results of the model, and monitors the working status of the water pump in real time. If an abnormal situation is detected (such as a flow rate of 0 or a flow rate exceeding a set maximum value), an alarm signal is quickly issued and corresponding protective measures are taken. In addition, the system sets a variety of alarm mechanisms, which can promptly detect and handle abnormal situations in the operation of the water pump to ensure the safety of the equipment.

[0057] The advantage of this design is that the system can automatically determine the state of the water pump during operation, realize intelligent control, reduce manual intervention, and improve the level of automation. At the same time, the combination of real-time monitoring and model prediction can quickly respond to changes under different working conditions to ensure that the water pump is always in the best operating state. Through the feedback mechanism, the system can continuously optimize and adjust the model to improve the overall economic benefits and operational safety. In short, the automatic exhaust control method and device for the chiller water pump of the present invention, by comprehensively considering efficiency and safety, realizes an intelligent water pump management method, which significantly improves the operating efficiency of the chiller and the safety of the equipment.

[0058] The present invention is not limited to the above-mentioned embodiments. No matter any changes are made in the shape or material composition, any structural design provided by the present invention is a variation of the present invention and should be considered to be within the protection scope of the present invention.

Claims

1. A chiller water pump automatic exhaust control method, characterized in that: The following steps are involved: S1: Collect historical data of the working status of the water pump, and construct a feature set after data preprocessing; S2: Divide the collected historical data into training set and test set, and use the training set data to train the model; S3: Obtain relevant data in real time during the operation of the water pump, input the real-time monitoring data into the trained model, and determine whether exhaust is needed; S4: Execute based on judgment and monitor real-time data.

2. The method for controlling the automatic exhaust of a water chiller pump according to claim 1, characterized in that: The step S2 comprises the following steps: S2.1: Use the dataset partitioning function to divide the feature set and label; S2.2: Use part of the data as the training set and the remaining data as the test set, and set a random seed; S2.3: Import the random forest classifier, create a random forest classifier object, and set parameters; S2.4: Input the training set features and labels into the model for training, and input the test set features into the model to obtain the prediction results; S2.5: By comparing the prediction results with the actual labels, calculate the various indicators of the model and evaluate the model effect.

3. The method for controlling the automatic exhaust of a chiller pump according to claim 1, characterized in that: The step S3 comprises the following steps: S3.1: Each detection unit detects relevant data in real time and calculates features; S3.2: Encapsulate the calculated features into an array or list and convert them into the input format of the model; S3.3: Use the prediction method to input the real-time features into the model and obtain the prediction results; S3.4: Determine whether exhaust is required based on the prediction results.

4. The method for controlling the automatic exhaust of a chiller pump according to claim 1, characterized in that: The step S4 comprises the following steps: S4.1: Determine whether to perform exhaust according to the situation. If yes, proceed to step S4.2; if not, return to step S3.1; S4.2: Start the solenoid valve C1 to exhaust, open the solenoid valve C2 to replenish water, and close the solenoid valves C1 and C2 according to the set time; S4.3: Each detection unit monitors relevant data in real time and determines whether there is any abnormality; S4.4: Record the results and compare with the actual state, adjust the model parameters or training data based on the feedback and return to step S3.

1.

5. A chiller water pump automatic exhaust control method according to claim 1 or 2 or 3 or 4, characterized in that: The historical data includes flow rate, pressure, temperature and water pump operation time; the feature set includes flow rate change rate, pressure change, temperature change and drainage times; the indicators of the model include model accuracy, precision and recall rate.

6. The method for controlling the automatic exhaust of a chiller pump according to claim 4, characterized in that: In step S4.3, if the flow rate is detected to be 0 for a long time after exhaust is performed, it is determined that the water pump has not started, and an alarm for water pump not starting is issued; if the flow sensor detects that the flow rate exceeds the set maximum value, an overcurrent alarm is issued and the water pump is controlled to shut down.

7. A chiller water pump automatic exhaust control method according to claim 4 or 6, characterized in that: In step S4.3, if exhaust is performed multiple times in a short period of time but the water pump flow rate is not restored, a continuous exhaust alarm is issued; if exhaust is performed multiple times in a short period of time and the flow rate is not detected to increase after exhaust, it is determined that the water pump is abnormal and a water pump abnormality alarm is issued.

8. The method for controlling the automatic exhaust of a chiller pump according to claim 2, characterized in that: In step S2.2, 70% to 80% of the data are used as a training set, and the remaining 20% ​​to 30% are used as a test set.

9. An automatic exhaust device for a chiller water pump, using an automatic exhaust control method for a chiller water pump according to any one of claims 1 to 8, characterized in that: The invention comprises an exhaust device and a control system, wherein the control system comprises a control module and a data detection module (4), wherein the data detection module (4) is installed in the exhaust device, wherein the data detection module is connected to an alarm module via the control module, and wherein the control module is connected to a human-computer interaction module, wherein the exhaust device comprises a water pump (3) and a first solenoid valve (2.1) and a second solenoid valve (2.2) located at both ends of the water pump (3).

10. The automatic exhaust device for a chiller water pump according to claim 9, characterized in that: The front end of the water pump (3) is connected to the second solenoid valve (2.2), and the rear end is connected to the first solenoid valve (2.1); the front end of the second solenoid valve (2.2) is connected to a water tank (1); the data detection module (4) is installed between the water pump (3) and the first solenoid valve (2.1); and the rear end of the first solenoid valve (2.1) is connected to an exhaust pipe.

Images