Energy-saving electric control system of air separation plant

Abstract

The invention relates to an energy-saving electric control system of air separation equipment, which comprises a data collection module used for collecting circuit operation data when the air separation equipment operates in different processing modes and operation data of sensor collection equipment, and transmitting the data to a data storage module in a corresponding storage time period; a data processing module; a wireless transmission module; an abnormity alarm module; a data storage module; and a scheduling module. According to circuit operation data of different storage time periods collected by a data collection module and operation data of equipment collected by a sensor, after the data are used for manufacturing a deep learning model, a circuit operation mode of the corresponding time period is formulated, and therefore main operation equipment in the corresponding time section is determined; the air component concentrations in different time periods are treated, the situation that low-content air components in the current small time section are treated through long-time high-efficiency operation is avoided, and the energy-saving effect is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of air separation equipment control and relates to an energy-saving electric control system for air separation equipment. Background Art

[0002] Air separation equipment is a device that uses air as raw material, converts the air into liquid through compression cycle deep freezing, and then gradually separates and produces inert gases such as oxygen, nitrogen and argon from the liquid air through distillation.

[0003] The concentration distribution of air components changes throughout the day due to factors such as temperature, human activities and natural phenomena. For example, the oxygen concentration increases slightly during the day due to plant photosynthesis, but decreases slightly at night due to the respiration of animals and plants. Carbon dioxide gradually accumulates at night due to lack of photosynthesis and the concentration increases. During the day, as plant photosynthesis proceeds, carbon dioxide is consumed and the concentration decreases. If the default electronic control mode is still used after the air component concentration changes, it is easy to cause unnecessary module power consumption and affect the energy saving effect.

[0004] Therefore, an energy-saving electric control system for air separation equipment is designed to overcome the above problems. Summary of the invention

[0005] The purpose of the present invention is to overcome the deficiencies of the prior art and to provide an energy-saving electric control system for an air separation plant which is simple and reasonable in structure and practical, safe and reliable.

[0006] The present invention is implemented by the following technical scheme: an energy-saving electric control system for an air separation device, comprising: A data collection module is used to collect circuit operation data of the air separation equipment when it is operated in different processing modes, as well as operation data of the sensor collection equipment, and transmit the above data to the data storage module in a corresponding storage time period; The data processing module is used to process the historical data stored in the data storage module, and after making a deep learning model based on the historical data, formulate the circuit operation mode for the corresponding time period; The wireless transmission module is responsible for sending the data from the data collection module to the remote monitoring module; The abnormal alarm module is responsible for receiving the abnormal data collected by the data collection module and triggering the alarm module to remind the user of the occurrence of abnormal situations; Data storage module, storing various operation data of the equipment; The scheduling module is used to schedule the operating status of different modules in the air separation equipment under different modes, and is also used to respond to the abnormal alarm module and shut down the operation of abnormal circuits.

[0007] Preferably, the storage time period of the data storage module includes small time segments of 8-12, 12-18, 18-22, and 22-8.

[0008] Preferably, the data collection module uses the local four seasons as intervals, utilizes a deep learning model to process historical circuit data, and formulates an operation mode.

[0009] Preferably, the data collection module collects data information collected by the sensor collection device when the meteorological conditions change according to the changes in meteorological conditions, and uses the data collection module to collect the current circuit operation data, puts it into the deep learning model for processing, and then formulates the air separation equipment operation mode according to the short-term meteorological changes.

[0010] As a preference, the abnormal alarm module can also send the circuit area where the abnormal data is collected to the scheduling module and the user terminal, so as to accurately determine the maintenance area for subsequent maintenance planning.

[0011] As a preference, the specific implementation of the data collection module includes: Step 1: Data preprocessing: standardize the circuit operation data, convert the circuit operation data into a format suitable for LSTM processing, that is, 3D format, and resample the time series data. The specific processing method is to use the storage time period as the data sampling interval based on one hour; Step 2: Normalize the sequence data, rescale the data from the original range so that all values ​​are within the range of 0 to 1; Step 3: Build a stacked LSTM model; Step 4: Model compilation; Step 5: Model training and validation: Use the preprocessed data to train the defined LSTM model, set the corresponding loss function and optimizer, and determine the batch size and number of iterations. Step 6: Model evaluation and optimization: Evaluate the performance of the LSTM model on an independent test set and obtain the evaluation results. Step 7: Deploy the application. The trained and optimized LSTM model can be saved and deployed in actual applications for real-time or batch circuit data analysis.

[0012] Preferably, a Dropout layer may be introduced in step 5; the initial selection of the Dropout rate is set at 0.6.

[0013] Preferably, the evaluation result of step six is ​​used to further adjust and optimize the model, specifically, repeatedly adjusting the parameters of steps three and four.

[0014] The beneficial effects of the present invention are as follows: 1. The present invention uses the circuit operation data of different storage time periods collected by the data collection module and the operation data of the sensor collection device, and uses these data to make a deep learning model, and then formulates the circuit operation mode of the corresponding time period, so as to determine the main operating equipment in the corresponding time section, so as to process the concentration of air components in different time periods, avoid long-term high-efficiency operation to process the air components with low content in the current small time section, and increase the energy saving effect; 2. The present invention utilizes a deep learning model to continuously learn and continuously optimize the current circuit operation mode. It can also be used for real-time or batch circuit data analysis, such as fault diagnosis and performance degradation monitoring, to ensure that the circuit is repaired before a fault occurs. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a schematic diagram of the overall structure of the present invention. DETAILED DESCRIPTION

[0016] In order to enable those skilled in the art to more clearly understand the objectives, technical solutions and advantages of the present invention, the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0017] In the description of the present invention, it is necessary to understand that the orientation or position relationship indicated by the terms such as "up", "down", "left", "right", "inside", "outside", "horizontal" and "vertical" is based on the orientation or position relationship shown in the accompanying drawings, and is only for the convenience of describing the present invention, rather than indicating or implying that the device or original referred to must have a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0018] The present invention will be described in detail below with reference to the accompanying drawings: Figure 1 As shown, an energy-saving electric control system for an air separation device comprises: A data collection module is used to collect circuit operation data of the air separation equipment when it is operated in different processing modes, as well as operation data of the sensor collection equipment, and transmit the above data to the data storage module in a corresponding storage time period; The data processing module is used to process the historical data stored in the data storage module, and after making a deep learning model based on the historical data, formulate the circuit operation mode for the corresponding time period; The wireless transmission module is responsible for sending the data from the data collection module to the remote monitoring module; The abnormal alarm module is responsible for receiving the abnormal data collected by the data collection module and triggering the alarm module to remind the user of the occurrence of abnormal situations; Data storage module, storing various operation data of the equipment; The scheduling module is used to schedule the operating status of different modules in the air separation equipment under different modes, and is also used to respond to the abnormal alarm module and shut down the operation of abnormal circuits.

[0019] The storage time period of the data storage module includes small time segments of 8-12, 12-18, 18-22, and 22-8.

[0020] The data collection module uses the local four seasons as intervals, utilizes a deep learning model to process historical circuit data, and formulates an operating mode.

[0021] The data collection module collects data information collected by the sensor collection device when the meteorological conditions change according to the changes in meteorological conditions, and uses the data collection module to collect the current circuit operation data, puts it into the deep learning model for processing, and then formulates the air separation equipment operation mode according to the short-term meteorological changes.

[0022] The abnormal alarm module can also send the circuit area where the abnormal data is collected to the scheduling module and the user terminal for accurate determination of the maintenance area in the subsequent maintenance plan.

[0023] The specific implementation of the data collection module includes: Step 1: Data preprocessing: standardize the circuit operation data, convert the circuit operation data into a format suitable for LSTM processing, that is, 3D format, and resample the time series data. The specific processing method is to use the storage time period as the data sampling interval based on one hour; Step 2: Normalize the sequence data, rescale the data from the original range so that all values ​​are within the range of 0 to 1; Step 3: Build a stacked LSTM model; Step 4: Model compilation; Step 5: Model training and validation: Use the preprocessed data to train the defined LSTM model, set the corresponding loss function and optimizer, and determine the batch size and number of iterations. Step 6: Model evaluation and optimization: Evaluate the performance of the LSTM model on an independent test set and obtain the evaluation results. Step 7: Deploy the application. The trained and optimized LSTM model can be saved and deployed in actual applications for real-time or batch circuit data analysis.

[0024] A Dropout layer may also be introduced in step 5; the initial selection of the Dropout rate is set at 0.6.

[0025] The evaluation result of step six is ​​used to further adjust and optimize the model, specifically by repeatedly adjusting the parameters of steps three and four.

[0026] The specific embodiments described herein are merely illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical ideas disclosed by the present invention shall still be covered by the claims of the present invention.

[0027] Example 1 like Figure 1 As shown, an energy-saving electric control system for an air separation device in an embodiment of the first aspect of the present invention comprises: The data collection module is used to collect circuit operation data such as voltage, current, active power, reactive power, power factor, etc. of the air separation equipment when it is running in different processing modes, as well as operation data of the sensor collection equipment, such as temperature, pressure, flow, etc., and transmit the above data to the data storage module in the corresponding storage time period; Taking one day as a large interval, the small time segments of 8-12, 12-18, 18-22, and 22-8 are divided as the storage time periods of the above data. The above storage time periods are divided based on the fact that the concentration distribution of air components will change due to factors such as temperature, human activities and natural phenomena during the day, so they are divided based on the time range of human daily activities; The data processing module is used to process the historical data stored in the data storage module, and after making a deep learning model based on the historical data, formulate the circuit operation mode for the corresponding time period; Specific implementation methods include: Step 1: Data preprocessing: standardize the circuit operation data to facilitate better model learning and feature extraction. The key to this step is to convert the circuit operation data into a format suitable for LSTM processing, namely 3D format (samples, time steps, features). When processing sequential circuit data, the time series data needs to be resampled. The specific processing method is to use the storage time period as the data sampling interval based on one hour to ensure that all sequences have the same length. Step 2: Normalize the sequence data. Normalization is to rescale the data from the original range so that all values ​​are within the range of 0 to 1. Step 3: Build a stacked LSTM model, including the number of layers, the number of neurons in each layer, and the activation function; Step 4: Model compilation, including definition of loss function, optimizer and evaluation index. The loss function guides the model on how to optimize its weights to minimize the prediction error; the optimizer determines how the weights are updated; the evaluation index is used to measure the performance of the model. In sequence generation or classification tasks, cross entropy loss is used as the loss function; Step 5: Model training and verification. Use the preprocessed data to train the defined LSTM model. In this step, you also need to set the corresponding loss function and optimizer, and determine the batch size and number of iterations. In order to prevent overfitting, you also need to introduce the Dropout layer. The Dropout layer can reduce the complex co-adaptation relationship between neurons by randomly shutting down a certain proportion of neurons (i.e. setting their output to 0) in each training batch, thereby alleviating the overfitting phenomenon. In this embodiment, the initial selection of the Dropout rate is set at 0.6 to process the LSTM model with a large number of circuit data parameters; Step 6: Model evaluation and optimization, evaluate the performance of the LSTM model on an independent test set, focusing on key performance indicators such as prediction accuracy and error range. Based on the evaluation results, further adjust and optimize the model, and specifically, repeatedly adjust the parameters of steps 3 and 4 to achieve the best performance; Step 7: Deploy the application. The trained and optimized LSTM model can be saved and deployed in actual applications for real-time or batch circuit data analysis, such as fault diagnosis, performance degradation monitoring, and operation mode changes in different time periods. Specifically, based on historical data, the different working states of the equipment in the small time segments of 8-12, 12-18, 18-22, and 22-8 are determined, so as to determine the main operating equipment in the corresponding small time segments to process the air component concentrations in different time periods, and avoid long-term high-efficiency operation to process air components with low content in the current small time segment.

[0028] The wireless transmission module is responsible for sending the data from the data collection module to the remote monitoring module; The abnormal alarm module is responsible for receiving the abnormal data collected by the data collection module and triggering the alarm module to remind the user of the occurrence of abnormal conditions. It also sends the circuit area where the abnormal data is collected to the scheduling module and the user terminal, so that the subsequent maintenance plan can accurately determine the maintenance area. Data storage module, storing various operation data of the equipment; The scheduling module is used to schedule the operating status of different modules in the air separation equipment under different modes, and is also used to respond to the abnormal alarm module and shut down the operation of abnormal circuits.

[0029] Example 2 like Figure 1 As shown, an energy-saving electrical control system for an air separation device includes all the contents of Example 1. In addition, based on regional and seasonal changes, a whole year is used as an interval, and the local four seasons are used as intervals. A deep learning model is used to process historical circuit data and formulate a corresponding operation mode.

[0030] The technical effect achieved by the above embodiment is: according to the seasons in different regions, the air separation equipment processes the corresponding high-concentration air components, reduces the processing efficiency of low-concentration air components, and thus reduces power consumption.

[0031] Example 3 like Figure 1 As shown, an energy-saving electric control system for an air separation device includes all the contents of Example 2. In addition, the data collection module can also collect data information collected by the sensor collection device when the meteorological conditions change according to the changes in meteorological conditions, and use the data collection module to collect the current circuit operation data, put it into the deep learning model for processing, and then formulate the future air separation device operation mode according to the short-term meteorological changes.

Claims

1. An energy-saving electric control system for an air separation plant, characterized in that: include: A data collection module is used to collect circuit operation data of the air separation equipment when it is operated in different processing modes, as well as operation data of the sensor collection equipment, and transmit the above data to the data storage module in a corresponding storage time period; The data processing module is used to process the historical data stored in the data storage module, and after making a deep learning model based on the historical data, formulate the circuit operation mode for the corresponding time period; The wireless transmission module is responsible for sending the data from the data collection module to the remote monitoring module; The abnormal alarm module is responsible for receiving the abnormal data collected by the data collection module and triggering the alarm module to remind the user of the occurrence of abnormal situations; Data storage module, storing various operation data of the equipment; The scheduling module is used to schedule the operating status of different modules in the air separation equipment under different modes, and is also used to respond to the abnormal alarm module and shut down the operation of abnormal circuits.

2. The energy-saving electric control system for air separation equipment according to claim 1 is characterized in that: The storage time period of the data storage module includes small time segments of 8-12, 12-18, 18-22, and 22-8.

3. The energy-saving electric control system for air separation equipment according to claim 1 is characterized in that: The data collection module uses the local four seasons as intervals, utilizes a deep learning model to process historical circuit data, and formulates an operating mode.

4. The energy-saving electric control system for air separation equipment according to claim 1 is characterized in that: The data collection module collects data information collected by the sensor collection device when the meteorological conditions change according to the changes in meteorological conditions, and uses the data collection module to collect the current circuit operation data, puts it into the deep learning model for processing, and then formulates the air separation equipment operation mode according to the short-term meteorological changes.

5. The energy-saving electric control system for air separation equipment according to claim 1 is characterized in that: The abnormal alarm module can also send the circuit area where the abnormal data is collected to the scheduling module and the user terminal for accurate determination of the maintenance area in the subsequent maintenance plan.

6. The energy-saving electric control system for air separation equipment according to claim 1 is characterized in that: The specific implementation of the data collection module includes: Step 1: Data preprocessing: standardize the circuit operation data, convert the circuit operation data into a format suitable for LSTM processing, that is, 3D format, and resample the time series data. The specific processing method is to use the storage time period as the data sampling interval based on one hour; Step 2: Normalize the sequence data, rescale the data from the original range so that all values ​​are within the range of 0 to 1; Step 3: Build a stacked LSTM model; Step 4: Model compilation; Step 5: Model training and validation: Use the preprocessed data to train the defined LSTM model, set the corresponding loss function and optimizer, and determine the batch size and number of iterations. Step 6: Model evaluation and optimization: Evaluate the performance of the LSTM model on an independent test set and obtain the evaluation results. Step 7: Deploy the application. The trained and optimized LSTM model can be saved and deployed in actual applications for real-time or batch circuit data analysis.

7. The energy-saving electric control system for air separation equipment according to claim 6 is characterized in that: A Dropout layer may also be introduced in step 5; the initial selection of the Dropout rate is set at 0.

6.

8. The energy-saving electric control system for air separation equipment according to claim 6 is characterized in that: The evaluation result of step six is ​​used to further adjust and optimize the model, specifically by repeatedly adjusting the parameters of steps three and four.

Images