Industrial internet-oriented multi-source data fusion ventilation energy consumption prediction method and system

Abstract

The invention relates to the technical field of ventilation system control, in particular to an industrial internet-oriented multi-source data fusion ventilation energy consumption prediction method and system. According to the method, the state of the ventilation system is accurately identified, key parameters are monitored in real time through various high-precision sensors, and the state of the ventilation system can be dynamically identified and predicted in real time by combining a fluid mechanics principle, a mathematical modeling method, an intelligent algorithm and a model established through machine learning, so that an accurate basis is provided for subsequent control; the system can timely and accurately know the self operation condition, conditions are created for measures such as energy conservation and consumption reduction, and the overall performance and reliability of the system are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of ventilation system control, and particularly relates to a multi-source data fusion ventilation energy consumption prediction method and system for an industrial internet. BACKGROUND

[0002] Industrial ventilation refers to a comprehensive technical measure for controlling, removing or diluting harmful substances, residual heat and residual humidity in a workshop, adjusting air flow and temperature and humidity, and ensuring personnel health, improving working environment and meeting production process requirements by reasonably designing and using ventilation equipment, systems and technical means in the industrial production process.

[0003] The existing ventilation system has some deficiencies in state recognition and automatic control. For example, the traditional ventilation system state recognition method can only monitor a few key parameters such as wind speed, wind pressure, etc., and the precision is not high enough to achieve comprehensive and accurate recognition of the multi-dimensional state of the ventilation system. At the same time, the automatic adjustment and control function is relatively simple and cannot accurately and flexibly control according to different working conditions and environmental requirements, resulting in poor ventilation effect, energy waste and other problems.

[0004] Therefore, a technology capable of realizing multi-dimensional state recognition and accurate automatic control of the ventilation system is urgently needed to improve the performance and energy utilization efficiency of the ventilation system. SUMMARY

[0005] To achieve the above-mentioned purpose, the present application proposes a multi-source data fusion ventilation energy consumption prediction method and system for an industrial internet, including the following steps:

[0006] Step one: sensor installation, wind speed sensors are installed at different cross-section straight pipe sections of the ventilation duct, wind pressure sensors are installed at the flanges of the fan inlet and outlet, temperature and humidity sensors and air quality sensors are arranged in the ventilation area, and a sensor network is established;

[0007] Step two: data acquisition, set different data acquisition periods, perform high-frequency acquisition on wind speed and wind pressure data, perform low-frequency acquisition on temperature and humidity and air quality data, and use Kalman filtering and wavelet transform for signal noise reduction;

[0008] Step three: data processing, establish a temperature threshold abnormality filtering mechanism through a data cleaning module, and perform multi-source data normalization processing;

[0009] Step four: data fusion, construct a wind speed data fusion model based on Kalman filtering, establish a wind speed state space equation and perform prediction update recursive calculation to generate a fused wind speed data set;

[0010] Step five: energy consumption prediction, build LSTM neural network model to fuse data as input features, use BGD / SGD optimizer to train energy consumption prediction model, output future period total energy consumption and unit time energy consumption prediction value;

[0011] Step six: actuator control, generate frequency control curve of frequency converter and opening degree adjustment instruction of air valve based on energy consumption prediction value, dynamically adjust actuator parameters through PID algorithm;

[0012] Step seven: feedback and optimization, establish safe operation constraint rules of actuator, set speed and opening degree threshold boundary, real-time feedback running state deviation value and trigger self-correction mechanism.

[0013] In one example, in step two, for wind speed sensor and wind pressure sensor, the collection period is set to 1 to 10 times per second, for temperature sensor, humidity sensor and air quality sensor, the collection period is set to 1 to 5 times per minute.

[0014] In one example, in step four, for wind speed data, Kalman filtering algorithm is used for fusion, wind speed sensor data from different positions is taken as input observation value of Kalman filtering algorithm, state space model of wind speed is established, and fused wind speed data is obtained through recursive calculation.

[0015] In one example, in step five, energy consumption prediction includes prediction model selection and training, prediction data, energy consumption prediction calculation and prediction result storage and display.

[0016] In one example, the multi-source data fusion ventilation energy consumption prediction system for industrial internet includes perception layer, data layer, data fusion layer, energy consumption prediction layer, actuator control layer and feedback and optimization layer.

[0017] In one example, the perception layer includes sensor module and data collector module, the sensor module includes wind speed sensor submodule, wind pressure sensor submodule, temperature and humidity sensor submodule and air quality sensor submodule.

[0018] In one example, the data layer includes data storage submodule and data preprocessing submodule, the data storage submodule stores preprocessed data in the local storage medium of the data collector, the data preprocessing submodule includes data cleaning unit and data normalization unit, the data cleaning unit cleans the collected data, and the data normalization unit normalizes the cleaned data.

[0019] In one example, the data fusion layer includes data fusion module, and the data fusion module uses Kalman filtering algorithm to fuse wind speed data.

[0020] In one example, the energy consumption prediction layer comprises an energy consumption prediction module, which comprises a prediction model training unit, a prediction data reading unit, an energy consumption prediction calculation unit and a prediction result storage and display unit.

[0021] In one example, the execution mechanism control layer comprises a control instruction generation module, a control instruction output module, an execution mechanism module and a safety and fault processing module.

[0022] The industrial internet-oriented multi-source data fusion ventilation energy consumption prediction method and system can bring the following beneficial effects:

[0023] 1. The present application realizes accurate identification of the state of the ventilation system, real-time monitoring of key parameters by various high-precision sensors, and the establishment of a model combining fluid mechanics principles, mathematical modeling methods and intelligent algorithms and machine learning to dynamically identify and predict the state of the ventilation system in real time, providing accurate basis for subsequent control, enabling the system to accurately understand its own operating conditions in time, creating conditions for energy saving and other measures, and improving the overall performance and reliability of the system.

[0024] 2. The present application improves the operating efficiency and energy saving effect of the ventilation system, formulates and executes automatic adjustment control strategies according to the ventilation system state identification results, accurately controls the operating state of the ventilation equipment by the intelligent control system, realizes accurate adjustment and control, optimizes the operation of the ventilation system, avoids energy waste, improves energy utilization efficiency, reduces energy consumption and saves operating costs under the premise of meeting the ventilation requirements, and the control process has high automation degree, reduces manual intervention and improves operating stability. BRIEF DESCRIPTION OF DRAWINGS

[0025] The drawings described herein are used to provide a further understanding of the present application, and form a part of the present application. The illustrative embodiments of the present application and their descriptions serve to explain the present application, and do not constitute an improper limitation on the present application. In the drawings:

[0026] Fig. 1 The flowchart of the industrial internet-oriented multi-source data fusion ventilation energy consumption prediction method;

[0027] Fig. 2 The schematic diagram of the industrial internet-oriented multi-source data fusion ventilation energy consumption prediction system architecture; DETAILED DESCRIPTION

[0028] In order to more clearly explain the overall concept of the present application, the following will be described in detail in an exemplary manner with reference to the drawings.

[0029] In the description of the present application, it needs to be understood that the terms "center", "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "axial", "radial", "circumferential" and the like indicate the orientation or positional relationship shown in the drawings, which are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation of the present application.

[0030] In addition, the terms "first", "second" are only for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "multiple" is two or more, unless otherwise explicitly specified and limited.

[0031] In the present application, unless otherwise explicitly specified and limited, the terms "mounting", "connection", "connection", "fixing" and the like should be understood broadly, for example, it can be fixed connection, or detachable connection, or integral; it can be mechanical connection, or electrical connection, or communication; it can be direct connection, or indirect connection through intermediate medium, or internal communication of two elements or interaction relationship between two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0032] In the present application, unless otherwise explicitly specified and limited, the first feature is "on" or "under" the second feature, which can be direct contact between the first and second features, or indirect contact between the first and second features through intermediate medium. In the description of the present application, the description of the terms "one scheme", "some schemes", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the scheme or example are included in at least one scheme or example of the present application. In the present application, the illustrative description of the above terms does not necessarily refer to the same scheme or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more schemes or examples in a suitable manner.

[0033] As Figs. 1-2 shown, the present application proposes a multi-source data fusion ventilation energy consumption prediction method and system for industrial internet, including the following steps:

[0034] Step 1: Sensor installation, install wind speed sensors at different cross-sections of the ventilation duct. The specific location should be selected to be a straight pipe section with stable airflow, away from disturbance components such as bends and valves, to accurately monitor the wind speed changes at different positions.

[0035] Install wind pressure sensors at the inlet and outlet of the fan, close to the flange connection of the fan inlet and outlet, to accurately monitor the size and fluctuation of the wind pressure.

[0036] Install temperature and humidity sensors at different positions in the ventilation area, select representative positions such as areas with frequent personnel activities and areas with high heat generation of equipment, to monitor the temperature and humidity conditions in real time.

[0037] Install air quality sensors near the air inlet and outlet, at a certain distance from the air outlet to avoid the influence of airflow impact on the measurement accuracy of the sensor, for monitoring air quality parameters such as carbon dioxide concentration and dust concentration.

[0038] Connect the output of the sensor to the input of the data collector, use shielded cable for connection to reduce the influence of electromagnetic interference on signal transmission, and ensure the stability and reliability of data transmission.

[0039] Step 2: Data collection, set a reasonable data collection period according to the operating characteristics of the ventilation system and the energy consumption prediction requirements.

[0040] For wind speed sensors and wind pressure sensors, since the monitored parameter changes relatively quickly, the collection period is set to 1-10 times per second; for temperature sensors, humidity sensors and air quality sensors, the collection period can be set to 1-5 times per minute.

[0041] Turn on the power of the data collector, start the data collection program, and the data collector scans the output signals of each sensor in turn according to the preset collection period, and converts the analog signals to digital signals.

[0042] The data collector performs preliminary processing on the collected digital signals, including filtering and denoising, signal amplification, etc., using digital filtering algorithms such as Kalman filtering and wavelet transform filtering to remove high-frequency noise and random interference signals in the data, improving the accuracy of the data.

[0043] Step 3: Data preprocessing, including data cleaning and data normalization, store the preprocessed data in the local storage medium of the data collector, such as SD card, hard disk, etc., at the same time, store according to certain file naming rules and data format, convenient for subsequent data reading and analysis.

[0044] The data cleaning cleans the collected data, removes obvious error data and outliers, and marks and removes data beyond the threshold value by setting a reasonable data range threshold. For example, the normal range of temperature data is generally between -20℃ and 50℃. If the collected data exceeds this range, it is determined as an outlier and is removed.

[0045] The data normalization converts data of different dimensions and orders of magnitude to the same scale range, such as [0, 1] or [-1, 1]. The normalization formula is:

[0046] X_n=(X-X_min) / (X_max-X_min)

[0047] Where X_n is the normalized data, X is the original data, X_max and X_min are the maximum and minimum values of the original data, respectively.

[0048] Step four: data fusion. For wind speed data, Kalman filter algorithm is used for fusion. The data from wind speed sensors at different positions is used as the input observation value of Kalman filter algorithm, and the state space model of wind speed is established. The fused wind speed data is obtained by recursive calculation.

[0049] According to the state equation and observation equation of Kalman filter algorithm, the prediction step and update step are calculated. The prediction step predicts the state value at the current time according to the state estimation value at the last time. The update step corrects the predicted value according to the observation value at the current time, and obtains the optimal state estimation value at the current time, i.e. the fused wind speed data.

[0050] The fused wind speed data and other sensor monitoring data are stored in the data fusion result file, which provides data basis for subsequent energy consumption prediction.

[0051] Step five: energy consumption prediction, including the following steps:

[0052] S5.1 prediction model selection and training: select long short-term memory network (LSTM) model as the energy consumption prediction model. Collect a large amount of historical data, including fused multi-dimensional state data of ventilation system and historical energy consumption data, and train the LSTM model. Batch gradient descent method (BGD) or stochastic gradient descent method (SGD) is used to train the model, and the weight and bias parameters of the model are adjusted by minimizing the loss function (such as mean square error loss function).

[0053] S5.2 prediction data: read the latest fusion data from the data fusion result file, including wind speed, wind pressure, temperature, humidity, air quality and other parameters, as the input data of the energy consumption prediction model.

[0054] S5.3 Energy consumption prediction calculation: input data into the trained LSTM model, the model according to the input data characteristics, using time series analysis and machine learning algorithm to predict the future energy consumption of the ventilation system, output the predicted energy consumption value, including the total energy consumption of the ventilation system, unit time energy consumption and other indicators.

[0055] S5.4 Prediction results storage and display: store the prediction results in the energy consumption prediction results file, and display the prediction results in the form of charts through data visualization tools, which is convenient for users to view and analyze.

[0056] Step six: actuator control, including the following processes:

[0057] S6.1 Control instruction generation:

[0058] According to the energy consumption prediction results, combined with the current production demand and environmental conditions, formulate the corresponding energy-saving control strategy. For example, if the prediction result shows that the energy consumption will rise in the future period, and the air quality allows, the strategy of reducing the fan speed can be planned.

[0059] According to the control strategy, calculate the specific target parameters of the actuator, such as the target frequency of the fan frequency converter, the target opening of the air valve actuator, etc. This usually relies on pre-set control algorithms, such as PID (Proportional-Integral-Derivative) control algorithm, which calculates the deviation between current parameters and target parameters.

[0060] S6.2 Control instruction output:

[0061] Convert the formulated control strategy into signals that can be recognized by the actuator. For example, convert the digital signals of the control system into analog signals (such as 4-20mA current signals) or pulse signals required by the fan frequency converter.

[0062] Through wired (such as RS485 bus) or wireless communication (such as Wi-Fi, ZigBee), send the control signal to the actuator. Ensure the stability of signal transmission, avoid receiving incorrect instructions by the actuator due to line interference or signal attenuation.

[0063] S6.3 Actuator response:

[0064] After receiving the control signal, the fan frequency converter adjusts the output frequency to change the speed of the fan motor. For example, if the control signal requires to reduce the speed, the frequency converter reduces the output frequency, the fan speed decreases, the ventilation volume decreases, and the energy-saving purpose is achieved.

[0065] The damper actuator adjusts the valve opening based on the received signals. If the control signal requires an increase in opening, the actuator drives the valve to open, increasing the airflow passage area of the ventilation duct and increasing the ventilation volume; conversely, it reduces the opening and reduces the ventilation volume.

[0066] S6.4 Operation state monitoring:

[0067] During the operation of the actuator, its operating state is continuously monitored, such as the actual speed of the fan and the actual opening of the damper. This can be achieved through feedback sensors (such as encoders, position sensors) installed on the actuator.

[0068] The actual state monitored is compared and evaluated with the control target. If the deviation exceeds the allowed range, it is fed back to the control system in time to adjust the control strategy and instructions to ensure that the actuator accurately performs the control task.

[0069] S6.5 Safety and fault handling:

[0070] The safety operation limits of the actuator are set, such as the upper and lower limits of the fan speed and the extreme values of the damper opening. Once the limits are approached or exceeded, the actuator automatically takes protective measures to prevent equipment damage.

[0071] Real-time monitoring of the operating parameters of the actuator, if an abnormality occurs (such as motor overheating, valve sticking), quickly diagnose the fault and trigger the alarm mechanism to notify the maintenance personnel for repair, to ensure the safe and stable operation of the system.

[0072] Step seven: feedback and optimization process, during the operation of the system, real-time monitoring of the effect and energy consumption of the ventilation system. Through sensors and calculation models, collect the actual operation data of the ventilation system, such as actual energy consumption, ventilation volume, air quality improvement degree, etc. and compare and analyze it with the energy consumption prediction results.

[0073] According to the feedback data, optimize the energy consumption prediction model. Use model optimization methods in machine learning, such as adjusting the hyperparameters of the model, increasing the amount of training data, improving the model structure, etc. to improve the prediction accuracy and generalization ability of the model for energy consumption of the ventilation system.

[0074] According to the feedback data and actual operation effect, optimize each link in the data processing process. For example, adjust the data collection period, improve the data filtering algorithm, optimize the data fusion algorithm, etc. to improve the efficiency and quality of data processing, and provide more accurate data support for energy consumption prediction.

[0075] The system architecture for implementing the above method includes a perception layer, a data layer, a data fusion layer, an energy consumption prediction layer, an actuator control layer, and a feedback and optimization layer.

[0076] The perception layer comprises a sensor module and a data collector module, the sensor module comprises a wind speed sensor sub-module, a wind pressure sensor sub-module, a temperature and humidity sensor sub-module and an air quality sensor sub-module, the data collector module is connected with the output ends of various sensors through a shielded cable, converts analog signals output by the sensors into digital signals, scans output signals of various sensors in turn according to a preset collection period, and performs preliminary processing on the collected digital signals, including filtering and denoising, signal amplification and the like.

[0077] The wind speed sensor sub-module is installed at a straight pipe section at different sections of a ventilation duct, and is used for monitoring wind speed changes at various positions in real time.

[0078] The wind pressure sensor sub-module is installed at a flange connection at an inlet and an outlet of a fan, and is used for accurately monitoring wind pressure and fluctuations.

[0079] The temperature and humidity sensor sub-module is arranged at representative positions in ventilation areas such as frequently-visited areas by personnel and areas with large heat generation of equipment, and is used for obtaining environmental temperature and humidity data in real time.

[0080] The air quality sensor sub-module is installed at appropriate positions (a certain distance away from the air inlet and the air outlet to avoid air flow impact) near the air inlet and the air outlet, and is used for monitoring air quality parameters such as carbon dioxide concentration and dust concentration.

[0081] The data layer comprises a data storage sub-module and a data preprocessing sub-module, the data storage sub-module stores preprocessed data in a local storage medium of the data collector, and stores the data according to certain file naming rules and data formats, so as to facilitate subsequent reading and analysis.

[0082] The data preprocessing sub-module comprises a data cleaning unit and a data normalization unit, the data cleaning unit cleans the collected data, sets a reasonable data range threshold (such as -20℃ to 50℃ for temperature), and eliminates obviously erroneous data and outliers beyond the threshold, and the data normalization unit performs normalization processing on the cleaned data, converts the data to the same scale range (such as [0, 1] or [-1, 1]), and calculates by using a normalization formula X_n=(X-X_min) / (X_max-X_min).

[0083] The data fusion layer comprises a data fusion module, the data fusion module fuses wind speed data by using a Kalman filtering algorithm, takes wind speed sensor data at different positions as input observation values, establishes a wind speed state space model, obtains fused wind speed data through recursive calculation, completes calculation of a prediction step and an update step according to a state equation and an observation equation of the Kalman filtering algorithm, and obtains an optimal state estimation value at a current time.

[0084] The fused wind speed data and other sensor monitoring data are stored in a data fusion result file to provide a data basis for subsequent energy consumption prediction.

[0085] The energy consumption prediction layer includes an energy consumption prediction module, which includes a prediction model training unit, a prediction data reading unit, an energy consumption prediction calculation unit, and a prediction result storage and display unit.

[0086] The prediction model training unit selects a long short-term memory (LSTM) model as the energy consumption prediction model. A large amount of historical data (including fused multi-dimensional state data of the ventilation system and historical energy consumption data) is collected to train the LSTM model. Batch gradient descent (BGD) or stochastic gradient descent (SGD) is used to adjust the weight and bias parameters of the model by minimizing the loss function (such as mean square error loss function).

[0087] The prediction data reading unit reads the latest fusion data from the data fusion result file, including wind speed, wind pressure, temperature, humidity, air quality, etc. as input data for the energy consumption prediction model.

[0088] The energy consumption prediction calculation unit sends the input data to the trained LSTM model. The model uses time series analysis and machine learning algorithms to predict the future energy consumption of the ventilation system, and outputs the predicted energy consumption values, such as total energy consumption, unit time energy consumption, etc.

[0089] The prediction result storage and display unit stores the prediction results in the energy consumption prediction result file, and uses data visualization tools to display the prediction results in the form of charts, making it easy for users to view and analyze.

[0090] The actuator control layer includes a control instruction generation module, a control instruction output module, an actuator module, and a safety and fault handling module.

[0091] The control instruction generation module includes an energy-saving control strategy development unit and a target parameter calculation unit. The energy-saving control strategy development unit develops appropriate energy-saving control strategies based on the energy consumption prediction results, combined with current production needs and environmental conditions. For example, in the case of rising predicted energy consumption and acceptable air quality, a strategy to reduce fan speed is planned.

[0092] The target parameter calculation unit calculates the specific target parameters of the actuators, such as the target frequency of the fan frequency converter and the target opening of the air valve actuator, based on the control strategy and the pre-set control algorithm.

[0093] The control instruction output module includes a signal conversion unit and a communication unit. The signal conversion unit converts the formulated control strategy into a signal recognizable by the actuator, such as converting a digital signal of the control system into an analog signal or a pulse signal required by the fan frequency converter.

[0094] The communication unit sends the control signal to the actuator through wired or wireless communication, ensuring the stability of signal transmission.

[0095] The actuator module includes a fan frequency converter unit and a fan valve actuator unit. After receiving the control signal, the fan frequency converter unit adjusts the output frequency to change the speed of the fan motor, thereby controlling the ventilation volume and achieving energy-saving purposes.

[0096] The fan valve actuator unit adjusts the valve opening according to the received signal to control the airflow passage area of the ventilation duct, thereby adjusting the ventilation volume.

[0097] The running state monitoring module includes a state monitoring unit and a comparison and evaluation unit. The state monitoring unit continuously monitors the running state of the actuator during operation through the feedback sensors installed on the actuator, such as the actual speed of the fan and the actual opening of the valve.

[0098] The comparison and evaluation unit compares the monitored actual state with the control target and evaluates the deviation. If the deviation exceeds the allowed range, it is fed back to the control system in time to adjust the control strategy and instruction.

[0099] The safety and fault handling module includes a safety operation limiting unit and a fault diagnosis and alarm unit. The safety operation limiting unit sets the safety operation limits of the actuator, such as the upper and lower limits of the fan speed and the extreme value of the valve opening. Once the limits are approached or exceeded, the actuator automatically takes protective measures to prevent equipment damage.

[0100] The fault diagnosis and alarm unit monitors the running parameters of the actuator in real time. If an abnormality occurs (such as motor overheating or valve jamming), it quickly diagnoses the fault and triggers the alarm mechanism to notify the maintenance personnel for repair.

[0101] The feedback and optimization layer includes a feedback data collection module, a model optimization module, and a data processing flow optimization module.

[0102] The feedback data collection module monitors the effect of the ventilation system and the energy consumption in real time. Through sensors and calculation models, it collects actual running data of the ventilation system, such as actual energy consumption, ventilation volume, and air quality improvement degree, and compares and analyzes them with the energy consumption prediction results.

[0103] The model optimization module optimizes the energy consumption prediction model based on the feedback data using model optimization methods in machine learning, improving the prediction accuracy and generalization ability of the model for energy consumption of the ventilation system.

[0104] The data processing flow optimization module optimizes each link in the data processing flow according to the feedback data and the actual operation effect, improves the efficiency and quality of data processing, and provides more accurate data support for energy consumption prediction.

[0105] Each of the embodiments in the specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the difference from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the part of the method embodiment.

[0106] The above only describes the embodiments of the present application and is not used to limit the present application. The present application can have various changes and modifications for those skilled in the art. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application shall be included in the scope of claims of the present application.

Claims

1. A multi-source data fusion method for predicting ventilation energy consumption for the Industrial Internet, characterized by: Includes the following steps: Step 1: Sensor installation. Install wind speed sensors on straight pipe sections with different cross-sections in the ventilation duct, install wind pressure sensors at the inlet and outlet flanges of the fan, and deploy temperature and humidity sensors and air quality sensors in the ventilation area to establish a sensor network. Step 2: Data Acquisition. Set differentiated data acquisition cycles, perform high-frequency acquisition of wind speed and wind pressure data, and low-frequency acquisition of temperature, humidity and air quality data. Use Kalman filtering and wavelet transform for signal noise reduction. Step 3: Data processing. A temperature threshold anomaly filtering mechanism is established through the data cleaning module, and multi-source data normalization processing is performed. Step 4: Data fusion. Construct a wind speed data fusion model based on Kalman filtering, establish the wind speed state space equation and perform prediction update recursive calculation to generate a fused wind speed dataset. Step 5: Energy consumption prediction. Build an LSTM neural network model, use fused data as input features, train the energy consumption prediction model using BGD / SGD optimizer, and output the predicted values ​​of total energy consumption and energy consumption per unit time for future periods. Step 6: Actuator control. Based on the energy consumption prediction value, generate the frequency control curve of the frequency converter and the valve opening adjustment command, and dynamically adjust the actuator parameters through the PID algorithm; Step 7: Feedback and optimization. Establish safety operation constraint rules for the actuator, set speed and opening threshold boundaries, provide real-time feedback on operating status deviations, and trigger a self-correction mechanism.

2. The multi-source data fusion ventilation energy consumption prediction method for the Industrial Internet according to claim 1, characterized in that: In step two, the acquisition cycle for wind speed and wind pressure sensors is set to 1 to 10 times per second, and the acquisition cycle for temperature, humidity and air quality sensors is set to 1 to 5 times per minute.

3. The multi-source data fusion ventilation energy consumption prediction method for the Industrial Internet according to claim 1, characterized in that: In step four, the wind speed data is fused using the Kalman filter algorithm. Wind speed sensor data from different locations are used as input observations for the Kalman filter algorithm to establish a state-space model of the wind speed. The fused wind speed data is then obtained through recursive calculation.

4. The multi-source data fusion ventilation energy consumption prediction method for the Industrial Internet according to claim 1, characterized in that: Step five, energy consumption prediction, includes prediction model selection and training, prediction data, energy consumption prediction calculation, and prediction result storage and display.

5. A multi-source data fusion ventilation energy consumption prediction system for industrial internet as described in claims 1 to 4, characterized in that: It includes a perception layer, a data layer, a data fusion layer, an energy consumption prediction layer, an actuator control layer, and a feedback and optimization layer.

6. The multi-source data fusion ventilation energy consumption prediction system for the Industrial Internet according to claim 5, characterized in that: The sensing layer includes a sensor module and a data acquisition module. The sensor module includes a wind speed sensor submodule, a wind pressure sensor submodule, a temperature and humidity sensor submodule, and an air quality sensor submodule.

7. The multi-source data fusion ventilation energy consumption prediction system for the Industrial Internet according to claim 5, characterized in that: The data layer includes a data storage submodule and a data preprocessing submodule. The data storage submodule stores the preprocessed data in the local storage medium of the data collector. The data preprocessing submodule includes a data cleaning unit and a data normalization unit. The data cleaning unit cleans the collected data, and the data normalization unit normalizes the cleaned data.

8. The multi-source data fusion ventilation energy consumption prediction system for the Industrial Internet according to claim 5, characterized in that: The data fusion layer includes a data fusion module, which uses a Kalman filter algorithm to fuse wind speed data.

9. The multi-source data fusion ventilation energy consumption prediction system for the Industrial Internet according to claim 5, characterized in that: The energy consumption prediction layer includes an energy consumption prediction module, which includes a prediction model training unit, a prediction data reading unit, an energy consumption prediction calculation unit, and a prediction result storage and display unit.

10. The multi-source data fusion ventilation energy consumption prediction system for the Industrial Internet according to claim 5, characterized in that: The actuator control layer includes a control command generation module, a control command output module, an actuator module, and a safety and fault handling module.

Images