Intelligent oil fume machine and oil fume separation method thereof

Abstract

The process parameters can be preset, the quality problem is prevented in advance, automatic production is achieved, and the production efficiency is improved. The invention is applicable to the technical field of range hood control, and provides an intelligent oil fume machine and an oil fume separation method thereof, and the method is applied to an intelligent range hood oil fume separation control system. The system comprises an information acquisition module, a model construction module, a model training module, a model evaluation module, a parameter determination module, an application control module, a wireless communication module, a memory, an alarm, a processing center and an intelligent mobile terminal. The information acquisition module, the model construction module, the model training module, the model evaluation module, the parameter determination module, the application control module, the wireless communication module, the memory and the alarm are respectively connected with the processing center; the invention further provides an intelligent oil fume machine. Operation data of the range hood can be collected, the preset control parameters are matched according to the model, and automatic precise control is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of range hood control, and specifically to an intelligent range hood and its oil fume separation method in the field of artificial intelligence. Background Art

[0002] Existing range hoods perform manual operation control for oil fume separation at high, medium, and low grades. Sometimes, inaccurate manual judgment leads to incorrect gear control, resulting in unqualified oil separation degree. A large amount of oil will enter the interior of the range hood and adhere to key components such as the motor and impeller, causing an increase in the operating load of the motor, an increase in power consumption, and a shortening of the service life of the range hood. In addition, the unseparated oil will be discharged along the flue and accumulate and condense in the flue, which may cause a fire hazard and pollute the exterior wall, affecting aesthetics and cleaning. Therefore, thorough oil fume separation is an urgent task in the current industry. Summary of the Invention

[0003] In order to solve the deficiencies in the prior art, the purpose of the present invention is to provide an intelligent range hood and its oil fume separation method. By setting up various functional modules such as an information acquisition module, a model construction module, a model training module, a model evaluation module, a parameter determination module, and an application control module, the operating data of the range hood can be collected for model training, and the preset control parameters can be matched according to the model to achieve automatic and precise control, thereby thoroughly controlling oil fume separation and avoiding oil fume accumulation from blocking the flue.

[0004] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0005] An oil fume separation method for an intelligent range hood is applied to an oil fume separation control system of an intelligent range hood. The system includes an information acquisition module, a model construction module, a model training module, a model evaluation module, a parameter determination module, an application control module, a wireless communication module, a memory, an alarm, a processing center, and an intelligent mobile terminal; the information acquisition module, the model construction module, the model training module, the model evaluation module, the parameter determination module, the application control module, the wireless communication module, the memory, and the alarm are respectively connected to the processing center; the intelligent mobile terminal includes a smart phone, a tablet computer, and a smart remote control, and is wirelessly connected to the wireless communication module within the range of a wireless network or the Internet.

[0006] An oil fume separation method for an intelligent range hood provided by the present invention includes the following steps:

[0007] S10. Before oil fume separation, the information acquisition module acquires historical image data of normal and abnormal oil fume separation of similar range hoods and performs preprocessing for subsequent model construction, and transmits it to the model construction module;

[0008] S20. The model construction module constructs a model by determining the model architecture, designing the network hierarchical structure, adding auxiliary layers, and defining model parameters according to the characteristic information of oil fume separation for subsequent model training and verification, and then transfers it to the model training module;

[0009] S30. The model training module trains the model based on the collected data, adjusts the model parameters to optimize the model performance to obtain predetermined oil fume separation control parameters, ensures that the oil fume separation of the range hood can be automatically designed, and then transfers it to the model evaluation module;

[0010] S40. The model evaluation module conducts model verification and evaluation based on the trained model, adjusts and optimizes the oil fume separation model according to the verification results to improve the performance and accuracy of the model, and then transfers it to the parameter determination module;

[0011] S50. The parameter determination module conducts simulation scenario tests based on the oil fume separation control parameters predicted by the trained model to obtain the optimal oil fume separation control parameters to ensure the subsequent oil fume separation effect, and then transfers it to the application control module;

[0012] S60. The application control module controls the range hood to perform oil fume separation operation according to the oil fume separation control parameters, so that the actual oil fume separation effect reaches the highest national standard to ensure the quality effect of oil fume separation, and then transfers it to the processing center;

[0013] S70. The processing center compares the quality information of oil fume separation with the oil fume separation quality standard stored in the memory of this range hood: if it meets the standard, it notifies the customer that they can continue to use; if it does not meet the standard, it transfers it to the alarm and notifies for debugging or maintenance.

[0014] The system of the present invention further includes a computer-readable storage medium containing a memory; the memory stores a computer program, and when each functional module executes the computer program, it realizes the steps of an intelligent range hood oil fume separation method described in any one of the above; the computer program stored on the computer-readable storage medium, when executed by each functional module, realizes the steps of an intelligent range hood oil fume separation method described above; it further includes an intelligent range hood oil fume separation control device implemented by using an intelligent range hood oil fume separation method described above.

[0015] The present invention also provides an intelligent range hood implemented by using an intelligent range hood oil fume separation method described above.

[0016] The beneficial effects of the present invention compared with the prior art:

[0017] By setting up various functional modules, collecting various fume data, and constructing an AI model for fume separation, combined with AI algorithms to optimize and predict stir-frying scenarios and automatically switch to high-speed modes, etc., to achieve fume concentration prediction and dynamic adjustment of purification strategies. Through Internet of Things control, it collaborates with devices such as gas stoves and fresh air systems to build a closed-loop management of kitchen air quality. By using upgraded filter materials such as nanofiber filters and oil-repellent coatings to improve filtration efficiency and extend service life, and optimizing the electrode plate structure to improve electrostatic removal technology, reducing energy consumption and reducing the generation of ozone by-products. Implementing a composite purification system technology of electrostatic adsorption + photocatalytic oxidation + activated carbon filtration, taking into account particulate matter purification and odor elimination, and adopting a low-energy consumption design to reduce the operating energy consumption of equipment through variable-frequency motors and high-efficiency impellers. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for use in the embodiments or exemplary technical descriptions. The following drawings are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0019] Figure 1 Schematic diagram of the system module of the present invention;

[0020] Figure 2 Schematic diagram of the information acquisition module of the present invention;

[0021] Figure 3 Schematic diagram of the model construction module of the present invention;

[0022] Figure 4 Schematic diagram of the model training module of the present invention;

[0023] Figure 5 Schematic diagram of the model evaluation module of the present invention;

[0024] Figure 6 Schematic diagram of the parameter determination module of the present invention;

[0025] Figure 7 Schematic diagram of the application control module of the present invention;

[0026] Figure 8 Schematic diagram of the method process control program of the present invention;

[0027] Figure 9 Schematic diagram of the program of step S10 in the method process of the present invention;

[0028] Figure 10 Schematic diagram of the program of step S20 in the method process of the present invention;

[0029] Figure 11It is a schematic diagram of the procedure of step S30 in the method flow of the present invention;

[0030] Figure 12 It is a schematic diagram of the procedure of step S40 in the method flow of the present invention;

[0031] Figure 13 It is a schematic diagram of the procedure of step S50 in the method flow of the present invention;

[0032] Figure 14 It is a schematic diagram of the procedure of step S60 in the method flow of the present invention. Detailed implementation manners

[0033] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0034] The following describes in detail the specific implementation of the present invention with reference to specific embodiments:

[0035] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application. It should be noted that when a module is referred to as being "disposed on" another module, it can be directly on the other module or indirectly on the other module. When a module is referred to as being "connected to" another module, it can be directly connected to the other module or indirectly connected to the other module.

[0036] In the description of the present application, the meaning of "a plurality" is two or more, unless otherwise clearly and specifically defined. The meaning of "several" is one or more, unless otherwise clearly and specifically defined. In the present application, "oil fume separation of the range hood" and "oil fume separation" both refer to the oil fume separation technology of the range hood under the same conditions; "data" all refers to the operation data of the oil fume separation of the range hood; "model" all refers to the convolutional neural network model (AI) of the oil fume separation of the range hood;

[0037] In the description of the present application, it should be noted that unless otherwise clearly specified and defined, the terms "connected" and "coupled" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two components or the interaction relationship between two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances. The terms "comprising", "including", "having" and their variants all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0038] Please refer to Figure 1 As shown, the present invention provides an oil fume separation method for an intelligent range hood, which is applied to an oil fume separation control system of an intelligent range hood. The system includes an information acquisition module, a model construction module, a model training module, a model evaluation module, a parameter determination module, an application control module, a wireless communication module, a memory, an alarm, a processing center, and an intelligent mobile terminal; the information acquisition module, the model construction module, the model training module, the model evaluation module, the parameter determination module, the application control module, the wireless communication module, the memory, and the alarm are respectively connected to the processing center; the intelligent mobile terminal includes a smart phone, a tablet computer, and an intelligent remote control, and is wirelessly connected to the wireless communication module within the range of a wireless network or the Internet.

[0039] The wireless communication module is provided with a wireless network unit, which is responsible for the transceiver of wireless signals, and automatically forms a network connection with the intelligent mobile terminal within an effective network range, and is connected to other smart home devices in the kitchen such as an intelligent range hood, a camera, an air purifier, a gas / smoke alarm, and an automatic window opener, and implements networked interactive control operation to ensure the normal operation of the devices in the kitchen; the wireless signals include various Internet of Things signals such as MQTT, CoAP, HTTP, REST API, Zagbee, LoRaWAN, NB-IoT, Bluetooth, or 5G, 4G, Wifi network signals.

[0040] The alarm compares the actual evaluation score of the model with the model evaluation score standard stored in the memory. If the standard is not met, it will automatically emit a sound alarm and notify to continue training; and compares the oil fume separation quality information with the oil fume separation quality standard of the range hood stored in the memory. If the standard is not met, it will automatically emit a sound alarm and notify to re-adjust; and compares the quality information of the oil fume separation with the oil fume separation quality standard of the range hood stored in the memory. If the standard is not met, it will automatically emit a sound alarm and notify to adjust or repair.

[0041] The memory is responsible for storing information of the information acquisition module, model construction module, model training module, model evaluation module, solution determination module, application control module, wireless communication module, and alarm, as well as storing the model evaluation sub-criteria and the oil fume separation quality standard of the range hood.

[0042] The processing center is responsible for information transfer of the information acquisition module, model construction module, model training module, model evaluation module, parameter determination module, application control module, wireless communication module, alarm, and memory. It is the hub center of the system and compares the actual model evaluation score with the model evaluation sub-criteria stored in the memory: if it meets the standard, it is the predetermined oil fume separation control parameter; if it does not meet the standard, it is passed to the alarm and notifies to continue training. It also compares the oil fume separation quality information with the oil fume separation quality standard of the range hood stored in the memory: if it meets the standard, it is set as the formal oil fume separation control parameter; if it does not meet the standard, it is passed to the alarm and notifies to re-debug. It further compares the oil fume separation quality information with the oil fume separation quality standard of the range hood stored in the memory: if it meets the standard, it notifies the customer that they can continue to use; if it does not meet the standard, it is passed to the alarm and notifies to debug or repair.

[0043] Please refer to Figure 9 As shown, the information acquisition module includes a data acquisition unit, a data cleaning unit, a data enhancement unit, an integration data unit, a conversion data unit, a filtering and denoising unit, a grayscale conversion unit, and a feature extraction unit. It is responsible for acquiring historical image data of normal and abnormal oil fume separation similar to that of a range hood, preprocessing it, and passing it to the model construction module.

[0044] Furthermore, the data acquisition unit obtains historical data of oil fume emissions, equipment operation parameters, separation efficiency, and environmental maintenance of similar range hoods through an industry database, preprocesses it, and passes it to the data cleaning unit; the data cleaning unit cleans the acquired historical oil fume separation data to remove outliers, duplicate values, or missing values in the data, and passes it to the data enhancement unit; the data enhancement unit increases the quantity and diversity of oil fume separation training data through data enhancement methods such as rotation, scaling, flipping, cropping, and color transformation, and passes it to the integration data unit; the integration data unit obtains standardized oil fume separation data according to the data standardization processing formula "z=(x - μ) / σ", and passes it to the conversion data unit; the conversion data unit obtains normalized oil fume separation data according to the normalization calculation formula "x'=(x - min(x)) / (max(x) - min(x))", and passes it to the filtering and denoising unit; the filtering and denoising unit is based on the calculation formula of the Gaussian filtering method Obtain the denoised oil fume separation image and transfer it to the grayscale conversion unit; the grayscale conversion unit obtains the grayscale oil fume separation image according to the grayscale calculation formula "f(I,j) = max(R(I,j), G(I,j), B(I,j))" and transfers it to the feature extraction unit; the feature extraction unit converts the labeled data into feature vectors useful for model training and extracts the key features of multi-layer filtering, condensation patterns, separation traces, grease collection, and performance degradation to improve the model performance.

[0045] Please refer to Figure 3 As shown, the model construction module includes a network layer unit, a network parameter unit, a hierarchical optimization unit, and a data partitioning unit, which are responsible for determining the model architecture, designing the network layer structure, adding auxiliary layers, defining model parameters for modeling according to the feature information of oil fume separation, and transferring it to the model training module.

[0046] Furthermore, the network layer unit customizes and optimizes the network layer structure according to the selected convolutional neural network as the model architecture and specific application scenarios and requirements, and transfers it to the network parameter unit; the network parameter unit adjusts the number of network layers, convolutional kernel size, stride, padding method, and selects the loss function and optimization algorithm according to the network layer structure, and transfers it to the hierarchical optimization unit; the hierarchical optimization unit adjusts the size of the input oil fume separation data and performs normalization processing by adding zero-padding layers and normalization layers before and after the convolutional layer respectively, and transfers it to the data partitioning unit; the data partitioning unit divides the dataset with the extracted key feature vectors into a training set, a validation set, and a test set and establishes a convolutional neural network structure for subsequent model training and learning.

[0047] Please refer to Figure 4 As shown, the model training module includes an input padding unit, a convolutional input unit, a pooling conversion unit, a fully connected layer unit, a normalization output unit, an output conversion unit, and a backpropagation unit, which are responsible for training the model according to the collected data, adjusting the model parameters to optimize the model performance to the predetermined oil fume separation control parameters, and transferring it to the model evaluation module.

[0048] Furthermore, the input padding unit fills the size of the oil fume separation data according to the formula "Ph = [(Ho - 1)*Sh + Kh - Hi]

[0049] / 2, Pw = [(Wo - 1) * Sw + Kw - Wi] / 2” to obtain the oil fume separation filling data, and transfer it to the convolution input unit; the convolution input unit obtains the convolution input value according to the convolution input calculation formula “z(t) = ∫x(m)y(t - m)dm” and activates the function input, and transfers it to the pooling conversion unit; the pooling conversion unit obtains the maximum pixel value after pooling according to the average pooling calculation formula “Z(I, j) = mean(X[i*Ps(i + 1)*Ps, j*Ps(j + 1)*Ps])” to retain important feature information and enter the fully connected layer, and transfers it to the fully connected layer unit; the fully connected layer unit obtains the output data of the connection layer according to the fully connected layer calculation formula “y = f(∑(Wn*Xn) + b)” and enters the output layer, and transfers it to the normalization output unit; the normalization output unit obtains the output of Batch Norm according to the Batch Norm layer calculation formula “h = φ(Bn(Wx + b)), where h is the output of the fully connected layer, φ is the activation function, Bn is the operator of batch normalization, W is the weight parameter, X is the input of the fully connected layer, and b is the bias parameter”, and transfers it to the output conversion unit; the output conversion unit obtains the size of the output value after convolution according to the output layer conversion calculation formula “N = (P - F + 2C) / S + 1” and transfers it to the backpropagation unit; the backpropagation unit obtains the backpropagation of the convolution output value according to the backpropagation calculation formula “dz[l] = da[l] * g[l]′(z[l])” to optimize the performance of the network model.

[0050] Please refer to Figure 5 As shown, the model evaluation module includes a classification marking unit, a learning and training unit, a cost verification unit, and a model evaluation unit, which are responsible for validating and evaluating the model according to the trained model, adjusting and optimizing the oil fume separation model according to the verification results, and transferring it to the parameter determination module.

[0051] Furthermore, the classification marking unit classifies and marks the corresponding data sample categories according to the abnormal categories of the oil fume separation quality of the oil fume machine for subsequent machine learning of the model, and transfers it to the learning and training unit; the learning and training unit determines the training model according to the gradient descent data, inputs the trained oil fume separation data into the model for simulation training and continuously iterates and optimizes to ensure the accuracy of model recognition, and transfers it to the cost verification unit; the cost verification unit calculates according to the cost function formula “J(θ) = -1 / m∑m∑K[y k (i)log(h θ (x(i))k)+(1 - y k (i))log(1 - (h θThe cost function value after model training is obtained from “(x(i))k)] + λ / 2m∑L-1∑sl∑sl+1(θj,i(l))2” to ensure the accuracy of model evaluation and is passed to the model evaluation unit; the model evaluation unit obtains the comprehensive evaluation score F according to the model evaluation score calculation formula “F = 2(Ac*Re) / (Ac+Re)” for subsequent evaluation of the model training results and passes it to the processing center.

[0052] Please refer to Figure 6 As shown, the parameter determination module includes an oil-grease separation unit, a fume exhaust concentration unit, a wind pressure matching unit, a test application unit, and an anomaly identification unit. The optimal oil fume separation control parameters are obtained through simulation scenario tests based on the oil fume separation control parameters predicted by the trained model and are passed to the application control module.

[0053] Furthermore, the oil-grease separation unit obtains the oil-grease separation degree according to the oil fume machine oil-grease separation degree calculation formula “η = (M1 - M2) / M1*100%” for subsequent debugging and detection and passes it to the fume exhaust concentration unit; the fume exhaust concentration unit obtains the fume concentration according to the oil fume machine fume exhaust concentration calculation formula “Co = M / [(Q1 + Q2 + Q3) / ε*(tp - ti)]” for subsequent debugging and detection and passes it to the wind pressure matching unit; the wind pressure matching unit obtains the wind pressure matching value according to the oil fume machine wind pressure matching calculation formula “Po ≥ ρ*g*(No - no)*ho + ρv2 / 2” for subsequent debugging and detection and passes it to the test application unit; the test application unit conducts an operation test based on the oil fume separation control parameters matched according to the customer's needs and preferences by the trained model and obtains the quality information of oil fume separation to ensure the accuracy of the parameters and passes it to the processing center; the anomaly identification unit matches the real-time obtained oil fume separation quality detection image data of the oil fume machine with the corresponding image data in the trained convolutional neural network model and confirms its anomaly category for subsequent rectification of anomaly problems.

[0054] Please refer to Figure 7 As shown, the application control module includes a quality prevention unit, a separation setting unit, a condensation separation unit, a contact separation unit, a centrifugal suction separation unit, and a quality detection unit. It controls the oil fume machine to perform oil fume separation operation according to the oil fume separation control parameters, so that the actual oil fume separation effect reaches the highest national standard and is passed to the processing center.

[0055] Furthermore, the quality prevention unit confirms whether the quality abnormality prevention measures predetermined by the model are in place one by one, so as to ensure that the oil fume separation operation can achieve the predetermined effect, and transmits it to the separation setting unit; the separation setting unit implements the oil fume separation degree improvement prevention measures predetermined by the model on the range hood one by one, so as to ensure that the oil fume separation degree reaches the predetermined standard, and transmits it to the condensation separation unit; the condensation separation unit intercepts the high-temperature oil fume through the condensation plate, separates the grease adhesion, collects and guides the oil droplets, and condenses the residual oil to achieve oil fume condensation separation, so as to facilitate the subsequent oil net contact separation, and transmits it to the contact separation unit; The contact separation unit realizes oil net contact separation for the condensed and liquefied oil fume through interception and adhesion, multi-layer filtration, convergence and diversion, and residual interception, so as to facilitate subsequent centrifugal separation and transfer it to the separation unit; the separation unit realizes centrifugal separation for the oil fume after contact separation through impeller collision, centrifugal separation, convergence and diversion, purification and discharge, so as to ensure the quality of oil fume separation, and transfer it to the quality inspection unit; the quality inspection unit detects the quality information of oil separation degree, oil fume concentration, odor reduction, air volume / noise, and maximum static pressure of oil fume separation through oil fume separation detection equipment, so as to ensure the accuracy of oil fume separation.

[0056] System operation working principle:

[0057] Before oil fume separation, the information acquisition module obtains normal and abnormal historical image data similar to oil fume separation of range hoods and performs preprocessing for subsequent model construction, and passes it to the model construction module; then the model construction module determines the model architecture, designs the network hierarchy, adds auxiliary layers, and defines model parameters according to the characteristic information of oil fume separation to build a model for subsequent model training and verification, and passes it to the model training module; then the model training module trains the model according to the collected data and adjusts the model parameters to optimize the model performance to predetermine the oil fume separation control parameters to ensure that the oil fume separation of the range hood can be automatically designed, and passes it to the model evaluation module; then the model evaluation module verifies and evaluates the model according to the trained model, and verifies the oil fume separation model according to the verification results. The model is adjusted and optimized to improve the performance and accuracy of the model, and is passed to the parameter determination module; the parameter determination module performs a simulated scenario test based on the oil fume separation control parameters predicted by the trained model to obtain the best oil fume separation control parameters to ensure the subsequent oil fume separation effect, and passes them to the application control module; then the application control module controls the range hood to perform oil fume separation operation according to the oil fume separation control parameters, so that the actual oil fume separation effect reaches the highest national standard to ensure the quality effect of oil fume separation, and passes it to the processing center; the processing center compares the quality information of oil fume separation with the oil fume separation quality standard of the range hood stored in the memory: if the standard is met, the customer is notified that it can continue to be used; if the standard is not met, the alarm is passed to notify debugging or maintenance.

[0058] When the user or manager is nearby or away, they can use a smart mobile terminal to automatically form a network connection within a wireless network or the Internet through the wireless communication module, integrating with the smart mobile terminal, the Internet, and Internet of Things technologies. They can control the operation of the range hood closely or remotely through the APP software on the smart mobile terminal or the remote control software on the computer to meet the quality requirements of customers. The user or manager can remotely control and monitor the oil fume separation situation of the range hood to dynamically adjust the operating state, avoid long-term overloading, and can view the daily operating conditions, data change situations, equipment activity logs, etc. By means of prefabricated deployment codes or adding deployment codes, remote connection and control of the range hood operation are achieved. The oil fume separation efficiency of the remote range hood can be viewed in real time through both the smart mobile terminal and the computer, improving energy use efficiency and reducing energy costs.

[0059] Please refer to Figure 8 As shown, a method for separating oil fume of an intelligent range hood provided by the present invention includes the following steps:

[0060] S10. Before oil fume separation, the information acquisition module acquires historical image data of normal and abnormal oil fume separation of similar range hoods and performs preprocessing for subsequent model construction, and transmits it to the model construction module;

[0061] Please refer to Figure 9 As shown, the step S10 includes the following steps:

[0062] S11. The data acquisition unit acquires historical data of oil fume emissions, equipment operation parameters, separation efficiency, and environmental maintenance of similar range hoods through the industry database and performs preprocessing for subsequent data cleaning, and transmits it to the data cleaning unit;

[0063] Further explanation: According to the operating data of the customer's range hood, similar oil fume separation data such as oil fume emissions, equipment operating parameters, separation efficiency, and environmental maintenance, as well as the operating information of associated images or videos such as corresponding oil fume dynamic capture, equipment operating status monitoring, user behavior association, maintenance and cleaning verification, and data on different customer evaluation effects are collected from the domestic and foreign range hood operating databases through a data collector and classified and aggregated; the oil fume separation data includes, but is not limited to, historical data on different customer evaluation effects of oil fume separation of the same brand of range hoods, and also includes normal and abnormal quality information on oil fume separation of similar range hoods and historical data on different customer evaluation effects obtained through legal means such as web crawling, sensor collection, manual annotation, dataset purchase, and crowdsourcing from third-party shared data platforms such as the Internet, social media, professional testing institutions, kitchenware trading platforms, and public databases, so as to obtain historical data on similar oil fume separation quality with high quality, diversity, and richness, specifically including oil fume separation data such as oil separation degree, air volume and air pressure, odor reduction degree, noise control, energy consumption efficiency, filter clogging warning, oil stain deposition rate, and component durability.

[0064] S12. The cleaning data unit cleans the collected historical oil fume separation data to remove outliers, duplicate values, or missing values in the data, so as to improve the quality and accuracy of the oil fume separation data, and transmits it to the enhanced data unit;

[0065] Further explanation: Due to sensor failures, recording errors, or system anomalies occurring during oil fume separation or detection, abnormal values may appear, affecting the accuracy of subsequent analysis; handle missing values for data with missing parts in the oil fume separation data; detect each feature in the dataset by writing code to determine missing values, identify missing value patterns, and select to fill or delete samples or variables containing missing values based on the number and impact of the missing values by writing code. For time series data, forward filling or backward filling is used to fill in the missing values; verify the processing effect by comparing indicators such as data quality, model accuracy, and reliability before and after processing. If the processing effect is not good, reselect the processing method or re-clean the data until the best effect is achieved to remove duplicate, incorrect, and invalid data in the oil fume separation data to effectively handle missing values for subsequent model construction and analysis; the oil fume emission data in the oil fume separation includes pollutant concentration data such as oil fume concentration, particulate matter, and total non-methane hydrocarbons, environmental parameters such as temperature, humidity, and air pressure, and also includes smoke morphology images (capturing the density, flow direction, etc. of the rising oil fume in real time through a camera for evaluating the oil fume concentration), smoke diffusion trajectory images (capturing the smoke morphology image in real time through a camera and analyzing it as a smoke diffusion trajectory in combination with image processing technology for evaluating the oil fume separation efficiency), oil fume particle distribution images (using an infrared or high-definition camera to capture the oil fume particle distribution, and combining with a deep learning model to identify the aggregation state of submicron particles to assist in optimizing the centrifugal separation and electrostatic adsorption parameters), etc.

[0066] S13. The enhanced data unit uses data augmentation methods such as rotation, scaling, flipping, cropping, and color transformation to increase the quantity and diversity of the oil fume separation training data, improve the model generalization ability, and transfer it to the integrated data unit;

[0067] Furthermore, by rotating the oil fume separation image data, the LabelImg tool is used to annotate targets such as oil fume and cookware in the cleaned oil fume separation image data. After rotating by a certain angle, multiple pixels will be rotated and corresponding to the same pixel, resulting in the loss of pixels in the rotated oil fume separation image data. Therefore, a mapping is constructed using reverse thinking to map the pixel coordinates after rotation to the pixel coordinates of the original oil fume separation image data, so as to ensure that each pixel after rotation has a pixel value; the neighborhood interpolation algorithm is used to copy each original pixel in the rotated oil fume separation data unchanged to the corresponding four pixels after expansion, retaining all the information of the original oil fume separation image data; the oil fume separation image data with abnormal orientation after scaling is flipped 180 degrees around the central axis or symmetry axis of the original image to ensure the orientation consistency of the oil fume separation data; the flipped oil fume separation image data is cropped to ensure the integrity and clarity of the image; the equipment operation parameter data in the oil fume separation includes hardware operation indicators such as the operation status of the fan, the efficiency of the purifier, the impeller speed and the air volume, energy consumption and fault warnings such as power consumption data and abnormal alarms, and also includes the filter status image (regularly taking images of the filter surface through a camera, and judging the degree of blockage by extracting features such as the oil stain coverage area and texture changes, triggering self-cleaning or maintenance reminders), the separation plate status image (capturing the attachment of oil droplets on the surface of the centrifugal separation plate through a camera, evaluating the throwing and suction efficiency and dynamically adjusting the impeller speed), and the fan and motor operation images (monitoring the temperature distribution of the motor using a thermal imaging camera, and combining image analysis to identify abnormal heat generation such as bearing wear and overloading to prevent faults).

[0068] S14. The integrated data unit obtains the standardized oil fume separation data according to the data standardization processing formula "z = (x - μ) / σ, where z is the standardized data, x is the original data, μ is the mean of the data, and σ is the standard deviation of the data", to improve the stability of the model and transmit it to the conversion data unit;

[0069] Further explanation: Calibrate the original oil fume separation data collected, including time synchronization, unit conversion, range adjustment, etc. For example, ensure that all timestamps are in a unified format and synchronized with the system clock; convert the data collected by different sensors, detection devices, etc. to the same unit to ensure the accuracy of the oil fume separation data; use the above-mentioned oil fume separation data standardization processing formula "z = (x - μ) / σ" through the written code to convert data from different sources such as dates, values, texts, etc. into a standard normal distribution with a mean of 0 and a standard deviation of 1, subtract the mean and divide by the standard deviation to make it in a unified standard format to ensure the consistency and comparability of the oil fume separation data for subsequent processing; encode the unstructured oil fume separation data to convert it into structured data, or convert the text data into numerical representation for easy model processing and analysis, so as to process the original data into a form that conforms to a specific standard for subsequent model training, learning, and analysis; the separation efficiency data in the oil fume separation includes oil separation degree, odor reduction degree, exhaust air volume, wind pressure and static pressure values, total pressure efficiency, noise control, etc.

[0070] S15. The data conversion unit obtains the normalized oil fume separation data according to the normalization calculation formula "x' = (x - min(x)) / (max(x) - min(x)), where x' is the normalized data, x is the original data, and min(x) and max(x) are the minimum and maximum values of the data respectively" to improve the model performance and transfer it to the filtering and denoising unit;

[0071] Further explanation: Through the written code, use the above-mentioned normalization calculation formula "x' = (x - min(x)) / (max(x) - min(x))" to scale and map the cleaned data to the specified [0, 1] range, convert the data to a unified format or range for data normalization, and discretize the data such as converting continuous features to discrete features for data normalization to convert the oil fume separation data, so as to eliminate the total difference between different samples or features and the dimensionality impact on the oil fume separation data, make the data distribution consistent, and confirm whether the oil fume separation data has been normalized to the specified range or distribution as expected, and verify the normalization effect through statistical descriptions such as maximum and minimum values; the oil fume separation data includes various types such as texts, pictures, audios, videos, etc. to eliminate the dimensionality difference between different oil fume separation data, thereby improving the accuracy and efficiency of the oil fume separation data analysis to better adapt to subsequent analysis and processing; the environmental maintenance data in the oil fume separation includes maintenance and cleaning records such as the degree of filter clogging and self-cleaning frequency, user behavior data such as cooking time periods and intensities and purification mode switching records, and regularly take pictures of the oil accumulation images at parts such as the oil collection box and oil guide groove (to quantify the cleaning effect and optimize the maintenance cycle), and record the oil shedding process images during high-temperature dissolution or electrostatic plate heating (to verify the effectiveness of the cleaning strategy and iterate and optimize).

[0072] S16. The filtering and denoising unit obtains the oil fume separation image after denoising according to the Gaussian filtering method calculation formula where G(x, y) is the pixel of the two-dimensional Gaussian function, (x, y) is the pixel coordinate, and σ is the standard deviation, so as to perform grayscale image conversion and transfer it to the grayscale conversion unit;

[0073] Further explanation: Obtaining the Gaussian filtering output pixel value according to the Gaussian filtering method calculation formula is the weighted average of the input pixel values in its neighborhood, and the weight is given by the Gaussian function. Among them, the standard deviation σ determines the width of the Gaussian function, and the Gaussian function is discretized, that is, the weights are calculated within a certain window size, and then these weights are applied to the corresponding neighborhood pixel values of the input oil fume separation image to obtain the output pixel value. The pixels are arranged in ascending order according to the grayscale value and the median value is taken as the new grayscale value; check the sampling values in the input signal to determine whether they represent the signal itself. By using an observation window composed of an odd number of samplings, the values in the window are sorted, the median value is taken as the output, the earliest value is discarded, and new samplings are obtained. Repeat the above calculation process to reduce the random noise in the oil fume separation image and make the oil fume separation image smoother; the user behavior data in the oil fume separation includes: cooking action recognition image: collect video data of the stove area to identify actions such as the position of the user's hand and the movement trajectory of the cooking utensil, such as stir-frying and frying, to predict the intensity of oil fume explosion and start the high-speed mode in advance, and the oil fume emission port monitoring image: take real-time pictures of the smoke exhaust port image to analyze the transparency and colors such as blackening and graying of the emitted smoke, and judge whether the purification effect meets the environmental protection standards.

[0074] S17. The grayscale conversion unit obtains the grayscale oil fume separation image according to the grayscale conversion calculation formula "f(I, j) = max(R(I, j), G(I, j), B(I, j)), where f(I, j) is the image after grayscale, and R(I, j), G(I, j), and B(I, j) are the original images of the three colors" for subsequent feature extraction and transfers it to the feature extraction unit;

[0075] It is further explained that the grayscale oil fume separation image is obtained through the grayscale calculation formula, the sampling value in the input signal is checked to determine whether it represents the signal itself, and the values ​​in the window are sorted by using an observation window composed of an odd number of samples, the middle value is taken as the output, the earliest value is discarded, and a new sample is obtained. The above calculation process is repeated to remove the noise in the oil fume separation image or other signals, so as to subsequently extract the key features of the oil fume separation image, including multi-layer filtering structure, condensation and contact separation traces, dynamic separation process, grease collection system, electrostatic separation components, performance degradation and other characteristic information; then the standardized oil fume separation data is classified through the written code, and the classified data is automatically labeled to add data Add accurate labels or annotations as the target variables or features in subsequent model training, and use them as training sets and validation sets to ensure that the model has accurate reference to oil fume separation data during the training process, so that it can learn how to generate corresponding labels based on the characteristics of oil fume separation data to improve labeling efficiency and accuracy; for some data, customers are required to enter the system remotely through their mobile phones to manually add new labels or modify existing labels to ensure that they accurately reflect the essential characteristics of the oil fume separation data, and deploy the labeled oil fume separation data to the corresponding system platform for subsequent application or analysis to improve the accuracy and reliability of the labels, so as to facilitate the extraction of key features in the subsequent oil fume separation data, which is more conducive to model training, testing and verification.

[0076] S18. The feature extraction unit converts the labeled data into feature vectors useful for model training and extracts key features of multi-layer filtering, condensation patterns, separation traces, grease collection, and performance degradation to improve the training speed and performance of the model.

[0077] It is further explained that the most representative features and the most useful features for the model that can reflect the essential content of the data and have the degree of discrimination and description are selected from the pre-processed oil fume separation data to reduce the dimension of the features, reduce the computational complexity, and improve the performance and generalization ability of the model; after feature extraction, principal component analysis is used to reduce the dimension of the features to obtain a large amount of feature information in order to reduce the dimension of the features and reduce the computational complexity; after feature dimensionality reduction, the features are binary encoded to improve the interpretability of the features and the performance of the model; since some information of the original data may be lost after feature extraction and dimensionality reduction, principal component reconstruction or minimum The feature reconstruction such as square reconstruction is to restore the data information as much as possible; the evaluation is carried out through cross-validation, and the method and parameters of feature extraction are adjusted according to the evaluation results to evaluate the quality and effect of the extracted features, improve the training speed and performance of the model, and improve the efficiency of data analysis by reducing the amount or complexity of data while keeping the original appearance of the data as much as possible; the model AI technology is used to conduct in-depth analysis of the oil fume separation data to extract key information, and provide a basis for the subsequent generation of oil fume separation control parameters; the multi-layer filtering features include oil net morphological features (densely arranged metal grids or long strip grille structures in the image), separation plate layout Features (the separation plate with stepped or vortex design presents irregular surface texture in the image), turbine component features (serrated impeller structure with a diameter of ≥28 cm), etc.; the condensation pattern features are the grease condensation pattern features that are obvious in the image of the tempered glass smoke barrier and are formed by the liquefaction of high-temperature oil smoke when it is cooled; the separation trace features include grease adhesion layer features (the surface of the oil net presents flaky or drop-shaped grease deposition), negative pressure zone morphology features (the side-suction model image shows a three-dimensional low-pressure area wrapped around the cooker, and the oil smoke inhalation path is visualized through airflow lines), centrifugal oil rejection trajectory features (radial grease splash marks can be seen around the high-speed turbine), electrode plate distribution Characteristics (parallel arrangement of metal plates with fine oil droplets attached to their surface can be seen in the electrostatic separation technology image), electric field visualization characteristics (displaying the electric field strength gradient through special imaging technology); the grease collection characteristics include oil pipe structure characteristics (the metal straight-through pipe connecting the separation component and the oil collecting box presents a smooth inner wall in the image to reduce residue), oil collecting box design characteristics (transparent or semi-enclosed containers show stratified oil accumulation), etc.; the performance degradation characteristics include oil net blockage state characteristics (after long-term use, the image shows that the grid gap is reduced, and the grease deposition thickness is ≥0.5mm38), air duct pollution distribution characteristics (uneven oil adhesion inside the volute), etc.

[0078] S20, the model building module determines the model architecture, designs the network hierarchy, adds auxiliary layers, defines model parameters and models the model according to the characteristic information of oil fume separation, so as to facilitate subsequent model training and verification, and transmits the model to the model training module;

[0079] See also Figure 10As shown, step S20 includes the following steps:

[0080] S21. The network layer unit customizes and optimizes the network layer structure according to the selected convolutional neural network as the model architecture and based on specific application scenarios and requirements for subsequent defect detection and quality recognition, and transmits it to the network parameter unit;

[0081] Further explanation: Since the convolutional neural network (CNN) with deep neurons is suitable for processing data with a grid topology structure, it can automatically learn and extract useful features such as edges, lines, corners, and more complex combined features from images during the oil fume separation design of the range hood for subsequent defect detection, quality classification, etc. Since the basic structure of the CNN consists of an input layer, a convolutional layer, an activation function layer, a pooling layer, a fully connected layer, and an output layer, these layers can work together in the CNN model developed for the oil fume separation of the range hood to extract and process image features, including: The input layer can receive the image data of the oil fume separation of the range hood as input; The convolutional layer sets multiple convolutional layers, each containing multiple convolutional kernels, which can extract local features in the image. As the network structure deepens, it gradually becomes more abstract from the appearance; Adding a ReLU activation function layer after the convolutional layer can introduce non-linear factors and enhance the expressive ability of the model; Adding a pooling layer after the convolutional layer can reduce the dimension of the feature map, reduce the amount of calculation, and at the same time maintain the spatial invariance of important features; Adding a fully connected layer at the end of the network is used to convert the feature maps output by the convolutional layer and the pooling layer into classification results, that is, the categories or quality grades of the oil fume separation of the range hood.

[0082] S22. The network parameter unit adjusts the number of network layers, the size of the convolutional kernel, the stride, the padding method, and selects the loss function and optimization algorithm according to the network layer structure to ensure the performance and accuracy of the oil fume separation model, and transmits it to the hierarchical optimization unit;

[0083] Further explanation: Defining model parameters according to the network layer design, including the size of the convolutional kernel, the stride, the padding method, the activation function, etc., as well as the pooling method of the pooling layer, the size of the pooling window, etc., and the number of neurons in the fully connected layer and other layer parameters directly affect the performance and accuracy of the model; Then set hyperparameters related to model training such as the size of the input oil fume separation image, the number of label types, the total number of training cycles, the batch size, etc., and adjust them according to specific datasets and task requirements; Selecting optimization algorithms such as SGD, Adam, etc., and setting corresponding parameters such as the learning rate, momentum, etc., directly affect the training effect and convergence speed of the model to ensure that the model can accurately extract features from the oil fume separation image and perform effective classification or prediction.

[0084] S23. The hierarchical optimization unit adjusts the size of the input oil fume separation data and performs normalization processing by adding zero-padding layers and normalization layers before and after the convolutional layer respectively, so as to improve the training speed and stability of the oil fume separation model, and then transfers it to the data partitioning unit;

[0085] Further explanation: By setting the zero-padding layer to perform zero-value padding on the edges of the input oil fume separation image before the convolutional operation, the size of the input data is adjusted to facilitate the convolutional operation, maintain the image edge information, ensure that the convolutional kernel can be correctly applied to the image boundary, and at the same time help to maintain the image edge information and avoid information loss during the convolutional process; Through zero-padding, different-sized input images can be processed more effectively without complex preprocessing, which helps to control the size of the output feature map of the convolutional layer, making the network design more flexible and controllable, ensuring that the network can adapt to the oil fume separation changes of different sizes and resolutions in the oil fume machine, and accurately extract and process image features; Normalize the distribution of the input data to improve the training speed and stability of the model, accelerate the training process of the neural network and improve the convergence speed, while enhancing the stability of the oil fume separation model. Normalize the input of the activation function before the activation function of each layer of the network. Calculate the mean and variance of this batch for each small batch of data, and then perform batch normalization on the linear calculation results, that is, subtract the mean and divide by the standard deviation to ensure that the calculation results conform to the standard normal distribution with a mean of 0 and a variance of 1, and then perform translation and scaling operations to adapt to different data distribution requirements, making the input of the middle layer of the network relatively stable, which helps to solve the problem of gradient disappearance or gradient explosion during the training process, thereby accelerating the training and enhancing the stability of the model, improving the robustness of the model to the weight initialization method, and reducing the trouble brought by the weight initialization selection, so as to improve the generalization ability and training efficiency of the oil fume separation model.

[0086] S24. The data partitioning unit divides the dataset with key feature vectors extracted into a training set, a validation set, and a test set and builds a convolutional neural network structure for subsequent model training and learning.

[0087] Furthermore, according to the preselected segmentation strategy, determine the proportions of the training set, validation set, and test set, and ensure that the data distributions among the subsets are as consistent as possible to avoid introducing biases. In some application scenarios, due to the time series characteristics of the data, use the written code to divide the dataset into a training set for training the model, a validation set for adjusting the model parameters and selecting the best model, and a test set for evaluating the final performance of the model according to the time series to ensure the temporal consistency of the training and test sets. Establish the training set and test set required for predicting the abnormal recognition mode of the oil fume separation quality in different environments by passing the dataset through the deep learning of the convolutional neural network in a 9:1 relationship, thereby establishing a deep convolutional neural network structure: 3 convolutional layers, 3 max pooling layers, 2 fully connected layers, 1 softmax layer, and 1 output layer. Add zero-padding layers and normalization layers when necessary. Since the convolutional kernels in the convolutional layers contain weight coefficients while the pooling layers do not, the pooling layers are not independent layers. The convolutional layers are used to extract local features of the image, the pooling layers are used to reduce the dimension of the feature maps, and the fully connected layers are used to integrate features and perform classification.

[0088] S30. The model training module trains the model based on the collected data, adjusts the model parameters to optimize the model performance to the predetermined oil fume separation control parameters, ensures that the oil fume separation of the oil fume machine can be automatically designed, and transmits it to the model evaluation module.

[0089] Please refer to Figure 11 As shown, the step S30 includes the following steps:

[0090] S31. The input padding unit obtains the oil fume separation padding data according to the oil fume separation data padding size calculation formula "Ph = [(Ho - 1) * Sh + Kh - Hi] / 2, Pw = [(Wo - 1) * Sw + Kw - Wi] / 2, where Ph and Pw are the padding sizes in the height / width directions respectively, Hi and Wi are the height / width of the input feature map respectively, Kh and Kw are the height / width of the convolutional kernel respectively, Sh and Sw are the values of the stride in the height / width directions respectively, and Ho and Wo are the height and width of the output feature map respectively" for subsequent convolutional input, and transmits it to the convolutional input unit.

[0091] Further explanation: The standardized oil fume separation data first enters the input layer, and the padded data is obtained through the above-mentioned padding size calculation formula and enters the zero-padding layer. Zero values are padded to the edges of the input data of the convolutional layer to adjust the size of the input data for convolution operations, so as to maintain the same spatial dimension of the input and output, thereby preventing the loss of edge information of the oil fume separation image and enabling the spatial dimension of the input data to remain unchanged during subsequent convolution operations, so as to maintain the edge information of the oil fume separation image and the processing of subsequent layers; the number of padding is jointly determined by the size of the convolutional kernel, the stride, and the padding size. In order to keep the height and width of the input and output images the same, or to reduce the rapid decrease in size during consecutive convolution operations, the number is usually set to the convolutional kernel size minus; the height Hi / width Wi of the input feature map, the height Kh / width Kw of the convolutional kernel, the stride values Sh / Sw in the height / width direction, and the height Ho / width Wo of the output feature map are all obtained through model design parameters.

[0092] S32. The convolution input unit obtains the convolution input value according to the convolution input calculation formula "z(t) = ∫x(m)y(t - m)dm, where z(t) is the output function, x(t) is the input function, y(t) is the convolution kernel function, and dm is the integral microelement of the variable m", and activates the function input for subsequent pooling transformation and passes it to the pooling transformation unit.

[0093] Further explanation: The "width" or "increment" of the integral microelement dm when integrating each infinitesimal interval of the variable m is used to obtain the cumulative effect z(t) over the entire domain by accumulating the function values (i.e., x(m)y(t - m)) on these infinitesimal intervals. dm is the integral microelement of the variable m, that is, the infinitesimal quantity in the integration process; and the convolution calculation is performed to obtain the convolution value, which enters the convolutional layer and the activation function value is obtained through the ReLU function calculation formula "f(x) = max(0, x + Y)" and enters the activation function layer. In the formula, f(x) is the activated function, x is the eigenvalue output by the convolutional layer or the fully connected layer, and Y is a random variable. When the input x is greater than 0, x is output; when the input x is less than or equal to 0, 0 is output; when x = 0, it is non-differentiable, that is, when the input is positive, the original value is maintained, and when the input is negative, 0 is output; however, the ReLU activation function is used after the convolutional layer to increase the non-linearity of the network and promote the model to learn complex patterns.

[0094] S33. The pooling conversion unit obtains the maximum pixel value after pooling according to the average pooling calculation formula "Z(I,j) = mean(X[i*Ps(i+1)*Ps,j*Ps(j+1)*Ps]), where Z(I,j) is a pixel value in the output feature map after pooling, i and j are the positions in the output feature map respectively, mean is to obtain the average value of all pixel values in the input window, X is the input feature map, and Ps is the window size of the pooling operation", so as to retain important feature information and enter the fully connected layer, and transfer it to the fully connected layer unit;

[0095] Further explanation, calculating the average value of all values in each area through the above average pooling calculation formula as the output can retain more information in the feature map, which helps to improve the accuracy of the model, retains more detailed information, and is used in the deeper layers of the network to ensure the integrity of information; then, according to the maximum pooling calculation formula "Z(I,j) = max(X[i*Ps(i+1)*Ps,j*Ps(j+1)*Ps])", the maximum pixel value after average pooling is obtained. Z(i,j) is a pixel value in the output feature map after pooling, and max is to obtain the maximum value among all pixel values in the input window. According to this formula, the maximum value is selected from each area of the input feature map as the output, retaining the most significant features in the feature map, reducing the size of the feature map, and losing some useful information. Since the maximum value of each area is concerned, the grain size feature is retained, and the features with better classification recognition are selected.

[0096] S34. The fully connected layer unit obtains the output data of the connection layer according to the fully connected layer calculation formula "y = f(∑(Wn*Xn)+b), where y is the output result, f is the activation function, Wn is the weight of the nth input feature, Xn is the n input features, n is the dimension of the input features, and b is the bias", and enters the output layer for subsequent normalization output calculation, and transfers it to the normalization output unit;

[0097] Further explanation, the data after pooling is calculated by the fully connected layer according to the fully connected layer calculation formula "y = f(∑(Wn*Xn)+b)" to obtain the output value after convolution. In the formula, y is the output result, f is the activation function, Wn is the weight of the nth input feature, Xn is the n input features, n is the dimension of the input features, and b is the bias; the formula "y = f(∑(Wn*Xn)+b)" is deduced from "y = f(W*X+b)" to "y = f(W1*X1+W2*X2+…+Wn*Xn+b)" and transformed. In the formula, W is the weight and Xn is the input feature; through the fully connected layer, the features extracted by the convolutional layer and the pooling layer are integrated and classified, so that each neuron is connected to all neurons in the previous layer, and the output is generated through weighted summation and non-linear activation function.

[0098] S35. The normalization output unit obtains the output of Batch Norm according to the Batch Norm layer calculation formula "h = φ(Bn(Wx + b)), where h is the output of the fully connected layer, φ is the activation function, Bn is the operator for batch normalization, W is the weight parameter, X is the input of the fully connected layer, and b is the bias parameter", so as to perform subsequent data output conversion and transfer it to the output conversion unit;

[0099] For further explanation, the Batch Norm layer calculation formula is derived from "Bn(x) = γ⊙[(X - μb) / σb] + β", where μb is the mean, σb is the standard deviation, γ is the stretching parameter, β is the offset parameter, and (X - μb) / σb is the standardized normal distribution; the Batch Norm layer (abbreviated as "BN layer") is placed between the affine transformation and the activation function in the fully connected layer, standardizes the output values of each layer in the network to make them more conform to the normal distribution, which can accelerate the training speed of the neural network, help prevent gradient vanishing and gradient explosion; the BN layer reduces the correlation between input data by normalizing the data of each mini-batch, making the mean of each sample 0 and the variance 1, thus reducing the problem of internal covariate shift, helping to accelerate the training process, improve the stability and generalization ability of the model, and avoid the problem of reducing the representation ability of the neural network.

[0100] S36. The output conversion unit obtains the size of the output value after convolution according to the output layer conversion calculation formula "N = (P - F + 2C) / S + 1, where N is the size of the output after convolution, P is the size of the input before convolution, F is the size of the convolution kernel, C is the number of layers of 0 added around the image, and S is the step size", so as to perform subsequent model analysis and training and transfer it to the backpropagation unit;

[0101] Further explanation: The output size N is the size of the feature map after the convolution operation. The input size P is the width or height of the input image or feature map. By calculating through the above step S21, the padding size PH in the height direction and the padding size PW in the width direction are respectively obtained. The convolution kernel size F is obtained by calculating through the above step S22. The number of layers C of adding 0 around the image is the number of layers of adding 0 around the input image or feature map and is used to control the size of the output feature map, and is also obtained by the above network construction. The stride size S is the distance that the convolution kernel slides on the input image or feature map and is obtained by the above calculation. Then, according to the loss function calculation formula "L = max(0, m + y * (f(x) - b))", the loss function value is obtained. In the formula, L is the loss function value, m is a hyperparameter, y is the sample label (0 or 1), f(x) is the predicted value of the model, and b is the classification boundary. When f(x) ≤ b, then L = m + y * (f(x) - b); when f(x) > b, then L = 0. In machine learning or optimization problems, f(x) is an optimization target, b is a threshold, and m and y are parameters. When the target value f(x) is less than or equal to the threshold b, the value of L is equal to m + y * (f(x) - b); when the target value f(x) is greater than the threshold b, the value of L is 0. The model is trained using the training set data, and the model parameters are adjusted to minimize the loss function, so that the predicted value of the positive sample is greater than a certain threshold, and the predicted value of the negative sample is less than a certain threshold.

[0102] S37. The backpropagation unit performs backpropagation on the convolution output value according to the backpropagation calculation formula "dz[l] = da[l] * g[l]′(z[l])", where dz[l] is the input gradient, da[1] is the output gradient, g[l]′(z[l]) is the derivative of the activation function, and z[l] is the input of the activation function, so as to optimize the performance of the network model.

[0103] Further explanation: The input gradient dz[l] is the gradient of the input z[l] of the activation function of the l-th layer. During the backpropagation process, the dz[l] of each layer needs to be calculated in order to further calculate the gradients of the weight W[l] and the bias b[l]. The output gradient da[l] is the gradient of the output a[l] of the activation function of the l-th layer, which is passed down from the gradients of higher layers through the chain rule during the backpropagation process. The derivative of the activation function g[l]′(z[l]) is the derivative of the activation function g[l] of the l-th layer at z[l], which reflects how a small change in the input z[l] of the activation function affects the change in the output a[l] of the activation function. The input z[l] of the activation function is the input of the activation function of the l-th layer, which is obtained by linearly transforming the output a[l-1] of the activation function of the previous layer, the weight W[l] and the bias b[l] of the current layer during the forward propagation process. After being processed by the activation function g[l], the output a[l] of the activation function of the current layer is obtained. The convolution output values are obtained through the above backpropagation calculation formula as dw[l] = dz[l] * a[l-1], db[l] = dz[l], da[l-1] = W[l]T * dz[l]. The error term da[l] is passed to the previous layer through the convolution operation in the backpropagation of the convolutional layer. The convolutional kernel is flipped and applied to the error map. The local gradient is calculated through the derivative of the activation function and is realized through matrix operations. The input and the convolutional kernel are deformed into matrices for matrix multiplication, that is, the numbers in the convolution window are pulled into a row to form a column vector for matrix multiplication. Then, through the backpropagation of the fully connected layer, the gradients of the weights and the biases are calculated through the chain rule. For each node, the error term is passed to the nodes of the previous layer through the weight matrix, and the local gradient is calculated through the derivative of the activation function and is realized through matrix operations. The error term is multiplied by the output of the current layer and the input of the previous layer to update it in the direction of minimizing the loss function.

[0104] S40. The model evaluation module performs model verification and evaluation based on the trained model, adjusts and optimizes the oil fume separation model according to the verification results to improve the performance and accuracy of the model, and transmits it to the parameter determination module;

[0105] Please refer to Figure 12 As shown, the step S40 includes the following steps:

[0106] S41. The classification and marking unit classifies and marks the corresponding data sample categories according to the abnormal categories of the oil fume separation quality of the oil fume machine for subsequent machine learning of the model, and transmits it to the learning and training unit;

[0107] Further explanation: Classification is carried out according to the abnormal phenomena (and preventive measures) of oil fume separation in the range hood: "Incomplete oil fume separation (smoke leakage)" is category 1, "Decreased exhaust efficiency" is category 2, "Oil leakage or oil stain backflow" is category 3, and "Abnormal noise and vibration" is category 4; When category 1 is the positive sample, the remaining categories are negative samples, that is, the marked data is (1, 0, 0, 0). When category 2 is the positive sample, the remaining categories are negative samples, that is, the marked data is (0, 1, 0, 0). When category 3 is the positive sample, the remaining categories are negative samples, that is, the marked data is (0, 0, 1, 0). When category 4 is the positive sample, the remaining categories are negative samples, that is, the marked data is (0, 0, 0, 1); The above four abnormal categories are relatively typical abnormal classifications of oil fume separation in the range hood, including but not limited to the above four categories. One or more of these categories can also be reclassified into more categories and more detailed classification can be carried out according to the actual application results. The above classification is not fixed. After each category change, the model is retrained to adapt to the actual needs of the model; The preventive measures for the abnormal quality of oil fume separation in the range hood include: 1) Incomplete oil fume separation (smoke leakage): Select a low-suction range hood (distance from the stove top ≤ 40 cm) to reduce the suction attenuation and shorten the suction port distance, thereby optimizing the exhaust port structure, and use a silicone rubber sealing ring to reinforce the oil box interface to prevent oil leakage, thereby strengthening the sealing design; 2) Decreased exhaust efficiency: Clean the oil mesh / filter every month and disassemble and clean the impeller and air duct every six months for deep cleaning cycle, and check the rotation flexibility of the check valve every quarter. When rusted, replace the bearing shaft for check valve maintenance; 3) Oil leakage or oil stain backflow: The installation height of the side-suction type ≤ 40 cm, the top-suction type ≤ 75 cm to adapt to the suction range for height adjustment, and keep the exhaust hose horizontal or slightly upward to reduce bends (≤ 2 places) for flue pipe layout; 4) Abnormal noise and vibration: Turn on the range hood 1 minute before cooking and delay for 3 minutes after cooking to exhaust the remaining smoke for pre-start and delayed shutdown, and close the doors and windows during cooking to reduce the interference of crosswind on the stability of the negative pressure area for environmental air flow control; During model training, the above corresponding quality preventive measures are implemented to avoid similar quality abnormal problems.

[0108] S42. The learning and training unit determines the training model according to the gradient descent data, inputs the trained oil fume separation data into the model for simulation training, and continuously iterates and optimizes to ensure the accuracy of model recognition, and transmits it to the cost verification unit;

[0109] Furthermore, by continuously adjusting the model parameters to minimize the loss function, the prediction accuracy of the model is improved to ensure that the model can accurately identify the adaptability of oil fume separation in different environments; the model is trained with the data set in the training set to fit the data analysis law, that is, to determine various learning parameters such as the weights and biases of the model; and the model parameters and hyperparameters are adjusted during the model training process with the data set in the validation set to optimize the model performance, avoid overfitting, select the model, but do not participate in the determination of the learning parameters, but select the model parameters and hyperparameters with smaller model errors; then, the generalization ability of the model on unknown data is evaluated with the data set in the test set after the model training is completed to improve the accuracy and generalization ability of the model, and it is used once after training to evaluate the final model effect, and does not participate in the learning parameter process or the hyperparameter selection process; the oil fume separation data includes the quality data of oil fume separation, quality image data, user evaluation data, etc. Among them, the oil fume separation quality image data includes: 1) Oil fume dynamic capture images: smoke form images (capturing images of the density and flow direction of oil fume rising in real time through a camera to evaluate the oil fume concentration), smoke diffusion trajectories (trajectory images of smoke diffusion to evaluate the oil fume separation efficiency), oil fume particle distribution images (using an infrared or high-definition camera to capture the oil fume particle distribution and combining with a deep learning model to identify the aggregated distribution state image of sub-micron oil fume particles to assist in optimizing the centrifugal separation and electrostatic adsorption parameters), etc.; 2) Equipment operation status monitoring images: filter surface images (regularly capturing images of the oil stain coverage area and texture changes on the filter through a camera to judge the degree of blockage and trigger self-cleaning or maintenance reminders), separation plate status images (capturing images of the oil droplet attachment on the surface of the centrifugal separation plate through a camera to evaluate the throwing and suction efficiency and dynamically adjust the impeller speed), fan and motor operation images (using a thermal imaging camera to monitor the motor temperature distribution image, and combining image analysis to identify abnormal heat generation such as bearing wear and overloading to prevent failures), etc.; 3) User behavior association images: cooking action recognition images (collecting video data of the stove area through a camera to identify action images such as the hand position and pan movement trajectory of the user's stir-frying and frying to predict the intensity of oil fume explosion and start the high-speed mode in advance), oil fume exhaust port monitoring images (capturing the exhaust port image in real time through a camera to analyze the transparency and colors such as blackening and graying of the discharged smoke to judge whether the purification effect meets environmental protection standards such as GB 18483), etc.; 4) Maintenance and cleaning verification images: oil stain accumulation images (regularly capturing the oil stain accumulation in parts such as the oil collection box and oil guide groove through a camera, and using an image segmentation algorithm to quantify the cleaning effect and optimize the maintenance cycle), self-cleaning process record images (recording the oil stain shedding process during high-temperature dissolution or electrostatic plate heating to verify the effectiveness of the cleaning strategy and iteratively optimize it), etc.

[0110] S43. The cost verification unit obtains the value of the cost function after model training according to the cost function calculation formula "J(θ) = -1 / m ∑m ∑K [y k (i) log(h θ (x(i))k) + (1 - y k (i)) log(1 - (h θ (x(i))k)] + λ / 2m ∑L-1 ∑sl ∑sl+1 (θj,i(l))2, where J(θ) is the cost function, m is the total number of training samples, K is the total number of categories, y k (i) is the true label value of the sample, h θ (x(i))k is the predicted probability of the sample, λ is the regularization parameter, L is the total number of layers of the neural network, sl is the number of neurons in the l-th layer, and θj,i(l) is the weight parameter between connected neurons", to ensure the accuracy of model evaluation and transmit it to the model evaluation unit;

[0111] Further explanation: The cost function calculation formula obtains the minimized cost function of the model to measure the difference between the model prediction result and the true label, which can optimize the parameters of the model and improve the prediction accuracy of the model; θ is the model parameter, including weight and bias, which is continuously adjusted during training to minimize the cost function; the total number of categories K, in a multi-classification problem, each sample is assigned a category label; the true label value of the sample "y k (i)" is the value of the true label of the i-th sample on the k-th category. If it is a binary classification problem, its value is 0 or 1. If it is a multi-classification problem, its value will be a one-hot encoded vector, that is, the probability that the sample belongs to a certain category; the predicted probability of the sample "h θ (x(i))k" is the probability that the model predicts the i-th sample as the k-th category; the regularization parameter λ controls the intensity of the regularization term to prevent the model from overfitting, which is achieved by penalizing larger weight parameters; the weight parameter between connected neurons "θj,i(l)" is the weight parameter between the i-th neuron in the l-th layer and the j-th neuron in the l+1-th layer; the "-1 / m ∑m ∑K [y k (i) log(h θ (x(i))k) + (1 - y k (i)) log(1 - (h θ (x(i))k)]" is the cross-entropy loss function to measure the difference between the model prediction result and the true label; the "λ / 2m ∑L-1 ∑sl ∑sl+1 (θj,i(l))2" is the regularization term to prevent the model from overfitting. By minimizing the cost function, the parameters of the model can be optimized and the generalization ability of the model can be improved.

[0112] S44. The model evaluation unit obtains the comprehensive evaluation score F according to the model evaluation score calculation formula "F = 2(Ac * Re) / (Ac + Re), where F is the evaluation score of the trained model, Ac is the precision rate, and Re is the recall rate" for subsequent evaluation of the model training result and transmits it to the processing center;

[0113] Further explanation: After obtaining the verification of the cost function through the above step S43, the comprehensive evaluation score of the model is obtained according to the model evaluation score calculation formula. Among them, Ac and Re are calculated through "Ac = (Tp + Tn) / (Tp + Fp + Fn + Tn) and Re = Tp / (Tp + Fn)" respectively. In the formula, Tp is the number of true positive samples, Fp is the number of actually false positive samples, Fn is the number of false negative samples, and Tn is the number of true negative samples; After the model training is completed, a test is carried out to verify the accuracy and reliability of the model, and problems existing in the model may be found and optimized and adjusted during the test; The evaluation score of the trained model, that is, the F value, is the harmonic mean of the precision rate and the recall rate, which can evaluate the performance of the classification model. The value range is from 0 to 1, where 1 is the best performance and 0 is the worst performance. The higher the F value, the better the prediction effect of the model, and vice versa. Therefore, if the F value is close to 1, it indicates that the model performs well in maintaining the balance between the precision rate and the recall rate; Adjust the hyperparameters according to the performance of the model such as the learning rate and regularization coefficient to optimize the model performance.

[0114] S45. The processing center compares the actual evaluation score of the model with the model evaluation score standard stored in the memory: If it meets the standard, it is the predetermined oil fume separation control parameter; if it does not meet the standard, it is transmitted to the alarm and notifies to continue training until it meets the standard.

[0115] Further explanation: The test evaluation criteria for the abnormal category recognition training model of the oil fume separation quality of the range hood are best when greater than 0.8, which are specifically determined by factors such as equipment structure and design (oil mesh and filtration system, multi-layer separation plate design, impeller and air duct design, material and coating technology, etc.), operating status and performance (oil separation degree, air volume and air pressure matching, dynamic adjustment ability, etc.), use, maintenance and installation conditions (cleaning and maintenance frequency, installation specifications, environmental adaptability, etc.), external interference (smoke exhaust system resistance, cooking methods and frequencies, etc.); The oil fume separation control parameters include: 1) Core control parameters: oil separation degree (the ratio of the oil fume separation unit of the range hood to separate oil particles, which determines oil pollution deposition and environmental protection), exhaust air volume (the volume of air that can be discharged per minute, which determines the oil fume suction speed), air pressure (the anti-backflow ability of the range hood, which is directly related to the smoke exhaust resistance), noise control (national standard limit ≤ 74 dB, high-quality models can be optimized to 52 - 70 dB), odor reduction degree (the normal reduction degree is to reduce the abnormal odor by ≥ 90% within 30 minutes, and the instantaneous reduction degree is to reduce the peak abnormal odor by ≥ 50% within 3 minutes); 2) Additional control parameters: dynamic adjustment ability (real-time matching of air volume / air pressure through a variable frequency motor or AI algorithm to cope with instantaneous oil fume concentration changes), self-cleaning function (technologies such as high-temperature steam cleaning and high-pressure bubble cleaning reduce oil adhesion on the oil mesh / impeller and maintain the separation efficiency), material and structure optimization (non-stick oil coating inner cavity, multi-layer separation plate design improves the oil separation effect), and control parameters of other associated equipment, etc. These oil fume separation control parameters are automatically matched in the solution application through the trained model to avoid various quality abnormalities in the oil fume separation of the range hood.

[0116] S50. The parameter determination module conducts a simulation scenario test based on the oil fume separation control parameters predicted by the trained model to obtain the optimal oil fume separation control parameters, so as to ensure the subsequent oil fume separation effect and transmit them to the application control module;

[0117] Please refer to Figure 13 As shown, the step S50 includes the following steps:

[0118] S51. The oil separation unit obtains the oil separation degree according to the oil fume separation degree calculation formula of the range hood "η = (M1 - M2) / M1 * 100%, where η is the oil separation degree, M1 is the amount of oil entering the range hood, and M2 is the amount of oil discharged from the range hood" for subsequent debugging and detection, and transmits it to the smoke exhaust concentration unit;

[0119] Further explanation: The oil separation rate η is the ratio of the range hood separating oil from the cooking fumes, reflecting the purification ability of the device for oil. If the η value is higher, it indicates that more oil is intercepted in the oil net or the separation device, reducing the accumulation of internal oil stains. The amount of oil entering M1 is the total amount of oil entering the range hood under experimental conditions. Usually, the amount of oil generated during actual cooking is simulated through standardized tests. By heating a certain amount of oil to a specific temperature of 300°C ± 10°C, a real fume environment is simulated and obtained, which is determined by factors such as oil temperature, fume concentration, and device suction. The amount of oil discharged M2 is the amount of oil that is not separated, that is, the oil residue discharged to the outside through the range hood. If M2 is lower, it indicates that the separation efficiency of the device is higher and the environmental protection performance is better. The oil is intercepted through multiple filters such as a dynamic filter wall and centrifugal oil drainage. By measuring the initial mass of the range hood body, the random filter, and the independent filter, after the range hood operates for 30 minutes and then shuts down, the above components are weighed again, and the value is obtained by calculating the mass difference.

[0120] S52. The fume concentration unit obtains the fume concentration according to the range hood fume concentration calculation formula "Co = M / [(Q1 + Q2 + Q3) / ε * (tp - ti)]", where Co is the range hood fume concentration (mg / m 3 ), M is the fume emission mass (mg), Q1 is the heat dissipation of the device (W), Q2 is the heat dissipation of personnel (W), Q3 is the heat dissipation of lighting (W), ε is the comprehensive heat coefficient (W·°C / m 3 ), tp is the exhaust air temperature (°C), and ti is the indoor temperature (°C)" for subsequent debugging and detection, and transmits it to the air pressure matching unit;

[0121] Further explanation: The range hood fume concentration calculation formula is obtained by substituting "Q = Q1 + Q2 + Q3" into "L1 = Q / ε * (tp - ti)" to get "L1 = (Q1 + Q2 + Q3) / ε * (tp - ti)", and then substituting it into "Co = M / L1" to obtain Co. In the formula, L1 is the exhaust air volume (m 3 / h), Q is the total heat generation of the kitchen (W); the exhaust air volume L1 is the volume of air that the range hood needs to discharge per hour, which is affected by factors such as kitchen type and scene selection; the heat dissipation Q1 of the equipment is the heat generated by the operation of electrical appliances such as processing, cooking, and storage in the kitchen, which is obtained by calculating based on the power on the equipment nameplate or the measured power and combined with the load rate, and is usually estimated at 80% of the total power of the equipment; the heat dissipation Q2 of the personnel is the heat generated by the activities of the personnel in the kitchen, and is comprehensively valued by combining the maximum personnel density and the stay time, and is usually calculated at 102 W / person; the heat dissipation Q3 of the lighting is the heat generated by the lighting equipment in the kitchen, which is estimated by multiplying the unit area power or the total power of the lamps by the usage factor, or by estimating the unit area power; the exhaust air temperature tp is the temperature of the air discharged from the inner kitchen to the outside, which is obtained through a thermometer or a sensor; the indoor temperature ti is the target temperature maintained in the kitchen, which is affected by factors such as process requirements and human comfort, and is obtained through a thermometer or a sensor; the comprehensive coefficient ε is a coefficient derived from physical property parameters such as air density and specific heat capacity in the kitchen and unit conversion, and the typical value is 0.337; if there are very few electrical appliances and very few people in the kitchen, then the exhaust air volume of the range hood is obtained according to the formula "L2 = ω * P * H1", where L2 is the exhaust air volume based on the wind speed (m 3 / h), ω is the exhaust air coefficient (q·m / h), that is, the coefficient of the exhaust air efficiency per unit length and height of the exhaust hood, which is related to the suction speed or safety margin in the design specification, and generally takes a value of 1000 - 2000, and is obtained through historical experience or database query; P is the effective perimeter of the exhaust hood (m), that is, the total length of the open edge of the exhaust hood, which is used to calculate the effective suction area and does not include the side against the wall, and is obtained through the house construction drawing or measurement; H1 is the vertical height from the hood opening to the stove surface (m), which is obtained through the house construction drawing or measurement; if there are very few electrical appliances and very few people in the kitchen but the range hood is used frequently and thus frequent ventilation is required, then the exhaust air volume is calculated using "L3 = S * H2 * N", where L3 is the exhaust air volume based on the ventilation frequency (m 3 / h), H2 is the storey height of the kitchen (m), that is, the vertical height from the kitchen floor to the ceiling or ceiling (i.e., the net height of the kitchen), and both the kitchen area (㎡) S are obtained through the house construction drawing or measurement, and N is the ventilation frequency (q / h) which is the number of times the air is completely replaced per hour, and is determined by the exhaust capacity and the space volume comprehensively, and its reference values are 40 - 50 q / h for Chinese restaurants, 30 - 40 q / h for Western restaurants, and 25 - 35 q / h for staff canteens.

[0122] S53. The air pressure matching unit calculates according to the range hood air pressure matching formula "Po ≥ ρ * g * (No - no) * ho + ρv2 / 2, where Po is the air static pressure value (Pa), ρ is the air fluid density (kg / m 3 ), g is the acceleration due to gravity (m / s 2), where No is the total number of floors, no is the current floor number, ho is the height of a single floor (m), and v is the air velocity in the pipe (m / s)” to obtain the wind pressure matching value for subsequent debugging and testing, and transmit it to the test application unit;

[0123] Further explanation: The air static pressure value Po is obtained by measuring with a pressure sensor or a pitot tube; the air fluid density ρ is obtained by looking up a table, and it is generally 1.225 kg / m³ at standard atmospheric pressure and 15 °C 3 ; the value of the gravitational acceleration g is generally 9.8 m / s² 2 , but the actual value varies slightly due to geographical location; the total number of floors No is obtained from the building floor plan or construction drawings; the current floor number no is obtained by confirming through floor signs or property information; the height of a single floor ho is directly measured with a laser rangefinder or a tape measure. Generally, the height of a single floor in a conventional office building is 3 - 4 m / floor, and that of an industrial building is higher; the air velocity v in the pipe is obtained by measuring with an anemometer or a pitot tube, and it is generally 3 - 8 m / s.

[0124] S54. The test application unit conducts an operation test based on the oil fume separation control parameters matched according to the customer's needs and preferences by the trained model and obtains the quality information of oil fume separation to ensure the accuracy of the parameters and transmit them to the processing center;

[0125] It is further explained that after the model is trained, the installation confirmation of the standard prototype and the prevention confirmation of various measures are carried out according to the oil fume separation control parameters matched by the model and the corresponding preventive measures; then the oil fume separation simulation is run according to the predetermined control parameters, and the kitchen air quality is monitored in real time through the set gas sensors, PM2.5 detectors, CO sensors, etc., the ignition status of the stove is detected using door and window sensors or micro switches, and the oil fume flow rate and density are captured in real time through the built-in airflow sensor, and the oil fume diffusion trend is dynamically analyzed in combination with the AI ​​model algorithm; based on sensor data such as firepower size and PM2.5 value, the AI ​​model automatically matches the wind speed gear, automatically switches from the stir-fry gear to the high-speed mode, optimizes the airflow path through the centrifugal impeller serrated structure, reduces noise while improving separation efficiency, and automatically starts the range hood when the gas stove is ignited, and delays closing after it is turned off. When gas leaks or CO exceeds the standard, the range hood is automatically turned on and the gas valve is closed to implement security linkage; the discharged high-temperature oil fume first contacts the low-temperature condensation plate, and is Due to the temperature difference, part of the grease quickly liquefies and adheres to the surface to be intercepted by the condensation plate, and the temperature of the condensation plate is usually more than 50°C lower than the temperature of the oil smoke, thereby accelerating the conversion of gaseous grease into liquid and achieving oil smoke condensation separation; then when the oil smoke passes through the dense stainless steel oil net or the long strip grille, the grease particles are intercepted and adsorbed due to inertial collision to achieve multi-layer filtration, and some oil nets use oleophobic coatings to reduce grease adhesion residues and improve filtration efficiency, thereby achieving oil net contact separation; and the high-speed rotating serrated impeller (speed ≥ 800rpm) is used to generate centrifugal force to throw the remaining grease to the separation plate or the inner wall of the volute, and the radial groove design on the turbine surface can enhance the grease separation effect, and the separation rate is increased by 15%-20%, thereby achieving centrifugal separation; finally, the separated liquid grease flows into the oil collecting box along the oleophobic coating air duct, or a straight-through oil guide pipe is used to reduce residues, and the oil box is equipped with a silicone sealing ring to prevent grease backflow or leakage; the grease separation degree, oil fume concentration and other quality data of the range hood are obtained through grease separation detection equipment.

[0126] S55. The processing center compares the oil fume separation quality information with the oil fume separation quality standard of the range hood stored in the memory: if the standard is met, it is set as the official oil fume separation control parameter; if the standard is not met, it is transmitted to the alarm and notified to re-debugging.

[0127] Further explanation: the oil fume separation quality standards of the range hood include: 1) the new national standard (GB / T 17713-2025) oil separation degree ≥85% (old national standard ≥80%), some high-end models can reach more than 90%, while the brand flagship model is ≥90%, the mid-to-high-end model is 80%-90%, and the ordinary model is 70%-80%. Therefore, ordinary models with oil separation degrees of 70%-80% or mid-to-high-end models below 85% are eliminated, and the oil fume concentration is ≤2.0mg / m 3; 2) The normal odor reduction degree ≥ 90% and the continuous purification effect lasts within 30 minutes, while the instantaneous odor reduction degree ≥ 60% (old national standard ≥ 50%) and it has the ability to reduce kitchen odors within 3 minutes; 3) The air volume ≥ 7m 3 / min (old national standard ≥ 10m 3 / min), the mainstream models are 12 - 17m 3 / min. The high air volume improves the oil fume inhalation efficiency and can simulate the real air volume under the actual exhaust resistance (such as a shared flue); 4) When the air volume ≤ 12m 3 / min, the noise ≤ 72dB, when the air volume > 12m 3 / min, the noise ≤ 73dB; 5) The maximum static pressure ≥ 350Pa, cancel the old national standard "standard wind pressure ≥ 100Pa". The larger the value, the stronger the exhaust capacity and the better the anti - oil fume backflow effect; 6) Tested in a semi - anechoic chamber, the result is lower than the old standard, but the actual use noise will be slightly higher due to environmental reflection; 7) The working noise needs to be comprehensively evaluated in combination with the kitchen space and exhaust resistance. Low - noise models pay more attention to the duct design and motor matching; Then, compare these quality information with the calculation results of the above steps S51 - S53 and input them into the model for matching to obtain quality anomalies.

[0128] S56. The anomaly recognition unit matches the real - time obtained oil fume separation quality detection image data of the range hood with the corresponding image data in the trained convolutional neural network model and confirms its anomaly category for the subsequent rectification of anomaly problems.

[0129] Further explanation: Input the image into the model, use the trained model to predict the new monitored image data, and determine the abnormal type of the oil fume separation quality of the range hood according to the output result. If the model's judgment of the abnormal oil fume separation quality reaches a certain threshold, trigger the early warning system and take timely rectification measures; the abnormal types of the oil fume separation quality of the range hood include: 1) Incomplete oil fume separation (smoke leakage): The blunt body wake effect of oil fume dispersion caused by the air flow generating a reflux vortex when the air inlet is blocked by a human body or an object, or the boundary layer separation caused by the oil fume particles hitting the body wall to form a reflux bubble and the high-concentration oil fume diffusing into the environment, or the improper installation height caused by the traditional top-suction range hood being installed too high (>75 cm), resulting in significant suction attenuation and reduced separation efficiency. Therefore, select a low-suction range hood (≤40 cm from the stove top) to reduce suction attenuation and shorten the suction port distance, thereby optimizing the exhaust port structure, and use a silicone rubber sealing ring to reinforce the oil box interface to prevent oil leakage and strengthen the sealing design; 2) Decreased exhaust efficiency: The filter screen / oil net is blocked due to the grid gap being reduced by ≥0.5 mm due to oil scale deposition, hindering the passage of oil fume, or the exhaust path is blocked due to the hose being bent or the check valve rusting, causing backflow and increasing the flue resistance, or the motor power decays due to long-term high-load operation, resulting in a decrease in suction and motor aging. Therefore, clean the oil net / filter screen monthly and disassemble and clean the impeller and air duct every six months for deep cleaning, and check the rotational flexibility of the check valve every quarter. When rusted, replace the bearing shaft for check valve maintenance; 3) Oil leakage or oil stain backflow: The oil box seal fails due to the aging or deformation of the oil box sealing strip, causing grease to leak into the body, or the condensate backflows along the pipeline to the stove top due to the grease condensing after the oil guide pipe is not cleaned in time, resulting in condensate backflow. Therefore, adjust the height of the side-suction installation ≤40 cm and the top-suction ≤75 cm to match the suction range, and keep the exhaust hose horizontal or slightly upward to reduce bends (≤2 places) for flue pipe layout; 4) Abnormal noise and vibration: Vibration occurs during high-speed operation due to uneven weight distribution of the impeller caused by oil contamination, resulting in impeller imbalance, or foreign objects enter the air duct after the filter screen is damaged and rub against the turbine, causing noise and resulting in air duct foreign objects. Therefore, turn on the range hood 1 minute before cooking and delay for 3 minutes after cooking to exhaust the remaining smoke for pre-start and delayed shutdown, and close the doors and windows during cooking to reduce the interference of side wind on the stability of the negative pressure area for environmental air flow control; etc. Return to step S54 according to these abnormal problems to re-adjust the control parameters and make improvements until all quality standards are met.

[0130] S60. The application control module controls the range hood to perform oil fume separation operation according to the oil fume separation control parameters, so that the actual oil fume separation effect reaches the highest national standard, to ensure the quality effect of oil fume separation, and transmits it to the processing center;

[0131] Please refer to Figure 14 As shown, the step S60 includes the following steps:

[0132] S61. The quality prevention unit confirms whether the quality abnormality prevention measures predetermined by the model are in place item by item to ensure that the oil fume separation operation can achieve the predetermined effect, and transmits it to the separation setting unit;

[0133] It is further explained that the quality abnormality prevention measures predetermined by the model are: 1) Use a low-suction range hood, and its installation position should be ≤40cm away from the stove to reduce suction attenuation and shorten the distance from the air intake port, thereby optimizing the exhaust port structure; and use a silicone sealing ring to reinforce the oil box interface to prevent oil leakage and strengthen the sealing design. This can prevent the human body or objects from blocking the air inlet, causing the air flow to produce a reflux vortex and cause the fumes to escape. The blunt body wake effect can also be prevented, and the fumes particles can be prevented from hitting the body wall to form reflux bubbles, causing high-concentration fumes to diffuse into the environment and cause boundary layer separation. It can also avoid installing a traditional top-suction range hood, and installing a high 1) The distance from the stove to the stove is >75cm, which causes the suction force to attenuate significantly and the separation efficiency to decrease, resulting in improper installation height, thus avoiding the phenomenon of incomplete oil fume separation (smoke leakage); 2) Clean the oil net every 1-2 weeks, clean the filter once a month, disassemble and clean the impeller and air duct every 3 months, and check the rotation flexibility of the check valve once a quarter. If there is rust, replace the bearing shaft in time and inspect the check valve to ensure timely deep cleaning. When the grease thickness is ≥0.5mm, clean it immediately to avoid the mesh gap being reduced by ≥0.5mm due to grease deposition, which hinders the passage of oil fume and causes the filter / oil net to be blocked. This can prevent the exhaust path from being blocked due to hose bending or rusting of the check valve, causing backflow and increasing smoke duct resistance. It can also prevent the motor power from attenuating due to long-term high-load operation, which will cause the suction to drop and lead to motor aging, thereby preventing grease deposition from causing a decrease in oil fume separation efficiency; 3) Use side suction installation with a height of ≤40cm and top suction installation with a height of ≤75cm to adapt to the suction range, set the exhaust hose horizontally or slightly upward to reduce the number of bends to ≤2, and layout the smoke pipe. This can avoid aging or deformation of the oil collection box sealing strip, which may cause grease to leak into the inside of the fuselage, resulting in failure of the oil box seal, and can also prevent failure to timely Clean the oil guide pipe, so that the grease condenses and flows back to the stove along the pipe, causing the condensate to flow back, thus avoiding oil leakage or oil backflow; 4) Turn on the range hood 1 minute before cooking, and delay exhausting the remaining smoke for 3 minutes after the end, and perform pre-start and delayed shutdown, and close the doors and windows when cooking to reduce the side wind from interfering with the stability of the negative pressure area, and control the ambient airflow, so as to avoid uneven distribution of impeller weight due to oil stains, causing vibration during high-speed operation and causing impeller imbalance, and also prevent foreign matter from entering the air duct due to damage to the filter, causing noise caused by friction with the turbine, resulting in foreign matter in the air duct, thus avoiding abnormal noise and vibration. Before using the range hood, implement the above preventive measures one by one, and readjust them if they are not in place, to ensure that the quality of oil fume separation of the range hood meets the national standard.

[0134] S62. The separation setting unit implements the oil fume separation improvement preventive measures preset by the model item by item for the range hood to ensure that the oil fume separation degree reaches the preset standard and transmits it to the condensation separation unit;

[0135] Further explanation, the oil fume separation improvement preventive measures preset by the model include: 1) Structural optimization: Adopt a three-stage filtration system of "condensation plate + stainless steel oil mesh + centrifugal throwing and suction" to increase the grease interception efficiency to more than 90%. At the same time, optimize the structure of the centrifugal impeller. The centrifugal force generated by high-speed rotation throws the oil droplets to the inner wall to reduce the grease entering the motor and the exhaust duct. And use intelligent variable frequency technology to automatically adjust the air volume according to the oil fume concentration to balance the air pressure and separation efficiency, reduce the risk of grease escape under high load, and achieve dynamic air duct adjustment; 2) Material and coating upgrade: Spray oil-repellent coatings such as nano-level polytetrafluoroethylene on the surface of the oil mesh, inner cavity and impeller to reduce the grease adhesion rate by 30%-50%, reduce the cleaning frequency, and select a corrosion-resistant stainless steel filter mesh to avoid pore blockage caused by high-temperature oxidation. At the same time, it is equipped with an electric heating or steam self-cleaning function, which is started once every 3 months to dissolve the oil stains on the inner wall regularly and maintain the stability of the separation efficiency to achieve the integration of self-cleaning technology; 3) Operating parameter optimization: Match the air volume and air pressure. If it is a household kitchen, select a model with an air volume ≥ 15m 3 / min and an air pressure ≥ 300Pa to ensure rapid inhalation and effective separation of oil fume. If it is a high-rise residence, the air pressure ≥ 400Pa to avoid poor exhaust caused by the resistance of the common flue. At the same time, optimize the electric heating module to control the oil fume temperature ≤ 150°C, reduce the adhesion of high-temperature oil fume, and install a deflector in the exhaust duct to extend the residence time of the oil fume and increase the natural sedimentation rate of the oil droplets; 4) Maintenance and management: Check the tightness of the exhaust duct to avoid air leakage or condensate reflux interfering with the separation process, and configure an oil fume purifier with an electrostatic adsorption or UV photolysis module to additionally increase the grease separation efficiency by 5%-10%, especially for commercial kitchens. Before turning on the range hood, implement the above preventive measures item by item. If not in place, readjust. If all are in place, the user can be allowed to use. This can not only significantly improve the grease separation performance of the range hood, but also extend the equipment life and reduce the maintenance cost. Before using the range hood, after ensuring that all measures in the above step S61 are in place, and implement the above preventive measures item by item. If not in place, readjust to ensure that the oil fume separation degree of the range hood reaches the preset highest standard.

[0136] S63. The condensation separation unit realizes the condensation separation of high-temperature oil fume through condensation plate interception, grease adhesion separation, oil droplet collection and diversion, and residual condensation, so as to realize subsequent contact separation by the oil mesh and transmit it to the contact separation unit;

[0137] It is further explained that when the ignition switch of the gas stove is turned on according to the preset operating parameters and operating instructions, the range hood is automatically started, and the fire power size and PM2.5 value are obtained through the set sensor. The input system automatically matches the wind speed gear, and can automatically switch from the stir-fry gear to the high-speed mode. When the gas leaks or the CO exceeds the standard, the range hood is automatically turned on and the gas valve is closed in conjunction; then the high-temperature oil smoke discharged by the range hood rises to the smoke-proof condensation plate made of tempered glass or metal material under the suction action of the fan, and its surface temperature is significantly lower than the oil smoke temperature by more than 150°C, and some oil particles in the high-temperature oil smoke encounter the condensation plate with a temperature lower than the oil smoke temperature by more than 50°C, which accelerates the gas phase The grease is rapidly cooled and condensed into liquid grease, and adheres to the surface of the condensation plate; through the oleophobic coating such as nano-scale polytetrafluoroethylene or the smooth metal properties on the surface of the condensation plate, the condensed grease is quickly gathered into oil droplets and initially separated from the flue gas, thereby intercepting about 30%-50% of the grease; then the condensation plate is designed at a specific angle such as a 90° open screen smoke barrier to guide the oil droplets to slide along the inclined surface to the oil collecting tank or oil guide channel, or a guide groove is set under the condensation plate to further reduce the secondary evaporation of the oil droplets; finally, the fine oil mist that is not completely separated enters the subsequent separation structure such as the centrifugal impeller or metal filter, and realizes secondary condensation and capture through temperature gradient or physical interception.

[0138] S64, the contact separation unit realizes oil net contact separation of the condensed and liquefied oil smoke through interception and adhesion, multi-layer filtration, convergence and diversion, and residual interception, so as to facilitate subsequent centrifugal separation and transfer to the separation unit;

[0139] It is further explained that according to the predetermined operating parameters and operating commands, the condensed and liquefied oil fume enters the contact separation stage through the separation equipment, and the oil fume airflow passes through the multi-layer staggered arrangement of dense long grille stainless steel and the oil net with oleophobic coating, so as to maximize the contact area and collision probability of the oil mist, reduce the oil adhesion residue and improve the filtration efficiency, and force the liquid oil particles in the oil fume to directly adhere to the surface of the oil net; then, laser welding is used to form dense honeycomb pores with a pore size of 0.5-1.5mm, and larger oil droplets directly hit the oil net due to inertia and adhere to the surface, and tiny oil mist particles generate molecules with the surface of the oil net. Diffusion adsorption, some oil droplets naturally settle to the bottom of the oil net due to the reduction of flow rate, thus achieving physical interception and adhesion, which can intercept about 20-30% of grease; then the intercepted grease gathers into an oil film on the surface of the oil net, and the gravity is used to make the oil droplets slide down by setting the inclination angle of the oil net ≥30°, and the grease is guided into the oil collecting box through the V-shaped oil guide groove set at the edge of the oil net, thereby achieving oil droplet convergence and diversion; finally, the oil mist that is not completely separated enters the subsequent centrifugal impeller to enter the subsequent suction separation stage, but the oleophobic coating such as nano-ceramics on the surface of the oil net can reduce secondary evaporation to ensure stable interception efficiency.

[0140] S65. The centrifugal throwing and suction separation unit realizes centrifugal throwing and suction separation of the oil fume after contact separation through impeller collision, centrifugal separation, converging and guiding, and purification and discharge, so as to ensure the quality of oil fume separation and transmit it to the quality detection unit;

[0141] Further explanation: According to the predetermined operation parameters and operation commands, the residual oil fume after condensation and contact separation enters the radially grooved and serrated centrifugal impeller rotating at high speed along with the air flow, so that the rotation speed is ≥ 1000 r / min, and the turbine blades form a spiral negative pressure area, forcing the oil fume molecules to collide violently with the impeller surface; A strong centrifugal force field of more than 2000G is generated by the impeller rotating at high speed. Since the density of grease (0.9 g / cm 3 ) is significantly greater than that of air, the oil droplets are thrown towards the inner wall of the volute shroud of the volute, while the flue gas continues to flow towards the smoke exhaust port due to its smaller inertia, realizing the physical separation of oil and gas; Then the separated oil droplets form an oil film along the inner wall of the volute. By using the spiral oil guide groove or multi-layer eddy current separation plate arranged in the volute, the oil droplets are completely separated from the air flow after three collisions, and the accumulated oil droplets flow into the oil collection box along the oil nozzle or oil guide pipe of the volute under the action of gravity, or the oil droplets slide down after being accelerated through the oil repellent coating air duct; Finally, the purified gas with a grease content of ≤ 5% after centrifugal separation enters the common flue through the check valve for discharge to avoid oil backflow, and the separated liquid grease flows into the oil collection box along the oil repellent coating air duct or the straight-through oil guide pipe, and a silica gel sealing ring is equipped to prevent oil backflow or leakage. The radially grooved design on the surface of the turbine can enhance the oil separation effect, and the separation rate is increased by 15 - 20%. At the same time, the serrated design on the surface of the turbine can optimize the air flow path, reducing noise and improving the separation efficiency; The oil volume is monitored by the sensor arranged in the grease sedimentation tank to remind the user to clean the oil box, and the high-temperature electric heating technology is used to automatically soften the oil stain to achieve one-key self-cleaning. It can also be controlled by voice commands such as "increase the wind force" or gesture control in the air, reducing manual operation. The user can remotely view the oil stain accumulation status through the APP, and the cleaning suggestions are pushed to the user. When the user finishes using, the range hood will automatically delay closing within a predetermined time without manual shutdown.

[0142] S66. The quality detection unit detects the quality information of grease separation degree, oil fume concentration, odor reduction degree, air volume / noise, and maximum static pressure of oil fume separation through the oil fume separation detection equipment to ensure the accuracy of oil fume separation.

[0143] Further explanation: When the range hood is turned on, it can monitor the kitchen air quality in real time through the installed gas sensors, PM2.5 detectors, CO sensors, etc., and use the door and window sensors or micro switches to detect the ignition status or trigger signals of the stove. Then, it can capture the oil fume flow rate and density in real time through the built-in air flow sensor, and dynamically analyze the oil fume diffusion trend in combination with the AI algorithm. Then, when the automatic detection system is started according to the operation command, the self-check program of each sensor, data acquisition module and communication interface is run to ensure that the core components are fault-free. Then, it detects environmental parameters such as temperature, humidity, and air pressure and automatically calibrates the instruments to eliminate environmental interference. The full-automatic infrared oil analyzer and dust monitor installed at the air outlet of the exhaust duct can transmit data such as oil fume concentration and non-methane total hydrocarbons (NMHC) to the cloud platform in real time. And it can automatically collect oil fume samples using a metal filter cartridge sampling tube, extract the oil with ultrasonic extraction using tetrachloroethylene solvent, and automatically calculate the separation degree through the infrared spectrophotometry formula "separation degree = 1 - oil volume at the outlet / oil volume at the inlet", and simultaneously monitor the particulate matter concentration and noise level to evaluate the comprehensive purification performance.

[0144] S70. The processing center compares the oil fume separation quality information with the oil fume separation quality standard of this range hood stored in the memory: If it meets the standard, it notifies the customer that they can continue to use; if it does not meet the standard, it passes it to the alarm and notifies for debugging or repair.

[0145] Further explanation: In addition to the oil fume separation quality standard described in step S55 above, the oil fume separation quality standard of this range hood also includes: 1) The internal test environment includes the installation height of the range hood, the space size of the test room, the same stove, firepower, cookware, etc.; 2) The external test environment simulates the same exhaust capacity of the flue gas pipeline, such as the auxiliary fan on the top, creates the same amount of oil in the space, and after the range hood works for the same time, weighs the weight of the range hood before and after work, and obtains the final oil separation degree index through the calculation formula; automatically adjusts the air volume (≥7m 3 / min) and air pressure (≥350Pa) according to the real-time oil fume concentration to optimize the separation efficiency. The system automatically compares the inlet / outlet data, calculates the purification efficiency (required to be ≥90%), and determines whether it meets the GB18483 standard (oil fume concentration ≤ 2.0mg / m 3 ); when abnormal oil fume concentration or separation degree is detected, it automatically triggers an alarm and generates a fault report, which is pushed to the management terminal, and prompts to clean the filter or replace the filter cartridge according to the oil deposition data (such as the oil layer thickness ≥ 0.5mm) to ensure long-term separation performance.

[0146] Further explanation: The above steps are shown in sequence according to the arrows. However, these steps do not necessarily need to be executed in the order indicated by the arrows. Unless otherwise clearly stated in this document, there is no strict order restriction for the execution of these steps. These steps can also be executed according to other orders. Moreover, some steps may include multiple sub-steps or multiple stages. These sub-steps or stages do not necessarily need to be completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential either, but can be executed alternately or in rotation with at least a part of other steps or sub-steps or stages of other steps.

[0147] Further explanation: The present invention is described in accordance with the content implemented by the software program of an intelligent oil fume separation control system for an oil fume purifier. For each intelligent oil fume separation method, it is divided into several modules or units to implement the software program instructions generated by each step. The software program instructions include the above-mentioned intelligent oil fume separation method for an oil fume purifier.

[0148] The system of the present invention further includes a computer-readable storage medium containing a memory; the memory stores a computer program, and when each functional module executes the computer program, it realizes the steps of the intelligent oil fume separation method for an oil fume purifier described in any one of the above; the computer-readable storage medium stores a computer program, and when the computer program is executed by each functional module, it realizes the steps of the intelligent oil fume separation method for an oil fume purifier described above; it further includes an intelligent oil fume separation control device for an oil fume purifier, which is realized by using the intelligent oil fume separation method for an oil fume purifier described above.

[0149] Further explanation: The computer-readable storage medium includes a memory, which can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program and its instructions corresponding to the intelligent oil fume separation method for an oil fume purifier in the present invention, including the information transfer instructions of each module; the memory executes the respective functional applications and data processing of each module by running the stored non-volatile software programs and instructions, that is, it realizes the intelligent oil fume separation method for an oil fume purifier in the above embodiment; one or more units are stored in the memory, and when executed by the one or more modules, they execute the intelligent oil fume separation method for an oil fume purifier in any of the above process embodiments; the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by one or more modules, it can also be the intelligent oil fume separation method for an oil fume purifier in any of the above process embodiments.

[0150] Further explanation, the computer-readable medium can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two above. It can be any tangible medium that contains or stores a program, and this program can be used by or in conjunction with an instruction execution system, device, or component. This includes, but is not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. Among them, the computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, carrying computer-readable program code. The propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. It can also be any computer-readable medium other than a computer-readable storage medium. This computer-readable signal medium can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, device, or component. The program code it contains can be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0151] Further explanation, the computer program is divided into multiple modules / units, and the multiple modules / units are stored in the memory and executed by each module / unit to complete the present invention. The multiple modules / units can be a series of computer program instructions that can complete specific functions, and these instructions are used to describe the execution process of the computer program in each module / unit.

[0152] The present invention also provides an intelligent oil fume machine, which is implemented by using the above-mentioned intelligent oil fume separation method of an intelligent oil fume machine.

[0153] Further explanation, the control device is composed of the above-related modules or devices. According to the needs of the user using the range hood, it can be made into a complete set of automatic range hoods together with the range hood oil, or made into a unit device of each functional module belonging to each step or the entire control device. However, the detection device can be installed near the range hood according to the user's needs, or can be removed and taken back to the manufacturer after the manufacturer installs and verifies normal operation. When the range hood has an abnormality and needs to be repaired, the manufacturer's customer service will carry the automatic detection device to the door for detection and repair, and realize wired or wireless connection to each control device and the automatic range hood through wireless connection; it can also be used in a professional range hood detection agency; the structures of these devices are not described in detail here.

[0154] For those skilled in the art, it is obvious that the present application is not limited to the details of the above exemplary embodiments, and the present application can be implemented in other specific forms without departing from the spirit or basic characteristics of the present application. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present application is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be encompassed by the present application.

[0155] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit them. Although the present application has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present application can be modified or equivalently replaced, and all of them should be included in the protection scope of the present application.

Claims

1. A smart range hood oil fume separation method, characterized in that: The method is applied to an intelligent range hood oil fume separation control system, the system comprising an information acquisition module, a model construction module, a model training module, a model evaluation module, a parameter determination module, an application control module, a wireless communication module, a memory, an alarm, a processing center, and an intelligent mobile terminal; the information acquisition module, the model construction module, the model training module, the model evaluation module, the parameter determination module, the application control module, the wireless communication module, the memory, and the alarm are respectively connected to the processing center; the intelligent mobile terminal is respectively connected to the wireless communication module wireless network within the range of a wireless network or the Internet; the method comprises the following steps: S10, before oil fume separation, the information acquisition module obtains normal and abnormal historical image data similar to oil fume separation of range hoods and performs preprocessing for subsequent model construction, and passes it to the model construction module; S20, the model building module determines the model architecture, designs the network hierarchy, adds auxiliary layers, defines model parameters and models the model according to the characteristic information of oil fume separation, so as to facilitate subsequent model training and verification, and transmits the model to the model training module; S30, the model training module trains the model according to the collected data and adjusts the model parameters to optimize the model performance to predetermine the oil fume separation control parameters to ensure that the oil fume separation of the range hood can be automatically designed, and transmits it to the model evaluation module; S40, the model evaluation module performs model verification and evaluation based on the trained model, and adjusts and optimizes the oil fume separation model based on the verification results to improve the performance and accuracy of the model, and passes it to the parameter determination module, including the following steps: S41, the classification and labeling unit classifies the data samples into corresponding data sample categories according to the abnormal categories of oil fume separation quality of the range hood and labels them for the purpose of machine learning of the subsequent model, and transmits them to the learning and training unit; S42, the learning training unit determines the training model according to the gradient descent data and inputs the trained oil smoke separation data into the model for simulation training and continuous iteration and optimization to ensure the accuracy of model recognition, and passes it to the cost verification unit; S43, the cost verification unit calculates the cost function according to the formula "J(θ)=-1 / m∑ m∑K[y k (i)log(h θ (x(i))k)+(1-y k (i))log(1-(h θ (x(i))k)]+λ / 2m∑L-1∑sl∑sl+1(θj,i(l))2, J(θ) is the cost function, m is the total number of training samples, K is the total number of categories, y k (i) is the true label value of the sample, h θ (x(i))k is the sample prediction probability, λ is the regularization parameter, L is the total number of neural network layers, sl is the number of neurons in the lth layer, and θj,i(l) is the weight parameter between the connected neurons. The cost function value after model training is obtained to ensure the accuracy of model evaluation and is passed to the model evaluation unit; S44, the model evaluation unit obtains the comprehensive evaluation score F according to the model evaluation score calculation formula "F = 2 (Ac * Re) / (Ac + Re), F is the training model evaluation score, Ac is the precision rate, and Re is the recall rate", and transmits it to the processing center; S45, the processing center compares the actual evaluation score of the model with the evaluation score standard of the model stored in the memory: if it meets the standard, it is the predetermined oil fume separation control parameter; if it does not meet the standard, it is transmitted to the alarm and notified to continue training; S50, the parameter determination module performs a simulation scenario test based on the oil fume separation control parameters predicted by the trained model to obtain the best oil fume separation control parameters to ensure the subsequent oil fume separation effect, and transmits them to the application control module; S60, the application control module controls the range hood to perform the oil fume separation operation according to the oil fume separation control parameters, so that the actual oil fume separation effect reaches the highest national standard to ensure the quality effect of the oil fume separation, and transmits it to the processing center; S70, the processing center compares the oil fume separation quality information with the oil fume separation quality standard of the range hood stored in the memory: if it meets the standard, the customer is notified to continue using it; if it does not meet the standard, the alarm is transmitted to the alarm and the debugging or maintenance is notified.

2. The intelligent range hood oil fume separation method according to claim 1 is characterized in that: The system further comprises: the wireless communication module is provided with a wireless network unit, which is responsible for sending and receiving wireless signals and automatically networking with the intelligent mobile terminal within an effective network range; The alarm compares the actual evaluation score of the model with the evaluation score standard of the model stored in the memory, and automatically sounds an alarm and notifies to continue training if the standard is not met; and compares the oil fume separation quality information with the oil fume separation quality standard of the range hood stored in the memory, and automatically sounds an alarm and notifies to re-debugging if the standard is not met; and compares the oil fume separation quality information with the oil fume separation quality standard of the range hood stored in the memory, and automatically sounds an alarm and notifies to debug or repair if the standard is not met; The memory is responsible for storing information of the information acquisition module, the model construction module, the model training module, the model evaluation module, the solution determination module, the application control module, the wireless communication module, and the alarm, as well as the model evaluation sub-standard and the range hood oil fume separation quality standard; The processing center is responsible for information acquisition module, model construction module, model training module, model evaluation module, parameter determination module, application control module, wireless communication module, alarm, and information transmission of memory, and is the hub center of the system. The processing center compares the actual evaluation score of the model with the evaluation score standard of the model stored in the memory: if it meets the standard, it is the predetermined oil fume separation control parameter; if it does not meet the standard, it is passed to the alarm and notified to continue training; and compares the oil fume separation quality information with the oil fume separation quality standard of the range hood stored in the memory: if it meets the standard, it is set as the official oil fume separation control parameter; if it does not meet the standard, it is passed to the alarm and notified to re-debugging; and compares the oil fume separation quality information with the oil fume separation quality standard of the range hood stored in the memory: if it meets the standard, the customer is notified to continue using it; if it does not meet the standard, it is passed to the alarm and notified to debug or repair; The information acquisition module includes a data acquisition unit, a cleaning data unit, an enhancement data unit, an integration data unit, a conversion data unit, a filtering and denoising unit, a grayscale conversion unit, and a feature extraction unit, which is responsible for acquiring normal and abnormal historical image data similar to oil fume separation of range hoods, preprocessing them, and passing them to the model building module; The model building module includes a network hierarchy unit, a network parameter unit, a hierarchical optimization unit, and a data partitioning unit, which are responsible for determining the model architecture, designing the network hierarchy structure, adding auxiliary layers, defining model parameters, and modeling according to the characteristic information of oil fume separation, and passing them to the model training module; The model training module includes an input filling unit, a convolution input unit, a pooling conversion unit, a fully connected layer unit, a normalized output unit, an output conversion unit, and a back propagation unit, which is responsible for training the model according to the collected data and adjusting the model parameters to optimize the model performance to predetermine the oil fume separation control parameters and pass them to the model evaluation module; The model evaluation module includes a classification labeling unit, a learning training unit, a cost verification unit, and a model evaluation unit, which is responsible for model verification and evaluation based on the trained model, and adjusting and optimizing the oil fume separation model based on the verification results, and passing them to the parameter determination module; The parameter determination module includes a grease separation unit, a smoke concentration unit, a wind pressure matching unit, a test application unit, and an abnormality recognition unit. The simulated scene test is performed according to the oil smoke separation control parameters predicted by the trained model to obtain the optimal oil smoke separation control parameters, and the parameters are passed to the application control module; The application control module includes a quality prevention unit, a separation setting unit, a condensation separation unit, a contact separation unit, a suction separation unit, and a quality detection unit, and controls the range hood to perform oil fume separation operation according to the oil fume separation control parameters, so that the actual oil fume separation effect reaches the highest national standard and is transmitted to the processing center; The step S20 comprises the following steps: S21, the network level unit uses the selected convolutional neural network as the model architecture and customizes and optimizes the network level structure according to the specific application scenarios and requirements for subsequent defect detection and quality identification, and passes it to the network parameter unit; S22, the network parameter unit adjusts the number of network layers, convolution kernel size, step size, padding method and selects loss function and optimization algorithm according to the network hierarchical structure to ensure the performance and accuracy of the oil fume separation model, and passes it to the hierarchical optimization unit; S23, the hierarchical optimization unit adjusts the size and normalization of the input oil fume separation data by adding a zero padding layer and a normalization layer before and after the convolution layer, so as to improve the training speed and stability of the oil fume separation model, and transmits it to the data division unit; S24. The data partitioning unit divides the data set with key feature vectors into training set, validation set and test set according to the extracted data set and establishes a convolutional neural network structure for subsequent model training and learning. S30, the model training module trains the model according to the collected data and adjusts the model parameters to optimize the model performance to predetermine the oil fume separation control parameters to ensure that the oil fume separation of the range hood can be automatically designed and passed to the model evaluation module.

3. The intelligent range hood oil fume separation method according to claim 1 is characterized in that: The step S10 comprises the following steps: S11. The data acquisition unit obtains historical data of oil fume emission, equipment operation parameters, separation efficiency, and environmental maintenance of similar range hoods through the industry database and performs preprocessing for subsequent data cleaning and transfers it to the cleaning data unit; S12, the cleaning data unit cleans the collected oil fume separation historical data to remove abnormal values, duplicate values ​​or missing values ​​in the data to improve the quality and accuracy of the oil fume separation data, and transmits it to the enhanced data unit; S13, the enhanced data unit increases the quantity and diversity of the oil smoke separation training data through data enhancement methods such as rotation, scaling, flipping, cropping, and color transformation to improve the generalization ability of the model, and passes it to the integrated data unit; S14, the integrated data unit obtains the standardized oil fume separation data according to the data standardization processing formula "z = (x-μ) / σ, z is the standardized data, x is the original data, μ is the mean value of the data, σ is the standard deviation of the data" to improve the stability of the model, and passes it to the conversion data unit; S15, the conversion data unit obtains the normalized oil smoke separation data according to the normalization calculation formula "x'=(x-min(x)) / (max(x)-min(x)), x' is the normalized data, x is the original data, min(x) and max(x) are the minimum and maximum values ​​of the data respectively" to improve the model performance, and passes it to the filtering and denoising unit; S16, filtering and denoising unit is calculated according to the Gaussian filtering method formula G(x,y) is the two-dimensional Gaussian function pixel, (x,y) is the pixel coordinate, and σ is the standard deviation. Obtain the denoised smoke separation image for grayscale image conversion and pass it to the grayscale conversion unit; S17, the grayscale conversion unit obtains the grayscale oil smoke separation image according to the grayscale calculation formula "f(I,j)=max(R(I,j),G(I,j),B(I,j)), f(I,j) is the grayscale image, R(I,j), G(I,j), B(I,j) are the original images of three colors respectively", so as to facilitate the subsequent feature extraction, and transmits it to the feature extraction unit; S18, the feature extraction unit converts the labeled data into feature vectors useful for model training and extracts key features of multi-layer filtering, condensation patterns, separation traces, grease collection, and performance degradation to improve the training speed and performance of the model; The step S50 comprises the following steps: S51, the grease separation unit obtains the grease separation degree according to the range hood grease separation degree calculation formula "η=(M1-M2) / M1*100%, η is the grease separation degree, M1 is the amount of grease entering the range hood, and M2 is the amount of grease discharged from the range hood" for subsequent debugging and testing, and transmits it to the exhaust concentration unit; S52, exhaust concentration unit is calculated according to the range hood exhaust concentration formula "Co = M / [(Q1+Q2+Q3) / ε*(tp-ti)], Co is the range hood exhaust concentration (mg / m 3 ), M is the mass of oil fume emission (mg), Q1 is the heat dissipation of equipment (W), Q2 is the heat dissipation of personnel (W), Q3 is the heat dissipation of lighting (W), ε is the comprehensive heat coefficient (W·℃ / m 3 ), tp is the exhaust temperature (℃), ti is the indoor temperature (℃)” to obtain the oil smoke concentration for subsequent debugging and testing, and pass it to the wind pressure matching unit; S53, the wind pressure matching unit is calculated according to the range hood wind pressure matching formula "Po ≥ ρ*g*(No-no)*ho+ρv2 / 2, Po is the air static pressure value (Pa), ρ is the air fluid density (kg / m 3 ), g is the acceleration due to gravity (m / s 2 ), No is the total number of floors, no is the current number of floors, ho is the height of a single floor (m), and v is the wind speed in the duct (m / s)” to obtain the wind pressure matching value for subsequent debugging and testing, and pass it to the test application unit; S54. The test application unit uses the trained model to match the oil fume separation control parameters with customer needs and preferences to conduct operation tests and obtain the quality information of oil fume separation to ensure the accuracy of the parameters and transmit it to the processing center; S55. The processing center compares the oil fume separation quality information with the oil fume separation quality standard of the range hood stored in the memory: if the standard is met, it is set as the official oil fume separation control parameter; if the standard is not met, it is transmitted to the alarm and notified to re-debugging. S56. The abnormality recognition unit matches the range hood oil fume separation quality detection image data acquired in real time with the corresponding image data in the trained convolutional neural network model and confirms its abnormality category to facilitate subsequent rectification of abnormal problems.

4. The intelligent range hood oil fume separation method according to claim 1 is characterized in that: The step S30 comprises the following steps: S31. The input filling unit obtains the filling data for fume separation according to the filling size calculation formula "Ph = [(Ho-1)*Sh+Kh-Hi] / 2, Pw = [(Wo-1)*Sw+Kw-Wi] / 2, Ph and Pw are the filling sizes in the height / width direction, Hi and Wi are the height / width of the input feature map, Kh and Kw are the height / width of the convolution kernel, Sh and Sw are the values ​​of the step size in the height / width direction, Ho and Wo are the height and width of the output feature map" for subsequent convolution input, and passes it to the convolution input unit; S32, the convolution input unit obtains the convolution input value and activates the function input according to the convolution input calculation formula "z(t) = ∫x(m)y(tm)dm, z(t) is the output function, x(t) is the input function, y(t) is the convolution kernel function, and dm is the integral differential element of the variable m" for subsequent pooling conversion and passes it to the pooling conversion unit; S33, the pooling conversion unit obtains the maximum pixel value after pooling according to the average pooling calculation formula "Z(I,j)=mean(X[i*Ps(i+1)*Ps,j*Ps(j+1)*Ps], Z(I,j) is a pixel value in the output feature map after pooling, i and j are the positions in the output feature map, mean is the average value of all pixel values ​​in the input window, X is the input feature map, and Ps is the window size of the pooling operation" to retain important feature information to enter the fully connected layer and pass it to the fully connected layer unit; S34, the fully connected layer unit obtains the connection layer output data according to the fully connected layer calculation formula "y=f(∑(Wn*Xn)+b), y is the output result, f is the activation function, Wn is the weight of the nth input feature, Xn is the nth input feature, n is the dimension of the input feature, and b is the bias" and enters the output layer for subsequent normalized output calculation and is passed to the normalized output unit; S35, the normalized output unit obtains the output of Batch Norm according to the Batch Norm layer calculation formula "h = φ (Bn (Wx + b)), h is the output of the fully connected layer, φ is the activation function, Bn is the batch normalization operator, W is the weight parameter, X is the input of the fully connected layer, and b is the bias parameter" for subsequent data output conversion and passes it to the output conversion unit; S36, the output conversion unit obtains the output value size after convolution according to the output layer conversion calculation formula "N = (P-F + 2C) / S + 1, N is the output size after convolution, P is the input size before convolution, F is the convolution kernel size, C is the number of layers of 0 added around the image, and S is the step size" for subsequent model analysis and training, and passes it to the back propagation unit; S37, the back propagation unit obtains the convolution output value back propagation according to the back propagation calculation formula "dz[l] = da[l] * g[l]′(z[l]), dz[l] is the input gradient, da[1] is the output gradient, g[l]′(z[l]) is the activation function derivative, and z[l] is the activation function input" to optimize the performance of the network model; The step S60 comprises the following steps: S61. The quality prevention unit confirms whether the quality abnormality prevention measures predetermined by the model are in place item by item to ensure that the oil fume separation operation can achieve the predetermined effect, and transmits it to the separation setting unit; S62, the separation setting unit implements the preventive measures for improving the oil fume separation degree according to the model one by one on the range hood to ensure that the oil fume separation degree reaches the predetermined standard, and transmits it to the condensation separation unit; S63, the condensation separation unit intercepts the high-temperature oil smoke through the condensation plate, separates the grease adhesion, collects and guides the oil droplets, and condenses the residual oil smoke to achieve condensation separation of the oil smoke, so as to facilitate the subsequent oil net contact separation and transfer to the contact separation unit; S64, the contact separation unit realizes oil net contact separation of the condensed and liquefied oil smoke through interception and adhesion, multi-layer filtration, convergence and diversion, and residual interception, so as to facilitate subsequent centrifugal separation and transfer to the separation unit; S65, the separation unit realizes centrifugal separation of the oil fume after contact separation through impeller collision, centrifugal separation, convergence and diversion, purification and discharge to ensure the separation quality of the oil fume, and transmits it to the quality inspection unit; S66. The quality inspection unit detects the quality information of oil separation degree, oil fume concentration, odor reduction, air volume / noise, and maximum static pressure of oil fume separation through oil fume separation detection equipment to ensure the accuracy of oil fume separation.

5. The intelligent range hood oil fume separation method according to claims 1-4 is characterized in that: The system also includes a computer-readable storage medium containing a memory; the memory stores a computer program, and when the functional modules execute the computer program, the steps of an intelligent range hood oil fume separation method as described in any one of claims 1 to 4 are implemented; the computer-readable storage medium stores a computer program, and when the computer program is executed by the functional modules, the steps of an intelligent range hood oil fume separation method as described in any one of claims 1 to 4 are implemented; and also includes an intelligent range hood oil fume separation control device, which is implemented by the intelligent range hood oil fume separation method as described in any one of claims 1 to 4.

6. An intelligent range fumes machine, characterized in that: This is achieved by using an intelligent range hood oil fume separation method as described in any one of claims 1 to 4 above.

Images