Data processing method and device, electronic equipment and storage medium

Abstract

The invention discloses a data processing method and device, electronic equipment and a storage medium, and the data processing method comprises the steps: obtaining a real-time in-cavity image and real-time steam state data in a cooking cavity in a current period; estimating real-time steam quantity data in the cooking cavity based on the real-time in-cavity image and the real-time steam state data to obtain fused steam state data of the current period; determining first steam state data in the target steam state data of the last period based on the fused steam state data; determining second steam state data in the fused steam state data based on the target steam state data of the previous period; determining target steam state data of the current period in the first steam state data and the second steam state data based on the fault probability of the steam generation device; and predicting the fault probability of the steam generation device in the next period based on the target steam state data of the current period. The maintenance cost of the steam generation device is reduced.

Description

Technical Field

[0001] This application relates to the technical field of kitchen appliances, and particularly to a data processing method, apparatus, electronic device, and storage medium. Background Art

[0002] In modern kitchens, steam generation devices, as the core components of steam ovens and multifunctional cookers, are crucial for ensuring the quality of food and the efficiency of the cooking process. Such devices generate steam through heating to cook food, thereby being able to lock in the nutrients and flavors of ingredients in a short time. However, due to the long-term high-temperature and high-humidity working environment, steam generation devices are prone to problems such as aging of heating elements, scale deposition, and sensor malfunctions. If these problems are not addressed in a timely manner, they may lead to a decline in device performance or even failures.

[0003] In existing cooking equipment, the status monitoring of steam generation devices usually relies on simple sensors and fixed fault detection algorithms. These systems mainly use temperature sensors and humidity sensors to collect data during operation, and compare the collected data with preset safety thresholds for analysis to determine whether the device is in a normal working state. Once it is detected that the steam volume is below or exceeds the set range, the system will trigger an alarm mechanism to inform the user that there may be an abnormal situation with the device.

[0004] However, the above-mentioned fault detection methods have certain limitations, that is, they mainly issue alarms after a fault has occurred and cannot provide early warnings for potential faults. Therefore, this method may lead to a lag in equipment maintenance, not only increasing the maintenance cost, but also prolonging the downtime of the equipment, affecting the continuity and efficiency of production. Summary of the Invention

[0005] To solve the problems of the prior art, embodiments of this application provide a data processing method, apparatus, electronic device, and storage medium. The technical solutions are as follows:

[0006] On the one hand, a data processing method is provided, which is applied to a steam generation device for injecting steam into a cooking cavity. The method includes:

[0007] In the current cycle, obtain the real-time cavity image and real-time steam state data in the cooking cavity;

[0008] Based on the real-time cavity image and the real-time steam state data, estimate the real-time steam volume data in the cooking cavity to obtain the fused steam state data for the current cycle;

[0009] Determine first steam state data from the target steam state data of the previous cycle based on the influence degree of the target steam state data of the previous cycle on the fused steam state data; and determine second steam state data from the fused steam state data based on the correlation degree between the fused steam state data and the target steam state data of the previous cycle;

[0010] Determine the target steam state data of the current cycle from the first steam state data and the second steam state data based on the influence degree of the first steam state data and the second steam state data on the failure probability of the steam generating device;

[0011] Predict the failure probability of the steam generating device in the next cycle based on the target steam state data of the current cycle and the influence degree of the target steam state data of the current cycle on the failure probability of the steam generating device.

[0012] On the other hand, a data processing device is provided, which is applied to a steam generating device for injecting steam into a cooking cavity. The device includes:

[0013] A data acquisition module for acquiring the real-time in-cavity image and real-time steam state data in the cooking cavity in the current cycle;

[0014] A data fusion module for estimating the real-time steam quantity data in the cooking cavity based on the real-time in-cavity image and the real-time steam state data to obtain the fused steam state data of the current cycle;

[0015] A data screening module for determining first steam state data from the target steam state data of the previous cycle based on the influence degree of the target steam state data of the previous cycle on the fused steam state data; and determining second steam state data from the fused steam state data based on the correlation degree between the fused steam state data and the target steam state data of the previous cycle;

[0016] A target screening module for determining the target steam state data of the current cycle from the first steam state data and the second steam state data based on the influence degree of the first steam state data and the second steam state data on the failure probability of the steam generating device;

[0017] A failure prediction module for predicting the failure probability of the steam generating device in the next cycle based on the target steam state data of the current cycle and the influence degree of the target steam state data of the current cycle on the failure probability of the steam generating device.

[0018] In an exemplary embodiment, the data fusion module includes:

[0019] An image segmentation module for segmenting a steam region from the real-time cavity image;

[0020] A pixel processing module for determining the number of pixels corresponding to the steam region and the intensity of each pixel in the pixels corresponding to the steam region;

[0021] A steam amount estimation module for estimating the real-time steam amount data in the cooking cavity based on the number of pixels corresponding to the steam region and the intensity of each pixel in the pixels corresponding to the steam region;

[0022] A state fusion module for fusing the real-time steam state data and the real-time steam amount data to obtain the fused steam state data for the current cycle.

[0023] In an exemplary embodiment, the state fusion module includes:

[0024] A state prediction module for using the fused steam state data of the previous cycle as an input to a steam generation dynamics model, so that the steam generation dynamics model predicts the steam state of the current cycle to obtain a predicted value of the steam state of the current cycle;

[0025] A gain determination module for comparing the real-time steam state data and the real-time steam amount data as steam state observation values with the predicted value of the steam state to determine a target gain;

[0026] A prediction adjustment module for adjusting the predicted value of the steam state based on the target gain and the steam state observation values to obtain the fused steam state data for the current cycle.

[0027] In an exemplary embodiment, the data screening module includes:

[0028] A network acquisition module for acquiring a pre-trained memory network; the memory network includes a forgetting gate in a memory layer;

[0029] A forgetting gate processing module for using the target steam state data of the previous cycle as an input to the forgetting gate to obtain the first steam state data; the parameters of the forgetting gate indicate the influence degree of the target steam state data of the previous cycle on the fused steam state data;

[0030] The memory network further includes an input gate in the memory layer; correspondingly, the data screening module further includes:

[0031] An input gate processing module, configured to use the fused steam state data as the input of the input gate to obtain the second steam state data; parameters of the input gate indicate the degree of association between the fused steam state data and the target steam state data of the previous cycle;

[0032] The memory network further includes an output gate of the memory layer; correspondingly, the target screening module includes:

[0033] An output gate processing module, configured to use the first steam state data and the second steam state data as the input of the output gate to obtain the target steam state data of the current cycle; parameters of the output gate indicate the degree of influence of the first steam state data and the second steam state data on the failure probability of the steam generating device;

[0034] The memory network further includes a fully connected layer; correspondingly, the fault prediction module includes:

[0035] A fully connected layer processing module, configured to use the target steam state data of the current cycle as the input of the fully connected layer to obtain the failure probability of the steam generating device in the next cycle; parameters of the fully connected layer indicate the degree of influence of the target steam state data of the current cycle on the failure probability of the steam generating device.

[0036] In an exemplary embodiment, the device further includes a network training module for training an initial memory network, and the network training module includes:

[0037] A sample acquisition module, configured to acquire sample steam state data corresponding to historical cycles and the true fault state corresponding to the sample steam state data; the true fault state indicates whether the steam generating device fails;

[0038] A network prediction module, configured to input the sample steam state data into the initial memory network to predict the failure probability of the steam generating device to obtain a predicted failure probability;

[0039] A loss determination module, configured to determine the binary cross-entropy loss of the memory network based on the difference between the predicted failure probability and the true fault state;

[0040] A parameter adjustment module, configured to adjust the parameters of the initial memory network based on the binary cross-entropy loss until a preset training end condition is met to obtain the pre-trained memory network.

[0041] In an exemplary embodiment, the device further includes an adaptive control module for adjusting the target control signal of the steam generating device according to the steam amount during the operation of the steam generating device, and the adaptive control module includes:

[0042] A reference acquisition module for acquiring a reference control signal of the steam generating device in the current cycle;

[0043] A control simulation module for simulating the amount of steam generated by the steam generating device under the control of the reference control signal to obtain reference steam amount data;

[0044] An error determination module for determining the error between the steam amount data in the fused steam state data of the current cycle and the reference steam amount data;

[0045] A signal adjustment module for adjusting the reference control signal based on the error and the fused steam state data to obtain a target control signal of the steam generating device in the next cycle.

[0046] On the other hand, an electronic device is provided, including a processor and a memory. At least one instruction or at least one program segment is stored in the memory, and the at least one instruction or the at least one program segment is loaded and executed by the processor to implement the data processing method in any of the above aspects.

[0047] On the other hand, a computer-readable storage medium is provided. At least one instruction or at least one program segment is stored in the computer-readable storage medium, and the at least one instruction or the at least one program segment is loaded and executed by a processor to implement the data processing method in any of the above aspects.

[0048] On the other hand, a computer program product or a computer program is provided. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the electronic device executes the data processing method in any of the above aspects.

[0049] In the embodiments of the present application, by acquiring real-time images and steam state data in the cooking cavity, combining image analysis and data fusion technologies to estimate the real-time steam amount, and based on the data correlation analysis between the current cycle and the previous cycle, the first and second steam state data that have a significant impact on the failure probability of the steam generating device are screened out, and then the target steam state data of the current cycle is determined, and finally the accurate prediction of the failure probability of the steam generating device in the next cycle is realized. Effectively improves the accuracy and timeliness of fault warning, helps preventive maintenance, and reduces the possibility of sudden equipment failures. Description of the Drawings

[0050] To more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on these drawings.

[0051] Figure 1 is a schematic flowchart of a data processing method provided by an embodiment of the present application;

[0052] Figure 2 is a schematic flowchart of another data processing method provided by an embodiment of the present application;

[0053] Figure 3 is a structural block diagram of a data processing device provided by an embodiment of the present application;

[0054] Figure 4 is a hardware structural block diagram of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0055] The following will clearly and completely describe the technical solutions in the embodiments of the present application in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0056] It should be noted that the terms "first", "second", etc. in the description and claims of the present application and the above accompanying drawings are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence. It should be understood that such data used can be interchanged under appropriate circumstances so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0057] It can be understood that in the specific implementation manners of the present application, when it comes to data related to user information, etc., when the above embodiments of the present application are applied to specific products or technologies, user permission or consent needs to be obtained, and the collection, use, and processing of relevant data need to comply with relevant laws, regulations, and standards of relevant countries and regions.

[0058] Please refer to Figure 1, which shows a schematic flowchart of a data processing method provided by an embodiment of the present application. It should be noted that this specification provides method operation steps such as in the embodiment or flowchart, but based on routine or non-creative labor, there may be more or fewer operation steps. The step order listed in the embodiment is only one way among the execution orders of numerous steps and does not represent the only execution order. When the actual system or product is executed, it can be executed in the order shown in the embodiment or the drawings, or executed in parallel (for example, in an environment of parallel processors or multi-threaded processing). Specifically, as Figure 1 shown, this method is applied to a steam generating device, and the steam generating device is used to inject steam into the cooking cavity. This method may include:

[0059] S101, in the current cycle, obtain the real-time in-cavity image and real-time steam state data in the cooking cavity.

[0060] Specifically, in the current cycle, the system performs a complete data acquisition, processing, and prediction once.

[0061] Among them, the cooking cavity refers to the space inside the cooking device for placing food for heating or cooking.

[0062] Among them, the real-time in-cavity image refers to the image inside the cooking cavity taken in the current cycle, which is used to reflect the real-time steam state. In specific implementation, an image acquisition device is set to acquire the real-time image in the cooking cavity.

[0063] Among them, the real-time steam state data refers to the data obtained in real time for monitoring the working state of the steam generating device. Specifically, the real-time steam state data is usually provided by various sensors. For example, a temperature sensor is set to measure the temperature T(t) in the cooking cavity, a humidity sensor is set to measure the humidity H(t) in the cooking cavity, and a gas sensor is set to measure the steam concentration G(t) in the cooking cavity.

[0064] S103, based on the real-time in-cavity image and the real-time steam state data, estimate the real-time steam quantity data in the cooking cavity to obtain the fused steam state data for the current cycle.

[0065] Among them, the real-time steam quantity data is the rough data reflecting the steam quantity in the cooking cavity in the current cycle. Specifically, image processing technology is used to identify the features of the real-time in-cavity image to preliminarily estimate the steam quantity data.

[0066] Among them, the fused steam state data refers to the data formed by combining real-time steam state data and real-time steam volume data, which helps to more comprehensively evaluate the steam condition in the cavity and provides a richer information basis for subsequent analysis. For example, the Kalman filtering algorithm is used to fuse multi-sensor data to obtain a more accurate steam volume estimate. In specific implementation, the fused steam state data may include more accurate steam volume, temperature, humidity and other data.

[0067] Specifically, an image acquisition device, a temperature sensor, a humidity sensor and a gas sensor are used to collect multi-dimensional data in the cooking cavity in real time, and multi-sensor data fusion is carried out by combining the data of the temperature, humidity and gas sensors to solve the problem of single-sensor data dependence. More accurate steam state data in the cooking cavity helps to achieve higher-precision fault detection.

[0068] In an exemplary embodiment, the above step S103 may include the following steps:

[0069] Segment the steam area from the real-time cavity image;

[0070] Determine the number of pixels corresponding to the steam area, and the intensity of each pixel in the pixels corresponding to the steam area;

[0071] Estimate the real-time steam volume data in the cooking cavity based on the number of pixels corresponding to the steam area and the intensity of each pixel in the pixels corresponding to the steam area; [[ID=1�]]

[0072] Fuse the real-time steam state data and the real-time steam volume data to obtain the fused steam state data of the current cycle.

[0073] Specifically, an image processing algorithm is used to identify the image features of the real-time cavity image and calculate the real-time steam volume data V s (t). Specifically, preprocessing such as denoising and enhancement is performed on the real-time cavity image, edge detection algorithms such as Canny edge detection are used to extract image features, the Kmeans clustering algorithm is used to segment the image into a steam area and a non-steam area, and the steam volume is estimated by calculating the number of pixels and the average intensity of the steam area to obtain the real-time steam volume data. Formula (1) is the calculation formula for the real-time steam volume data:

[0074]

[0075] Among them, N is the number of pixels in the steam area, P i is the area of the i-th pixel, and I i is the intensity of the i-th pixel.

[0076] Specifically, the fused steam state data of the previous cycle is used as the input of the steam generation kinetic model, so that the steam generation kinetic model predicts the steam state of the current cycle to obtain the predicted value of the steam state of the current cycle; the real-time steam state data and the real-time steam volume data are used as the steam state observation values, compared with the predicted value of the steam state, and the target gain is determined; based on the target gain and the steam state observation values, the predicted value of the steam state is adjusted to obtain the fused steam state data of the current cycle.

[0077] In specific implementation, the Kalman filtering algorithm is used to fuse the multi-sensor data to obtain a more accurate steam volume estimation V f (t). Specifically, referring to formulas (2) to (8), where formula (2) is the state equation, formula (3) is the observation equation, formulas (4) and (5) are the prediction step equations, and formulas (6) to (8) are the update step equations:

[0078] X(t) = AX(t - 1) + BU(t) + W(t) (2)

[0079] Among them, X(t) is the state variable, including the real-time steam state data and the real-time steam volume data. The real-time steam state data includes temperature, humidity, etc. A is the state transition matrix, U(t) is the control input, B is the control matrix, and W(t) is the process noise.

[0080] Z(t) = HX(t) + V(t) (3)

[0081] Among them, Z(t) is the steam state observation value, including the real-time steam volume data V s (t), temperature T(t), humidity H(t), gas concentration G(t). H is the observation matrix, and V(t) is the observation noise.

[0082]

[0083] P(t|t - 1) = AP(t - 1|t - 1)A T + Q (5)

[0084] Among them, is the predicted value of the steam state, P(t|t - 1) is the prediction error covariance, and Q is the process noise covariance.

[0085] K(t) = P(t|t - 1)H T (HP(t|t - 1)H T + R) -1 (6)

[0086]

[0087] P(t|t) = (I - K(t)H)P(t|t - 1) (8)

[0088] Wherein, K(t) is the target gain, R is the observation noise covariance, is the updated state, that is, the fused steam state data, and P(t|t) is the updated error covariance.

[0089] By comparing the fused steam state data V f (t) with the preset steam volume threshold V th , it is determined whether the steam generating device is in a fault state. If V f (t) ≥ V th , the steam generating device is in a normal state. If V f (t) < V th , the steam generating device is in a fault state. Refer to formula (9):

[0090]

[0091] As can be seen from the above technical solutions of the embodiments of the present application, through multi-sensor data fusion and the Kalman filtering algorithm, the multi-dimensional data is comprehensively analyzed, the accuracy and reliability of steam volume estimation are improved, the situations of false alarms and missed alarms are effectively reduced, and thus a higher-precision fault detection is achieved.

[0092] S105. Based on the influence degree of the target steam state data of the previous cycle on the fused steam state data, determine the first steam state data in the target steam state data of the previous cycle; and based on the correlation degree between the fused steam state data and the target steam state data of the previous cycle, determine the second steam state data in the fused steam state data.

[0093] Wherein, the target steam state data is further screened from the first steam state data and the second steam state data in the corresponding cycle.

[0094] Wherein, the first steam state data reflects how the actual steam state of the current cycle is affected by past operations. Specifically, the influence degree of the first steam state data in the target steam state data of the previous cycle on the fused steam state data is higher than that of other data in the target steam state data of the previous cycle on the fused steam state data.

[0095] Wherein, the second steam state data reflects the actual steam state of the current cycle and has established a connection with historical data, which is helpful for predicting future trends. Specifically, the correlation degree between the second steam state data in the fused steam state data and the target steam state data of the previous cycle is higher than that between other data in the fused steam state data and the target steam state data of the previous cycle.

[0096] Specifically, by considering the target steam state data of the previous cycle and analyzing its influence on the actual steam state of the current cycle, the most representative state data is selected as the first steam state data; based on the correlation between the actual steam state of the current cycle and the target steam state data of the previous cycle, the key state information in the current cycle is screened out as the second steam state data.

[0097] Specifically, the fused steam state data includes the steam volume V(t), temperature T(t), humidity H(t), gas concentration G(t), etc. during the operation of the steam generating device in the current cycle. Preprocessing is performed on these data, including steps such as denoising, normalization, and outlier processing. Assume that the preprocessed dataset is D = {(V i , T i , H i , G i , y i )}, where y i represents the state (normal or faulty) of the steam generating device. Features are extracted from it to construct the feature vector φ(t), including the steam volume, temperature, humidity, gas concentration, and their change rates, etc. The expression of the feature vector φ(t) can refer to formula (10):

[0098] φ(t) = [V(t), T(t), H(t), G(t), ΔV(t), ΔT(t), ΔH(t), ΔG(t)] (10)

[0099] Among them, the expressions of the steam volume change rate ΔV(t), temperature change rate ΔT(t), humidity change rate ΔH(t), and steam concentration change rate ΔG(t) can refer to formulas (11) - (14): T

[0100] ΔV(t) = V(t) - V(t - 1) (11)

[0101] ΔT(t) = T(t) - T(t - 1) (12)

[0102] ΔH(t) = H(t) - H(t - 1) (13)

[0103] ΔG(t) = G(t) - G(t - 1) (14)

[0104] Based on the interdependent relationship between the processed fused steam state data and the target steam state data of the previous cycle, the first steam state data and the second steam state data are determined.

[0105] S107. Based on the influence degrees of the first steam state data and the second steam state data on the failure probability of the steam generating device, the target steam state data of the current cycle is determined from the first steam state data and the second steam state data.

[0106] Specifically, the influence degree of the target steam state data in the first steam state data and the second steam state data on the failure probability of the steam generating device is higher than that of other data in the first steam state data and the second steam state data on the failure probability of the steam generating device.

[0107] S109. Predict the failure probability of the steam generating device in the next cycle based on the target steam state data in the current cycle and the influence degree of the target steam state data in the current cycle on the failure probability of the steam generating device.

[0108] Specifically, by ignoring irrelevant data, analyze historical data and real-time data in the time series to predict future failure probability.

[0109] As can be seen from the above technical solutions of the embodiments of the present application, the embodiments of the present application estimate the real-time steam volume by acquiring the real-time image and steam state data in the cooking cavity, combining image analysis and data fusion technologies, and based on the data correlation analysis between the current cycle and the previous cycle, screen out the first and second steam state data that have a significant impact on the failure probability of the steam generating device, and then determine the target steam state data in the current cycle, and finally achieve accurate prediction of the failure probability of the steam generating device in the next cycle. Effectively improve the accuracy and timeliness of failure warning, contribute to preventive maintenance, and reduce the possibility of sudden equipment failure.

[0110] In an exemplary embodiment, as Figure 2 shown, the above steps S105 to S109 may include:

[0111] S201. Obtain a pre-trained memory network.

[0112] Wherein, the memory network at least includes a memory layer and a fully connected layer, and the memory layer at least includes a forgetting gate, an input gate, and an output gate.

[0113] Specifically, a prediction model is established through machine learning algorithms (such as time series prediction, deep learning), and potential failures are identified in advance according to the output of the prediction model, and preventive measures are taken to achieve predictive maintenance.

[0114] In specific implementation, the memory network can select LSTM (Long Short-Term Memory). Specifically, the deep learning algorithm Long Short-Term Memory (LSTM) is used to construct a prediction model. LSTM is suitable for processing time series data and can capture the laws of device state changes over time. LSTM is a special type of Recurrent Neural Network (RNN), which is designed to solve the problem of gradient disappearance or gradient explosion encountered by traditional RNNs when processing long sequence data. LSTM controls the flow of information over time by introducing "memory cells" and three gating mechanisms (forget gate, input gate, output gate), enabling the network to learn long-term dependencies.

[0115] The following are the key components of LSTM and their functions: The input layer receives the external feature vector φ(t) as input; the LSTM layer captures long-term dependencies in time series data, including multiple LSTM units. Each unit contains a memory cell and three gating mechanisms inside. The forget gate determines which information needs to be discarded from the memory cell, the input gate determines which new information needs to be stored in the memory cell, the candidate memory cell is responsible for generating new candidate memory values, the memory cell updates the state of the memory cell according to the results of the forget gate and the input gate, and the output gate determines which part of the memory cell state will be passed as output to the next cycle; the fully connected layer maps the output of the LSTM layer to the final prediction result, that is, predicts the failure probability of the future device; the output layer gives the final output, which is the probability P f (t) of the device failing at a certain future time point.

[0116] S203: Use the target steam state data of the previous cycle as the input of the forget gate to obtain the first steam state data.

[0117] Among them, the parameters of the forget gate indicate the influence degree of the target steam state data of the previous cycle on the fused steam state data, including at least the weight W f and the bias parameter b f . Among them, the target steam state data of the previous cycle is the hidden state output by the LSTM layer in the previous cycle.

[0118] Specifically, the forget gate receives the fused steam state data of the current cycle and the target steam state data of the previous cycle, and outputs a vector. Each element of this vector is a number between 0 and 1, indicating the degree of retaining the old state. Formula (15) is the calculation formula of the forget gate:

[0119] f t =σ(W f ·[h t-1 , φ(t)] + bf ) (15)

[0120] Among them, f t is the output of the forget gate, which controls which information in the state C of the old memory unit t-1 should be forgotten, h t-1 is the hidden state of the previous cycle, that is, the target steam state data of the previous cycle, and σ is the sigmoid activation function.

[0121] S205: Use the fused steam state data as the input of the input gate to obtain the second steam state data.

[0122] Among them, the parameters of the input gate indicate the degree of association between the fused steam state data and the target steam state data of the previous cycle, and at least include the weight W i and the bias parameter b i

[0123] Specifically, the input gate receives the fused steam state data of the current cycle and the target steam state data of the previous cycle. First, it determines which values need to be updated through a sigmoid layer, then creates a candidate memory unit vector through a tanh layer, and finally multiplies these two vectors to obtain the information that actually needs to be added to the memory unit, that is, the second steam state data. Formula (16) is the calculation formula of the input gate:

[0124] i t =σ(W i ·[h t-1 , φ(t)] + b i ) (16)

[0125] Among them, i t is the output of the input gate, which controls which information in the new candidate memory unit C t should be stored.

[0126] Specifically, the candidate memory unit generates a new candidate memory value in the current cycle, and this value will be controlled by the input gate to determine whether it should be written into the memory unit. The candidate memory unit is generated through the tanh activation function to ensure that the generated value is between 1 and +1, which is convenient for subsequent weighted operations. Combine the output of the forget gate and the output of the input gate to update the state of the memory unit. This step is actually multiplying the state of the old memory unit by the output of the forget gate and then adding the output of the input gate. Formula (17) is the calculation formula of the candidate memory unit, and formula (18) is the calculation formula for updating the memory unit:

[0127]

[0128]

[0129] Among them, is the candidate memory cell value of the current cycle, W C is the weight of the candidate memory cell, b C is the bias parameter of the candidate memory cell, and tanh is the hyperbolic tangent function, which is used to limit the output of the candidate memory cell between -1 and 1; C t is the memory cell state of the current cycle.

[0130] S207: Use the first steam state data and the second steam state data as the input of the output gate to obtain the target steam state data of the current cycle.

[0131] Among them, the parameters of the output gate indicate the degree of influence of the first steam state data and the second steam state data on the failure probability of the steam generating device, including the weight W o and the bias parameter b o .

[0132] Among them, the target steam state data of the current cycle is the hidden state output by the LSTM layer in the current cycle.

[0133] Specifically, the output gate uses the Sigmoid layer to determine which memory cell states need to be output, then processes these states through the tanh layer, and finally multiplies them by the output of the Sigmoid layer to obtain the final output and update the hidden state. Formula (19) is the calculation formula of the output gate, and formula (20) is the formula for updating the hidden state:

[0134] o t = σ(W o ·[h t-1 , φ(t)] + b o ) (19)

[0135] h t = o t ·tanh (C t ) (20)

[0136] Among them, h t is the hidden state of the current cycle, that is, the target steam state data of the current cycle, and o t is the output of the output gate of the current cycle.

[0137] Specifically, the gating mechanism in the LSTM network structure is controlled by the sigmoid activation function, which generates a value between 0 and 1 to represent the degree of opening of the gate. The tanh activation function is used to generate the new state of the candidate memory cell. The weights and bias parameters of these gating mechanisms will be adjusted through the backpropagation algorithm during the training process.

[0138] S209. Use the target steam state data of the current cycle as the input of the fully connected layer to obtain the failure probability of the steam generator in the next cycle.

[0139] Among them, the parameters of the fully connected layer indicate the influence degree of the target steam state data of the current cycle on the failure probability of the steam generator.

[0140] Specifically, the fully connected layer is a part of the LSTM architecture and is located after the LSTM layer. It is mainly responsible for mapping the output of the LSTM layer to the final prediction result. The fully connected layer connects the output of the LSTM layer to all neurons in this layer, integrates the information provided by the LSTM layer, and converts this information into the required output form, such as classification probability or regression value. Specifically, it receives the output from the LSTM layer, which is usually a multi-dimensional vector. Each neuron has a weight matrix that maps the output of the previous layer to the output of this layer, and each neuron has a bias vector used to adjust the activation threshold of the neuron. The output layer maps the output of the fully connected layer to a probability value between 0 and 1, or converts the output into the probability distribution of each category.

[0141] In specific implementation, during the operation of the steam generator, the current feature vector φ(t) is input into the trained LSTM model in real time, and the failure probability P f (t + k) in the future for a period of time is predicted according to formula (21):

[0142] P f (t + k) = LSTM(φ(t), φ(t - 1), …, φ(t - n)) (21)

[0143] Among them, k is the prediction step size, and n is the time window size.

[0144] According to the predicted failure probability P f (t + k), set the failure probability threshold P th . When P f (t + k) ≥ P th , trigger a maintenance alarm to prompt the user for preventive maintenance, specifically referring to formula (22):

[0145]

[0146] Assume that the set failure probability threshold P th = 0.8, the time window size n = 10, and the prediction step size k = 5. At a certain moment t, the real-time collected feature vector φ(t) = [0.75, 100, 50, 0.8, 0.05, 0.5, -0.2, 0.03]. Input the feature vector φ(t) into the LSTM model to obtain the failure probability P of the future 5 step sizes shown in formula (23) f(t + 5):

[0147] P f (t + 5) = LSTM(φ(t), φ(t - 1), …, φ(t - 10)) (23)

[0148] Assume the predicted failure probability is P f (t + 5) = 0.85, due to P f (t + 5) ≥ P th , trigger a maintenance alarm.

[0149] As can be seen from the above technical solutions of the embodiments of the present application, by establishing a memory network through a machine learning algorithm, the embodiments of the present application can predict possible failures in advance, achieve predictive maintenance, not only reduce the unplanned downtime and maintenance costs of the steam generation device, but also improve the reliability and service life of the steam generation device, and can help users take preventive measures in advance to optimize equipment management and maintenance strategies.

[0150] In an exemplary embodiment, the training of the above memory network may include the following steps:

[0151] Obtain the sample steam state data corresponding to the historical period, and the true failure state corresponding to the sample steam state data; the true failure state indicates whether the steam generation device fails;

[0152] Input the sample steam state data into the initial memory network to predict the failure probability of the steam generation device, and obtain the predicted failure probability;

[0153] Based on the difference between the predicted failure probability and the true failure state, determine the binary cross-entropy loss of the memory network;

[0154] Based on the binary cross-entropy loss, adjust the parameters of the initial memory network until the preset training end condition is met, and obtain the pre-trained memory network.

[0155] Among them, the sample steam state data corresponding to the historical period includes the steam volume V(t), temperature T(t), humidity H(t) during the operation of the steam generation device, and the working state of the steam generation device. Preprocess these data, including steps such as denoising, normalization, and outlier processing. The preprocessed data set is D = {(V i , T i , H i , G i , y i )}, where y i is the state (normal or faulty) of the steam generation device. Extract features from it to construct a feature vector φ(t), including steam volume, temperature, humidity, gas concentration and its change rate, etc., referring to formulas (10) to (14).

[0156] Use the historical dataset D for model training and adopt the cross-validation method to evaluate the performance of the model. The loss function uses the cross-entropy loss:

[0157]

[0158] where y i is the true fault state, and P f (t) is the predicted fault probability.

[0159] The optimization algorithm uses the Adam optimizer to update the model parameters, referring to formula (25):

[0160]

[0161] where θ is the model parameter and α is the learning rate.

[0162] It can be seen from the above technical solutions of the embodiments of the present application that by training the initial memory network, the embodiments of the present application obtain a model that can accurately predict the future state of the device, which can accurately capture the time-series characteristics of the device operation state, thereby realizing the early prediction of future faults of the device.

[0163] In an exemplary embodiment, the following steps may further be included:

[0164] Obtain the reference control signal of the steam generating device in the current cycle;

[0165] Simulate the steam quantity generated by the steam generating device under the control of the reference control signal to obtain the reference steam quantity data;

[0166] Determine the error between the steam quantity data in the fusion steam state data of the current cycle and the reference steam quantity data;

[0167] Based on the error and the fusion steam state data, adjust the reference control signal to obtain the target control signal of the steam generating device in the next cycle.

[0168] Specifically, the control signal of the steam generating device, as the control input of the steam generating device, may include specific physical parameters such as the steam quantity and temperature to be adjusted, so that the steam generating device can perform corresponding operations. The control signal may include one parameter or multiple parameters.

[0169] Specifically, let the target control signal of the steam generating device be u(t), and the fusion steam state data X(t) includes the steam quantity data V(t), temperature T(t), and humidity H(t), which can be expressed as formula (26):

[0170] X(t) = [V(t), T(t), H(t)] T (26)

[0171] The model reference adaptive control (MRAC) algorithm is adopted. By adjusting the target control signal u(t), the output V(t) of the actual system is made to follow the output of the reference model, that is, the reference steam quantity data V m (t). The reference model can be expressed as:

[0172] V m (t) = A m V m (t - 1) + B m r(t) (27)

[0173] where V m (t) is the reference steam quantity data, r(t) is the reference control signal, and A m and B m are the reference model parameters.

[0174] u(t) is determined by formula (28):

[0175] u(t) = θ(t) T φ(t) (28)

[0176] where θ(t) is the adaptive parameter vector and φ(t) is the feature vector, including the system state and the reference input:

[0177] φ(t) = [X(t) T , r(t)] T (29)

[0178] The update formula for the adaptive parameter vector θ(t) is:

[0179] ˙

[0180] θ(t) = -γφ(t)e(t) (30)

[0181] where γ is the learning rate and e(t) is the error, defined as the difference between the steam quantity data in the fused steam state data of the current cycle and the reference steam quantity data:

[0182] e(t) = V(t) - V m (t) (31)

[0183] Specifically, the initial adaptive parameter θ(0) and the reference model parameters A m , B m , V m, obtain the steam volume V(t), temperature T(t), humidity H(t) and reference control signal r(t) in the cooking cavity in real time, calculate the error e(t) between the actual output and the reference output, update the adaptive parameter θ(t) according to the error e(t) and the eigenvector φ(t), calculate the new target control signal u(t), and adjust the output parameters of the steam generating device. Repeat the above steps to achieve real-time adaptive control.

[0184] Exemplarily, assume the reference model parameters are A m = 0.9, B m = 0.1, and the learning rate γ = 0.01. The initial adaptive parameter θ(0) = [0, 0, 0, 0] T . At a certain moment t, assume the current system state is X(t) = [0.8, 100, 50] T , and the reference input r(t) = 1.

[0185] Calculate the eigenvector:

[0186] φ(t) = [0.8, 100, 50, 1] T (32)

[0187] Calculate the reference output:

[0188] V m (t) = 0.9V m (t - 1)+0.1r(t) (33)

[0189] Assume the reference output V m (t - 1) = 0.7, then:

[0190] V m (t) = 0.9·0.7 + 0.1·1 = 0.73 (34)

[0191] Assume the actual steam volume V(t) = 0.75, then the error:

[0192] e(t) = 0.75 - 0.73 = 0.02 (35)

[0193] Update the adaptive parameter:

[0194]

[0197] Calculate the new target control signal:

[0198] u(t) = θ(t) T φ(t)(37)

[0199] Assume the updated adaptive parameter is θ(t) = [-0.001, 0.02, 0.01, 0.001] T, then:

[0200] u(t) = [-0.001, 0.02, 0.01, 0.001] T ·[0.8, 100, 50, 1] T = 2.5002(38)

[0201] Adjust the target control signal u(t) of the steam generating device to 2.5002.

[0202] As can be seen from the above technical solutions of the embodiments of the present application, through the multi-dimensional data obtained in real time, the embodiments of the present application adaptively adjust the output parameters of the steam generating device, such as the steam volume, temperature, etc. With the model reference adaptive control algorithm, the output parameters of the steam generating device are dynamically adjusted, enabling the system to dynamically optimize the control parameters according to the actual operating state, achieving the best operating state, which not only improves the operating efficiency and stability of the steam generating device, but also reduces the need for human intervention.

[0203] Corresponding to the data processing methods provided in the above several embodiments, the embodiments of the present application also provide a data processing device. Since the data processing device provided in the embodiments of the present application corresponds to the data processing methods provided in the above several embodiments, the implementation manners of the foregoing data processing methods are also applicable to the data processing device provided in this embodiment and will not be described in detail in this embodiment.

[0204] Please refer to Figure 3 , which shows a structural schematic diagram of a data processing device provided by an embodiment of the present application. The device has the function of implementing the data processing method in the above method embodiment. This function can be implemented by hardware or by hardware executing corresponding software. The device is applied to a steam generating device, and the steam generating device is used to inject steam into the cooking cavity, such as Figure 3 shown, the device may include:

[0205] A data acquisition module 310, configured to acquire the real-time in-cavity image and real-time steam state data in the cooking cavity in the current cycle;

[0206] A data fusion module 320, configured to estimate the real-time steam volume data in the cooking cavity based on the real-time in-cavity image and real-time steam state data, and obtain the fused steam state data of the current cycle;

[0207] A data screening module 330, configured to determine the first steam state data in the target steam state data of the previous cycle based on the influence degree of the target steam state data of the previous cycle on the fused steam state data; and determine the second steam state data in the fused steam state data based on the correlation degree between the fused steam state data and the target steam state data of the previous cycle;

[0208] A target screening module 340, configured to determine target steam state data for the current cycle from the first steam state data and the second steam state data based on the influence degrees of the first steam state data and the second steam state data on the failure probability of the steam generating device;

[0209] A failure prediction module 350, configured to predict the failure probability of the steam generating device in the next cycle based on the target steam state data for the current cycle and the influence degree of the target steam state data for the current cycle on the failure probability of the steam generating device.

[0210] In an exemplary embodiment, the data fusion module includes:

[0211] An image segmentation module, configured to segment a steam area from a real-time cavity image;

[0212] A pixel processing module, configured to determine the number of pixels corresponding to the steam area and the intensity of each pixel in the pixels corresponding to the steam area;

[0213] A steam quantity estimation module, configured to estimate real-time steam quantity data in the cooking cavity based on the number of pixels corresponding to the steam area and the intensity of each pixel in the pixels corresponding to the steam area;

[0214] A state fusion module, configured to fuse the real-time steam state data and the real-time steam quantity data to obtain fused steam state data for the current cycle.

[0215] In an exemplary embodiment, the state fusion module includes:

[0216] A state prediction module, configured to use the fused steam state data for the previous cycle as an input to a steam generation dynamics model, so that the steam generation dynamics model predicts the steam state for the current cycle to obtain a predicted value of the steam state for the current cycle;

[0217] A gain determination module, configured to use the real-time steam state data and the real-time steam quantity data as steam state observation values, compare them with the predicted value of the steam state, and determine a target gain;

[0218] A prediction adjustment module, configured to adjust the predicted value of the steam state based on the target gain and the steam state observation values to obtain the fused steam state data for the current cycle.

[0219] In an exemplary embodiment, the data screening module includes:

[0220] A network acquisition module, configured to acquire a pre-trained memory network; the memory network includes a forgetting gate of a memory layer;

[0221] The forgetting gate processing module is used to take the target steam state data of the previous cycle as the input of the forgetting gate to obtain the first steam state data; the parameter of the forgetting gate indicates the influence degree of the target steam state data of the previous cycle on the fused steam state data.

[0222] The memory network further includes an input gate of the memory layer; correspondingly, the data screening module further includes:

[0223] The input gate processing module is used to take the fused steam state data as the input of the input gate to obtain the second steam state data; the parameter of the input gate indicates the correlation degree between the fused steam state data and the target steam state data of the previous cycle.

[0224] The memory network further includes an output gate of the memory layer; correspondingly, the target screening module includes:

[0225] The output gate processing module is used to take the first steam state data and the second steam state data as the input of the output gate to obtain the target steam state data of the current cycle; the parameter of the output gate indicates the influence degree of the first steam state data and the second steam state data on the failure probability of the steam generating device.

[0226] The memory network further includes a fully connected layer; correspondingly, the fault prediction module includes:

[0227] The fully connected layer processing module is used to take the target steam state data of the current cycle as the input of the fully connected layer to obtain the failure probability of the steam generating device in the next cycle; the parameter of the fully connected layer indicates the influence degree of the target steam state data of the current cycle on the failure probability of the steam generating device.

[0228] In an exemplary embodiment, the device further includes a network training module for training the initial memory network, and the network training module includes:

[0229] The sample acquisition module is used to acquire the sample steam state data corresponding to the historical cycle and the true fault state corresponding to the sample steam state data; the true fault state indicates whether the steam generating device fails.

[0230] The network prediction module is used to input the sample steam state data into the initial memory network to predict the failure probability of the steam generating device to obtain the predicted failure probability.

[0231] The loss determination module is used to determine the binary cross-entropy loss of the memory network based on the difference between the predicted failure probability and the true fault state.

[0232] The parameter adjustment module is used to adjust the parameters of the initial memory network based on the binary cross-entropy loss until the preset training end condition is met to obtain the pre-trained memory network.

[0233] In an exemplary embodiment, the device further includes an adaptive control module for adjusting the target control signal of the steam generating device according to the amount of steam during the operation of the steam generating device. The adaptive control module includes:

[0234] A reference acquisition module for acquiring the reference control signal of the steam generating device in the current cycle;

[0235] A control simulation module for simulating the amount of steam generated by the steam generating device under the control of the reference control signal to obtain reference steam amount data;

[0236] An error determination module for determining the error between the steam amount data in the fused steam state data in the current cycle and the reference steam amount data;

[0237] A signal adjustment module for adjusting the reference control signal based on the error and the fused steam state data to obtain the target control signal of the steam generating device in the next cycle.

[0238] It should be noted that for the device provided in the above embodiment, when implementing its functions, only the above-mentioned division of each functional module is used for illustration. In practical applications, the above functions can be allocated to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device provided in the above embodiment and the method embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0239] The embodiment of the present application provides an electronic device, which includes a processor and a memory. At least one instruction or at least one program segment is stored in the memory, and the at least one instruction or the at least one program segment is loaded and executed by the processor to implement any one of the data processing methods provided in the above method embodiment.

[0240] The memory can be used to store software programs and modules. The processor runs the software programs and modules stored in the memory to execute various functional applications and data processing. The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for functions, etc.; the data storage area can store data created according to the use of the device, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices. Correspondingly, the memory may further include a memory controller to provide the processor with access to the memory.

[0241] The method embodiments provided in the embodiments of the present application can be executed in a computer terminal, a server, or a similar computing device, that is, the above-mentioned electronic device may include a computer terminal, a server, or a similar computing device. Figure 4 is a hardware block diagram of a computer device for running a data processing method provided by an embodiment of the present invention. As Figure 4 shown, the internal structure of the computer device may include, but is not limited to: a processor, a network interface, and a memory. Among them, the processor, network interface, and memory in the computer device can be connected through a bus or other means. In the embodiments shown in this specification Figure 4 take the connection through the bus as an example.

[0242] Among them, the processor (or CPU (Central Processing Unit, central processing unit)) is the computing core and control core of the computer device. The network interface may optionally include a standard wired interface, a wireless interface (such as WIFI, a mobile communication interface, etc.). The memory (Memory) is a memory device in the computer device, used to store programs and data. It can be understood that the memory here can be a high-speed RAM storage device, or a non-volatile storage device (nonvolatile memory), such as at least one disk storage device; optionally, it can also be at least one storage device located far from the aforementioned processor. The memory provides a storage space, and this storage space stores the operating system of the electronic device, which may include, but is not limited to: Windows system (an operating system), Linux (an operating system), Android (Android, a mobile operating system) system, IOS (a mobile operating system) system, etc. The present invention does not limit this; and, one or more instructions suitable for being loaded and executed by the processor are also stored in this storage space, and these instructions can be one or more computer programs (including program codes). In the embodiments of this specification, the processor loads and executes one or more instructions stored in the memory to implement the data processing method provided by the above-mentioned method embodiments.

[0243] The embodiments of the present application also provide a computer-readable storage medium. The computer-readable storage medium can be set in the electronic device to store at least one instruction or at least one segment of a program related to implementing a data processing method. The at least one instruction or the at least one segment of the program is loaded and executed by the processor to implement any one of the data processing methods provided by the above-mentioned method embodiments.

[0244] Optionally, in this embodiment, the above storage medium may include, but is not limited to: various media that can store program codes such as USB flash drives, read-only memories (ROM, Read Only Memory), random access memories (RAM, Random Access Memory), mobile hard disks, magnetic disks, or optical discs.

[0245] It should be noted that: the above sequence of embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above specific embodiments of this specification have been described. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be executed in a different order than in the embodiments and still achieve the desired result. Additionally, the processes depicted in the drawings do not necessarily require the specific order or sequential order shown to achieve the desired result. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0246] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the apparatus embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiments.

[0247] Those of ordinary skill in the art can understand that all or part of the steps to implement the above embodiments can be completed by hardware, or can be completed by a program instructing relevant hardware. The program can be stored in a computer-readable storage medium. The above-mentioned storage medium can be a read-only memory, a magnetic disk, or an optical disc, etc.

[0248] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A data processing method, characterized in that, Applied to a steam generating device for injecting steam into a cooking cavity, the method includes: In the current cycle, obtain the real-time in-cavity image and real-time steam state data in the cooking cavity; Based on the real-time in-cavity image and the real-time steam state data, estimate the real-time steam quantity data in the cooking cavity to obtain the fused steam state data for the current cycle; Based on the influence degree of the target steam state data of the previous cycle on the fused steam state data, determine the first steam state data in the target steam state data of the previous cycle; and based on the correlation degree between the fused steam state data and the target steam state data of the previous cycle, determine the second steam state data in the fused steam state data; Based on the influence degree of the first steam state data and the second steam state data on the failure probability of the steam generating device, determine the target steam state data for the current cycle in the first steam state data and the second steam state data; Based on the target steam state data for the current cycle and the influence degree of the target steam state data for the current cycle on the failure probability of the steam generating device, predict the failure probability of the steam generating device in the next cycle.

2. The data processing method according to claim 1, wherein The estimating the real-time steam quantity data in the cooking cavity based on the real-time in-cavity image and the real-time steam state data to obtain the fused steam state data for the current cycle includes: Segment the steam area from the real-time in-cavity image; Determine the number of pixels corresponding to the steam area and the intensity of each pixel in the pixels corresponding to the steam area; Based on the number of pixels corresponding to the steam area and the intensity of each pixel in the pixels corresponding to the steam area, estimate the real-time steam quantity data in the cooking cavity; Fuse the real-time steam state data and the real-time steam quantity data to obtain the fused steam state data for the current cycle.

3. The data processing method according to claim 2, wherein The fusing the real-time steam state data and the real-time steam quantity data to obtain the fused steam state data for the current cycle includes: Take the fused steam state data of the previous cycle as the input of the steam generation dynamics model, so that the steam generation dynamics model predicts the steam state of the current cycle to obtain the predicted value of the steam state of the current cycle; Take the real-time steam state data and the real-time steam quantity data as the steam state observation values, compare them with the predicted value of the steam state, and determine the target gain; Based on the target gain and the steam state observation value, adjust the predicted value of the steam state to obtain the fused steam state data for the current cycle.

4. The data processing method according to claim 1, characterized in that The determining the first steam state data in the target steam state data of the previous cycle based on the influence degree of the target steam state data of the previous cycle on the fused steam state data includes: Obtain a pre-trained memory network; the memory network includes a forgetting gate in the memory layer; Using the target steam state data of the previous cycle as the input of the forget gate to obtain the first steam state data; the parameter of the forget gate indicates the influence degree of the target steam state data of the previous cycle on the fused steam state data; The memory network further includes an input gate of the memory layer; correspondingly, determining the second steam state data from the fused steam state data based on the correlation degree between the fused steam state data and the target steam state data of the previous cycle includes: Using the fused steam state data as the input of the input gate to obtain the second steam state data; the parameter of the input gate indicates the correlation degree between the fused steam state data and the target steam state data of the previous cycle; The memory network further includes an output gate of the memory layer; correspondingly, determining the target steam state data of the current cycle from the first steam state data and the second steam state data based on the influence degrees of the first steam state data and the second steam state data on the failure probability of the steam generating device includes: Using the first steam state data and the second steam state data as the input of the output gate to obtain the target steam state data of the current cycle; the parameter of the output gate indicates the influence degree based on the first steam state data and the second steam state data on the failure probability of the steam generating device; The memory network further includes a fully connected layer; correspondingly, predicting the failure probability of the steam generating device in the next cycle based on the target steam state data of the current cycle and the influence degree of the target steam state data of the current cycle on the failure probability of the steam generating device includes: Using the target steam state data of the current cycle as the input of the fully connected layer to obtain the failure probability of the steam generating device in the next cycle; the parameter of the fully connected layer indicates the influence degree of the target steam state data of the current cycle on the failure probability of the steam generating device.

5. The data processing method according to claim 4, wherein The method further includes: Obtaining the sample steam state data corresponding to the historical cycle and the real failure state corresponding to the sample steam state data; the real failure state indicates whether the steam generating device fails; Inputting the sample steam state data into the initial memory network to predict the failure probability of the steam generating device to obtain the predicted failure probability; Determining the binary cross-entropy loss of the memory network based on the difference between the predicted failure probability and the real failure state; Adjusting the parameters of the initial memory network based on the binary cross-entropy loss until the preset training end condition is met to obtain the pre-trained memory network.

6. The data processing method according to claim 3, wherein The method further includes: Obtaining the reference control signal of the steam generating device in the current cycle; Simulating the steam amount generated by the steam generating device under the control of the reference control signal to obtain the reference steam amount data; Determining the error between the steam amount data in the fused steam state data of the current cycle and the reference steam amount data; Adjust the reference control signal based on the error and the fused steam state data to obtain the target control signal of the steam generating device in the next cycle.

7. A data processing device, characterized in that, Applied to a steam generating device for injecting steam into a cooking cavity, the device includes: A data acquisition module for acquiring the real-time in-cavity image and real-time steam state data in the cooking cavity in the current cycle; A data fusion module for estimating the real-time steam quantity data in the cooking cavity based on the real-time in-cavity image and the real-time steam state data to obtain the fused steam state data in the current cycle; A data screening module for determining first steam state data from the target steam state data in the previous cycle based on the influence degree of the target steam state data in the previous cycle on the fused steam state data; and determining second steam state data from the fused steam state data based on the correlation degree between the fused steam state data and the target steam state data in the previous cycle; A target screening module for determining the target steam state data in the current cycle from the first steam state data and the second steam state data based on the influence degree of the first steam state data and the second steam state data on the failure probability of the steam generating device; A fault prediction module for predicting the failure probability of the steam generating device in the next cycle based on the target steam state data in the current cycle and the influence degree of the target steam state data in the current cycle on the failure probability of the steam generating device.

8. An electronic device, characterized in that, Comprising a processor and a memory, wherein at least one instruction or at least one program segment is stored in the memory, and the at least one instruction or the at least one program segment is loaded and executed by the processor to implement the data processing method according to any one of claims 1 to 6.

9. A computer-readable storage medium, wherein at least one instruction or at least one program segment is stored in the storage medium, and the at least one instruction or the at least one program segment is loaded and executed by a processor to implement the data processing method according to any one of claims 1 to 6.

10. A computer program, characterized in that, When the computer program is executed by a processor, it implements the data processing method according to any one of claims 1 to 6.

Images