Method, device and equipment for monitoring residual amount of material in crucible of perovskite vacuum evaporation equipment and medium

Abstract

The application discloses a perovskite vacuum evaporation equipment crucible material residual amount monitoring method, device, equipment and medium, and relates to the technical field of computers, which comprises the following steps: obtaining historical production data related to the evaporation process of the material in the crucible based on the historical production record database of the perovskite vacuum evaporation equipment, preprocessing the historical production data, and constructing a sample data set based on the processed production data and the material consumption amount in the crucible per unit time; dividing the data set to obtain a training set, constructing a neural network comprising an input layer, an LSTM hidden layer, a Dropout regularization layer and a fully connected output layer, training the neural network based on the training set and using a back propagation algorithm; determining the material consumption amount prediction value per unit time by using the obtained target neural network, and determining the material residual amount based on the material consumption amount prediction value per unit time. The accuracy and efficiency of the perovskite vacuum evaporation equipment crucible material residual amount monitoring are improved.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to a method, apparatus, equipment, and medium for monitoring the amount of material remaining in the crucible of a perovskite vacuum evaporation equipment. Background Technology

[0002] Perovskite solar cells possess advantages such as low fabrication temperature, high photoelectric conversion efficiency, and low cost, making them a core development direction for next-generation photovoltaic technology. Vacuum evaporation is a key process for preparing the functional layer, transport layer, and electrode layer of perovskite. It involves heating organic / inorganic materials within an evaporation source crucible under high vacuum, causing material atoms / molecules to deposit onto a substrate to form a uniform and dense thin film. The amount of material remaining in the evaporation source crucible directly determines the continuity of the evaporation process, the consistency of the film thickness, and the production cycle time; these are core parameters that must be precisely controlled in the large-scale production of perovskite solar cells.

[0003] Currently, perovskite vacuum evaporation production lines generally manage crucible materials by estimating based on manual experience and periodically breaking the vacuum to weigh them. This has the following insurmountable technical defects: (1) It is impossible to know the material consumption and remaining amount in real time: The evaporation process is in a high vacuum, high temperature and closed environment. It is impossible to directly observe or weigh the remaining material in the crucible online. It can only rely on historical experience to make a rough estimate. The error can reach more than 20%, and it is easy to have material interruption in the middle, resulting in batch scrap. (2) Frequent breaking of the vacuum seriously damages the stability of the process: In order to confirm the remaining amount, it is necessary to stop the machine, break the vacuum, open the lid to weigh, add / change the material, and then re-evacuate and heat up to stabilize the temperature. A single operation takes 24 hours, which greatly reduces the equipment uptime. Repeated breaking of the vacuum will introduce water vapor and oxygen, which will cause the residual material in the crucible to oxidize and deteriorate, affecting the film quality and battery efficiency. (3) The consumption model cannot be accurately matched under varying operating conditions: the production time, evaporation rate, evaporation source temperature and vacuum degree of different batches are all fluctuating. The traditional fixed formula for calculating consumption cannot be adapted to dynamic process parameters, resulting in large prediction deviations and making it impossible to accurately determine the production time that the remaining materials can support.

[0004] As can be seen from the above, improving the accuracy and efficiency of monitoring the remaining material in the crucible of perovskite vacuum evaporation equipment is an urgent problem to be solved. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a method, apparatus, equipment, and medium for monitoring the remaining material in the crucible of a perovskite vacuum evaporation equipment, which can improve the accuracy and efficiency of monitoring the remaining material in the crucible of the perovskite vacuum evaporation equipment. The specific solution is as follows: In a first aspect, this application provides a method for monitoring the amount of material remaining in the crucible of a perovskite vacuum evaporation equipment, including: Historical production data related to the material evaporation process in the crucible is obtained from the historical production record database of the perovskite vacuum evaporation equipment. This historical production data is preprocessed to obtain processed production data. A sample dataset is constructed based on this processed production data and the corresponding material consumption per unit time in the crucible. The historical production data includes the initial total weight of the material, the total weight of the material after evaporation, the total consumption time, the evaporation rate sequence, and the temperature sequence. The preprocessing operations include data cleaning, missing value imputation, and data standardization. The sample dataset is divided to obtain corresponding training, validation, and test sets. A neural network including an input layer, an LSTM hidden layer, a Dropout regularization layer, and a fully connected output layer is constructed. The neural network is trained based on the training set using the backpropagation algorithm to obtain the trained neural network. The trained neural network is validated and tested using the validation and test sets to obtain the target neural network. The target neural network is used to determine the predicted value of material consumption in the crucible per unit time corresponding to real-time production data, and the real-time remaining amount of material in the crucible is determined based on the predicted value of material consumption in the crucible per unit time.

[0006] Optionally, the historical production record database based on the perovskite vacuum evaporation equipment is used to obtain historical production data related to the material evaporation process in the crucible, including: Read the historical production record database of the perovskite vacuum evaporation equipment, and obtain the initial total weight of the material, the total weight of the material after evaporation, the total consumption time, the evaporation rate sequence recorded by the film thickness monitoring device, and the temperature sequence related to the material evaporation process in the crucible based on the historical production record database. Wherein, the initial total weight of the material is the initial weight of the material in the crucible after each feeding; the total weight of the material after evaporation is the weighing weight of the material in the crucible after the vacuum is broken at the end of the production cycle; and the total time consumed is the cumulative production time between the initial weight and the weighing weight.

[0007] Optionally, the preprocessing operation on the historical production data to obtain processed production data includes: Abnormal outliers in the historical production data are removed to obtain production data after removal; the abnormal outliers are data in the historical production data that exceed the target range. The evaporation rate sequence and temperature sequence in the removed production data are resampled and linearly interpolated based on a preset time interval to obtain the supplemented production data. The supplemented production data is standardized to obtain processed production data.

[0008] Optionally, the step of constructing a sample dataset based on the processed production data and the corresponding material consumption per unit time in the crucible includes: Based on the processed production data, the target evaporation rate sequence and target temperature sequence within each time window are determined. The amount of material consumed in the crucible per unit time is determined based on the initial total weight of the material, the total weight of the material after evaporation, and the total consumption time in the production cycle corresponding to each time window. A sample dataset is constructed using the target evaporation rate sequence and target temperature sequence for each time window, as well as the corresponding material consumption per unit time in the crucible.

[0009] Optionally, the sample dataset is divided to obtain corresponding training, validation, and test sets; a neural network including an input layer, an LSTM hidden layer, a Dropout regularization layer, and a fully connected output layer is constructed; and the neural network is trained based on the training set using the backpropagation algorithm to obtain the trained neural network, including: The sample dataset is divided according to a preset division ratio to obtain corresponding training set, validation set and test set; A neural network is constructed comprising an input layer, an LSTM hidden layer, a Dropout regularization layer, and a fully connected output layer; the number of layers corresponding to the LSTM hidden layer is greater than the target number. The neural network is trained based on the training set using the mean squared error loss function, backpropagation algorithm, and Adam optimizer to obtain the trained neural network.

[0010] Optionally, the step of using the target neural network to determine the predicted value of material consumption in the crucible per unit time corresponding to real-time production data, and determining the real-time remaining amount of material in the crucible based on the predicted value of material consumption in the crucible per unit time, includes: The real-time production data corresponding to the target perovskite vacuum evaporation equipment is obtained, and the real-time production data is input into the target neural network to obtain the predicted value of material consumption in the crucible per unit time. The cumulative material consumption is determined based on the predicted value of material consumption in the crucible per unit time and the target cumulative production time corresponding to the real-time production data. The current real-time material remaining amount in the crucible is determined based on the difference between the initial total weight of the target material in the real-time production data and the cumulative material consumption.

[0011] Optionally, after determining the real-time remaining material in the crucible based on the predicted material consumption per unit time, the method further includes: If the real-time remaining material amount is lower than the target remaining threshold, a corresponding early warning is triggered, and the crucible is added or replaced based on the early warning.

[0012] Secondly, this application provides a device for monitoring the amount of material remaining in the crucible of a perovskite vacuum evaporation equipment, comprising: A dataset construction module is used to acquire historical production data related to the material evaporation process in the crucible based on the historical production record database of the perovskite vacuum evaporation equipment. The historical production data is preprocessed to obtain processed production data. A sample dataset is constructed based on the processed production data and the corresponding material consumption per unit time in the crucible. The historical production data includes the initial total weight of the material, the total weight of the material after evaporation, the total consumption time, the evaporation rate sequence, and the temperature sequence. The preprocessing operations include data cleaning, missing value imputation, and data standardization. The model training module is used to divide the sample dataset to obtain corresponding training, validation, and test sets, construct a neural network including an input layer, an LSTM hidden layer, a Dropout regularization layer, and a fully connected output layer, train the neural network based on the training set and using the backpropagation algorithm to obtain the trained neural network, and validate and test the trained neural network using the validation set and the test set to obtain the target neural network. The remaining quantity determination module uses the target neural network to determine the predicted value of material consumption in the crucible per unit time corresponding to the real-time production data, and determines the real-time remaining quantity of material in the crucible based on the predicted value of material consumption in the crucible per unit time.

[0013] Thirdly, this application provides an electronic device, comprising: Memory, used to store computer programs; A processor is used to execute the computer program to implement the aforementioned method for monitoring the amount of material remaining in the crucible of a perovskite vacuum evaporation equipment.

[0014] Fourthly, this application provides a computer-readable storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the aforementioned method for monitoring the remaining amount of material in the crucible of a perovskite vacuum evaporation apparatus.

[0015] This application acquires historical production data related to the material evaporation process in the crucible from a historical production record database of perovskite vacuum evaporation equipment. The historical production data is preprocessed to obtain processed production data. A sample dataset is constructed based on the processed production data and the corresponding material consumption per unit time in the crucible. The historical production data includes the initial total weight of the material, the total weight after evaporation, the total consumption time, the evaporation rate sequence, and the temperature sequence. The preprocessing operations include data cleaning, missing value imputation, and data standardization. The sample dataset is divided to obtain corresponding training, validation, and test sets. A neural network is constructed, including an input layer, an LSTM hidden layer, a Dropout regularization layer, and a fully connected output layer. The neural network is trained using the training set and the backpropagation algorithm to obtain a trained neural network. The trained neural network is validated and tested using the validation and test sets to obtain a target neural network. The target neural network is used to determine the predicted value of material consumption per unit time in the crucible corresponding to the real-time production data. Based on the predicted value of material consumption per unit time in the crucible, the current real-time material remaining amount in the crucible is determined.

[0016] As can be seen from the above, this application obtains data based on the existing historical production record database of the vapor deposition equipment without adding any sensors or hardware modifications, thus ensuring data security. It preprocesses multi-dimensional data on material consumption within the covered crucible and constructs a sample dataset based on the processed data and the corresponding material consumption per unit time. This dataset is then used to train the constructed time-series prediction network, and a backpropagation algorithm is used to obtain the target time-series network. Real-time production data is then acquired and input into the target time-series network to obtain the predicted value of material consumption per unit time within the crucible. In this way, the cumulative consumption is calculated based on the predicted consumption per unit time, effectively improving the stability of the vapor deposition process and the material utilization rate. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0018] Figure 1 This application discloses a flowchart of a method for monitoring the amount of material remaining in the crucible of a perovskite vacuum evaporation equipment. Figure 2 This is a schematic diagram of a neural network model structure disclosed in this application; Figure 3This is a schematic diagram of a deployment architecture and data flow disclosed in this application; Figure 4 This is a schematic diagram of a monitoring device for the remaining material in the crucible of a perovskite vacuum evaporation equipment disclosed in this application. Figure 5 This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Currently, perovskite vacuum evaporation production lines generally manage crucible materials using manual experience estimation and periodic vacuum breaking and weighing. However, this method cannot provide real-time information on material consumption and remaining quantities. The evaporation process takes place in a high-vacuum, high-temperature, and sealed environment, making it impossible to directly observe or weigh the remaining material in the crucible online. Only rough estimates based on historical experience are possible, which can easily lead to material shortages and batch scrapping. Furthermore, frequent vacuum breaking severely disrupts process stability, resulting in large prediction errors and an inability to accurately determine the production time that the remaining material can support. Therefore, this application provides a method for monitoring the remaining material in the crucible of perovskite vacuum evaporation equipment. Based on the predicted consumption per unit time, the cumulative consumption is calculated, effectively improving the stability of the evaporation process and material utilization.

[0021] See Figure 1 As shown in the figure, an embodiment of the present invention discloses a method for monitoring the amount of material remaining in the crucible of a perovskite vacuum evaporation equipment, comprising: Step S11: Obtain historical production data related to the material evaporation process in the crucible from the historical production record database of the perovskite vacuum evaporation equipment. Perform preprocessing operations on the historical production data to obtain processed production data. Construct a sample dataset based on the processed production data and the corresponding material consumption per unit time in the crucible. The historical production data includes the initial total weight of the material, the total weight of the material after evaporation, the total consumption time, the evaporation rate sequence, and the temperature sequence. The preprocessing operations include data cleaning, missing value imputation, and data standardization.

[0022] In this embodiment, production data from multiple batches completing production cycles are extracted from the historical production record database of the perovskite vacuum evaporation equipment. This includes the initial total weight of the material, the total weight of the material after evaporation, the total consumption time, the evaporation rate sequence, and the temperature sequence. The initial total weight of the material is the initial weight of the material in the crucible after each feeding, determined by weighing. The total weight of the material after evaporation is the weight of the material in the crucible after the vacuum is broken at the end of a certain production cycle. The total consumption time is the cumulative production time between two weighings. The evaporation rate sequence is the evaporation rate data recorded in real time by the film thickness monitoring device, typically expressed as... The unit is angstroms per second; the temperature sequence is the heating temperature data of the evaporation source crucible, usually in °C. It is worth noting that historical production data may also include vacuum chamber pressure, substrate temperature, and material type code.

[0023] Specifically, the acquisition of historical production data related to the evaporation process of materials in the crucible based on the historical production record database of the perovskite vacuum evaporation equipment includes: reading the historical production record database of the perovskite vacuum evaporation equipment, and acquiring, based on the historical production record database, the initial total weight of the material, the total weight of the material after evaporation, the total consumption time, the evaporation rate sequence recorded by the film thickness monitoring device, and the temperature sequence related to the evaporation process of materials in the crucible; wherein, the initial total weight of the material is the initial weight of the material in the crucible after each feeding; the total weight of the material after evaporation is the weighing weight of the material in the crucible after the vacuum is broken at the end of the production cycle; and the total consumption time is the cumulative production time between the initial weight and the weighing weight.

[0024] It is understandable that after obtaining the historical production data, outlier data in the historical production data is removed, such as data points where the evaporation rate abruptly drops to 0 in a certain second or exceeds the normal range by 10 times. Then, the evaporation rate sequence and the temperature sequence are resampled and linearly interpolated at fixed time intervals to form a regular time series. The obtained production data is then standardized to obtain processed production data. Specifically, the preprocessing operation of the historical production data to obtain processed production data includes: removing outlier data from the historical production data to obtain removed production data; the outlier data refers to data in the historical production data that exceeds the target range; resampling and linearly interpolating the evaporation rate sequence and temperature sequence in the removed production data based on a preset time interval to obtain interpolated production data; and standardizing the interpolated production data to obtain processed production data.

[0025] Furthermore, based on the processed production data, a tag value is determined for the corresponding production cycle. This tag value represents the amount of material consumed in the crucible per unit time, and the corresponding formula is as follows: ; in, This refers to the amount of material consumed in the crucible per unit time. The initial total weight of the material; The total weight of the material after evaporation; The total consumption time is given. Based on the material consumption per unit time in the crucible, a sample dataset is constructed using the target evaporation rate sequence, the target temperature sequence, and the corresponding material consumption per unit time in the crucible.

[0026] Specifically, constructing a sample dataset based on the processed production data and the corresponding material consumption per unit time in the crucible includes: determining the target evaporation rate sequence and target temperature sequence within each time window based on the processed production data; determining the material consumption per unit time in the crucible based on the initial total weight of material, the total weight of material after evaporation, and the total consumption time in the production cycle corresponding to each time window; and constructing a sample dataset using the target evaporation rate sequence and target temperature sequence within each time window and the corresponding material consumption per unit time in the crucible.

[0027] Step S12: Divide the sample dataset to obtain corresponding training set, validation set and test set, construct a neural network including input layer, LSTM hidden layer, Dropout regularization layer and fully connected output layer, train the neural network based on the training set and using backpropagation algorithm to obtain trained neural network, and use the validation set and the test set to verify and test the trained neural network to obtain target neural network.

[0028] In this embodiment, the sample dataset is randomly divided according to a preset partitioning ratio, such as 70% for training, 15% for validation, and 15% for testing, to construct a deep neural network containing an input layer, at least two LSTM hidden layers, a Dropout regularization layer, and a fully connected output layer. Figure 2This is a schematic diagram of a neural network model structure. The first LSTM layer can have 64-128 units, and the second LSTM layer can have 32-64 units, used to extract temporal features. Dropout layers can be added between LSTM layers to prevent overfitting. The training set data is input into the network for forward propagation to calculate predicted values. The mean-square error (MSE) loss function is used to calculate the error between the predicted and true label values, and the network weight parameters are updated through backpropagation and the Adam optimizer. The loss function is monitored on the validation set. If the loss no longer decreases after N consecutive training rounds, an early stopping mechanism is triggered to terminate training, and the weight file of the best-performing model on the validation set is saved. The trained model is deployed to a local computer in the production environment. Deployment can use a lightweight deep learning inference framework (such as ONNX Runtime, TensorFlow Lite, etc.) to reduce the demand for computing resources. It is deployed to a local industrial control computer or ordinary PC located on the same local area network as the vapor deposition equipment. A data reading service program and an inference engine are installed on this computer.

[0029] Specifically, the process of dividing the sample dataset to obtain corresponding training, validation, and test sets, constructing a neural network including an input layer, an LSTM hidden layer, a Dropout regularization layer, and a fully connected output layer, and training the neural network based on the training set using the backpropagation algorithm to obtain the trained neural network includes: dividing the sample dataset according to a preset partitioning ratio to obtain corresponding training, validation, and test sets; constructing a neural network including an input layer, an LSTM hidden layer, a Dropout regularization layer, and a fully connected output layer; wherein the number of layers corresponding to the LSTM hidden layer is greater than the target number; and training the neural network based on the training set using the mean squared error loss function, the backpropagation algorithm, and the Adam optimizer to obtain the trained neural network.

[0030] Step S13: Use the target neural network to determine the predicted value of material consumption in the crucible per unit time corresponding to the real-time production data, and determine the real-time remaining amount of material in the crucible based on the predicted value of material consumption in the crucible per unit time.

[0031] In this embodiment, Figure 3This diagram illustrates a deployment architecture and data flow. Based on a preset communication protocol, it communicates with the film thickness monitoring system and temperature control system of the evaporation equipment to read the current evaporation rate and evaporation source temperature data in real time. The preset communication protocol can be OPC UA, Modbus TCP / IP, or RS232 serial communication. A prediction process can be triggered at preset time intervals, which can be set according to actual production conditions. Preprocessed evaporation rate and temperature data from a preset time window (e.g., the past 60 minutes) are input into the target neural network to obtain a predicted value of material consumption per unit time in the crucible. The cumulative material consumption is determined based on the predicted value of material consumption per unit time in the crucible and the target cumulative production time corresponding to the real-time production data, using the following formula: ; in, This represents the cumulative material consumption corresponding to the current forecast. This represents the cumulative consumption since the last addition of fuel; This is the predicted value of the material consumption in the crucible per unit time corresponding to the current prediction. The prediction interval is used. The real-time remaining material in the crucible is determined based on the difference between the initial total weight of the target material in the real-time production data and the cumulative material consumption. Additionally, assuming the process parameters remain stable for a period of time, the estimated remaining usable production time at the current evaporation rate and temperature can be determined based on the consumption rate corresponding to the most recent prediction. The corresponding formula is: Remaining time = Remaining material / Current predicted consumption rate.

[0032] Specifically, the step of using the target neural network to determine the predicted value of material consumption in the crucible per unit time corresponding to real-time production data, and determining the real-time remaining amount of material in the crucible based on the predicted value of material consumption in the crucible per unit time, includes: acquiring real-time production data corresponding to the target perovskite vacuum evaporation equipment, and inputting the real-time production data into the target neural network to obtain the predicted value of material consumption in the crucible per unit time; determining the cumulative material consumption based on the predicted value of material consumption in the crucible per unit time and the target cumulative production time corresponding to the real-time production data; and determining the real-time remaining amount of material in the crucible based on the difference between the initial total weight of the target material in the real-time production data and the cumulative material consumption.

[0033] In one specific implementation, if the predicted remaining material amount is lower than the target remaining threshold, an early warning is automatically triggered. This is communicated to the operator via a pop-up window on the human-machine interface, an audible alarm, or an instant messaging message, indicating that "the crucible material is about to run out; please prepare to add more material." Regardless of whether an early warning is triggered, the system continuously updates and displays the current remaining material amount, remaining available time, and the consumption trend curve over a past period on the monitoring interface. The target remaining threshold can be 10% or 20% of the initial total material weight, or it can be adjusted according to actual conditions. Specifically, after determining the real-time remaining material amount in the crucible based on the predicted material consumption per unit time, the system further includes: if the real-time remaining material amount is lower than the target remaining threshold, a corresponding early warning is triggered, and the crucible is added or replaced based on the early warning. When the production task ends, the operator performs vacuum breaking, removes the crucible, and weighs it to record the actual remaining weight.

[0034] Understandably, the data for the complete cycle (initial weight, actual remaining weight, total consumption time, and all historical time-series data) is packaged and stored in the updated data pool. When the number of complete cycles accumulated in the updated data pool reaches a preset threshold K (e.g., K=10), the system automatically, or after operator confirmation, uses this new data to incrementally fine-tune the deployed target neural network, generating an updated model weight file and replacing the old model, thus achieving adaptive optimization of the model. Furthermore, for multi-evaporation source equipment, a multi-output neural network model can be established to simultaneously predict the material consumption of multiple crucibles.

[0035] As can be seen from the above, this application obtains data based on the existing historical production record database of the vapor deposition equipment without adding any sensors or hardware modifications, thus ensuring data security. It preprocesses multi-dimensional data on material consumption within the covered crucible and constructs a sample dataset based on the processed data and the corresponding material consumption per unit time. This dataset is then used to train the constructed time-series prediction network, and a backpropagation algorithm is used to obtain the target time-series network. Real-time production data is then acquired and input into the target time-series network to obtain the predicted value of material consumption per unit time within the crucible. In this way, the cumulative consumption is calculated based on the predicted consumption per unit time, effectively improving the stability of the vapor deposition process and the material utilization rate.

[0036] Accordingly, see Figure 4 As shown, this application also provides a device for monitoring the amount of material remaining in the crucible of a perovskite vacuum evaporation equipment, comprising: The dataset construction module 11 is used to obtain historical production data related to the material evaporation process in the crucible based on the historical production record database of the perovskite vacuum evaporation equipment, perform preprocessing operations on the historical production data to obtain processed production data, and construct a sample dataset based on the processed production data and the corresponding material consumption in the crucible per unit time. The historical production data includes the initial total weight of the material, the total weight of the material after evaporation, the total consumption time, the evaporation rate sequence, and the temperature sequence. The preprocessing operations include data cleaning, missing value imputation, and data standardization. The model training module 12 is used to divide the sample dataset to obtain corresponding training set, validation set and test set, construct a neural network including input layer, LSTM hidden layer, Dropout regularization layer and fully connected output layer, train the neural network based on the training set and using backpropagation algorithm to obtain trained neural network, and use the validation set and the test set to verify and test the trained neural network to obtain target neural network; The remaining quantity determination module 13 uses the target neural network to determine the predicted value of material consumption in the crucible per unit time corresponding to the real-time production data, and determines the real-time remaining quantity of material in the crucible based on the predicted value of material consumption in the crucible per unit time.

[0037] In some specific embodiments, the dataset construction module 11 may specifically include: The sequence acquisition unit is used to read the historical production record database of the perovskite vacuum evaporation equipment, and based on the historical production record database, acquire the initial total weight of the material, the total weight of the material after evaporation, the total consumption time, the evaporation rate sequence recorded by the film thickness monitoring device, and the temperature sequence related to the material evaporation process in the crucible.

[0038] In some specific embodiments, the dataset construction module 11 may specifically include: An outlier data removal unit is used to remove outlier data from the historical production data to obtain production data after removal; the outlier data is data in the historical production data that exceeds the target range. The sequence completion unit is used to resample and linearly interpolate the evaporation rate sequence and temperature sequence in the removed production data based on a preset time interval to obtain the completed production data. The supplemented production data is standardized to obtain processed production data.

[0039] In some specific embodiments, the dataset construction module 11 may specifically include: A sequence determination unit is used to determine the target evaporation rate sequence and the target temperature sequence within each time window based on the processed production data. The consumption determination unit is used to determine the amount of material consumed in the crucible per unit time based on the initial total weight of material, the total weight of material after evaporation, and the total consumption time in the production cycle corresponding to each time window. The dataset construction unit is used to construct a sample dataset using the target evaporation rate sequence and target temperature sequence in each time window, as well as the corresponding material consumption in the crucible per unit time.

[0040] In some specific embodiments, the model training module 12 may specifically include: The dataset partitioning unit is used to partition the sample dataset according to a preset partitioning ratio to obtain corresponding training set, validation set and test set; A neural network building unit is used to construct a neural network including an input layer, an LSTM hidden layer, a Dropout regularization layer, and a fully connected output layer; the number of layers corresponding to the LSTM hidden layer is greater than the target number. The neural network training unit is used to train the neural network based on the training set using the mean squared error loss function, the backpropagation algorithm, and the Adam optimizer to obtain the trained neural network.

[0041] In some specific embodiments, the remaining quantity determination module 13 may specifically include: The prediction value determination unit is used to acquire the real-time production data corresponding to the target perovskite vacuum evaporation equipment and input the real-time production data into the target neural network to obtain the predicted value of material consumption in the crucible per unit time. The cumulative consumption determination unit is used to determine the cumulative material consumption based on the predicted value of material consumption in the crucible per unit time and the target cumulative production time corresponding to the real-time production data. The remaining quantity determination unit is used to determine the real-time remaining quantity of material in the crucible based on the difference between the initial total weight of the target material in the real-time production data and the cumulative material consumption.

[0042] In some specific embodiments, the monitoring device for the remaining material in the crucible of the perovskite vacuum evaporation equipment may further include: The early warning triggering unit is used to trigger a corresponding early warning if the real-time material remaining amount is lower than the target remaining threshold, so as to add or replace the material in the crucible based on the early warning.

[0043] Furthermore, embodiments of this application also disclose an electronic device, Figure 5This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application. Specifically, the electronic device 20 may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the crucible material residue monitoring method of the perovskite vacuum evaporation equipment disclosed in any of the foregoing embodiments. Furthermore, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0044] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0045] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0046] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the method for monitoring the amount of material remaining in the crucible of the perovskite vacuum evaporation equipment disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.

[0047] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned method for monitoring the remaining material in the crucible of a perovskite vacuum evaporation apparatus. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0048] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0049] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0050] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0051] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0052] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for monitoring the amount of material remaining in the crucible of a perovskite vacuum evaporation equipment, characterized in that, include: Historical production data related to the material evaporation process in the crucible is obtained from the historical production record database of the perovskite vacuum evaporation equipment. This historical production data is preprocessed to obtain processed production data. A sample dataset is constructed based on this processed production data and the corresponding material consumption per unit time in the crucible. The historical production data includes the initial total weight of the material, the total weight of the material after evaporation, the total consumption time, the evaporation rate sequence, and the temperature sequence. The preprocessing operations include data cleaning, missing value imputation, and data standardization. The sample dataset is divided to obtain corresponding training, validation, and test sets. A neural network including an input layer, an LSTM hidden layer, a Dropout regularization layer, and a fully connected output layer is constructed. The neural network is trained based on the training set using the backpropagation algorithm to obtain the trained neural network. The trained neural network is validated and tested using the validation and test sets to obtain the target neural network. The target neural network is used to determine the predicted value of material consumption in the crucible per unit time corresponding to real-time production data, and the real-time remaining amount of material in the crucible is determined based on the predicted value of material consumption in the crucible per unit time.

2. The method for monitoring the remaining material in the crucible of the perovskite vacuum evaporation equipment according to claim 1, characterized in that, The historical production record database based on the perovskite vacuum evaporation equipment acquires historical production data related to the material evaporation process in the crucible, including: Read the historical production record database of the perovskite vacuum evaporation equipment, and obtain the initial total weight of the material, the total weight of the material after evaporation, the total consumption time, the evaporation rate sequence recorded by the film thickness monitoring device, and the temperature sequence related to the material evaporation process in the crucible based on the historical production record database. Wherein, the initial total weight of the material is the initial weight of the material in the crucible after each feeding; the total weight of the material after evaporation is the weighing weight of the material in the crucible after the vacuum is broken at the end of the production cycle; and the total time consumed is the cumulative production time between the initial weight and the weighing weight.

3. The method for monitoring the remaining material in the crucible of the perovskite vacuum evaporation equipment according to claim 1, characterized in that, The preprocessing operation on the historical production data to obtain processed production data includes: Abnormal outliers in the historical production data are removed to obtain production data after removal; the abnormal outliers are data in the historical production data that exceed the target range. The evaporation rate sequence and temperature sequence in the removed production data are resampled and linearly interpolated based on a preset time interval to obtain the supplemented production data. The supplemented production data is standardized to obtain processed production data.

4. The method for monitoring the remaining material in the crucible of the perovskite vacuum evaporation equipment according to claim 1, characterized in that, The construction of the sample dataset based on the processed production data and the corresponding material consumption per unit time in the crucible includes: Based on the processed production data, the target evaporation rate sequence and target temperature sequence within each time window are determined. The amount of material consumed in the crucible per unit time is determined based on the initial total weight of the material, the total weight of the material after evaporation, and the total consumption time in the production cycle corresponding to each time window. A sample dataset is constructed using the target evaporation rate sequence and target temperature sequence for each time window, as well as the corresponding material consumption per unit time in the crucible.

5. The method for monitoring the remaining material in the crucible of the perovskite vacuum evaporation equipment according to claim 1, characterized in that, The sample dataset is divided to obtain corresponding training, validation, and test sets. A neural network is constructed, comprising an input layer, an LSTM hidden layer, a Dropout regularization layer, and a fully connected output layer. Based on the training set, the neural network is trained using the backpropagation algorithm to obtain the trained neural network, including: The sample dataset is divided according to a preset division ratio to obtain corresponding training set, validation set and test set; A neural network is constructed comprising an input layer, an LSTM hidden layer, a Dropout regularization layer, and a fully connected output layer; the number of layers corresponding to the LSTM hidden layer is greater than the target number. The neural network is trained based on the training set using the mean squared error loss function, backpropagation algorithm, and Adam optimizer to obtain the trained neural network.

6. The method for monitoring the remaining material in the crucible of the perovskite vacuum evaporation equipment according to claim 1, characterized in that, The step of using the target neural network to determine the predicted value of material consumption in the crucible per unit time corresponding to real-time production data, and determining the real-time remaining amount of material in the crucible based on the predicted value of material consumption in the crucible per unit time, includes: The real-time production data corresponding to the target perovskite vacuum evaporation equipment is obtained, and the real-time production data is input into the target neural network to obtain the predicted value of material consumption in the crucible per unit time. The cumulative material consumption is determined based on the predicted value of material consumption in the crucible per unit time and the target cumulative production time corresponding to the real-time production data. The current real-time material remaining amount in the crucible is determined based on the difference between the initial total weight of the target material in the real-time production data and the cumulative material consumption.

7. The method for monitoring the amount of material remaining in the crucible of the perovskite vacuum evaporation equipment according to any one of claims 1 to 6, characterized in that, After determining the real-time remaining material in the crucible based on the predicted material consumption per unit time, the method further includes: If the real-time remaining material amount is lower than the target remaining threshold, a corresponding early warning is triggered, and the crucible is added or replaced based on the early warning.

8. A device for monitoring the remaining material in the crucible of a perovskite vacuum evaporation equipment, characterized in that, include: A dataset construction module is used to acquire historical production data related to the material evaporation process in the crucible based on the historical production record database of the perovskite vacuum evaporation equipment. The historical production data is preprocessed to obtain processed production data. A sample dataset is constructed based on the processed production data and the corresponding material consumption per unit time in the crucible. The historical production data includes the initial total weight of the material, the total weight of the material after evaporation, the total consumption time, the evaporation rate sequence, and the temperature sequence. The preprocessing operations include data cleaning, missing value imputation, and data standardization. The model training module is used to divide the sample dataset to obtain corresponding training, validation, and test sets, construct a neural network including an input layer, an LSTM hidden layer, a Dropout regularization layer, and a fully connected output layer, train the neural network based on the training set and using the backpropagation algorithm to obtain the trained neural network, and validate and test the trained neural network using the validation set and the test set to obtain the target neural network. The remaining quantity determination module uses the target neural network to determine the predicted value of material consumption in the crucible per unit time corresponding to the real-time production data, and determines the real-time remaining quantity of material in the crucible based on the predicted value of material consumption in the crucible per unit time.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor is configured to execute the computer program to implement the method for monitoring the amount of material remaining in the crucible of the perovskite vacuum evaporation apparatus as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Used to store a computer program, wherein the computer program, when executed by a processor, implements the method for monitoring the amount of material remaining in the crucible of the perovskite vacuum evaporation apparatus as described in any one of claims 1 to 7.

Images