CN118381415A - Independent frequency converter system and control method thereof - Google Patents

Abstract

The invention provides an independent frequency converter system and a control method thereof, and an optimization control strategy is realized through integrated deep learning. The system consists of a frequency converter main control unit, a rotating speed adjusting module, a voltage adjusting module and a frequency adjusting module, and realizes accurate motor control. The man-machine interface is convenient to operate, and displays the state and fault codes. The fault detection unit monitors and identifies faults in real time by deep learning, generates solutions and stores the solutions in the unit. The deep learning module analyzes the historical data to continuously optimize control and fault processing. The system is characterized in that the model can be updated according to a new fault solution or codes, and the intelligent level is improved. Through the collaborative work of each module, the high-efficiency stable control of the motor of the textile equipment is realized, the intelligent fault diagnosis processing capability is realized, the production efficiency and the equipment reliability are improved, the maintenance cost is reduced, and the intelligent fault diagnosis processing device has important practical value for the automatic upgrading of the textile industry.

Description

Independent frequency converter system and control method thereof

Technical Field

The invention relates to the technical field of frequency converters, in particular to an independent frequency converter system and a control method thereof.

Background

In the modern textile industry, the stability and efficiency of the equipment are directly related to the productivity benefits of the enterprise. Conventional textile equipment is usually driven by a fixed-frequency motor, and the speed of the equipment is regulated by a mechanical speed change device. This design has the following major problems:

the belt is severely worn: because of frequent starting, stopping and speed regulation, the transmission belt bears larger stress, so that the service life of the transmission belt is shortened, the replacement frequency is high, the maintenance cost is increased, and the production continuity is influenced.

The working strength of the machine repair is high: the adjustment of the mechanical speed changing device is not flexible enough, and manual adjustment is needed, so that the workload of machine repair personnel is increased, and the mechanical speed changing device can not respond in time in emergency situations, so that the production efficiency is affected.

Speed regulation inaccuracy: the mechanical speed changing device has limited adjusting range and low precision, and is difficult to meet the strict requirements of certain fine textile production on speed control.

Lack of intelligent management: the traditional textile equipment lacks an effective fault diagnosis and prevention mechanism, and once the traditional textile equipment fails, the traditional textile equipment often needs to be stopped for a long time for inspection, so that production interruption is caused, and the production plan is influenced.

The energy utilization rate is low: the motor with fixed frequency has low running efficiency under light load or no-load state, and the energy waste is serious.

The existing frequency converter can basically realize simple frequency conversion control and fault code display, and still lacks an effective fault diagnosis and prevention mechanism, and once a fault occurs, the existing frequency converter also needs to be stopped for a long time for inspection, so that production interruption is caused, and the production plan is influenced.

Therefore, it is important to develop an intelligent variable frequency control system which can overcome the above drawbacks and is suitable for textile equipment. The system can realize accurate speed regulation, reduce belt abrasion, reduce the working strength of machine repair, simultaneously have the fault diagnosis function, improve the energy utilization rate, and finally improve the overall production efficiency and the product quality. Therefore, the invention provides an intelligent variable frequency control system based on deep learning, which aims to solve the problems and bring creative changes to the textile industry.

Disclosure of Invention

The embodiment of the invention provides an independent frequency converter system and a control method thereof, which aim at a plurality of problems existing in the prior art.

The core technology of the invention is mainly an intelligent fault diagnosis and prediction system combined with deep learning, which can monitor the running state of the motor of the textile equipment in real time, analyze historical fault data through a deep learning model, continuously optimize a control strategy and a fault processing scheme, and realize the efficient control of the motor and the accurate prediction and processing of faults.

In a first aspect, the present invention provides an independent frequency converter system for motor control of textile equipment, comprising:

The frequency converter main control unit is electrically and communicatively connected with the textile equipment and is used for receiving and processing input signals, controlling the working state of the frequency converter and integrating a deep learning model for optimizing a control strategy;

the rotating speed adjusting module is used for adjusting the rotating speed of the motor of the textile equipment according to the set vehicle speed requirement;

the voltage adjusting module is used for adjusting the motor input voltage of the textile equipment according to the motor load condition and optimizing the working efficiency of the motor;

The frequency adjusting module is used for adjusting the working frequency of a motor of the textile equipment so as to realize stable speed control;

The man-machine interface is used for displaying the current working state and fault codes, providing corresponding operation interfaces and inputting data;

the fault detection unit is used for monitoring the running state of the motor in real time, identifying and recording faults through the deep learning model, and generating fault codes and corresponding solutions;

the storage unit is used for storing fault codes, corresponding solutions, historical fault data and a deep learning model;

The historical fault data comprises a historical fault code, a historical fault time and a historical solution;

The deep learning module is used for analyzing the historical fault data to continuously optimize the control strategy and the fault processing scheme so as to update the deep learning model;

The fault detection unit and the man-machine interface collect operation data and fault data of the textile equipment as training data;

the deep learning module performs training through training data to optimize a rotating speed, voltage and frequency adjustment strategy;

And updating the deep learning model according to the newly input fault solution or the new fault code in the human-computer interface.

Further, the fault detection unit monitors the running state of the motor in real time by acquiring the current, voltage, temperature and vibration frequency of the motor as monitoring data.

Further, the training steps of the deep learning model are as follows:

Acquiring historical monitoring data and preprocessing the historical monitoring data;

Extracting fault diagnosis features in the historical detection data;

selecting the most discriminant features using statistical methods or machine learning algorithms;

Based on a machine learning method, corresponding historical fault data is used as a training set, different types of fault modes are learned, and generalization capability and prediction accuracy of an evaluation model are verified to optimize model parameters, so that a trained deep learning model is obtained.

Further, the specific steps of the fault detection unit for detecting a fault are as follows:

inputting real-time monitoring data into a deep learning model after training is completed, and outputting whether the motor running state of the current textile equipment is normal or the existing fault type;

and generating a corresponding fault code according to the type of the fault, calling a historical solution corresponding to the fault code from the storage unit, and outputting the historical solution to a human-computer interface for display.

Further, the solutions include trouble codes, interpretation of trouble codes, trouble shooting content and/or video content, trouble shooting price costs, and trouble shooting time consumption.

Further, the solution is provided with at least two kinds.

Further, a solution selection module is also included that weights the scores based on the solution's troubleshooting price cost and the troubleshooting time spent, and selects an appropriate solution to execute based on the scores.

In a second aspect, the present invention provides a control method of an independent frequency converter system, comprising the steps of:

s00, operating the textile equipment, and monitoring a motor of the textile equipment in real time through a fault detection unit;

s10, when the man-machine interface displays faults, selecting whether to stop according to the fault type;

S20, giving an alarm for the failure of the immediate stop;

For fault types not subjected to shutdown, preventive maintenance is arranged after the production task is stopped;

s30, acquiring an actual fault solution, and analyzing whether a fault code corresponding to the solution is changed or not;

S40, if the fault code is unchanged, a solution is newly recorded under the fault code; if a change occurs, a new fault code is assigned or the correct fault code is recorded.

Further, the specific steps of S30-S40 are as follows:

Carrying out standardized treatment on the actual solution of the fault to enable the actual solution of the fault to be consistent with the format of the historical solution;

Content similarity comparison is carried out on the actual solution after the standardized processing and the solution corresponding to the fault code;

If the similarity is greater than the threshold, judging a consistent scheme; if the similarity is smaller than the threshold value, the method is considered to be a different scheme;

Judging an actual fault code corresponding to the actual solution according to the content of the actual solution;

if the corresponding actual fault code exists, changing the actual fault code into the corresponding fault code; if no corresponding actual fault code exists, a new fault code is allocated.

Further, the actual fault codes corresponding to the actual solutions are analyzed through a large language model.

The main contributions and innovation points of the invention are as follows:

1. Deep learning driven intelligent fault diagnosis: according to the invention, the running state of the motor is monitored in real time through the deep learning model, the faults are automatically identified and recorded, and the fault codes and the solutions are generated, so that the intelligent fault prediction and diagnosis are realized, and the accuracy and efficiency of fault processing are remarkably improved.

2. Optimizing an adaptive control strategy: based on historical fault data, the deep learning module continuously optimizes a control strategy, including a rotating speed, voltage and frequency adjusting strategy, so that the frequency converter system can be dynamically adjusted according to the motor load, and the optimal working efficiency and energy utilization efficiency are achieved.

3. Fault solution library and intelligent selection: the storage unit stores rich fault codes and solutions, including not only text interpretation, but also text and video tutorials, and cost and time estimation, ensuring quick and accurate solutions to the problem. The solution selection module intelligently recommends the best solution based on the cost and time weighted scores.

4. Real-time monitoring and early warning mechanism: by monitoring key indexes such as current, voltage, temperature and vibration frequency of the motor, the system can early warn potential faults in real time, reduce unplanned downtime and ensure production continuity and stability.

5. Closed loop feedback mechanism: by analyzing the similarity between the actual solution and the historical solution, the system can automatically update the fault code library, continuously improve the deep learning model and form a continuously self-perfected closed loop feedback system.

6. Human-computer interaction interface: the friendly man-machine interface not only displays the current working state and fault codes, but also provides data input and operation guidance, enhances user experience, and simplifies fault processing flow.

The details of one or more embodiments of the invention are set forth in the accompanying drawings and the description below to provide a more thorough understanding of the other features, objects, and advantages of the invention.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention and do not constitute a limitation on the invention. In the drawings:

FIG. 1 is a block diagram of an independent frequency converter system according to an embodiment of the invention;

fig. 2 is a training flow of a deep learning model according to an embodiment of the invention.

Detailed Description

Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with one or more embodiments of the present specification. Rather, they are merely examples of apparatus and methods consistent with aspects of one or more embodiments of the present description as detailed in the accompanying claims.

It should be noted that: in other embodiments, the steps of the corresponding method are not necessarily performed in the order shown and described in this specification. In some other embodiments, the method may include more or fewer steps than described in this specification. Furthermore, individual steps described in this specification, in other embodiments, may be described as being split into multiple steps; while various steps described in this specification may be combined into a single step in other embodiments.

Example 1

The present invention aims to propose an independent frequency converter system, in particular with reference to fig. 1, comprising:

The main control unit of the frequency converter: the core component is responsible for receiving and processing input signals, controlling the working state of the frequency converter, and integrating the deep learning model to optimize the control strategy.

In this embodiment, the connection with the textile device may be achieved by the following interfaces:

And (3) a power interface: ensuring that the frequency converter system is able to obtain the required power from the power supply system of the textile equipment.

Signal interface: and a standard industrial communication interface such as RS485, modbus, profibus and the like is used for realizing data communication between the frequency converter and textile equipment.

Control interface: direct control of the textile machine motor, such as PWM (pulse width modulation) control signals, digital/analog control signals, etc., is achieved through an input/output interface (I/O interface).

Wherein the communication protocol is selected to be compatible with the textile equipment, such as Modbus RTU, CANopen, etherCAT, etc. These protocols are widely used in the field of industrial automation and ensure reliable communication of the frequency converter system with existing equipment. And the data format is selected according to the textile equipment selection, and the data format which is communicated with the existing equipment is selected, wherein the data format comprises transmission formats of commands, state information and fault information, so that both sides can be ensured to be correctly resolved and responded. The description is not intended to be limiting.

The rotating speed adjusting module is used for: and the motor rotating speed of the textile equipment is regulated according to the set vehicle speed requirement, so that the accurate control of the production speed is ensured.

And a voltage regulating module: according to the motor load condition, the input voltage is automatically adjusted, the working efficiency of the motor is optimized, and the energy waste is reduced.

And a frequency adjusting module: and the working frequency of the motor is regulated, so that smooth speed change is realized, and the stability and the product quality of the production process are improved.

Human-machine interface: displaying the current working state, fault codes and providing an operation interface, allowing a user to input data and view fault solutions.

A fault detection unit: and monitoring the running state of the motor in real time, identifying and recording faults through a deep learning model, and generating fault codes and solutions.

And a storage unit: and storing fault codes, solutions, historical fault data and a deep learning model, and providing data support for fault diagnosis.

And the deep learning module is used for: and analyzing historical fault data, continuously optimizing a control strategy and a fault processing scheme, and updating the deep learning model.

The solution selection module: based on the cost and the time weighted score, the most appropriate solution is intelligently recommended, and the fault solving efficiency is improved.

Preferably, prior to actual deployment, compatibility testing is performed on the frequency converter system and textile equipment, ensuring that each control signal and data communication can work properly. System debugging and parameter calibration are carried out through a human-machine interface (HMI), so that the frequency converter system can accurately adjust the working state of the textile equipment according to a set control strategy.

It should be noted that, the automatic operations of the rotation speed adjustment, the voltage adjustment and the frequency adjustment according to the present invention are adjusted according to preset parameters, and then are optimized by a machine learning algorithm, and the rotation speed adjustment, the voltage adjustment, the frequency adjustment, etc. are already mature technical means in the art, so the present invention does not need any more detailed process.

Practical operation example:

it is assumed that in a textile mill, a textile apparatus using the independent frequency converter system suddenly decelerates, and the production efficiency decreases.

1. Real-time monitoring and alarming:

the failure detection unit detects an abnormal decrease in the motor rotational speed.

The system displays an alarm through a human-computer interface to prompt an operator that the motor rotation speed is reduced and check equipment.

2. And (3) fault primary diagnosis:

The deep learning module analyzes information such as motor load data, a rotating speed curve and the like, and judges possible reasons (such as motor overheat and mechanical component abrasion).

A fault code "MTR-007" is generated indicating "abnormal drop in motor rotation speed".

3. Solution recommendation:

the solution selection module retrieves historical cases from the storage unit based on the "MTR-007" code, evaluating the cost and time consumption of various repair solutions.

The operator is recommended to take the "cool motor and check mechanical connection" approach as a preliminary measure.

4. Operator response:

the operator pauses the operation of the apparatus according to the advice, performing cooling and inspection.

It was found that the motor did overheat, but the mechanical connection was not significantly problematic.

5. And (3) subsequent processing and learning:

The operator feeds back the result of the process (i.e., cooling the motor alone, returning to normal) to the system.

The deep learning module analyzes the case, optimizes the priority of the solution under the 'MTR-007' fault code, and increases the recommended weight of the cooling motor.

The storage unit updates the solution library, including the experience of this actual operation.

Results and benefits: through the series of automatic and intelligent auxiliary processes, the textile factory can rapidly identify and solve equipment faults, avoid long-time production interruption and simultaneously reduce maintenance cost. In addition, the system gradually improves the fault diagnosis capability and the accuracy of solution recommendation by continuously learning and improving the processing experience of each fault, and achieves the intellectualization and high efficiency of equipment maintenance.

Preferably, during daily operation of the textile equipment, the frequency converter and its associated components may be subject to various failures including, but not limited to, cooling system failure, power electronics aging, control circuit board failure, etc. These faults, if not handled in time, can lead to equipment downtime, severely affecting production efficiency. As shown in fig. 2, the following is a specific scheme of how to use the intelligent variable frequency control system for fault diagnosis and prediction:

1. data acquisition and preprocessing

And (3) real-time monitoring: the key parameters of the frequency converter and the motor, such as current, voltage, temperature, vibration frequency and the like, are continuously monitored.

Pretreatment: and cleaning the collected original data, including removing noise, filling missing values, detecting abnormal values and the like, so as to ensure the data quality.

2. Feature extraction and selection

Feature extraction: features that aid in fault diagnosis, such as peak value, root mean square value, harmonic content of current, average value, maximum value, rate of change of temperature, and spectral features of vibration, etc., are extracted from the raw data.

Feature selection: the most discriminant features are selected using statistical methods (e.g., correlation analysis, principal component analysis PCA) or machine learning algorithms (e.g., recursive feature elimination RFE, LASSO regression).

3. Establishing a fault prediction model

Model training: and selecting a proper machine learning algorithm, such as Random Forest (Random Forest), support Vector Machine (SVM), neural network (Neural Network) and the like, and learning the modes of different types of faults by using historical fault data as a training set.

Cross-validation: and evaluating the generalization capability and the prediction accuracy of the model through K-fold cross validation, and optimizing the model parameters.

4. Fault prediction and warning

And (3) real-time prediction: and inputting the real-time monitoring data into the trained model, and predicting whether the running state of the current equipment is normal or a certain fault mode exists.

An alarm system: when the model predicts that the equipment is likely to fail, the system immediately gives an alarm, and meanwhile, the specific failure type, possible reasons and recommended processing steps are displayed on the man-machine interaction interface.

Preventive maintenance: based on the failure prediction results, preventive maintenance, such as replacement of aged components in advance, is scheduled to avoid downtime due to sudden failures.

5. Dynamic optimization and learning

Model updating: along with the accumulation of the equipment operation data, the model is periodically retrained to adapt to a new fault mode caused by factors such as equipment aging, environmental change and the like.

Deep learning optimization: and the complex dependency relationship in the time sequence data is captured by utilizing a deep learning technology, such as a cyclic neural network (RNN), a long-short-time memory network (LSTM) and the like, so that the accuracy and timeliness of fault prediction are further improved.

Preferably, the operation example of the solution selection module of the present invention is as follows:

Three different solutions (A, B, C) are assumed to solve the problem of failure of the textile equipment, each with its specific cost and predicted solving time. The goal is to find a balance point that allows for both cost minimization and speed of solution. Data:

solution A: cost 1000 yuan, and the expected solving time is 4 hours.

Solution B: cost 1500 yuan, expected solution time 2 hours.

Solution C: cost 800 yuan, and the expected solving time is 6 hours.

Weight distribution:

Cost weight: 0.7 (cost is a more important consideration)

Time weight: 0.3 (the solution speed is also important but the specific gravity is small)

Calculating a score:

the weighted score for each solution is calculated as follows:

The specific calculation comprises the following steps:

scoring of solution a: 0.7 x 800/1000+0.3 x 2/4=0.71

Scoring of solution B: 0.7 x 800/1500+0.3 x 2/2 ≡0.673

Scoring of solution C: 0.7 x 800/800+0.3 x 2/6=0.8

From the above calculations, solution C gets the highest overall score (0.8) and is therefore the best choice from a cost and time combination point of view.

Example two

Based on the first embodiment, the present embodiment predicts faults by random forest faults. In this example, the process of equipment failure prediction using a Python programming language and scikit-learn machine learning libraries is implemented with a random forest algorithm. The following are specific steps and code implementations:

Step 1: data preparation

First, data on the operating state of the device, typically including sensor readings of current, voltage, temperature, vibration, etc., needs to be collected. To simplify the example, analog data will be used instead of truly acquired data. The code is as follows:

import pandas as pd

import numpy as np

# create DATAFRAME that contains analog data

data = {

'Current': np.random.normal(10, 2, 100),

'Voltage': np.random.normal(220, 10, 100),

'Temperature': np.random.normal(40, 5, 100),

'Vibration': np.random.normal(0.1, 0.05, 100),

'Fault': np.random.choice([0, 1], 100)

}

df = pd.DataFrame(data)

# Set the first 90 data as training set and the last 10 as test set

train_data = df.iloc[:90]

test_data = df.iloc[90:]

Step 2: feature engineering

Next, the data is preprocessed, key features are extracted, and feature matrix X and tag vector Y are constructed. The code is as follows:

feature matrix X and label vector Y

X_train = train_data.drop('Fault', axis=1)

y_train = train_data['Fault']

X_test = test_data.drop('Fault', axis=1)

y_test = test_data['Fault']

Step 3: model training

Training a model by using a random forest algorithm, and setting super parameters such as the number of trees, a feature selection strategy and the like. The code is as follows:

from sklearn.ensemble import RandomForestClassifier

Creating random forest classifier

clf = RandomForestClassifier(n_estimators=100, max_features='sqrt', random_state=42)

Training model #

clf.fit(X_train, y_train)

Step 4: model evaluation

And evaluating the accuracy, recall rate and F1 fraction of the model on the verification set, so as to ensure good performance of the model. The code is as follows:

from sklearn.metrics import accuracy_score, recall_score, f1_score

# prediction on test set

y_pred = clf.predict(X_test)

Calculation of evaluation index

accuracy = accuracy_score(y_test, y_pred)

recall = recall_score(y_test, y_pred)

f1 = f1_score(y_test, y_pred)

print(f"Accuracy: {accuracy:.2f}")

print(f"Recall: {recall:.2f}")

print(f"F1 Score: {f1:.2f}")

Step 5: fault prediction

And inputting the real-time monitoring data into the model, and outputting a fault prediction result.

Step 6: dynamic adjustment

And continuously optimizing the model and the maintenance strategy according to the prediction result and the actual maintenance effect. This step typically involves retraining of the model and adjustment of the hyper-parameters, which can be done periodically to accommodate changes in the data distribution.

Example III

Based on embodiment one, this embodiment uses the Python and Scikit-Learn libraries to construct an SVM-based overload detection model. The resultant data will be used to simulate current, voltage and temperature data for the motor under normal and overload conditions. The following are specific steps and code implementations:

Step 1: data preparation

First, data needs to be generated or loaded. Using synthetic data to simplify the example:

import numpy as np

from sklearn.model_selection import train_test_split

from sklearn.preprocessing import StandardScaler

from sklearn.svm import SVC

from sklearn.metrics import accuracy_score, recall_score, f1_score

# generating synthetic data

np.random.seed(42)

Normal_data=np.random.random (500, 3) 0.1 + [10, 220, 40] # current, voltage, temperature in normal state

Overload_data=np.range.randn (100, 3) 0.2 + [15, 230, 50] # current, voltage, temperature in overload state

Construction tag #

labels = np.concatenate((np.zeros(normal_data.shape[0]), np.ones(overload_data.shape[0])))

# Merging data

data = np.vstack((normal_data, overload_data))

# Division training set and test set

X_train, X_test, y_train, y_test = train_test_split(data, labels, test_size=0.2, random_state=42)

Step 2: data normalization

To ensure that all features are on the same order, STANDARDSCALER is used to normalize the data.

scaler = StandardScaler()

X_train = scaler.fit_transform(X_train)

X_test = scaler.transform(X_test)

Step 3: model training

Next, training is performed using an SVM model, where RBF kernel functions are selected.

svm_model = SVC(kernel='rbf', C=1, gamma='scale')

svm_model.fit(X_train, y_train)

Step 5: overload detection

The trained model is now used to detect overload conditions in real-time data.

Suppose this is a set of data monitored in real time

Real_time_data= [ [14.5, 228, 49] ] # is assumed here to be an example of an overload condition

real_time_data_scaled = scaler.transform(real_time_data)

Model predictive # use

predicted_overload = svm_model.predict(real_time_data_scaled)

if predicted_overload == 1:

print("Overload Detected!")

else:

print("Normal Operation.")

This example demonstrates the basic flow of how an overload detection model is built using SVMs.

Example IV

Based on the same conception, the invention also provides a control method of the independent frequency converter system, which comprises the following steps:

S00: running textile equipment and monitoring in real time

The operation is as follows: the textile equipment is started, the system automatically enters a monitoring mode, and the fault detection unit starts to collect motor operation data in real time, including but not limited to key indexes such as current, voltage and temperature.

The purpose is as follows: any potential faults can be found in time, so that corresponding measures can be taken, and equipment damage or production interruption can be avoided.

S10: fault display and shutdown decision

The operation is as follows: when the man-machine interface detects an abnormal condition, such as motor overload, excessive temperature or other preset fault conditions, the system can display a corresponding fault code.

Decision making: based on the severity and urgency of the fault, an operator or system automatically determines whether an immediate shutdown is required. For example, for a fault that may cause serious damage to the equipment, the machine should be shut down immediately; while maintenance at production gaps may be selected for minor faults that do not affect immediate operation.

S20: fault response

Failure of immediate shutdown: the system automatically sends an alarm to instruct the operator to stop immediately, preventing further damage to the equipment.

Failure of non-immediate shutdown: the system records fault information, but the equipment continues to operate, and maintenance and inspection are carried out after the production task is finished.

S30: solution normalization and alignment

And (3) standardization treatment: and unifying the formats of the collected fault solutions, including fault codes, interpretation, solving steps, cost, time and other information.

Content similarity comparison: the new solution is compared to the content in the historical database using a text similarity algorithm (e.g., cosine similarity, jaccard similarity, or edit distance, etc.).

Threshold judgment: setting a threshold, for example 0.8, and considering the new solution as the same solution if the similarity of the new solution to the historical solution exceeds the threshold; otherwise, the method is considered as a new scheme.

The existing large language model can be used for comparison, so that the method is flexible and convenient to operate.

S40: fault code update

Large language model analysis: the solution text is analyzed, fault codes are identified and matched using advanced natural language processing techniques, such as pre-trained language models.

Updating codes: if the new solution is matched with the historical fault codes, updating a fault code library; if the fault is a new fault, a new fault code is allocated and entered into the database.

The criteria of the solution are fault code, interpretation of fault code, fault resolution text and/or video content, fault resolution price cost and fault resolution time consumption. I.e. the following table:

preferably, the normalization process referred to in the present invention is to fill out the solution on demand.

It is assumed that during a production run, the system detects an abnormal rise in motor temperature, but does not reach the threshold for immediate shutdown.

Monitoring and display (S00-S10): the system displays a fault code "TEMP-001" indicating that the motor temperature is abnormal, but the current production task continues.

Response and recording (S20): after the production is finished, an operator checks the motor, finds that the cooling fin is blocked, and returns to normal after cleaning.

Normalization and alignment (S30): the "clean up fin" was used as a solution, and the normalized data was compared with the historical data, and the similarity was 0.9 (assuming that the threshold was 0.8).

Code update (S40): the large language model analysis confirms (only need tell the large language model all the historical fault data first, then directly say the solution of "clean up cooling fins" tells the large language model, the large language model will automatically make analysis confirm), this solution corresponds to the fault code "TEMP-001". The system automatically updates the fault code base and records the maintenance details.

Through the series of steps, the independent frequency converter system can monitor the state of equipment in real time, can intelligently analyze faults, optimize a solution, improve maintenance efficiency and ensure continuity of a production process and long-term stability of the equipment.

Embodiments of the invention may be implemented by computer software executable by a data processor of a mobile device, such as in a processor entity, or by hardware, or by a combination of software and hardware. Computer software or programs (also referred to as program products) including software routines, applets, and/or macros can be stored in any apparatus-readable data storage medium and they include program instructions for performing particular tasks. The computer program product may include one or more computer-executable components configured to perform embodiments when the program is run. The one or more computer-executable components may be at least one software code or a portion thereof. In addition, in this regard, it should be noted that any blocks of the logic flows as illustrated may represent program steps, or interconnected logic circuits, blocks and functions, or a combination of program steps and logic circuits, blocks and functions. The software may be stored on a physical medium such as a memory chip or memory block implemented within a processor, a magnetic medium such as a hard disk or floppy disk, and an optical medium such as, for example, a DVD and its data variants, a CD, etc. The physical medium is a non-transitory medium.

It should be understood by those skilled in the art that the technical features of the above embodiments may be combined in any manner, and for brevity, all of the possible combinations of the technical features of the above embodiments are not described, however, they should be considered as being within the scope of the description provided herein, as long as there is no contradiction between the combinations of the technical features.

The foregoing examples illustrate only a few embodiments of the invention, which are described in greater detail and are not to be construed as limiting the scope of the invention. It should be noted that it will be apparent to those skilled in the art that various modifications and improvements can be made without departing from the spirit of the invention, which are within the scope of the invention. Accordingly, the scope of the invention should be assessed as that of the appended claims.

Claims (10)

1. Independent frequency converter system for motor control of textile equipment, characterized by comprising:

The frequency converter main control unit is electrically and communicatively connected with the textile equipment and is used for receiving and processing input signals, controlling the working state of the frequency converter and integrating a deep learning model for optimizing a control strategy;

the rotating speed adjusting module is used for adjusting the rotating speed of the motor of the textile equipment according to the set vehicle speed requirement;

the voltage adjusting module is used for adjusting the motor input voltage of the textile equipment according to the motor load condition and optimizing the working efficiency of the motor;

The frequency adjusting module is used for adjusting the working frequency of a motor of the textile equipment so as to realize stable speed control;

The man-machine interface is used for displaying the current working state and fault codes, providing corresponding operation interfaces and inputting data;

the fault detection unit is used for monitoring the running state of the motor in real time, identifying and recording faults through the deep learning model, and generating fault codes and corresponding solutions;

the storage unit is used for storing fault codes, corresponding solutions, historical fault data and a deep learning model;

wherein, the historical fault data comprises a historical fault code, a historical fault time and a historical solution;

The deep learning module is used for analyzing the historical fault data to continuously optimize a control strategy and a fault processing scheme so as to update the deep learning model;

the fault detection unit and the man-machine interface collect operation data and fault data of textile equipment as training data;

the deep learning module performs training through the training data to optimize a rotating speed, voltage and frequency adjustment strategy;

And updating the deep learning model according to the newly input fault solution or the new fault code in the man-machine interface.

2. The independent frequency converter system according to claim 1, wherein the fault detection unit monitors the operating state of the motor in real time by acquiring the current, voltage, temperature, and vibration frequency of the motor as the monitoring data.

3. The independent frequency converter system of claim 2, wherein the training step of the deep learning model is as follows:

Acquiring historical monitoring data and preprocessing the historical monitoring data;

Extracting fault diagnosis features in the historical detection data;

selecting the most discriminant features using statistical methods or machine learning algorithms;

Based on a machine learning method, corresponding historical fault data is used as a training set, different types of fault modes are learned, and generalization capability and prediction accuracy of an evaluation model are verified to optimize model parameters, so that a trained deep learning model is obtained.

4. The independent frequency converter system according to claim 3, wherein the specific steps of the fault detection unit to detect a fault are as follows:

inputting real-time monitoring data into a deep learning model after training is completed, and outputting whether the motor running state of the current textile equipment is normal or the existing fault type;

and generating a corresponding fault code according to the type of the fault, calling a historical solution corresponding to the fault code from the storage unit, and outputting the historical solution to a human-computer interface for display.

5. The independent frequency converter system according to any of claims 1-4, wherein the solutions comprise fault codes, interpretation of fault codes, fault resolution teletext content and/or video content, fault resolution price costs and fault resolution time consumption.

6. The independent frequency converter system of claim 5, wherein said solution is provided in at least two types.

7. The independent frequency converter system of claim 6, further comprising a solution selection module that weights scores based on a solution's troubleshooting price cost and troubleshooting time spent, and selects an appropriate solution to execute based on the scores.

8. A method of controlling an independent frequency converter system according to any one of claims 1-7, characterized by the steps of:

s00, operating the textile equipment, and monitoring a motor of the textile equipment in real time through a fault detection unit;

s10, when the man-machine interface displays faults, selecting whether to stop according to the fault type;

S20, giving an alarm for the failure of the immediate stop;

For fault types not subjected to shutdown, preventive maintenance is arranged after the production task is stopped;

s30, acquiring an actual fault solution, and analyzing whether a fault code corresponding to the solution is changed or not;

S40, if the fault code is unchanged, a solution is newly recorded under the fault code; if a change occurs, a new fault code is assigned or the correct fault code is recorded.

9. The control method according to claim 8, wherein the specific steps of S30 to S40 are:

Carrying out standardized treatment on the actual solution of the fault to enable the actual solution of the fault to be consistent with the format of the historical solution;

Content similarity comparison is carried out on the actual solution after the standardized processing and the solution corresponding to the fault code;

If the similarity is greater than the threshold, judging a consistent scheme; if the similarity is smaller than the threshold value, the method is considered to be a different scheme;

Judging an actual fault code corresponding to the actual solution according to the content of the actual solution;

if the corresponding actual fault code exists, changing the actual fault code into the corresponding fault code; if no corresponding actual fault code exists, a new fault code is allocated.

10. The control method of claim 8, wherein the actual fault code corresponding to the actual solution is analyzed by a large language model.