Roller adaptive temperature measurement system and roller adaptive prediction algorithm

Abstract

The invention discloses a roller self-adaptive temperature measurement system and a roller self-adaptive prediction algorithm, and the system comprises an intelligent control module which is internally integrated with a machine learning algorithm and is used for predicting a bearing fault and generating state prediction information and maintenance suggestions; the multi-sensor module is mounted on a roller and comprises a temperature sensor, a vibration sensor and a pressure sensor; the multi-sensor module is used for collecting running state data of a roller; the display panel is used for displaying the state prediction information and the maintenance suggestions; the automatic alarm module comprises a sound alarm device and a light display alarm device; the automatic alarm module is used for reminding personnel that the roller breaks down; the wireless transmission module is responsible for signal connection among the multi-sensor module, the intelligent control module, the automatic alarm module and the display panel; therefore, by detecting the temperature, the pressure and the vibration of the bearing, possible faults and reasons of the roller are predicted, and suggestions are generated; therefore, the prompt accuracy of the existing monitoring system is effectively improved.

Description

Technical Field

[0001] The invention belongs to the technical field of metallurgy, and in particular relates to a roll adaptive temperature measurement system and a roll adaptive prediction algorithm. Background Art

[0002] In the production of bars and wires in the metallurgical industry, during the roughing and medium rolling stages, the red steel is deformed due to the rolling mill controlling the rotation of the rolls. However, since the rolls at this stage have a high temperature of about 1000°C, the rotation of the rolls can easily lead to burning of the bearings due to reasons such as excessive temperature, inadequate oiling, and bearing damage. This can then lead to equipment failure and production stagnation.

[0003] At present, the existing technology relies on manual temperature measurement or single temperature detection, which has the following problems:

[0004] 1. Lag: Alarm is only conducted through visual inspection, and faults cannot be predicted in advance;

[0005] 2. Singleness: Lack of support for multi-dimensional data (such as vibration, pressure, etc.) and high false alarm rate. Summary of the invention

[0006] The technical problem to be solved by the present invention is to address the deficiencies of the above-mentioned prior art and to provide a roller adaptive temperature measurement system and a roller adaptive prediction algorithm. Through multi-dimensional detection of bearing temperature, pressure, vibration, etc., in conjunction with the prediction algorithm, the possible roller failures and the causes of the failures can be predicted in a timely manner, and solution suggestions can be generated. In this way, the present invention can effectively improve the prompt accuracy of the existing monitoring system and avoid false alarms.

[0007] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0008] A roll adaptive temperature measurement system, comprising:

[0009] Intelligent control module, which integrates machine learning algorithms to predict bearing failures and generate condition prediction information and maintenance recommendations;

[0010] A multi-sensor module is installed on the roller, and includes a temperature sensor, a vibration sensor, and a pressure sensor; the multi-sensor module is used to collect roller operation status data in real time;

[0011] A display panel, used to display the status prediction information and maintenance suggestions;

[0012] An automatic alarm module, which includes a sound alarm device and a light display alarm device; the automatic alarm module is used to remind personnel that a roller failure has occurred;

[0013] The wireless transmission module is responsible for the signal connection between the multi-sensor module, the intelligent control module, the automatic alarm module and the display panel.

[0014] The temperature sensors are arranged in a circular array around the central axis of the bearing of each roller.

[0015] The wireless transmission module supports LoRa / Wi-Fi6 dual-mode communication.

[0016] S1, obtaining roller status information of bearing temperature, bearing vibration frequency, bearing vibration amplitude and bearing load pressure value, and preprocessing the roller status information;

[0017] S2, extracting features from the preprocessed roller status information, obtaining normal working condition data and fault working condition data of the roller, and marking the fault type in the fault working condition data;

[0018] S3, constructing a roll state prediction model and obtaining prediction results;

[0019] S4. Based on the prediction results, the roller status is determined by comparing with the fault condition data; and maintenance recommendations are generated based on the fault type.

[0020] The process of generating maintenance recommendations in S4 includes: building a rule base, matching the prediction results with the rule base to generate specific maintenance recommendations.

[0021] In S1, the process of preprocessing the roller status information includes the following steps:

[0022] S1.1, filtering the roller state information to remove noise in the sensor data;

[0023] S1.2, normalizing the acquired roll status information data to the same dimension in the range of 0-1;

[0024] S1.3. Extract temperature change rate, vibration spectrum characteristics, and pressure fluctuation amplitude from the original data.

[0025] Building a roll condition prediction model includes the following steps:

[0026] T1. Use LSTM neural network to build a time series prediction model, and capture the temporal change law of the roll status by learning historical data;

[0027] T2. Use the random forest algorithm to classify and regress the feature data to identify the characteristic patterns under different working conditions;

[0028] T3. Evaluate the prediction effect of the model through cross-validation and model performance indicators, and adjust and optimize the model parameters;

[0029] T4. Deploy the trained model to the roll adaptive temperature measurement system, receive new roll status data in real time, and perform status prediction and fault diagnosis.

[0030] In step T3, the prediction steps of the roll state prediction model include:

[0031] Input the preprocessed roll state data and extracted feature data into the trained LSTM neural network and random forest model;

[0032] The LSTM neural network uses its internal memory unit and gating mechanism to calculate the input time series data and output the roll health index and failure probability;

[0033] The random forest model classifies or regresses feature data through the voting or averaging mechanism of multiple decision trees to generate corresponding prediction results;

[0034] The prediction results of the LSTM neural network and the random forest model are fused to obtain the final roll status prediction results, including the health index and failure probability.

[0035] In step T4, the step of using the roll state prediction model to judge the roll state includes:

[0036] Compare the roll state prediction results with the preset fault condition data to analyze the similarity between the current roll state and the known fault mode;

[0037] According to the comparison results and the fault probability threshold, it is determined whether the roller is in a normal working condition, a warning state or a fault state;

[0038] When it is determined that the roller is in a faulty state or the predicted failure probability is high, the graded alarm mechanism is triggered to notify relevant personnel to take corresponding measures.

[0039] The present invention has the following beneficial effects:

[0040] The present invention obtains multi-dimensional information such as bearing temperature, bearing vibration frequency, bearing vibration amplitude and bearing load pressure value through a multi-sensor module, and obtains prediction results through a roll state prediction model; based on the prediction results, the roll state is judged, and possible roll failures and causes of the failures are predicted in time, and solution suggestions are generated; thereby, the present invention can effectively improve the prompt accuracy of the existing monitoring system and avoid false alarms. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a structural diagram of the roll adaptive temperature measurement system provided by the present invention.

[0042] Description of the drawings: 1. Display panel; 2. Intelligent control module; 3. Light display alarm device; 4. Temperature sensor; 5. Roller. DETAILED DESCRIPTION

[0043] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific preferred embodiments.

[0044] In the description of the present invention, it should be understood that the terms "left side", "right side", "upper part", "lower part" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. "First", "second" and the like do not indicate the importance of the components, and therefore cannot be understood as limiting the present invention. The specific dimensions used in this embodiment are only for illustrating the technical solution, and do not limit the scope of protection of the present invention.

[0045] The present invention provides a roll adaptive temperature measurement system and a roll adaptive prediction algorithm; Figure 1 , including an intelligent control module 2, which integrates a machine learning algorithm to predict bearing failures and generate state prediction information and maintenance suggestions; a multi-sensor module, which is installed on the roller 5 and includes a temperature sensor 4, a vibration sensor, and a pressure sensor; the multi-sensor module is used to collect the operating state data of the roller 5 in real time; a display panel 1, which is used to display the state prediction information and maintenance suggestions; an automatic alarm module, which consists of a sound alarm device and a light display alarm device 3; the automatic alarm module is used to remind personnel that the roller 5 has a failure; a wireless transmission module, which is responsible for the signal connection between the multi-sensor module, the intelligent control module 2, the automatic alarm module and the display panel 1. Among them, the temperature sensor 4 monitors the bearing temperature in real time; the vibration sensor monitors the bearing vibration frequency and amplitude; the pressure sensor monitors the bearing load.

[0046] The temperature sensors 4 are provided in a circular array around the central axis of the bearing of each roller 5 .

[0047] The wireless transmission module supports LoRa / Wi-Fi 6 dual-mode communication.

[0048] The present application also provides an adaptive prediction algorithm for a roll 5, which is applied to the above-mentioned roll adaptive temperature measurement system, and comprises the following steps:

[0049] Acquire the roller 5 status information of the bearing temperature, bearing vibration frequency, bearing vibration amplitude and bearing load pressure value, and pre-process the roller 5 status information;

[0050] Performing feature extraction on the preprocessed state information of the roller 5, obtaining normal working condition data and fault working condition data of the roller 5, and marking the fault type in the fault working condition data;

[0051] Constructing a state prediction model for roller 5 and obtaining prediction results;

[0052] Based on the prediction results, the state of the roller 5 is judged by comparing with the fault condition data; and maintenance suggestions are generated in combination with the fault type.

[0053] The process of generating maintenance recommendations includes: building a rule base based on expert experience and historical data, and matching the prediction results with the rule base to generate specific maintenance recommendations.

[0054] Wherein, the preprocessing comprises the following steps:

[0055] Filtering the roller 5 status information to remove noise in the sensor data;

[0056] Normalize the acquired roller 5 status information data to the same dimension in the range of 0-1;

[0057] The temperature change rate, vibration spectrum characteristics, and pressure fluctuation amplitude are extracted from the raw data.

[0058] Building a roll condition prediction model includes the following steps:

[0059] Use LSTM neural network to build a time series prediction model, and capture the temporal change pattern of roll status by learning historical data;

[0060] Among them, the LSTM neural network model includes:

[0061] Input layer: The input data dimension is (T, D), where T is the time step and D is the number of features.

[0062] Hidden layer: contains multiple LSTM layers, and the number of hidden units in each layer is n.

[0063] Input Gate:

[0064] i t =σ(W i ·[h t-1 ,x t ]+b i )

[0065] Forget Gate:

[0066] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0067] Memory unit:

[0068]

[0069] Output Gate:

[0070] o t =σ(W o ·[h t-1 ,x t ]+b o )

[0071] Hidden state:

[0072] h t =o t ⊙tanh(C t )

[0073] Among them, W i , W f , W C , W o is the weight matrix, b i 、b f 、b C 、b o is the bias vector, σ is the Sigmoid activation function, h t-1 is the hidden state of the previous time step, x t is the input feature of the current time step, C t-1 is the state of the memory unit at the previous time step, tanh is the hyperbolic tangent activation function, and ⊙ represents element-by-element multiplication.

[0074] Output layer: Output two nodes, one for predicting the health index of the roll and the other for predicting the probability of roll failure. The calculation formula of the output layer is:

[0075]

[0076] Among them, Dense represents the fully connected layer, h T is the hidden state at the last time step.

[0077] Then, the random forest algorithm is used to classify and regress the feature data to identify the characteristic patterns under different working conditions;

[0078] Among them, the random forest model includes:

[0079] Input features: The same feature vector as the LSTM model.

[0080] Decision tree: Construct multiple decision trees. For the kth decision tree T k, input feature vector x and target variable y, and construct a decision tree by recursively splitting the feature space. The splitting criterion of the decision tree is the Gini index or information gain. Among them,

[0081] Node splitting: For each node j, select a feature vector x i and a threshold θ to divide the data into two subnodes:

[0082] Node j ={x|x i ≤θ} and Node j ={x|x i >θ}

[0083] Among them, the eigenvector x i The selection of the threshold θ is based on the Gini index or information gain.

[0084] Random feature selection:

[0085] When constructing each decision tree, from the feature set {x 1 ,x 2 ,x 3 ,…,x D} randomly select m features for node splitting, where m≤D.

[0086] Voting or averaging:

[0087] For classification tasks, the final prediction result is determined by majority voting; specifically, as follows:

[0088] For input x, each decision tree T k Output a class prediction The final prediction result is the majority vote result:

[0089]

[0090] For regression tasks, the final prediction result is determined by the average value; specifically, as follows:

[0091] For input x, each decision tree T k Output a numerical prediction The final prediction result is the average value:

[0092]

[0093] in:

[0094] K is the number of decision trees;

[0095] m is the number of features randomly selected by each decision tree when splitting a node; for classification tasks, usually For regression tasks, usually

[0096] D is the total number of input features;

[0097] θ is the threshold for node splitting, which is used to divide the data into two sub-nodes;

[0098] T k is the kth decision tree, used to predict the input data;

[0099] Mode is the majority voting function, which is used for the final prediction of the classification task;

[0100] is the prediction result of the kth decision tree;

[0101] The prediction results of the LSTM neural network and the random forest algorithm are integrated and outputted.

[0102] Evaluate the prediction effect of the model through cross-validation and model performance indicators, and adjust and optimize the model parameters;

[0103] The trained model is deployed to the roll adaptive temperature measurement system to receive new roll 5 status data in real time for status prediction and fault diagnosis.

[0104] When constructing a data set for model training, historical data is collected, including sensor data under normal operating conditions and fault conditions; and the fault type (such as bearing wear, insufficient lubrication, temperature exceeding the limit, etc.) is labeled.

[0105] When training the model, the historical data is divided into a training set and a test set; then, the training set is used to train the model and optimize the model parameters, such as the number of hidden nodes of LSTM and the tree depth of random forest; finally, the test set is used to evaluate the model performance.

[0106] The prediction steps of the roll 5 state prediction model include:

[0107] Input the preprocessed roller 5 state data and extracted feature data into the trained LSTM neural network and random forest model;

[0108] The LSTM neural network uses its internal memory unit and gating mechanism to calculate the input time series data and output the health index and failure probability of roller 5;

[0109] The random forest model classifies or regresses feature data through the voting or averaging mechanism of multiple decision trees to generate corresponding prediction results;

[0110] The prediction results of the LSTM neural network and the random forest model are fused to obtain the final prediction result of the state of roller 5, including the health index and failure probability.

[0111] The steps of the roll 5 state prediction model for judging the roll 5 state include:

[0112] Compare the prediction result of the state of the roller 5 with the preset fault condition data, and analyze the similarity between the current state of the roller 5 and the known fault mode;

[0113] According to the comparison result and the fault probability threshold, it is determined whether the roller 5 is in a normal working condition, a warning state or a fault state;

[0114] When it is determined that the roller 5 is in a fault state or the predicted fault probability is high, the graded alarm mechanism is triggered to notify relevant personnel to take corresponding measures.

[0115] Working principle:

[0116] Data entry is done in real time;

[0117] Input the current sensor data into the trained model;

[0118] The model outputs prediction results, such as bearing health index and failure probability;

[0119] Conduct fault diagnosis;

[0120] Determine the bearing status based on the prediction results, such as normal, warning, or fault;

[0121] Generate maintenance recommendations based on the fault type. For example, if the frame bearing temperature is too high, it is recommended to add lubricating oil.

[0122] The generation of maintenance suggestions is achieved through the following steps:

[0123] Rules Engine:

[0124] Based on expert experience and historical data, a rule base is built. For example, when the temperature is greater than 900°C and the vibration is greater than 2kHz, it is classified as insufficient lubrication.

[0125] Match the model prediction results with the rule base to generate specific maintenance recommendations.

[0126] Dynamic Optimization:

[0127] Based on actual maintenance effect feedback, the rule base and model parameters are dynamically adjusted to improve prediction accuracy.

Claims

1. A roller adaptive temperature measurement system, characterized in that: include An intelligent control module (2) having an integrated machine learning algorithm for predicting bearing failures and generating condition prediction information and maintenance recommendations; A multi-sensor module is installed on the roller (5), comprising a temperature sensor (4), a vibration sensor, and a pressure sensor; the multi-sensor module is used to collect operating status data of the roller (5) in real time; A display panel (1) for displaying the status prediction information and maintenance suggestions; An automatic alarm module, comprising a sound alarm device and a light display alarm device (3); the automatic alarm module is used to remind personnel that a roller (5) has a fault; The wireless transmission module is responsible for signal connection between the multi-sensor module, the intelligent control module (2), the automatic alarm module and the display panel (1).

2. The roll adaptive temperature measurement system according to claim 1, characterized in that: Three temperature sensors (4) are arranged in a circular array around the central axis of the bearing of each roller (5).

3. The roll adaptive temperature measurement system according to claim 1, characterized in that: The wireless transmission module supports LoRa / Wi-Fi 6 dual-mode communication.

4. A roll adaptive prediction algorithm, applied to the roll adaptive temperature measurement system as claimed in any one of claims 1 to 3, characterized in that: The following steps are involved: S1, obtaining roller (5) status information including bearing temperature, bearing vibration frequency, bearing vibration amplitude and bearing load pressure value, and preprocessing the roller (5) status information; S2, extracting features from the preprocessed state information of the rolling mill (5), obtaining normal working condition data and fault working condition data of the rolling mill (5), and marking the fault type in the fault working condition data; S3, constructing a roller (5) state prediction model and obtaining prediction results; S4. Based on the prediction result, the state of the roller (5) is judged by comparing it with the fault condition data; and maintenance suggestions are generated in combination with the fault type.

5. The roll adaptive prediction algorithm according to claim 4, characterized in that: The process of generating maintenance recommendations in S4 includes: building a rule base, matching the prediction results with the rule base to generate specific maintenance recommendations.

6. The roll adaptive prediction algorithm according to claim 4, characterized in that: In S1, the process of preprocessing the state information of the roller (5) comprises the following steps: S1.1, filtering the state information of the roller (5) to remove noise in the sensor data; S1.2, normalizing the acquired roller (5) status information data to the same dimension in the range of 0-1; S1.

3. Extract temperature change rate, vibration spectrum characteristics, and pressure fluctuation amplitude from the original data.

7. The roll adaptive prediction algorithm according to claim 5, characterized in that: Constructing the roll (5) state prediction model includes the following steps: T1. Use LSTM neural network to build a time series prediction model, and capture the time series change law of the roller (5) state by learning historical data; T2. Use the random forest algorithm to classify and regress the feature data to identify the characteristic patterns under different working conditions; T3. Evaluate the prediction effect of the model through cross-validation and model performance indicators, and adjust and optimize the model parameters; T4. Deploy the trained model to the roller (5) adaptive temperature measurement system, receive new roller (5) status data in real time, and perform status prediction and fault diagnosis.

8. The roll adaptive prediction algorithm according to claim 7, characterized in that: In step T4, the prediction step of the roller (5) state prediction model includes: Input the preprocessed roller (5) state data and the extracted feature data into the trained LSTM neural network and random forest model; The LSTM neural network uses its internal memory unit and gating mechanism to calculate the input time series data and output the health index and failure probability of the roller (5); The random forest model classifies or regresses feature data through the voting or averaging mechanism of multiple decision trees to generate corresponding prediction results; The prediction results of the LSTM neural network and the random forest model are fused to obtain the final prediction results of the roller (5) state, including the health index and the failure probability.

9. The roll adaptive prediction algorithm according to claim 8, characterized in that: In step T4, the step of using the roller (5) state prediction model to judge the state of the roller (5) includes: Comparing the prediction result of the state of the roller (5) with the preset fault condition data, and analyzing the similarity between the current state of the roller (5) and the known fault mode; Judging whether the roller (5) is in a normal working condition, a warning state or a fault state according to the comparison result and the fault probability threshold value; When it is determined that the roller (5) is in a fault state or the predicted fault probability is high, a graded alarm mechanism is triggered to notify relevant personnel to take corresponding measures.

Images