if you meet EVA sensor to 6610v2 to Cloud problem , you can use this dashboard AI bot to ask.
https://dashboard-operation-ensaas.na.wise-ifactory.com/d/dLmRyPZNz/agent-builder?orgId=67
Regarding AI responses, please give me both positive and negative feedback. If reply mistake , please leave comment, I will try to fix LLM model.
reply example looks like below
2410-6610-PHM FAQ from PHM FAQ 20250108v1.3
Q: What underlying model or architecture is used for PHM?
A: Our AI service leverages the meta-learning model to discern patterns in multi-modal sensor data of machines. This model is complemented by domain expertise and international standards (ISO 10816) for machine vibration information. It’s been trained to assess and predict a machine’s health score based on its degradation level.
Q: Do PHM use machine learning or deep learning techniques?
A: Yes, we employ machine learning, specifically deep learning techniques, in conjunction with domain expertise and international standardization.
Q: Does the PHM require supervised, unsupervised, semi-supervised, or reinforcement learning?
A: Our approach primarily utilizes unsupervised learning methods for health score detection and future prediction.
Q: What is PHM accuracy?
A: While we do not rely on the term “accuracy” as a metric for health score detection and prediction, our AI’s performance is evaluated based on its ability to meet international standards and provide reliable assessments. Reliable assessment can refer to the successful cases.
Q: Which kind of PHM algorithm is used?
A: As previously mentioned, our model makes use of meta-learning in combination with other deep learning techniques, reinforced by domain knowledge.
Q: How do you ensure the PHM remains up-to-date and relevant over time?
A: Our AI service is adaptable and can be adjusted over time to cater to different types of machines through our retraining feature. We maintain its relevance by adhering to international standardization as a fundamental benchmark and assessment criterion, ensuring its applicability across changing conditions.
Q: How is its PHM performance compared to other similar models or solutions?
A: Our model depends only on healthy machine data, which can effectively shorten the data collection. It’s also faster in building the model of each machine. These features benefit from the pre-train and meta-learning techs.
Q: What are the features involved in the PHM AI calculation?
A: PHM AI involve sensor three-axis velocity root mean square, acceleration root mean square, peak acceleration, kurtosis, crest factor, skewness, displacement, and standard deviation. Temperature is also included as one of the criteria for assessing equipment health.
Q: What is ISO 10816 standard? How does it relate to the PHM AI model?
A: ISO 10816 is an international standard set by the International Organization for Standardization (ISO) for measuring and evaluating vibrations in mechanical and rotating machinery. It provides guidelines for assessing the condition and health of industrial equipment, helping industries monitor machinery performance, predict potential issues, and plan maintenance to prevent downtime and ensure safety. This standard is widely applied across various industries to ensure machinery reliability and longevity.
In the initial stages of implementing the model to the machine, we apply the ISO 10816 standard to ensure that the machine is in its normal state. This ensures that the AI engine is collecting accurate and reliable data to train the machine model.
Q: How many WISE2410 sensors can each 6610v2 system connect?
In the field, a single 6610v2 system can be configured to support up to 500 EVA sensors or WISE2410 devices. If data is uploaded every 15 minutes, and considering that the presence of machinery or a large amount of metal may affect the signal, the actual performance should be tested on-site. Testing can be conducted by testing 10 sensors at a time. In this case, there are 100 WISE2410 devices in the field.
Q: Can the WISE2410 measure current?
A: No, it detects the health of the motor.
Q: what is 6610 and 2410 setting process?
A: Mobile phone and PC in the same network → mobile phone scan ifactory QRcode to login iMobile service ->” iMobile service” Scan 6610 qrcode->” iMobile service” scan 2410 qrcode → download ENC file, import ENC file into 6610 ifactory service → enable ifactory service → click sync gateway in cloud → see gateway connection show green in cloud → see 2410 was added into 6610 → check 2410 current time & three axis setting-> check 6610 NTP/time-> check 2410 &6610 operation region ex:US → check RTM’s PHM tag name format & need to see total 27 parameter → check RTM’s PHM tag name and device log time if every 10 seconds(depends on 2410 time interval setting) → create PHM AI model → check PHM result on dashboard.
Q: How long does PHM collect data for training ?
A: Fixed Speed : To transmit 200 data points, if the upload time for 2410 is fixed at 10 seconds, it will take approximately 40 minutes to upload 200 data points.
Q: When PHM will complete training ?
A: AI Model Processing : Once the AI model receives 200 data points, it will show 99%. Training is performed at the top of each hour, and after completion, it will show 100%.
Q: Why RTM can’t see PHM’s health score parameter ?
A: The health score doesn’t appear because there are no values for displacement, kurtosis, crest factor, skewness, and deviation. Once these values are available and the training reaches 100% at the top of the hour, the health score will appear.
Q: why PHM time stamp show 2019 year?
A: The health score must have a value, and the device time format must be correct. If 2410 + 6610 have no time calibration, the timestamp 1552585093 will appear, which corresponds to 2019.
Q: why PHM parameter only import 20 parameter?
A: This is License Parameters issue , If only 20 measurement points are imported, it means the license parameters are insufficient. If it’s a shared license, PHM will be for the iFactory BU3. If it’s a PHM license issued by Hubert, it will be Hubert’s case.
Q: Why can ’t create PHM task?
A: Remote Access & PHM Task cannot be created remotely; you need to log in desk.
Q: What are the PHM typical vibration characteristics? Why is PHM deep learning needed to infer a health score that makes it easier for users to understand?
The typical vibration characteristics include the following (for WISE-2410/2460 vibration characteristics, please refer to the characteristic value file):
It is difficult for general users to quickly understand the equipment status by looking at so many characteristics. However, using Advantech’s AI and deep learning, multiple characteristics can be interpreted into a health score that is easier for users to understand, allowing them to quickly grasp the equipment’s condition.
The reason for using AI deep learning to infer a health score is that deep learning models can automatically extract complex features from large amounts of vibration data and combine multiple characteristics to assess the health status of the equipment. This converts complex technical indicators into simple, easy-to-understand health scores, making it more convenient for users to understand and monitor equipment conditions.
Q: Why is vibration typically used for AI in the industry, while temperature and pressure are less commonly used for AI?
A: Vibration signals can directly reflect the operational status and health condition of mechanical equipment, especially for rotating machinery such as motors. Vibration signals can reveal issues like bearing wear, imbalance, misalignment, and other problems. In contrast, temperature and pressure changes are usually slower and may be influenced by environmental factors, making it difficult to provide real-time and accurate fault information. Therefore, vibration signals have a greater advantage in mechanical fault diagnosis.
Q3: What is the difference between using ISO 10816 for motor fault detection and using deep learning PHM to detect anomalies based on multiple features?
The ISO 10816 standard primarily relies on the root mean square (RMS) of vibration velocity to assess the operational condition of mechanical equipment and determine if there is an anomaly based on predefined thresholds. This method is simple and straightforward, but it may overlook subtle fault characteristics.
In contrast, deep learning PHM can simultaneously consider multiple vibration features and automatically identify complex fault patterns through model training. This approach allows for more accurate anomaly detection and provides more detailed fault diagnosis information.
Q: What are the PHM benefits of motor vibration analysis, and how does it help maintenance personnel?
A: Motor vibration analysis offers the following benefits to maintenance personnel:
Q: What is the impact of PHM on motor anomaly detection? Why predict the motor status for the next seven days?
Predictive models can forecast future operating conditions based on historical data and the current status, which is very helpful for planning maintenance in advance. Predicting the motor status for the next seven days gives maintenance personnel sufficient time to prepare and schedule repairs, preventing unexpected failures and minimizing potential losses.
Q: Why is it not suitable to include temperature and pressure along with vibration features in PHM analysis?
A: Including temperature and pressure along with vibration features in PHM analysis may increase the complexity of the model, and the correlation between these parameters might not be strong, which could negatively impact the model’s accuracy. Additionally, temperature and pressure changes are typically slower, providing limited assistance for real-time fault detection.
Q: What is the PHM packet format for uploading data from 6610 Gateway to the RTM (Real-Time Messaging) system?
A: The packet consists of two main examples: the Heartbeat Topic and the Data Topic. The payload format of each example and its purpose will be detailed in the following sections.
iot-2/evt/waconn/fmt/scada_DSCjl9N54Q7b
{
"d": {
"wamqtttest_scada": {
"Hbt": 1
}
},
"ts": "2022-01-11T12:38:25Z"
}
Explanation: This packet is used to send a heartbeat signal from the device to ensure that the device’s connection with the RTM system remains active.
Hbt : Represents the heartbeat signal, where the value 1 indicates a heartbeat confirmation.ts : The timestamp indicating the time the packet was sent.iot-2/evt/wadata/fmt/scada_A5eXamesoV8I
{
"d": {
"W2410-5E8039": {
"Val": {
"W2410-5E8039:GwRssi": -39,
"W2410-5E8039:LstRecv": 1734935319,
"W2410-5E8039:TempHumi_Range": 0,
"W2410-5E8039:TempHumi_Status": 0,
"W2410-5E8039:TempHumi_Event": 0,
"W2410-5E8039:TempHumi_SenVal": 23875,
"W2410-5E8039:X-Axis_SenEvent": 0,
"W2410-5E8039:X-Axis_OAVelocity": 5,
"W2410-5E8039:X-Axis_Peakmg": 5,
"W2410-5E8039:X-Axis_RMSmg": 4,
"W2410-5E8039:X-Axis_Kurtosis": -14,
"W2410-5E8039:X-Axis_CrestFactor": 771,
"W2410-5E8039:X-Axis_Skewness": 55,
"W2410-5E8039:X-Axis_Deviation": 0,
"W2410-5E8039:X-Axis_Peak-to-Peak_Displacement": 1,
"W2410-5E8039:Y-Axis_SenEvent": 0,
"W2410-5E8039:Y-Axis_OAVelocity": 4,
"W2410-5E8039:Y-Axis_Peakmg": 5,
"W2410-5E8039:Y-Axis_RMSmg": 4,
"W2410-5E8039:Y-Axis_Kurtosis": -26,
"W2410-5E8039:Y-Axis_CrestFactor": 568,
"W2410-5E8039:Y-Axis_Skewness": 11,
"W2410-5E8039:Y-Axis_Deviation": 0,
"W2410-5E8039:Y-Axis_Peak-to-Peak_Displacement": 0,
"W2410-5E8039:Z-Axis_SenEvent": 0,
"W2410-5E8039:Z-Axis_OAVelocity": 6,
"W2410-5E8039:Z-Axis_Peakmg": 5,
"W2410-5E8039:Z-Axis_RMSmg": 4,
"W2410-5E8039:Z-Axis_Kurtosis": 17,
"W2410-5E8039:Z-Axis_CrestFactor": 885,
"W2410-5E8039:Z-Axis_Skewness": -8,
"W2410-5E8039:Z-Axis_Deviation": 0,
"W2410-5E8039:Z-Axis_Peak-to-Peak_Displacement": 1,
"W2410-5E8039:LogIndex_LogIndex": 0,
"W2410-5E8039:Timestamp_Timestamp": 1734935305,
"W2410-5E8039:Device_Events": 0,
"W2410-5E8039:Device_PowerSrc": 1,
"W2410-5E8039:Device_BatteryVolt": 0,
"W2410-5E8039:Device_Time": 1734935310
}
}
},
"ts": "2024-12-23T06:28:39Z"
}
Explanation:
W2410-5E8039 : This is the device ID, and the data represents various sensor readings, including signal strength (GwRssi ), temperature and humidity range (TempHumi_Range ), accelerometer readings (X-Axis , Y-Axis , Z-Axis ), and device status information.Val : Contains various measurement values related to the device’s status, sensor data, and diagnostics.ts : The timestamp of the packet, in ISO 8601 format.Purpose:
Advantech WISE-2460 PHM Feature Value Description
| Module | PHM Parameters | Definition | Application in Motor Anomaly Detection |
|---|---|---|---|
| WISE-2410 | Timestamp_Timestamp | Records the date and time of data | Used for time series analysis, helping to identify the exact time of anomaly occurrence. |
| WISE-2410 | Device_Time | The time of the device | Ensures data synchronization, facilitating long-term monitoring and analysis. |
| WISE-2410 | Device_BatteryVolt | The battery voltage of the device | Monitors the device’s battery status, ensuring data reliability. |
| WISE-2410 | X-Axis_Skewness | Skewness of the X-axis frequency distribution, measures the asymmetry of the distribution | Skewness anomalies may indicate imbalance or bearing failure. |
| WISE-2410 | X-Axis_RMSmg | The root mean square (RMS) value of the X-axis, in milligrams | High RMS values may indicate excessive vibration, a potential sign of mechanical failure. |
| WISE-2410 | X-Axis_Peakmg | The peak value of the X-axis, in milligrams | Peak value anomalies may indicate instantaneous shocks or failures. |
| WISE-2410 | X-Axis_Peak-to-Peak_Displacement | Peak-to-peak displacement of the X-axis | Large displacement may indicate bearing wear or misalignment. |
| WISE-2410 | X-Axis_OAVelocity | Overall acceleration velocity of the X-axis | High OAVelocity may indicate imbalance or mechanical failure. |
| WISE-2410 | X-Axis_Kurtosis | Kurtosis of the X-axis frequency distribution, measures the sharpness of the distribution | High kurtosis may indicate shocks or bearing failure. |
| WISE-2410 | X-Axis_Deviation | The standard deviation of the X-axis | High deviation may indicate unstable vibration, a potential sign of failure. |
| WISE-2410 | X-Axis_CrestFactor | Crest factor of the X-axis, the ratio of peak value to RMS value | High crest factor may indicate shocks or imbalance. |
| WISE-2410 | Y-Axis_Skewness | Skewness of the Y-axis frequency distribution | Skewness anomalies may indicate imbalance or bearing failure. |
| WISE-2410 | Y-Axis_RMSmg | The root mean square (RMS) value of the Y-axis | High RMS values may indicate excessive vibration, a potential sign of mechanical failure. |
| WISE-2410 | Y-Axis_Peakmg | The peak value of the Y-axis | Peak value anomalies may indicate instantaneous shocks or failures. |
| WISE-2410 | Y-Axis_Peak-to-Peak_Displacement | Peak-to-peak displacement of the Y-axis | Large displacement may indicate bearing wear or misalignment. |
| WISE-2410 | Y-Axis_OAVelocity | Overall acceleration velocity of the Y-axis | High OAVelocity may indicate imbalance or mechanical failure. |
| WISE-2410 | Y-Axis_Kurtosis | Kurtosis of the Y-axis frequency distribution | High kurtosis may indicate shocks or bearing failure. |
| WISE-2410 | Y-Axis_Deviation | The standard deviation of the Y-axis | High deviation may indicate unstable vibration, a potential sign of failure. |
| WISE-2410 | Y-Axis_CrestFactor | Crest factor of the Y-axis | High crest factor may indicate shocks or imbalance. |
| WISE-2410 | Z-Axis_Skewness | Skewness of the Z-axis frequency distribution | Skewness anomalies may indicate imbalance or bearing failure. |
| WISE-2410 | Z-Axis_RMSmg | The root mean square (RMS) value of the Z-axis | High RMS values may indicate excessive vibration, a potential sign of mechanical failure. |
| WISE-2410 | Z-Axis_Peakmg | The peak value of the Z-axis | Peak value anomalies may indicate instantaneous shocks or failures. |
| WISE-2410 | Z-Axis_Peak-to-Peak_Displacement | Peak-to-peak displacement of the Z-axis | Large displacement may indicate bearing wear or misalignment. |
| WISE-2410 | Z-Axis_OAVelocity | Overall acceleration velocity of the Z-axis | High OAVelocity may indicate imbalance or mechanical failure. |
| WISE-2410 | Z-Axis_Kurtosis | Kurtosis of the Z-axis frequency distribution | High kurtosis may indicate shocks or bearing failure. |
| WISE-2410 | Z-Axis_Deviation | The standard deviation of the Z-axis | High deviation may indicate unstable vibration, a potential sign of failure. |
| WISE-2410 | Z-Axis_CrestFactor | Crest factor of the Z-axis | High crest factor may indicate shocks or imbalance. |
| WISE-2410 | TempHumi_SenVal | Temperature and humidity sensor value | Monitors environmental conditions, ensuring data accuracy. |
| Module | PHM Parameters | Definition | Application in Motor Anomaly Detection |
|---|---|---|---|
| WISE-2460 | Time | Time | Used for time series analysis, helping to identify the exact time of anomaly occurrence. |
| WISE-2460 | Zoav | Overall acceleration velocity of the Z-axis | High OAVelocity may indicate imbalance or mechanical failure. |
| WISE-2460 | Zpeak | Peak value of the Z-axis | Peak value anomalies may indicate instantaneous shocks or failures. |
| WISE-2460 | Zrms | Root mean square (RMS) value of the Z-axis | High RMS values may indicate excessive vibration, a potential sign of mechanical failure. |
| WISE-2460 | Zkurtosis | Kurtosis of the Z-axis frequency distribution | High kurtosis may indicate shocks or bearing failure. |
| WISE-2460 | Zcrestfactor | Crest factor of the Z-axis | High crest factor may indicate shocks or imbalance. |
| WISE-2460 | Zskewness | Skewness of the Z-axis frequency distribution | Skewness anomalies may indicate imbalance or bearing failure. |
| WISE-2460 | Zdeviation | The standard deviation of the Z-axis | High deviation may indicate unstable vibration, a potential sign of failure. |
| WISE-2460 | Zdis | Displacement of the Z-axis | Large displacement may indicate bearing wear or misalignment. |
| WISE-2460 | Zclearancefactor | Clearance factor of the Z-axis | High clearance factor may indicate bearing wear. |
| WISE-2460 | Zshapefactor | Shape factor of the Z-axis | Shape factor anomalies may indicate mechanical failure. |
| WISE-2460 | Zimpulsefactor | Impulse factor of the Z-axis | High impulse factor may indicate instantaneous shocks or failures. |
| WISE-2460 | Temperature | Temperature | Monitors environmental conditions, ensuring data accuracy. |
| WISE-2460 | Zoavband01 | Frequency band 01 of overall acceleration velocity (Z-axis) | Used for spectral analysis, identifying anomalies in the 5 Hz to 1000 Hz frequency range. |
| WISE-2460 | Zoavband02 | Frequency band 02 of overall acceleration velocity (Z-axis) | Used for spectral analysis, identifying anomalies in the 1001 Hz to 2000 Hz frequency range. |
| WISE-2460 | Zoavband03 | Frequency band 03 of overall acceleration velocity (Z-axis) | Used for spectral analysis, identifying anomalies in the 2001 Hz to 3000 Hz frequency range. |
| WISE-2460 | Zoavband04 | Frequency band 04 of overall acceleration velocity (Z-axis) | Used for spectral analysis, identifying anomalies in the 3001 Hz to 4000 Hz frequency range. |
| WISE-2460 | Zoavband05 | Frequency band 05 of overall acceleration velocity (Z-axis) | Used for spectral analysis, identifying anomalies in the 4001 Hz to 5000 Hz frequency range. |
| WISE-2460 | Zoavband06 | Frequency band 06 of overall acceleration velocity (Z-axis) | Used for spectral analysis, identifying anomalies in the 5001 Hz to 6000 Hz frequency range. |
| WISE-2460 | Zoavband07 | Frequency band 07 of overall acceleration velocity (Z-axis) | Used for spectral analysis, identifying anomalies in the 6001 Hz to 7000 Hz frequency range. |
| WISE-2460 | Zoavband08 | Frequency band 08 of overall acceleration velocity (Z-axis) | Used for spectral analysis, identifying anomalies in the 7001 Hz to 8000 Hz frequency range. |
| WISE-2460 | Zoavband09 | Frequency band 09 of overall acceleration velocity (Z-axis) | Used for spectral analysis, identifying anomalies in the 8001 Hz to 9000 Hz frequency range. |
| WISE-2460 | Zoavband10 | Frequency band 10 of overall acceleration velocity (Z-axis |