The on-site oil monitoring system is a key technical means for industrial equipment status monitoring and fault diagnosis. By real-time collection and analysis of performance parameters and pollutant information of mechanical equipment lubrication or hydraulic oil, it achieves accurate assessment of equipment health status. This system integrates sensor technology, data acquisition and transmission technology, oil analysis algorithms, and intelligent diagnostic models to provide data support for preventive maintenance. The following provides a detailed explanation from three aspects: system composition, core functional modules, and working principles.
1、 System composition
The on-site oil monitoring system consists of five main parts: sensor array, data acquisition and transmission unit, central processing and analysis platform, alarm and control module, and auxiliary support system. Each module works together to form a closed-loop monitoring system.
1. Sensor array
Sensors are the data source of the system, responsible for capturing the physical, chemical, and abrasive characteristics of the oil, mainly including:
-Abrasive Particle Sensor: Using inductive, capacitive, or optical principles, it detects the concentration, size distribution, and morphology of metal abrasive particles in oil (such as ferromagnetic abrasive particle sensors that can distinguish between normal wear and abnormal wear).
-Pollution sensor: measures the total number of particles in the oil using light scattering or ultrasonic attenuation methods (such as ISO cleanliness level).
-Physical and chemical characteristic sensors: monitor key indicators such as viscosity, acid value (pH), moisture content, and oxidation degree of oil, using dielectric constant detection, electrochemical probes, or infrared spectroscopy technology.
-Temperature and pressure sensors: Real time acquisition of oil temperature, flow rate, and system pressure to assist in determining equipment operating conditions.
2. Data collection and transmission unit
-Data acquisition card (DAQ): converts sensor signals into digital signals, supports multi-channel synchronous sampling (sampling rate up to MHz level), and has anti-interference capability.
-Wireless communication module: enables remote data transmission through LoRa, Wi Fi, or 4G/5G, suitable for harsh industrial environments.
-Edge computing node: a microprocessor is pre installed on the device side to preliminarily filter, compress or extract features of the original data to reduce the transmission bandwidth requirements.
3. Central processing and analysis platform
-Data storage and management module: using a time-series database (such as InfluxDB) to store historical data, supporting efficient retrieval of massive amounts of data.
-Oil analysis algorithm library: including abrasive feature recognition (such as neural network classification), pollution trend prediction (ARIMA model), correlation analysis of physical and chemical indicators, etc.
-Intelligent diagnostic engine: Based on machine learning (such as SVM, random forest) or expert rule libraries, it outputs device fault types, remaining life prediction, and maintenance recommendations.
4. Alarm and Control Module
-Threshold alarm: Set the warning value according to ISO standards or historical data (such as triggering an alarm when the abrasive particle concentration exceeds 1000/mL).
-Linkage control: integrated with the equipment control system (PLC) to achieve automatic oil change, filtration, or shutdown protection.
-Visual interface: Display the oil status, trend curve, and fault location results through the SCADA system.
5. Auxiliary support system
-Calibration device: Regularly calibrate sensors (such as standard abrasive suspension, known pollution level oil samples).
-Oil circulation circuit: Install a sampling valve in the equipment bypass to ensure the representativeness of the oil sample (such as using online dilution or centrifugal separation technology).
-Self cleaning system: Clean the sensor probe by blowing back gas or ultrasonic waves to prevent blockage.
2、 Working principle
The on-site oil monitoring system achieves dynamic evaluation of equipment status through a closed-loop process of "perception transmission analysis decision". The core workflow is as follows:
1. Oil sampling and sensing
The sensor array continuously collects parameters of equipment lubricating oil or hydraulic oil:
-Grinding particle detection: When the oil flows through the grinding particle sensor, the electromagnetic field changes caused by metal particles are converted into electrical signals, and the grinding particle size is distinguished through pulse height analysis (such as grinding particles larger than 5 μ m indicating severe wear).
-Pollution monitoring: The light scattering sensor emits a laser beam, and the scattered light intensity of particles is inversely proportional to the particle size. The number and distribution of particles are calculated using Mie theory.
-Physical and chemical analysis: Infrared spectrometer scans the molecular bond vibration characteristics of oil to identify oxide (such as carboxylate) or water absorption peaks.
2. Data transmission and preprocessing
The edge computing node performs noise reduction (such as wavelet packet decomposition), feature extraction (such as RMS value extraction of wear particle signal) and data compression (such as LZ77 algorithm) on the original data, and transmits them to the cloud platform through 5G or industrial Ethernet.
3. Data analysis and diagnosis
The central platform is based on multi-dimensional data fusion analysis:
-Traceability of abrasive particles: Using principal component analysis (PCA) combined with equipment records (such as gearbox bearing models) to determine the source of wear (such as journal wear or gear pitting).
-Deterioration trend prediction: Using LSTM neural network to analyze the time series of oil acid value, moisture, and abrasive particle concentration, predict the inflection point of lubrication performance degradation.
-Fault correlation analysis: Combining equipment vibration data (such as high-frequency resonance peaks) with oil abrasive particle characteristics, locate early fatigue cracks.
4. Decision making and Control
The system outputs a graded response based on the diagnostic results:
-Level 1 warning: The cleanliness of the oil has decreased to ISO 18/15 level, indicating the need to strengthen filtration.
-Level 2 alarm: Copper alloy abrasive particles above 5 μ m detected, it is recommended to check the clearance of the sliding bearing.
-Emergency shutdown: If the moisture content exceeds 1% or the abrasive particle concentration increases tenfold, the equipment protection program will be triggered.