Intelligent filter element blockage early warning method and system

Abstract

The invention discloses an intelligent filter element blockage early warning method. The method comprises the steps that collected pressure difference data representing the filter element blockage state is compared with a pressure difference threshold value, and an early warning is output according to a comparison result; the pressure difference threshold value is calculated according to the collected flow at the filter element and the environment temperature. The method has the advantages that by dynamically setting the blockage threshold value of the filter element, blockage early warning and control can be more accurately carried out, and the accuracy of filter element blockage early warning is improved.

Description

Technical Field

[0001] This invention relates to the field of maintenance and monitoring, and in particular to an intelligent filter clogging early warning method and system. Background Technology

[0002] With the rapid development of industrial equipment, automobiles, household water purification systems, and HVAC (Heating, Ventilation, and Air Conditioning) industries, filter elements (such as air filters, oil filters, fuel filters, and water treatment filters) are key filtration components, and their performance directly affects system efficiency, energy consumption, and equipment lifespan. Traditional filter element maintenance mainly relies on periodic replacement or manual inspection, which has the following problems:

[0003] (1) Over-maintenance: prematurely replacing still usable filter elements increases costs.

[0004] (2) Insufficient maintenance: The filter blockage was not detected in time, which led to a decline in equipment performance or even damage.

[0005] (3) Lack of data support: Relying on experience to make judgments makes it difficult to accurately predict the condition of the filter element.

[0006] With the development of automotive technology, automated reminder and warning systems have also been developed, such as setting thresholds to determine whether the filter is currently clogged. However, the thresholds used in traditional clogging detection are all fixed empirical values. Although using fixed values ​​can meet the monitoring, identification and alarm requirements for clogging to a certain extent, fixed thresholds bring about technical problems of low accuracy. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of the prior art and provide an intelligent filter clogging early warning method and system. By dynamically setting the clogging threshold of the filter element, clogging early warning and control can be performed more accurately, thereby improving the accuracy of filter clogging early warning.

[0008] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0009] A smart filter clogging early warning method compares the collected differential pressure data characterizing the filter clogging state with a differential pressure threshold and outputs an early warning based on the comparison result; the differential pressure threshold is calculated based on the collected flow rate at the filter and the ambient temperature.

[0010] The baseline pressure difference is calculated by collecting the flow rate, and then the pressure difference threshold is obtained by compensating the baseline pressure difference with temperature.

[0011] A basic relationship model between flow rate and pressure difference is established, using a power function formula, with the flow rate as the input parameter to obtain the theoretical benchmark pressure difference value under the current flow rate conditions.

[0012] The theoretical baseline pressure difference is calculated using the baseline pressure difference calculation formula: Baseline pressure difference = K1 × (Flow rate value / 10) 1.8 Where K1 is the proportionality coefficient, which is a constant.

[0013] The compensation coefficient K2 is calculated based on temperature. The temperature compensation coefficient K2 is multiplied by the reference pressure difference to obtain the compensated pressure difference threshold.

[0014] The compensation coefficient K2 is negatively correlated with temperature; the lower the temperature, the larger the compensation coefficient should be.

[0015] The formula for calculating the compensation coefficient K2 is:

[0016] K2 = 1 + (reference temperature - actual temperature) * C; where C is a constant coefficient.

[0017] A tiered early warning system is used to issue different warning messages for different ranges exceeding the benchmark pressure difference.

[0018] Based on a fuzzy logic evaluation method that integrates multiple features, the system receives three sets of time-series data: differential pressure, flow rate, and vibration. It then calculates the comprehensive health index of the filter element and displays this comprehensive health index to the user as part of the early warning information.

[0019] A smart filter clogging early warning system includes a sensor, a processor, and an early warning module;

[0020] The sensor is used to collect differential pressure data characterizing the filter element clogging status, and its output is connected to the processor. The output of the processor is connected to the early warning module. The processor compares the differential pressure data with the differential pressure threshold and outputs an early warning based on the comparison result. The differential pressure threshold in the processor is calculated based on the flow rate and ambient temperature at the filter element.

[0021] The advantages of this invention are: by dynamically setting the clogging threshold of the filter element, clogging warning and control can be performed more accurately, improving the accuracy of filter element clogging warning. Based on accurate warning, accurate replacement and maintenance of the filter element can be achieved, optimizing maintenance strategies and reducing filter element replacement costs. Attached Figure Description

[0022] The following is a brief explanation of the contents of each of the accompanying drawings and the markings in the drawings:

[0023] Figure 1 This is a schematic diagram illustrating the principle of the congestion early warning method of the present invention. Detailed Implementation

[0024] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and the description of the preferred embodiments.

[0025] This embodiment provides an intelligent filter clogging early warning method that uses sensors to collect monitoring data and dynamically adjusts the differential pressure threshold. The accuracy and reliability of the early warning are improved through dynamic differential pressure threshold and temperature compensation. Simultaneously, a tiered early warning system is employed, allowing users to obtain the current filter status promptly and accurately.

[0026] like Figure 1 As shown, this embodiment of an intelligent filter clogging early warning method compares the collected differential pressure data characterizing the filter clogging state with a differential pressure threshold and outputs an early warning based on the comparison result; the differential pressure threshold is calculated based on the collected flow rate at the filter and the ambient temperature.

[0027] The core of this embodiment lies in the use of a dynamic threshold algorithm for calculating the differential pressure threshold. Specifically, a baseline differential pressure is calculated based on the collected flow rate, and then the baseline differential pressure is compensated for by temperature to obtain the differential pressure threshold. Then, the real-time filter cartridge differential pressure is compared with the differential pressure threshold to determine the current warning status and provide a corresponding warning alert.

[0028] The calculation of the baseline pressure difference includes: establishing a basic relationship model between flow rate and pressure difference, using a power function formula, and obtaining the theoretical baseline pressure difference value under the current flow rate conditions using the flow rate as the input parameter. The theoretical baseline pressure difference is then calculated using the following formula: Baseline Pressure Difference = K1 × (Flow Rate Value / 10) 1.8 Where K1 is the proportionality coefficient, which is a constant and is taken as 2.5 in this application.

[0029] In this embodiment, temperature is used to compensate for the reference pressure difference to obtain the final pressure difference threshold. A compensation coefficient K2 is calculated based on temperature, and multiplying K2 by the reference pressure difference yields the compensated pressure difference threshold. The compensation coefficient K2 is negatively correlated with temperature; the lower the temperature, the larger the compensation coefficient. A reference temperature is set, and compensation is performed when the ambient temperature is lower or higher than the reference temperature. When the ambient temperature is equal to the reference temperature, the compensation coefficient is 1. The lower the temperature, the larger the compensation coefficient, and the higher the reference pressure difference. In high-temperature environments, high-temperature compensation is activated, but in the opposite direction, to reflect the effect of decreased fluid viscosity at high temperatures.

[0030] The formula for calculating the compensation coefficient K2 in this embodiment is as follows:

[0031] K2 = 1 + (reference temperature - actual temperature) * C; where C is a constant coefficient, which is set to 0.015 in this embodiment. The specific value can be fine-tuned or calibrated according to the actual situation.

[0032] After obtaining the differential pressure threshold, a tiered early warning system is used to issue different warning messages for different ranges exceeding the baseline differential pressure. Different warning levels are set based on safety margin settings. Three-level early warning is configured:

[0033] Mild warning: When the real-time differential pressure exceeds the baseline value by 15% and the duration exceeds the time threshold, a warning is triggered. At this time, a scheduled maintenance reminder is sent to the user. This indicates that the blockage has started but is not serious, and maintenance needs to be carried out according to the plan.

[0034] Medium warning: A medium warning is triggered when the value exceeds the baseline by 30% and the duration exceeds the time threshold. At this time, a replacement reminder is issued, and users are advised to perform maintenance and upkeep immediately.

[0035] Critical warning: A critical alarm is triggered when the value exceeds the baseline by 50% or 70%, at which point the machine should be stopped immediately for inspection.

[0036] In a preferred embodiment, the collected sensor information is monitored in real time, and the amplitude of the sensor signal is calculated. When the amplitude is less than a threshold and the duration is greater than a set threshold T2, the sensor is determined to be faulty. At this time, an alarm is issued or a backup sensor channel is switched to obtain accurate data. The threshold T2 is a time threshold that can be adjusted and set according to actual needs.

[0037] In this embodiment, a fuzzy logic evaluation method based on multi-feature fusion is used to receive three sets of time-series data: differential pressure, flow rate, and vibration. The comprehensive health index of the filter element is then calculated and displayed to the user as a component of the early warning information. This embodiment not only displays clogging warning information to the user but also shows the filter element's health status and lifespan information, allowing the user to have a more comprehensive monitoring and understanding of the filter element's condition and providing basic data information for maintenance.

[0038] This embodiment also provides an intelligent filter clogging early warning system, including a sensor, a processor, and an early warning module. The sensor is used to collect differential pressure data characterizing the filter clogging state, and its output is connected to the processor. The processor's output is connected to the early warning module. The processor compares the differential pressure data with a differential pressure threshold and outputs an early warning based on the comparison result. The differential pressure threshold in the processor is calculated based on the collected flow rate and ambient temperature at the filter. The processor can use an on-board sensor to collect data such as flow rate, temperature, and differential pressure. Based on the data collected by the sensor, the system monitors and issues early warnings for the filter, as well as predicts its health status and lifespan.

[0039] The intelligent filter clogging early warning system provided in this embodiment monitors the working status of the filter in real time, and uses sensor data analysis and algorithm prediction to give early warning of filter clogging, helping users to replace the filter in time and ensuring the normal operation of the equipment.

[0040] Sensor installation location:

[0041]

[0042] Special installation requirements:

[0043] The differential pressure sensor must ensure that:

[0044] The measuring point should be at least 5 times the pipe diameter away from the bend, and the pressure tap should be perpendicular to the fluid direction. The differential pressure in this application refers to the pressure difference across the filter element, acquired by a differential pressure sensor.

[0045] Vibration-sensitive components should be installed on a rigid structure with a natural frequency >1kHz, avoiding the engine resonance frequency range (typically 80-120Hz).

[0046] Signal processing pipeline: Raw signal — Hardware filtering (10Hz low-pass) — Software FIR filtering — Moving average — Outlier removal (3σ principle). The acquired sensor data is processed according to the raw signal. After initialization processing such as filtering, smoothing, and outlier removal, it provides the basic data for subsequent early warning monitoring.

[0047] Fault diagnosis logic table:

[0048]

[0049] System architecture design:

[0050] 1. Hardware layer

[0051] In this embodiment, the hardware layer includes multiple sensors that collect data and process the data to provide a filter cartridge warning. Specific hardware components include:

[0052] Pressure sensor: monitors the pressure difference between the inlet and outlet of the filter element;

[0053] Flow sensor: detects liquid / gas flow rate;

[0054] Temperature sensor: monitors the ambient temperature.

[0055] Turbidity sensor: detects the cleanliness of the filter medium;

[0056] Microcontroller: Data acquisition and processing unit;

[0057] Communication module: Wi-Fi / Bluetooth / 4G / 5G data transmission.

[0058] 2. Software layer

[0059] Data acquisition module: collects sensor data in real time;

[0060] Data processing module: data cleaning, filtering, and feature extraction;

[0061] Analysis algorithm module: congestion level assessment and prediction;

[0062] Early warning module: threshold judgment and alarm triggering;

[0063] User interface: Displayed on mobile apps / web platforms.

[0064] 3. Core Functions

[0065] 1) Real-time monitoring

[0066] It continuously collects key parameters such as differential pressure, flow rate, and temperature, and visualizes the data through instruments, vehicle displays, etc.

[0067] 2) Congestion level assessment

[0068] The degree of clogging is assessed based on the differential pressure-flow relationship model and a multi-parameter comprehensive evaluation algorithm. Dynamic thresholds are used to assess the clogging of the filter element and issue early warning information.

[0069] 3) Predictive maintenance

[0070] The remaining service life prediction function, which uses historical data trend analysis, predicts the lifespan of the current filter element and sends a notification.

[0071] 4) Intelligent early warning

[0072] A multi-level early warning mechanism (mild / moderate / severe congestion) and multi-channel alarms (APP push, SMS, email) are used to notify users through multiple channels.

[0073] 5) Data recording and analysis

[0074] Core control strategy:

[0075] 1) Development of Dynamic Threshold Algorithm

[0076] Algorithm principle: To establish an adaptive threshold calculation model based on operating parameters, two real-time measurement parameters need to be received: flow rate and ambient temperature.

[0077] Implementation steps:

[0078] Based on the principles of fluid mechanics, a basic relationship model between flow rate and pressure difference is first established.

[0079] Baseline differential pressure calculation: A power function formula is used, with flow rate as the main input parameter. In this application, the flow rate for the air filter is measured at the outlet side of the filter element; for the oil filter and fuel filter, the measurement location should be the inlet side of the filter element or the flow rate in the main circuit of the system.

[0080] Basic calculation formula: Baseline pressure difference = 2.5 × (flow rate / 10) 1.8 .

[0081] This formula simulates the inherent characteristic curve of an air filter to obtain the theoretical reference pressure difference value under the current flow conditions.

[0082] Temperature Compensation Mechanism: Considering the significant impact of temperature on fluid properties, the algorithm introduces a temperature compensation mechanism. Temperature refers to the fluid temperature, i.e., the temperature of the working medium of the filter element. For oils, the oil temperature must be measured; for air, ambient temperature or intake air temperature is a feasible and common alternative, ultimately to characterize the fluid temperature.

[0083] Compensation is activated when the ambient temperature is below 20℃.

[0084] Compensation coefficient = 1 + (20 - actual temperature) × 0.015

[0085] The lower the temperature, the larger the compensation coefficient, and the higher the reference pressure difference. In high-temperature environments, high-temperature compensation is activated, but the compensation direction is reversed to reflect the effect of reduced fluid viscosity at high temperatures.

[0086] Safety margin settings:

[0087] Warning threshold = Baseline pressure difference × 1.3 (i.e., a warning is triggered when the pressure exceeds the baseline value by 30%)

[0088] Severe threshold = Baseline differential pressure × 1.7 (i.e., a severe alarm is triggered when the differential pressure exceeds the baseline value by 70%)

[0089] Output: Returns the warning threshold and severity threshold after adaptive adjustment based on operating conditions.

[0090] The above settings include two levels of warnings, but in reality, three or even more levels can be set, which can be adjusted according to actual needs.

[0091] 2). Health assessment model

[0092] In this embodiment, the warning information issued during monitoring includes the filter element monitoring evaluation results, including not only clogging warnings but also health status information, which allows users to better monitor the current status of the filter element.

[0093] Algorithm Principle: Based on a fuzzy logic evaluation method using multi-feature fusion. The model requires three sets of time-series data: pressure differential, flow rate, and vibration. The time window is typically the most recent hour. The model outputs a comprehensive health index: a score ranging from 0 to 100, with higher scores indicating better filter health. Three parameters are calculated within the time window: the integral area of ​​the pressure differential curve over time, the ratio of the standard deviation to the mean of the flow rate data, and the proportion of times the vibration amplitude exceeds 0.5. Then, a health assessment is calculated based on these parameters. The health assessment assigns weights k1, k2, and k3 (where k1 + k2 + k3 = 1) to the three parameters, corresponding to their respective proportions. The health level is then calculated using the formula:

[0094] Health Index = 100 - (k1×pressure differential integral + k2×flow fluctuation rate + k3×abnormal vibration percentage) ×100.

[0095] The specific steps include:

[0096] Step 1: Multidimensional Feature Extraction

[0097] Pressure differential integral characteristic: Calculate the integral area of ​​the pressure differential curve over time to reflect the cumulative degree of blockage.

[0098] Flow volatility: The ratio of the standard deviation to the mean of flow data, characterizing the stability of the system.

[0099] Abnormal vibration percentage: The proportion of time with vibration amplitude exceeding 0.5 indicates abnormal mechanical condition.

[0100] Step 2: Feature Normalization Processing

[0101] Because the dimensions and numerical ranges of different features vary greatly, the extracted features need to be normalized. Each feature value is divided by its corresponding upper limit of the normal range, mapping all features to a uniform range of 0 to 1. This process ensures the comparability between different features, laying the foundation for subsequent weighted fusion.

[0102] Step 3: Weight Allocation and Comprehensive Evaluation

[0103] Weight of differential pressure integral characteristic: 60% (main influencing factor)

[0104] Traffic volatility weight: 30% (secondary influencing factor)

[0105] Abnormal vibration weighting: 10% (auxiliary judgment factor)

[0106] Health Calculation:

[0107] Health Index = 100 - (0.6 × pressure differential integral + 0.3 × flow fluctuation rate + 0.1 × abnormal vibration ratio) × 100.

[0108] Output result: Health score from 0 to 100, the higher the score, the better the condition of the filter.

[0109] 3. Remaining life prediction model

[0110] Model architecture: LSTM (Long Short-Term Memory) neural network model based on deep learning.

[0111] The model requires three sets of trend data over 30 days: pressure differential trend (the daily average pressure differential change trend), flow rate decay (the daily decline in flow performance), and temperature gradient (the diurnal variation characteristics of ambient temperature). The model output is the remaining service life (the predicted number of days the filter element can still be used normally, ranging from 0 to 100 days).

[0112] Step 1: Data Preprocessing and Standardization

[0113] First, the three input trend data points are standardized by subtracting their respective means and calculating their standard deviations to eliminate the influence of different parameter units. Then, the three standardized features are combined into a multi-dimensional time series input. To capture time dependencies, the algorithm uses a sliding window technique to construct the model's input sequence from 30 consecutive days of historical data.

[0114] Step 2: Deep Feature Learning;

[0115] The first LSTM layer has 32 neurons and returns the complete time series.

[0116] The second LSTM layer has 16 neurons and outputs the final time step features.

[0117] Output layer: a fully connected layer with a single neuron, using the ReLU activation function.

[0118] The optimization process uses the Adam optimizer with an adaptive learning rate to ensure that the model can converge efficiently.

[0119] Input feature dimensions:

[0120] Pressure differential change trend (time series data);

[0121] Flow decay rate (quantization performance degradation rate);

[0122] Temperature gradient (the effect of environmental changes).

[0123] Output result:

[0124] Predict the remaining effective days of the filter cartridge, ranging from 0 to 100 days;

[0125] A higher value indicates a longer remaining lifespan, while 0 indicates that immediate replacement is required.

[0126] Training requirements:

[0127] At least three months of historical operational data are required for model training;

[0128] Time series cross-validation is used to ensure prediction accuracy;

[0129] The model needs to be updated regularly to adapt to changes in the aging characteristics of the equipment;

[0130] These three core algorithms together constitute the decision-making core of the intelligent filter blockage early warning system. Through working condition adaptation, multi-feature fusion, and deep learning prediction, a complete technology chain from real-time monitoring to accurate early warning has been realized.

[0131] The system implements a multi-level early warning mechanism: mild congestion (60% threshold) is detected by APP notification, moderate congestion (75% threshold) is detected by APP + SMS reminder, and severe congestion (90% threshold) is detected by audible and visual alarm.

[0132] Technical effects of this application

[0133] 1. Precise early warning to avoid sudden failures.

[0134] Early warning: Issues graded alarms (mild / moderate / severe) before the filter element becomes completely clogged (e.g., when it reaches 60%-80% clogging) to prevent sudden equipment shutdown. Reduced false alarms: By using multi-sensor fusion (differential pressure + flow rate + temperature) and AI algorithms, it reduces false alarms caused by fluctuations in a single parameter (such as brief differential pressure changes during rapid engine acceleration).

[0135] 2. Optimize maintenance strategies and reduce costs.

[0136] Replace as needed: Replace filter cartridges based on actual clogging levels, avoiding premature replacement (saving 20%-40% on filter cartridge costs) or equipment damage caused by premature replacement. Predictive maintenance: Predict remaining service life (RUL) through trend analysis, supporting planned maintenance scheduling.

[0137] 3. Extend equipment lifespan and improve energy efficiency.

[0138] 4. Protect critical equipment:

[0139] To prevent insufficient air intake due to clogged air filters and reduce abnormal wear of piston rings.

[0140] 5. Prevents pump cavitation caused by oil filter blockage in the hydraulic system, extending pump life by more than 30%.

[0141] Industrial air compressor filter clogging warnings can reduce energy consumption by 5%-8% (maintaining optimal pressure differential).

[0142] 6. Adaptability to special scenarios

[0143] Handling complex operating conditions:

[0144] For automobiles: Identifying moisture buildup (in the cabin air filter) caused by short-distance driving, prompting for drying maintenance instead of direct replacement. Core solution: Multi-sensor fusion diagnostic algorithm.

[0145] To distinguish between "permanent dust blockage" and "reversible moisture buildup," additional sensing dimensions and intelligent algorithms are needed.

[0146] New or utilized sensor data:

[0147] Existing data: Pressure difference ΔP across the air conditioning filter, blower airflow (which can be estimated using speed or damper sensor).

[0148] Key new / reused data:

[0149] T_Evap: Evaporator temperature (obtained from the vehicle's air conditioning control system);

[0150] T_in-car: In-vehicle ambient temperature / humidity (obtained from the vehicle domain controller);

[0151] Trip_Duration: Duration of a single trip (obtained from the vehicle's CAN bus);

[0152] A / C_Runtime: The running time of the air conditioning compressor during a single trip (obtained from the CAN bus);

[0153] B. Diagnostic Logic and Algorithm Flow:

[0154] Step 1: Identify risk scenarios

[0155] The system continuously monitors vehicle operating data. When both of the following conditions are met simultaneously, the system will mark the trip as a "high-risk trip due to moisture accumulation."

[0156] Trip duration too short: The single trip time is less than the set threshold (e.g., 15 minutes). This is the primary condition that moisture cannot be fully evaporated.

[0157] The air conditioner operates at full capacity: During this short trip, the air conditioner compressor runs for a sufficiently long time (e.g., more than half the trip time). This ensures that the evaporator has ample opportunity to produce a large amount of condensate.

[0158] This step is a prerequisite for early warning, as it identifies the vehicle usage scenarios that are most likely to cause the filter to become damp.

[0159] Step 2: Monitor "characteristic signals of moisture accumulation"

[0160] After each "high-risk trip" ends, the system monitors the following characteristics:

[0161] High resistance signal: ΔP reading consistently exceeds the dynamic threshold model prediction.

[0162] Environmental evidence: The humidity inside the car, H_In-Car, remains consistently high (e.g., >70%).

[0163] Recovery characteristic: During a subsequent long drive, as the air conditioning system operates for an extended period, the ΔP reading shows a slow downward trend (as moisture is gradually dried). This is the most critical difference from permanent blockage (where ΔP simply flattens or continues to rise).

[0164] Step 3: Diagnosis and Decision-Making Logic

[0165] Logical description:

[0166] The system employs a comprehensive decision-making model to diagnose the condition of the air conditioning filter. This model receives four key input signals:

[0167] Filter element ventilation resistance increasing trend: A value indicating the degree to which the current resistance has increased relative to the normal level.

[0168] High-risk trip indicator for moisture buildup: A "yes / no" Boolean sign indicates whether the trip is a short-distance, high-air conditioning-use scenario that is prone to moisture buildup.

[0169] In-vehicle humidity: A numerical value indicating the level of humidity inside the vehicle.

[0170] Indications of reversible drag recovery: A Boolean "yes / no" sign indicating whether a reversible decrease in drag was observed during long-distance driving.

[0171] The decision-making process is as follows:

[0172] First, the system assigns different importance weights to these four input signals:

[0173] The increasing trend of filter element ventilation resistance was given the highest weight (50%) because it is the most direct evidence of performance degradation.

[0174] High-risk travel markers for moisture buildup are given significant weight (30%) because they provide crucial contextual information.

[0175] In-vehicle humidity was assigned a secondary weight (10%) as supporting environmental evidence.

[0176] The signs of reversible resistance recovery are given a negative weight (-10%), and their appearance strongly suggests that the problem stems from reversible moisture rather than permanent blockage.

[0177] Next, the system calculates a comprehensive "moisture buildup confidence score". The calculation method is as follows: multiply the values ​​of the first three signals by their respective weights and add them together, then subtract the value of the fourth signal (signs of recovery) multiplied by its own weight.

[0178] Finally, the system makes a final decision based on the calculated confidence level:

[0179] If the confidence level exceeds 0.6 (60%), the problem is identified as "moisture buildup".

[0180] If the confidence level is below 0.2 (20%), and the increasing trend of filter element ventilation resistance is already very obvious (above 0.7), then the problem is determined to be "solid particle blockage".

[0181] If neither of the above two conditions is met, the system returns a "normal status" conclusion and continues monitoring.

[0182] Step 4: Execution and Interaction

[0183] Once the system determines that there is "moisture buildup," it will not trigger a "replace filter" alarm, but will instead send a detailed "drying maintenance reminder" to the user.

[0184] The prompt will clearly state the cause: explaining that the problem is due to the filter becoming damp from a short trip, rather than permanent damage. It will also provide specific operating instructions: guiding users to perform a simple drying procedure, such as "In a safe environment, turn on the external air circulation and the highest heating setting, and run the air conditioner for 15-20 minutes."

[0185] Managing user expectations: Informing users that airflow is expected to recover after this operation avoids an unnecessary replacement, saving costs and improving the user experience.

[0186] Industrial applications: Differentiating between dust load changes (e.g., in mining equipment) and actual filter aging. Core solution: Diagnostic algorithms based on trend analysis and condition recognition.

[0187] Key identification: The "reversibility" and "trend characteristics" of pressure differential changes.

[0188] A. Newly added or utilized sensor data:

[0189] Existing data:

[0190] ΔP: Pressure difference across the filter element

[0191] Flow: System flow

[0192] Key new data:

[0193] Dust_Concentration: Dust concentration before the filter element inlet (monitored in real time using a laser dust sensor or electrostatic induction sensor).

[0194] Work_Cycle_Phase: The phase of the equipment's working cycle (obtained from the PLC, such as "drilling", "loading", "driving", "standby").

[0195] B. Diagnostic Logic and Algorithm Flow:

[0196] Step 1: Establish a dynamic response model for "dust load-pressure difference". The system first learns the normal response range of pressure difference under different dust concentrations when the filter element is healthy.

[0197] Expected pressure differential increment = f(dust concentration, flow rate) #Pseudocode logic

[0198] When the real-time differential pressure increment significantly exceeds the model's expectations, it may indicate that the filter element is aging.

[0199] Step Two: Identifying the "Reversibility" Feature. This is the core of the identification process. After detecting an increase in differential pressure, the system will not immediately trigger an alarm, but will instead activate an "observation window."

[0200] Scenario: After a period of high-dust operation, the equipment enters a low-dust operation or standby state.

[0201] monitor:

[0202] If it is a change in dust load: After entering a low-dust environment, as clean air continues to pass through, some surface dust will be blown off, and the pressure difference (ΔP) will show a plateau period or a slow downward trend, and eventually stabilize at a new level that is slightly higher than before.

[0203] If it's genuine filter aging: After entering a low-dust environment, the pressure differential (ΔP) shows no decreasing trend, but remains high and continues to rise in any subsequent operation. The filter has lost its "self-cleaning" ability.

[0204] Step 3: Quantifying the "Permanent Blockage Increment" The system uses an algorithm to calculate the unrecoverable pressure differential increment after each high dust event.

[0205] Permanent blockage increment = Post-event steady-state pressure difference - Pre-event baseline pressure difference

[0206] The system continuously tracks the cumulative value of this "permanent blockage increment," which is the key indicator reflecting the true health status of the filter element.

[0207] Step 4: The intelligent diagnostic system based on the working cycle links the equipment's working cycle with the filter element's status, enabling more accurate judgment.

[0208] Example of a rule:

[0209] If the equipment is in the "crusher feeding" stage AND dust concentration > 50mg / m³ AND ΔP rises sharply: → judge as "normal high load" and suppress alarm.

[0210] If the equipment is in the "transfer travel" stage AND dust concentration < 5mg / m³ AND ΔP remains 30% higher than the historical benchmark: → judged as "true aging", triggering an early warning.

[0211] 3. User Interaction and Execution Plan

[0212] After system diagnosis, completely different prompts will be given:

[0213] When the condition is determined to be "dust load change":

[0214] The system displays the message: "Note: The current environment is high-dust. The filter element is working efficiently, and the increased pressure difference is normal. The system will automatically assess the filter element's recovery status once the environment improves."

[0215] System status: Alarms are suppressed, and the UI displays a "High load operation" status icon.

[0216] When determined to be "genuine filter element aging":

[0217] Warning message: "Warning: Irreversible degradation of filter performance has occurred. Despite a clean environment, the flow resistance remains high, and permanent clogging has reached XX%. Replacement is recommended at the next scheduled maintenance."

[0218] System status: Triggered a planned maintenance alert and provided a detailed performance degradation report.

[0219] Multi-filter collaborative management:

[0220] Fuel system coarse and fine filter linkage monitoring optimizes the maintenance schedule for both filtration stages. Core solution: Three-stage differential pressure monitoring and intelligent linkage logic.

[0221] The core of this solution is to install differential pressure sensors on the two-stage filter elements respectively, and to establish a linkage model between the two through an algorithm.

[0222] A. System sensor deployment:

[0223] First monitoring point: ΔP_coarse - pressure difference before and after coarse filtration.

[0224] Second monitoring point: ΔP_fine - pressure difference before and after fine filtration.

[0225] Third monitoring point (optional but recommended): P_in - Inlet pressure before coarse filtration, used to monitor the oil pump's oil supply capacity.

[0226] B. Diagnostic Logic and Maintenance Optimization Algorithm:

[0227] Step 1: Establish a two-stage filter "health baseline" system. When all filters are brand new, record the baseline values ​​of ΔP_coarse and ΔP_fine at different flow rates, and establish the initial proportional relationship between the two: R_base = ΔP_fine / ΔP_coarse.

[0228] Step Two: The real-time status diagnosis and fault location system performs a comprehensive diagnosis based on the three pressure readings to accurately pinpoint the source of the problem.

[0229]

[0230] Step 3: Optimizing the collaborative strategy for maintenance rhythm. This is key to achieving cost savings. The system no longer suggests "simultaneous replacement," but instead provides precise recommendations based on the actual status of each component:

[0231] When ΔP_coarse reaches the warning threshold first:

[0232] Action: The system prompts "Replace the coarse filter".

[0233] Logic: After replacing the coarse filter, the total pressure loss of the system is significantly reduced, the burden on the fine filter is lessened, and its pressure difference ΔP_fine may decrease slightly, thereby extending the service life of the fine filter. In this case, replacing the fine filter is absolutely not recommended.

[0234] When ΔP_fine reaches the warning threshold first, while ΔP_coarse remains normal:

[0235] Action: The system prompts "Replace fine filter" and simultaneously issues a prompt "Check coarse filter effectiveness".

[0236] Logic: It is unusual for the fine filter to clog before the coarse filter, which may mean:

[0237] The coarse filter is damaged and has lost its protective function, resulting in a large number of pollutants directly attacking the fine filter.

[0238] Abnormal contaminants (such as microbial slime and gum) were found in the fuel. These substances passed through the coarse filter but clogged the fine filter.

[0239] Maintenance strategy: When replacing the fine filter, the coarse filter must be checked for damage. If the coarse filter is damaged, it must be replaced as well.

[0240] Proportional Relationship Warning:

[0241] The system calculates the ratio R_current = ΔP_fine / ΔP_coarse in real time.

[0242] If R_current deviates significantly from R_base, it indicates that the clogging rate of the two-stage filter elements is abnormal. The system will issue an early warning, suggesting that there may be oil quality issues or mismatched filter element selection.

[0243] 3. User Interaction and Execution Plan

[0244] The system provides users with highly actionable intelligent suggestions:

[0245] Scenario 1 (Replacing the coarse filter is recommended):

[0246] Tip title: "Replace the fuel filter as recommended"

[0247] Scenario 2 (Recommended to replace the fine filter and check the coarse filter):

[0248] The message reads: "It is recommended to replace the fuel filter and check the coarse filter."

[0249] Obviously, the specific implementation of this invention is not limited to the above-described methods. Any non-substantial improvements made using the inventive concept and technical solution of this invention are within the protection scope of this invention.

Claims

1. A method for early warning of filter clogging, comparing the collected differential pressure data representing the clogging state of the filter with a differential pressure threshold value and outputting a warning according to the comparison result; characterized in that: The pressure difference threshold is calculated according to the collected flow at the filter element and the ambient temperature.

2. The intelligent filter element blockage early warning method of claim 1, wherein: The reference pressure difference is calculated by the collected flow, and then the temperature is used to compensate the reference pressure difference to obtain the pressure difference threshold.

3. The intelligent filter cartridge clogging pre-warning method of claim 2, wherein: A basic relationship model of flow and pressure difference is established, a power function relationship is used, and the flow value is used as an input parameter to obtain a theoretical reference pressure difference value under the current flow condition.

4. The intelligent filter element blockage early warning method of claim 2 or 3, characterized in that: The theoretical reference pressure difference is calculated by using a reference pressure difference calculation formula, and the reference pressure difference calculation formula is: reference pressure difference = K1 x (flow value / 10) 1.8 ; K1 is a proportional coefficient, which is a constant.

5. The intelligent filter element clogging pre-warning method of claim 2 or 3, wherein: A compensation coefficient K2 is calculated based on the temperature, the compensation coefficient K2 is multiplied by the reference pressure difference to obtain the compensated pressure difference threshold.

6. The intelligent filter cartridge clogging pre-warning method of claim 5, wherein: The compensation coefficient K2 is negatively correlated with the temperature, and the lower the temperature, the larger the compensation coefficient is set.

7. The intelligent filter element clogging pre-warning method of claim 5 or 6, wherein: The calculation formula of the compensation coefficient K2 is: K2 = 1 + (reference temperature - actual temperature) * C; C is a constant coefficient.

8. The intelligent filter cartridge clogging pre-warning method of claim 1, wherein: A hierarchical early warning method is used to send different early warning information for different ranges of reference pressure difference.

9. The intelligent filter element clogging pre-warning method of any one of claims 1-8, wherein: A fuzzy logic evaluation method based on multi-feature fusion receives three groups of time series data of pressure difference, flow, and vibration, calculates a comprehensive health index of the filter element, and displays the comprehensive health index as a part of the early warning information to the user.

10. An intelligent filter cartridge clogging early warning system characterized by: It comprises a sensor, a processor, and an early warning module. The sensor is used to collect pressure difference data representing the blockage state of the filter element, and the output end is connected to the processor; the output end of the processor is connected to the early warning module, and the processor compares the pressure difference data with the pressure difference threshold and outputs the early warning according to the comparison result, wherein the pressure difference threshold in the processor is calculated according to the collected flow at the filter element and the ambient temperature.

Images