A method for generating and visualizing health index of electromechanical equipment based on multi-dimensional sensor data fusion

Abstract

The application discloses a kind of based on multi-dimensional sensing data fusion electromechanical equipment health index dynamic generation and visualization method, belong to wisdom water affairs and electromechanical equipment intelligent monitoring technical field.The present water affairs equipment monitoring data is single, lacks quantitative health evaluation system and the problem of weak visualization ability, the present application is by deploying multiple types of sensors synchronous acquisition equipment vibration, temperature, noise, ultrasonic, magnetic flux and electrical parameter etc.Multiple dimensional sensing data is preprocessed after data, and multidimensional depth fusion analysis is carried out and adaptive fusion weight is distributed, constructs health score model, and the fusion result is dynamically mapped as the equipment health index of 0 to 100 interval and is divided into four state threshold values, finally, the integrated dynamic display and interaction of equipment health state is realized through visualization rendering module.

Description

Technical Field

[0001] This invention relates to the field of intelligent health monitoring technology for electromechanical equipment, and in particular to a method for dynamically generating and visualizing the health index of electromechanical equipment based on multi-dimensional sensor data fusion. Background Technology

[0002] Waterworks and wastewater treatment plants are core infrastructure for urban water supply and wastewater treatment. Their internal electromechanical equipment operates continuously 24 / 7 in complex environments characterized by high humidity, dust, corrosive media, and frequent load fluctuations, resulting in a significantly higher equipment failure rate than in conventional industrial settings. The stable operation of waterworks electromechanical equipment directly impacts urban water supply security, wastewater treatment compliance, and the protection of municipal and public welfare. Unplanned shutdowns or malfunctions of core equipment not only incur high costs for repair and replacement but also easily lead to water supply interruptions, wastewater retention, and substandard treated water quality, causing severe economic losses and social impacts.

[0003] Currently, the operation and maintenance (O&M) models for electromechanical equipment in domestic waterworks and sewage treatment plants remain relatively traditional. Most sites still rely on manual periodic inspections, planned maintenance, and emergency repairs after a failure. Manual inspections are limited by factors such as personnel experience, inspection frequency, and subjective judgment, making it impossible to accurately detect early-stage minor faults in real time. This often results in a focus on emergency repairs rather than proactive prevention, easily leading to the accumulation of minor equipment problems into major malfunctions. Meanwhile, planned maintenance often employs fixed-cycle maintenance strategies, which cannot adapt to the actual operating conditions of the equipment. This can easily lead to over-maintenance and wasted O&M costs, or untimely maintenance resulting in equipment operating with defects, failing to meet the development needs of refined and intelligent O&M in smart water management.

[0004] Existing water equipment monitoring technologies have significant shortcomings, making it difficult to achieve accurate quantification and intuitive control of equipment health status. Most current monitoring systems employ a single-sensor data acquisition and analysis model, only issuing threshold alarms for single parameters such as equipment vibration, temperature, and pressure, failing to achieve integrated analysis of multi-dimensional sensor data including vibration, temperature, current, voltage, noise, operating load, and media conditions. Water field equipment is diverse in brand and communication protocols, resulting in independent operational data from different devices and prominent data silos. Single-dimensional data cannot comprehensively reflect the overall health status of the equipment, easily leading to false alarms and missed alarms, low fault identification accuracy, and poor applicability.

[0005] More critically, existing technologies cannot achieve dynamic quantitative assessment of equipment health status. Current monitoring methods can only provide early warnings of parameter exceeding limits, lacking a unified and standardized equipment health index evaluation system. They cannot transform complex equipment operation data into intuitive and comparable quantitative health indicators, making it difficult for maintenance personnel to accurately determine equipment health degradation trends and predict potential failure risks. Furthermore, existing monitoring platforms offer limited data display, mostly consisting of raw data listings and simple curves, lacking customized visualization models for water utilities' electromechanical equipment maintenance scenarios. This prevents layered display of equipment health status, fault tracing, and trend projection, resulting in a lack of intuitive data support for maintenance decisions and a severe deficiency in intelligent and visualized management capabilities.

[0006] With the continuous development of smart water management and the ongoing upgrading of safety control standards in the municipal water supply and sewage treatment industries, the industry urgently needs an intelligent operation and maintenance method that can adapt to the complex operating conditions of waterworks and sewage treatment plants, achieve multi-dimensional sensor data fusion analysis, dynamically quantify and generate equipment health indices, and visually and intuitively display equipment operating status. This would break down data silos, enable the transformation of equipment status from passive alarm to proactive prediction, improve the operation and maintenance efficiency of water electromechanical equipment, reduce failure risks and operation and maintenance costs, and ensure the safe and stable operation of the water system.

[0007] The information disclosed in this background section is intended only to enhance the understanding of the overall background of the invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0008] The purpose of this invention is to provide an intelligent operation and maintenance method that can realize multi-dimensional sensor data fusion analysis, dynamically quantify and generate equipment health index, and visually and intuitively display the equipment operating status.

[0009] To achieve the above objectives, the present invention provides the following solution: A method for dynamically generating and visualizing the health index of electromechanical equipment based on multi-dimensional sensor data fusion includes the following steps: Multi-source heterogeneous sensor data acquisition and preprocessing: For electromechanical equipment in water treatment scenarios, multiple types of sensor terminals are deployed to simultaneously collect multi-dimensional sensor data on the operating status of the equipment. The multi-dimensional sensor data includes vibration data, temperature data, noise data, ultrasonic data, magnetic flux data, and electrical parameter data. The collected multi-source raw sensor data is subjected to standardized preprocessing, including data cleaning, outlier removal, time sequence alignment, unit unification, and missing data completion, to obtain preprocessed multi-dimensional sensor data. Multidimensional sensor data fusion modeling: Based on the equipment failure mechanism and operating condition characteristics, feature extraction and fusion weights are performed on different sensing dimensions in the preprocessed multidimensional sensor data. The fusion weights of each monitoring parameter are dynamically adjusted according to the equipment type, years of operation, load fluctuation pattern and environmental corrosion degree. The feature data of each dimension after weight adjustment are subjected to deep fusion correlation analysis to generate multidimensional fusion analysis results. Dynamic calculation of equipment health index: Construct a health scoring model adapted to water electromechanical equipment, map the multi-dimensional fusion analysis results to an equipment health index in the range of 0 to 100, and divide the equipment health index into four levels of state thresholds: healthy, sub-healthy, hidden danger, and faulty; adopt a time-series rolling calculation method, combine the historical operating baseline data of the equipment with the real-time operating deviation, and dynamically update the real-time health index and health decay rate of the equipment. Visualized display of equipment health status: Based on dynamically generated equipment health indices, the visualization rendering module enables integrated visualization of equipment operating parameter curves, health index change trends, equipment health status topology distribution maps, fault location markings, and anomaly tracing information. It also supports interactive operations such as viewing the status of a single device, providing an overview of the health status of all equipment on the site, and tracing back historical operation and maintenance data.

[0010] Optionally, in the multi-source heterogeneous sensing data acquisition and preprocessing step, the multi-type sensor terminals include vibration sensors, temperature sensors, noise sensors, ultrasonic sensors, and magnetic flux sensors. The multi-dimensional sensing data acquired simultaneously also includes equipment operating current, operating voltage, operating power, load conditions, operating time, and environmental parameters such as temperature and humidity and media corrosion conditions in the plant area.

[0011] Optionally, in the multi-dimensional sensor data fusion modeling step, the feature extraction based on the equipment fault mechanism and operating condition characteristics for different sensing dimensions includes: extracting bearing fault characteristic frequencies, imbalance characteristics, and misalignment characteristics from vibration data; extracting temperature rise rate and heat accumulation characteristics from temperature data; extracting abnormal acoustic patterns and high-frequency sound wave characteristics from noise and ultrasonic data; and extracting motor magnetic field operating state characteristics and electrical fault characteristics from magnetic flux data.

[0012] Optionally, in the multidimensional sensor data fusion modeling step, the fusion weights are allocated differently based on the contribution of different monitoring parameters to the health status of the equipment, and the fusion weights are dynamically adjusted as the operating status of the equipment changes, so as to achieve fusion analysis with mutual verification of five parameters.

[0013] Optionally, in the dynamic calculation step of the equipment health index, the classification method of the four-level state thresholds of health, sub-health, hidden danger, and fault is as follows: a health index ≥ 80 is a healthy level, and the equipment is operating normally; a health index of 60 to 80 is a sub-health level, and the vibration or temperature increases by 20% to 50% compared to the baseline value; a health index of 40 to 60 is a hidden danger level, and the vibration or temperature increases by 50% to 100% compared to the baseline value; a health index < 40 is a fault level, and the vibration or temperature exceeds the baseline value by more than 100% or exceeds the set upper limit.

[0014] Optionally, in the dynamic calculation step of the equipment health index, the step of adopting a time-series rolling calculation method to dynamically update the real-time health index and health decay rate of the equipment is as follows: with a predetermined time window as the period, the current health index is continuously calculated based on the real-time sensor data of the current period, and combined with the historical operating baseline data, the rate of change of the health index over time is calculated to characterize the trend of equipment performance degradation, thereby realizing the transformation of equipment health status from qualitative judgment to quantitative digital assessment.

[0015] Optionally, in the equipment health status visualization display step, the equipment health status topology distribution map uses the process flow diagram or equipment layout diagram as the base map, and overlays health status color indicators at the corresponding positions of each equipment, where green represents the healthy level, yellow represents the sub-healthy level, orange represents the potential danger level, and red represents the fault level, forming an integrated management view of the health status of all equipment in the station.

[0016] Optionally, in the equipment health status visualization display step, the abnormality tracing information includes fault type diagnosis results and fault location information. The fault types include bearing faults, imbalance faults, misalignment faults, mechanical loosening faults, motor rotor faults, seal leakage, cavitation, and insufficient bearing lubrication. The fault location information includes the equipment number where the fault occurred and the location of the measuring point.

[0017] Optionally, the method further includes a step of generating an equipment health assessment report: based on the historical trend data of the health index of each device, automatically generating a health assessment report for each device, the report including the current equipment health score, health trend chart and recommended maintenance plan; and performing comprehensive scoring statistics on all devices in the station according to equipment type and / or operating area to generate a comprehensive health score for all devices in the station.

[0018] Optionally, the method further includes a warning and work order linkage step: a four-level warning system is set based on the four-level status thresholds. When the equipment health index drops to the corresponding threshold range, a warning signal of the corresponding level is triggered and pushed to the maintenance personnel terminal. When a hidden danger level or fault level warning is triggered, a maintenance work order is automatically triggered. The maintenance work order includes the triggering equipment information, alarm type, suggested handling measures and expected completion time. After the work order is completed, the maintenance record is automatically archived to the equipment history.

[0019] Compared with the prior art, the present invention has the following beneficial effects: This invention overcomes the shortcomings of existing technologies that rely solely on a single sensing parameter (such as vibration or temperature) for threshold alarms by deploying a multi-dimensional intelligent sensor that integrates vibration, temperature, noise, ultrasound, and magnetic flux on the equipment. The five parameters are interconnected and mutually reinforcing, simultaneously covering all types of faults, including mechanical faults (bearing wear, rotor imbalance, misalignment, mechanical loosening), electrical faults (motor demagnetization, rotor bar breakage, air gap eccentricity), and hydraulic faults (seal leakage, cavitation). This significantly improves fault identification accuracy and effectively avoids the false alarms and missed alarms common in traditional single-parameter monitoring. The fault diagnosis accuracy is significantly higher than existing single-sensor solutions.

[0020] This invention constructs a health scoring model adapted to water utilities' electromechanical equipment scenarios, mapping complex multidimensional sensor data into a unified equipment health index ranging from 0 to 100, and dividing it into four quantitative threshold levels: healthy, sub-healthy, potential hazards, and faults. Through a time-series rolling calculation method, it continuously tracks the rate of equipment health degradation, fundamentally shifting from traditional qualitative judgments (such as "normal operation / abnormal operation") to quantitative digital assessments (such as "current health index 75 points, decaying at a rate of 0.3 points per day, expected to enter the potential hazard zone in 45 days"). Water utility managers can use this model to accurately determine the health trend of each piece of equipment and predict fault risks in advance, replacing the traditional method of relying on manual experience and improving the scientific rigor and accuracy of operation and maintenance decisions.

[0021] This invention, through its built-in AI edge computing chip, enables the complete data processing and analysis, as well as anomaly identification, to be performed locally on the device. This achieves second-level fault warnings without relying on the cloud, resulting in a response speed far faster than traditional pure cloud-based analysis methods. Combined with dynamic health indices and remaining lifespan predictions, maintenance personnel can identify risks 1-3 months in advance during the initial stages of a fault and formulate spare parts procurement and maintenance plans. This transforms the traditional "post-fault repair" and "fixed-cycle over-maintenance" models into "state-based predictive maintenance," effectively solving the problem of both over-maintenance and under-maintenance in planned maintenance. It significantly reduces the probability of unplanned downtime, emergency repair costs, and spare parts inventory pressure.

[0022] This invention establishes a visual rendering module for equipment status adapted to water management scenarios. Using a process flow diagram as the base image, it identifies the health status of each piece of equipment using four colors: green, yellow, orange, and red, achieving an integrated display of "process monitoring + equipment health management." Maintenance personnel can quickly locate problematic equipment, view fault tracing information, and health trends on a single screen, without needing to switch between multiple platforms or possess specialized vibration analysis knowledge to complete daily inspections. Simultaneously, the system automatically generates quarterly health assessment reports that can be submitted to regulatory authorities, making equipment operating status and effluent water quality compliance transparent and verifiable, effectively meeting the compliance requirements of water management regulations.

[0023] This invention incorporates an AI algorithm model, supporting continuous upgrades and iterations of both the firmware and the algorithm model. As operational data accumulates, the fusion weight model can adaptively optimize and adjust. By establishing differentiated weight allocation strategies for different equipment types (e.g., blowers vs. water pumps) and dynamically adjusting the fusion parameters based on equipment age, load fluctuation patterns, and environmental corrosion levels, this invention exhibits excellent scenario adaptability and generalization capabilities. It is widely applicable to various water-related scenarios such as waterworks, sewage treatment plants, and pumping stations, as well as to electromechanical equipment of different brands and service lives. This avoids the drawbacks of traditional fixed threshold schemes, which have poor applicability under different operating conditions, thus maximizing the value of equipment. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a schematic diagram of the method for dynamically generating and visualizing the health index of electromechanical equipment based on multi-dimensional sensor data fusion, provided in an embodiment of the present invention. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] The purpose of this invention is to provide an intelligent operation and maintenance method that can realize multi-dimensional sensor data fusion analysis, dynamically quantify and generate equipment health index, and visually and intuitively display the equipment operating status.

[0028] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0029] Example 1: like Figure 1 As shown in the figure, this embodiment uses a rural sewage treatment plant (designed treatment capacity of 400m³) as an example. 3 Using the MBR process as an application scenario, a method for dynamically generating and visualizing the health index of electromechanical equipment based on multi-dimensional sensor data fusion was implemented for the station's four blowers (5.5kW, 1455rpm) and three water pumps (lift pump, return pump, and sludge return pump).

[0030] I. Data Acquisition and Preprocessing of Multi-Source Heterogeneous Sensors: A GX1-P12-RS485 TxMaster multidimensional intelligent health diagnostic device is installed on both the drive end (DE) and non-drive end (NDE) of the drive motor of each blower. This sensor integrates a triaxial accelerometer, temperature sensor, microphone, ultrasonic sensor, and magnetic flux sensor, and can simultaneously collect five-dimensional data: vibration (triaxial, range ±16g, sampling rate 3200Hz), temperature (resolution 1℃), noise (100Hz~80kHz), ultrasound (software configurable frequency bands, including 0~250kHz, 351~815kHz, 1.024~2.475MHz, 3.072~4.8MHz), and magnetic flux (x / y axis ±1300μT, z axis ±2500μT). Simultaneously, it reads electrical parameters such as operating current, voltage, and power of the equipment via an RS-485 interface, and accesses environmental temperature and humidity sensor data from the factory area via a 4G gateway.

[0031] Each sensor continuously acquires vibration time-domain signals at a sampling rate of 3200Hz, temperature at a sampling rate of 1Hz, magnetic flux at a sampling rate of 250Hz, and ultrasonic signals at a sampling rate of 100kHz. The acquired raw data undergoes preliminary feature extraction on the AI ​​edge computing chip built into the sensor before being transmitted to the cloud-based PHM platform via a 4G network.

[0032] In the cloud, preprocessing operations are performed on the received raw sensor data from multiple sources: (1) Data cleaning: Remove null values ​​caused by abnormal power loss of the sensor and communication interruption, as well as outlier values ​​that are significantly beyond the range (e.g., vibration values ​​exceeding ±20g are considered outlier values).

[0033] (2) Outlier removal: Using the 3σ criterion, for each measuring point's effective vibration value sequence, points that deviate from the mean by more than 3 times the standard deviation are removed.

[0034] (3) Time alignment: Data with different sampling rates (vibration 3200Hz, temperature 1Hz, magnetic flux 250Hz) are resampled according to a unified timestamp (millisecond level). With 10 seconds as the time window, the effective value of vibration, peak value, kurtosis, average temperature, ultrasonic energy value, and average magnetic flux value in each window are calculated to form an aligned feature vector.

[0035] (4) Dimensional unification: All physical quantities are normalized to the range of [0,1], with the effective value of vibration based on the limit value of ISO 10816-3 standard, and the temperature based on the maximum allowable temperature of the equipment (e.g., the allowable temperature rise of the blower motor winding is 80℃).

[0036] (5) Missing data completion: For data points missing due to brief communication interruption, linear interpolation is used to complete them; if the missing data is missing for more than 30 seconds, the data in that period is marked as invalid and will not be included in subsequent fusion calculations.

[0037] After preprocessing, a set of standardized feature vectors F = [V] is generated every 10 seconds for each device measurement point. x V y V z [, T,N, U, M, I, P], where V x / V y / V z denoted as RMS value of triaxial vibration (mm / s), T as temperature (°C), N as noise sound pressure level (dB), U as ultrasonic energy characteristic value, M as magnetic flux intensity (μT), I as current (A), and P as active power (kW).

[0038] II. Multidimensional sensor data fusion modeling: For the two different types of equipment in the sewage treatment plant, blowers and water pumps, a fusion weight allocation model is established separately.

[0039] (1) Blower weight model: Based on the analysis of blower failure mechanisms, the most common failures are: fatigue spalling of the drive-end bearing (accounting for approximately 35% of failures), rotor imbalance (25%), belt loosening (20%), motor winding overheating (10%), and broken motor rotor bars (10%). Therefore, the initial weight allocation is as follows: Vibration (DE end): weight 0.35 (covering bearing failure and imbalance); Vibration (NDE end): weight 0.15 (mainly monitors bearing thermal failure); Temperature: weight 0.15 (for monitoring overheating and lubrication failure); Noise: Weight 0.10 (for monitoring abnormal noises); Ultrasound: weight 0.10 (monitoring early bearing lubrication deficiencies and seal leaks); Magnetic flux: weight 0.10 (monitoring broken bars and air gap eccentricity of motor rotor); Electrical parameters (current, power): weight 0.05 (to aid in judging load changes); Meanwhile, the weights are dynamically adjusted according to the equipment's operating years: for example, for blowers that have been in operation for more than 5 years, the probability of bearing failure increases, so the vibration weight is increased to 0.40 and the temperature weight is increased to 0.18; for newly commissioned equipment (<1 year), the magnetic flux weight (used to establish an electrical baseline) and the ultrasonic weight (used to establish a lubrication baseline) are increased.

[0040] (2) Pump weight model: For booster pumps and return pumps, the most common failures are drive-end bearing wear (40%), impeller erosion / cavitation (30%), mechanical seal leakage (20%), and shaft misalignment (10%). Weighting: Vibration (pump body bearing housing): weight 0.40; Vibration (motor drive end): weight 0.20; Ultrasonic testing: weight 0.20 (focus on monitoring seal leaks and cavitation); Temperature: weight 0.10; Noise: weight 0.05; Magnetic flux: weight 0.03; Electrical parameters: weight 0.02; During the fusion calculation, a weighted summation method is used to obtain the initial fusion value. Meanwhile, a working condition correction factor λ is introduced, which is corrected based on the current load rate (actual power / rated power) and ambient humidity: when the load rate is below 60%, the vibration weight is appropriately reduced (because the vibration reference value is low under low load), and when the humidity is >80%, the temperature weight is increased (because high humidity environment accelerates bearing corrosion).

[0041] The final output is the multidimensional fusion analysis result R = λ × S.

[0042] III. Dynamic Calculation of Equipment Health Index: A health scoring model is constructed, and the results of multidimensional fusion analysis R are mapped to a health index H in the range of 0 to 100.

[0043] First, establish an equipment baseline: During the first week after new equipment is put into operation or after a major overhaul, collect multidimensional fusion values ​​R_baseline under normal operating conditions as a health benchmark (corresponding to a health index of 100). Then, calculate the health index in real time using the Exponentially Weighted Moving Average (EWMA) method. Where α is the attenuation coefficient, set according to the equipment type: α=3.0 for blowers, α=2.5 for water pumps. When R(t) equals Rbaseline When H(t) = 100; when R(t) exceeds R baseline When the value is 50%, H(t)≈60; when it exceeds 100%, H(t)≈37.

[0044] Meanwhile, a time-series rolling window (window length 30 days, sliding 1 day at a time) is used to calculate the health decay rate ΔH / Δt, which is the slope of the H value over the most recent 30 days, expressed in minutes per day. For example, if the health index of a blower drops from 85 to 75 within 30 days, the decay rate is -0.33 minutes per day, indicating accelerated bearing wear.

[0045] Four-level state thresholds are defined based on the H value: Health Level (Green): H ≥ 80. For example, for blower No. 2, H=92, effective vibration value 1.2mm / s (baseline 1.0mm / s), temperature 42℃ (baseline 40℃), the system is judged to be healthy.

[0046] Sub-health level (yellow): 60 ≤ H < 80. For example, for pump No. 1 with H=68, the vibration at its drive end increased from 1.5mm / s to 2.1mm / s (a 40% increase from the baseline), and the temperature rose from 38℃ to 46℃. The system indicates that the frequency of inspections needs to be increased.

[0047] Hidden Danger Level (Orange): 40 ≤ H < 60. For example, for blower No. 3, H=53, the vibration value increased by 75% compared to the baseline, and a clear BPFO (outer ring pass frequency) sideband appeared in the velocity spectrum. The system diagnosed it as early pitting corrosion of the bearing outer ring, pushed an alarm, and requested an on-site inspection to be arranged within 24 hours.

[0048] Fault level (red): H < 40. For example, if a high-pressure pump has H = 28, the vibration value exceeds the baseline by 120%, and the temperature exceeds the upper limit of 85°C, the system will immediately trigger an emergency shutdown recommendation and push it to all responsible persons.

[0049] IV. Visual display of equipment health status: On the PHM platform's equipment health overview screen, with the wastewater treatment process flow as the background (displaying "collection well → equalization tank → biological treatment tank → MBR → filter → disinfection → effluent"), a circular health status indicator is superimposed above the corresponding equipment icon.

[0050] Specific interactive functions include: (1) Detailed viewing of a single device: Click on any device card, and a sidebar will pop up to display the device's real-time vibration time-domain waveform, FFT spectrum (automatically labeled with bearing characteristic frequencies BPFO / BPFI / BSF / FTF), temperature trend curve (last 7 days), and health index change curve (last 30 days). At the same time, the device's most recent diagnostic conclusion will be displayed, for example: "The characteristic frequency of the outer ring of the No. 1 blower drive end bearing is BPFO=89.2Hz, with an amplitude of 0.012g, which is 3 times higher than the baseline. It is diagnosed as early pitting corrosion of the bearing outer ring. The remaining life is predicted to be about 45 days. It is recommended to replace the bearing within 30 days." (2) Overview of the overall health status of all equipment: The main screen uses green / yellow / orange / red to indicate the health status of each equipment, counts the number of currently healthy equipment (≥80 points), the number of sub-healthy equipment, the number of equipment with potential hazards, and the number of equipment with malfunctions, and displays the overall health score of the entire station (the weighted average of the H values ​​of all equipment, with the importance of the equipment as the weight).

[0051] (3) Historical maintenance data retrospective: Users can select any time range (such as January to June 2026) to view the health index change curve of each device during that period, and can overlay maintenance events that occurred during the period (such as replacing bearings and adding grease) to evaluate the maintenance effect.

[0052] (4) Fault risk pop-up warning: When the equipment health index drops to the hidden danger level threshold, the system will automatically pop up a warning window, displaying the fault type, suggested handling measures, and responsible person information, and push it simultaneously through APP and SMS.

[0053] V. Early Warning and Work Order Linkage: The system is configured with a four-level early warning mechanism, corresponding to health status thresholds: When H drops below 80 (sub-health level), a yellow alert is generated, and a notification is pushed to the app to the on-duty maintenance personnel, reminding them to strengthen inspections.

[0054] When H drops below 60 (hazard level), an orange alert is generated, and a push notification is sent to the operation and maintenance manager via APP and SMS, requiring a response within 24 hours.

[0055] When H drops below 40 (fault level), a red alarm is generated, and a push notification is sent via APP, SMS, and automated voice call to the operations manager and water bureau supervisors, requiring immediate action.

[0056] Once an alert is triggered, the system automatically generates a maintenance work order. The work order includes: the triggering equipment number (e.g., BL-01-DE), the alarm type ("bearing outer ring failure"), suggested remedial measures ("suggest checking bearing lubrication status, replacing the bearing if necessary"), and the expected completion time (based on remaining lifespan prediction). It is then automatically assigned to the personnel most recently responsible for the equipment. The maintenance personnel fill out the maintenance record on the mobile app (e.g., "replaced SKF 6206 bearing, added 20g of grease, time taken 2.5 hours"). After submission, the work order status updates to "resolved," and the system automatically archives the maintenance record into the equipment's full lifecycle file for future maintenance reference.

[0057] VI. Equipment Health Assessment Report Generation: At the end of each quarter, the system automatically generates a health assessment report (PDF format) for each device. The report includes: Basic equipment information (name, model, commissioning date, cumulative running time).

[0058] Health index curve for this quarter (monthly average, fluctuation range).

[0059] Current health score and status level (e.g., "Blower No. 3, current H=72, sub-health level").

[0060] A list of potential failure risks identified (e.g., “the vibration value of the drive end bearing increases month by month, and the amplitude of the BPFO sideband increases, indicating that the bearing wear continues to develop”).

[0061] Recommended maintenance plan (e.g., "It is recommended to replace the bearing during the next quarter's overhaul, and spare parts should be purchased in advance").

[0062] A quarterly comprehensive score comparison chart of all equipment on the site (showing the health ranking of each device).

[0063] The report is automatically sent to the operator's email address and can be exported and submitted to regulatory authorities (such as the Mentougou District Water Resources Bureau) for reporting on the results of smart operation and maintenance.

[0064] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0065] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for dynamically generating and visualizing the health index of electromechanical equipment based on multi-dimensional sensor data fusion, characterized in that, Includes the following steps: Multi-source heterogeneous sensor data acquisition and preprocessing: For electromechanical equipment in water treatment scenarios, multiple types of sensor terminals are deployed to simultaneously collect multi-dimensional sensor data on the operating status of the equipment. The multi-dimensional sensor data includes vibration data, temperature data, noise data, ultrasonic data, magnetic flux data, and electrical parameter data. The collected multi-source raw sensor data is subjected to standardized preprocessing, including data cleaning, outlier removal, time sequence alignment, unit unification, and missing data completion, to obtain preprocessed multi-dimensional sensor data. Multidimensional sensor data fusion modeling: Based on the equipment failure mechanism and operating condition characteristics, feature extraction and fusion weights are performed on different sensing dimensions in the preprocessed multidimensional sensor data. The fusion weights of each monitoring parameter are dynamically adjusted according to the equipment type, years of operation, load fluctuation pattern and environmental corrosion degree. The feature data of each dimension after weight adjustment are subjected to deep fusion correlation analysis to generate multidimensional fusion analysis results. Dynamic calculation of equipment health index: Construct a health scoring model adapted to water electromechanical equipment, map the multi-dimensional fusion analysis results to an equipment health index in the range of 0 to 100, and divide the equipment health index into four levels of state thresholds: healthy, sub-healthy, hidden danger, and fault. The system adopts a time-series rolling calculation method, combining historical operating baseline data of the equipment with real-time operating deviations to dynamically update the real-time health index and health decay rate of the equipment; Visualized display of equipment health status: Based on dynamically generated equipment health indices, the visualization rendering module enables integrated visualization of equipment operating parameter curves, health index change trends, equipment health status topology distribution maps, fault location markings, and anomaly tracing information. It also supports interactive operations such as viewing the status of a single device, providing an overview of the health status of all equipment on the site, and tracing back historical operation and maintenance data.

2. The method for dynamic generation and visualization of electromechanical equipment health index based on multi-dimensional sensor data fusion according to claim 1, characterized in that, In the multi-source heterogeneous sensing data acquisition and preprocessing step, the multi-type sensor terminals include vibration sensors, temperature sensors, noise sensors, ultrasonic sensors, and magnetic flux sensors. The multi-dimensional sensing data acquired simultaneously also includes equipment operating current, operating voltage, operating power, load conditions, operating time, and environmental parameters such as temperature and humidity and media corrosion conditions in the plant area.

3. The method for dynamic generation and visualization of electromechanical equipment health index based on multi-dimensional sensor data fusion according to claim 1, characterized in that, In the multi-dimensional sensing data fusion modeling step, the feature extraction based on equipment fault mechanism and operating condition characteristics for different sensing dimensions includes: extracting bearing fault characteristic frequency, imbalance characteristics and misalignment characteristics from vibration data; extracting temperature rise rate and heat accumulation characteristics from temperature data; extracting abnormal sound patterns and high-frequency sound wave characteristics from noise and ultrasonic data; and extracting motor magnetic field operating state characteristics and electrical fault characteristics from magnetic flux data.

4. The method for dynamic generation and visualization of electromechanical equipment health index based on multi-dimensional sensor data fusion according to claim 1, characterized in that, In the multidimensional sensor data fusion modeling step, the fusion weights are allocated differently according to the contribution of different monitoring parameters to the health status of the equipment, and the fusion weights are dynamically adjusted as the operating status of the equipment changes, so as to achieve fusion analysis with mutual verification of five parameters.

5. The method for dynamic generation and visualization of electromechanical equipment health index based on multi-dimensional sensor data fusion according to claim 1, characterized in that, In the dynamic calculation step of the equipment health index, the classification method of the four-level state thresholds of health, sub-health, hidden danger, and fault is as follows: a health index ≥ 80 is a healthy level, and the equipment is operating normally; a health index of 60 to 80 is a sub-health level, and the vibration or temperature increases by 20% to 50% compared with the baseline value; a health index of 40 to 60 is a hidden danger level, and the vibration or temperature increases by 50% to 100% compared with the baseline value; a health index < 40 is a fault level, and the vibration or temperature exceeds the baseline value by more than 100% or exceeds the set upper limit.

6. The method for dynamic generation and visualization of electromechanical equipment health index based on multi-dimensional sensor data fusion according to claim 1, characterized in that, In the dynamic calculation step of the equipment health index, the use of a time-series rolling calculation method to dynamically update the real-time health index and health decay rate of the equipment is as follows: with a predetermined time window as the period, the current health index is continuously calculated based on the real-time sensor data of the current period, and combined with the historical operating baseline data, the rate of change of the health index over time is calculated to characterize the trend of equipment performance degradation, thereby realizing the transformation of equipment health status from qualitative judgment to quantitative digital assessment.

7. The method for dynamic generation and visualization of electromechanical equipment health index based on multi-dimensional sensor data fusion according to claim 1, characterized in that, In the equipment health status visualization display step, the equipment health status topology distribution map uses the process flow diagram or equipment layout diagram as the base map, and overlays health status color indicators at the corresponding positions of each equipment, where green represents the healthy level, yellow represents the sub-healthy level, orange represents the potential danger level, and red represents the fault level, forming an integrated management view of the health status of all equipment in the station.

8. The method for dynamic generation and visualization of electromechanical equipment health index based on multi-dimensional sensor data fusion according to claim 1, characterized in that, In the equipment health status visualization display step, the abnormal source tracing information includes fault type diagnosis results and fault location information. The fault types include bearing faults, imbalance faults, misalignment faults, mechanical loosening faults, motor rotor faults, seal leakage, cavitation, and insufficient bearing lubrication. The fault location information includes the equipment number where the fault occurred and the location of the measuring point.

9. The method for dynamic generation and visualization of electromechanical equipment health index based on multi-dimensional sensor data fusion according to claim 1, characterized in that, The method also includes a step of generating an equipment health assessment report: based on the historical trend data of the health index of each device, an automatic health assessment report is generated for each device, and the report includes the current equipment health score, health trend graph and recommended maintenance plan; The system performs comprehensive scoring and statistics on all equipment in the station according to equipment type and / or operating area, generating a comprehensive health score for all equipment in the station.

10. The method for dynamically generating and visualizing the health index of electromechanical equipment based on multi-dimensional sensor data fusion according to claim 1, characterized in that, The method also includes a warning and work order linkage step: a four-level warning system is set based on the four-level status thresholds. When the equipment health index drops to the corresponding threshold range, a warning signal of the corresponding level is triggered and pushed to the maintenance personnel terminal. When a hidden danger level or fault level warning is triggered, a maintenance work order is automatically triggered. The maintenance work order includes the triggering equipment information, alarm type, suggested handling measures and expected completion time. After the work order is completed, the maintenance record is automatically archived to the equipment history.

Images