CN-121976956-A - Chemical engineering leakage-free pump fault prediction management system

CN121976956ACN 121976956 ACN121976956 ACN 121976956ACN-121976956-A

Abstract

The invention discloses a chemical engineering leakage-free pump fault prediction management system, which relates to the field of fault detection, and comprises an auxiliary sensor pump body data acquisition module, a pump body operation condition maintenance quantification module, a pump body real-time condition-maintenance condition coupling evaluation module, a pump body state range migration prediction alarm module, a multi-maintenance pump body state grade mapping marking module and a data convergence center, wherein the aims to acquire real-time operation data of a pump body in advance, judge multiple types of working conditions, acquire the maintenance data of the pump body at the same time, perform real-time condition-maintenance condition to evaluate the state of the pump body, construct an intelligent fusion model capable of synthesizing multiple-source heterogeneous data and outputting a unified quantitative health state index, identify a pump body state grade track, and perform accurate, early and intelligent state monitoring and fault early warning on the leakage-free pump.

Inventors

SU FENGBO
WANG GUANCHAO
HOU WEIDA
MA CHAO
LIU BIN
SUN RONGLIANG

Assignees

山东博承泵业制造有限公司

Dates

Publication Date: 20260505
Application Date: 20260120

Claims (9)

1. The system is characterized by comprising an auxiliary sensor pump body data acquisition module, a pump body operation condition maintenance quantization module, a pump body real-time condition-maintenance state coupling evaluation module, a pump body state range migration prediction alarm module, a multi-maintenance pump body state grade mapping marking module and a data convergence center; The auxiliary sensor pump body data acquisition module is used for installing a plurality of monitoring sensors at key parts of the pump body to acquire real-time data, extracting vibration frequency band characteristics and temperature characteristics of the pump body in the acquired data, and carrying out preprocessing error calibration on the extracted data; the pump body operation condition maintenance quantization module segments the real-time working condition of the pump body according to a pre-constructed working condition mode database, establishes independent data packets for each working condition segment, acquires historical execution maintenance state data of the pump body, and performs pump body quality data quantization summarization based on the working condition and the maintenance state data; The pump body real-time working condition-maintenance state coupling evaluation module respectively carries out multi-maintenance evaluation on pump body working condition data and maintenance history data, carries out normalization processing on the evaluation data, constructs a pump body health index frame and divides a corresponding grade interval of the health index of the pump body; The pump body state range migration prediction alarm module continuously samples and segments the real-time data of the pump body, analyzes the pump body state grades in different time periods, acquires the pump body state grades in any two adjacent sampling time periods, carries out state grade track pattern recognition, and carries out early warning on the rapidly deteriorated pump body health index; The multi-maintenance pump body state grade mapping marking module gives a basic state grade based on the health index of the pump body, and meanwhile, identifies the fluctuation track of the range of the pump body state interval in a plurality of sampling points to dynamically correct the basic state grade, so that the comprehensive state grade is generated for marking.
2. The system is characterized in that the auxiliary sensor pump body data acquisition module comprises a sensor layout trigger signal synchronous acquisition sub-module and a dynamic data error calibration sub-module, the sensor layout trigger signal synchronous acquisition sub-module comprises a plurality of vibration sensors and distributed infrared temperature sensors, the vibration sensors and the distributed infrared temperature sensors are arranged at key parts of the non-leakage pump, the key parts are set and arranged by a technician, the vibration sensors acquire high-frequency vibration signals of the pump body, the distributed infrared temperature sensors acquire temperature values of different positions of the pump body, the sensors for monitoring acquisition are connected with synchronous trigger signals, uniform time stamps are made on the acquired data, the dynamic data error calibration sub-module preprocesses the sensor acquisition data, whether the acquired data of the vibration sensors are continuous pump body rotation periods or not is searched, the acquired signals under the discontinuous pump body rotation periods are simultaneously acquired in real time, the current environment temperature is taken as a reference, the temperature values of the different positions of the acquired pump body are compared with the environment temperature values of the environment temperature values, the temperature values of the different positions of the acquired pump body are lower than the environment temperature values of the environment temperature values, and the residual data of the pump body are summarized, and the residual data of the collected data are eliminated.
3. The system is characterized in that the pump body operation condition maintenance quantification module comprises a pump body operation condition segmentation automatic identification parameter recording sub-module and a multi-condition health benchmark maintenance database construction sub-module, the pump body operation condition segmentation automatic identification parameter recording sub-module is used for collecting operation data in a set time threshold interval after pump body installation and debugging are completed and normal operation is confirmed, the operation data are marked as current pump body initial health data benchmark, the pre-constructed operation mode database uploaded by an administrator in a data convergence center is used for carrying out standard operation condition type division on the initial health data of the current pump body according to different operation condition databases, the standard operation condition type division comprises a pump body idle operation condition, a rated operation condition, a load operation condition and an abnormal operation condition, initial health data corresponding to different operation condition types of the current pump body are obtained, and the initial health data corresponding to the different operation condition types of the pump body are marked as independent benchmark operation data packets under the different operation condition types of the pump body.
4. The system for predicting and managing the failure of the chemical non-leakage pump according to claim 3, wherein the multi-working-condition health reference maintenance database construction sub-module obtains maintenance event data of a current pump body, encodes each maintenance event data once, collects key data of an execution process of each maintenance event, including maintenance date, maintenance trigger reasons, maintenance working hours, executive personnel information, maintenance spare part information and post-maintenance pump body test data, and gathers each maintenance event data corresponding to different working condition types of the pump body into a health maintenance database under different working condition types of the pump body.
5. The system for predicting and managing the failure of the chemical leakage-free pump according to claim 1, wherein the pump body real-time working condition-maintenance state coupling evaluation module comprises a pump body multi-dimensional data evaluation sub-module and a multi-level pump body state range dividing sub-module, wherein the pump body multi-dimensional data evaluation sub-module sets a fixed sampling period to collect real-time operation data of the pump body, and according to the real-time operation data, the real-time operation data in the current sampling period are respectively set to be respectively according to different working condition types identified by matching analysis of the real-time operation data Screening out the most-accordant working condition type according to the working condition data of different types corresponding to the operation data, acquiring an independent reference operation data packet under the most-accordant working condition type, extracting the reference operation data under the most-accordant working condition type, and setting the reference operation data as Three real-time operation data of the sampling period are randomly extracted respectively as follows Wherein Three real-time operation data are extracted and respectively endowed with fixed weights according to the types of the correspondingly acquired sensors Calculating the working condition type of pump body real-time operation data matching in the current sampling period according to a formula: calculated, in the above formula Three real-time operation data in the current sampling period are arbitrarily extracted at present And when the different matching values of the three real-time operation data and the most-coincident working condition type which are screened at present are smaller than the set threshold, randomly extracting a plurality of real-time operation transportation in the real-time operation data of the pump body in the current sampling period, respectively carrying out one-by-one matching with reference values of different working condition types until the real-time operation data extracted by the pump body in the current sampling period are all greater than or equal to the set threshold, extracting the working condition type which is matched at present, marking the working condition type as the pump body in the current sampling period, extracting the real-time operation data in the current sampling period, respectively carrying out matching with the independent reference operation data packet of the working condition type which is marked at present, respectively calculating the mean value of the real-time operation data in the current sampling period, extracting the independent reference operation data packet of the corresponding working condition type, respectively carrying out mean value deviation analysis on the real-time operation data in the current sampling period and the independent reference operation data packet of the corresponding working condition type, and simultaneously carrying out mean value deviation analysis on the independent reference operation data packet of the working condition type in the current sampling period.
6. The system for predicting and managing the failure of the leakage-free pump in the chemical industry as set forth in claim 5, wherein the pump body multidimensional data evaluation submodule obtains a health maintenance database of the current pump body under different working condition types, obtains execution process key data of each maintenance event, an administrator matches corresponding quality weight coefficients for each maintenance execution process data through a data convergence center, and simultaneously, the administrator uploads corresponding scoring standards for each maintenance execution process data in advance, calculates a single maintenance quality comprehensive score of each maintenance event, and sets the single maintenance quality comprehensive score of each maintenance event as Any execution data of the current maintenance execution process data is extracted, the execution data is compared with a scoring standard which is uploaded in advance and corresponds to the execution data, and the scoring of the execution data in the current maintenance execution process data is set as The corresponding quality weight coefficient of the execution data in the current maintenance execution process data is that Calculating the single maintenance quality comprehensive score of the current maintenance event according to the formula: and acquiring the single maintenance quality comprehensive score of each maintenance event of the current pump body, and simultaneously calculating the average value of the single maintenance quality comprehensive scores of each maintenance event of the current pump body, and marking the average value as the comprehensive maintenance quality comprehensive score of the current pump body.
7. A failure prediction management system for a chemical non-leakage pump, as set forth in claim 5, wherein the multi-level pump state range dividing sub-module obtains the mean deviation of real-time operation data of the pump body in different sampling periods and independent reference operation data packets of corresponding working condition types and the comprehensive maintenance quality comprehensive score of the pump body respectively, extracts the pump body with the mean deviation of the independent reference operation data packets of any sampling period and corresponding working condition types greater than a set deviation threshold or lower than a set quality score threshold, performs failure marking on the pump body, performs failure prediction alarm on the extracted pump body, presets a multi-level deviation threshold interval and a multi-level comprehensive maintenance quality score threshold through a data convergence center respectively, wherein the multi-level threshold comprises an excellent level interval threshold, a good level interval value, a qualified level interval threshold and a key attention level interval threshold, compares the mean deviation of the current pump body in different sampling periods with the set level interval threshold respectively, and performs failure prediction alarm on the extracted pump body with a higher level corresponding to the current pump body in different sampling periods, and performs failure prediction alarm on the pump body state in different sampling periods and lower than the set level deviation of the same level in the same level.
8. The system is characterized in that the pump body state range migration prediction alarm module comprises a pump body state point track pattern recognition sub-module and a pump body state rapid degradation prediction alarm sub-module, the pump body state point track pattern recognition sub-module obtains continuous sampling periods of the pump body, the continuous sampling periods of the pump body are divided into a plurality of time slices according to fixed time lengths, real-time working condition data and maintenance data of a plurality of sampling periods under each fixed time length are counted, the grade corresponding to the average value of the real-time running data of the sampling periods and the comprehensive maintenance quality score of the pump body under each time slice is analyzed, the pump body state grade under the continuous time slices is marked as the pump body state grade of the time slice, the pump body state grade under any two adjacent time slices is collected and identified by the pump body state rapid degradation prediction alarm sub-module, if the corresponding state grade interval under the two adjacent time slices is larger than one grade, the corresponding state grade direction of the pump body under the subsequent time slices is the degradation direction, the current pump body rapid degradation condition is judged, early warning is carried out on the current pump body, and the collected and the pump body and the collected time slices and the manual pump body and the collected time slices and the collected to the working condition data and the real-time working condition data and the maintenance data of the subsequent time slices are sent to the pump body.
9. The system is characterized in that the multi-dimensional pump body state level mapping marking module comprises a basic state level giving correction submodule and a pump body comprehensive state aggregation uploading submodule, wherein the basic state level giving correction submodule obtains corresponding time segment pump body state levels under different time segments, extracts pump body state levels corresponding to a plurality of sampling periods of a current pump body under the last time segment divided, marks the pump body state levels as the current pump body basic state levels, simultaneously searches a pump body state level track under continuous time segments, judges whether the condition of rapid degradation of the continuous time segments exists or not, if the condition exists, extracts and detects the frequency of rapid degradation, when the frequency is larger than a set threshold value, marks the corresponding basic state levels, and the pump body comprehensive state aggregation uploading submodule aggregates the pump body state levels corresponding to the detected pump bodies and uploads the pump body state levels to a data aggregation center for manual examination.

Description

Chemical engineering leakage-free pump fault prediction management system Technical Field The invention relates to the field of fault detection, in particular to a chemical leak-free pump fault prediction management system. Background The leakage-free pump, in particular the magnetic pump and the shielding pump, is a core conveying device in the fields of high safety and high environmental protection requirements such as chemical industry, medicine, nuclear energy, electronics and the like because of the unique advantages of complete sealing and leakage-free, and currently, vibration acceleration, effective speed value or displacement amplitude is monitored by installing a vibration sensor on a pump body bearing seat, and an alarm is triggered when a measured value exceeds a preset fixed threshold value. Most of the existing systems rely on a single vibration signal for judgment, the evaluation dimension is single, the fault missing report and false alarm rate are high, the state of equipment is not only dependent on the current operation data, but also is further influenced by the maintenance history and the current operation working condition. The prior art systematically ignores these two key background information. The quality of maintenance activities is not recorded and quantified, the current data cannot be dynamically matched with the health standard under the corresponding working condition, so that the diagnosis result is separated from reality, accurate and deep insight cannot be provided for maintenance decision, and the traditional faults are reported one by one and are subjected to cutting treatment, so that operators cannot easily grasp the overall health level and risk trend of the equipment rapidly from complicated alarm information; The application aims to obtain real-time operation data of a pump body in advance, judge multiple types of working conditions, collect maintenance data of the pump body at the same time, evaluate the state of the pump body by carrying out real-time working condition-maintenance state, construct an intelligent fusion model which can synthesize multiple-source heterogeneous data and output unified quantitative health state indexes, identify the state grade track of the pump body, and accurately, early and intelligently monitor the state and early warn faults of the leak-free pump. Disclosure of Invention The invention aims to provide a chemical leak-free pump fault prediction management system, which aims to solve the problems in the prior art. In order to achieve the above purpose, the present invention provides the following technical solutions: The system comprises an auxiliary sensor pump body data acquisition module, a pump body operation condition maintenance quantization module, a pump body real-time condition-maintenance state coupling evaluation module, a pump body state range migration prediction alarm module, a multi-dimensional pump body state grade mapping marking module and a data convergence center; The auxiliary sensor pump body data acquisition module is used for installing a plurality of monitoring sensors at key parts of the pump body to acquire real-time data, extracting vibration frequency band characteristics and temperature characteristics of the pump body in the acquired data, and carrying out preprocessing error calibration on the extracted data; the pump body operation condition maintenance quantization module segments the real-time working condition of the pump body according to a pre-constructed working condition mode database, establishes independent data packets for each working condition segment, acquires historical execution maintenance state data of the pump body, and performs pump body quality data quantization summarization based on the working condition and the maintenance state data; The pump body real-time working condition-maintenance state coupling evaluation module respectively carries out multi-maintenance evaluation on pump body working condition data and maintenance history data, carries out normalization processing on the evaluation data, constructs a pump body health index frame and divides a corresponding grade interval of the health index of the pump body; The pump body state range migration prediction alarm module continuously samples and segments the real-time data of the pump body, analyzes the pump body state grades in different time periods, acquires the pump body state grades in any two adjacent sampling time periods, carries out state grade track pattern recognition, and carries out early warning on the rapidly deteriorated pump body health index; The multi-maintenance pump body state grade mapping marking module gives a basic state grade based on the health index of the pump body, and meanwhile, identifies the fluctuation track of the range of the pump body state interval in a plurality of sampling points to dynamically correct the basic state grade, so that the comprehensive state grade is generated for mar