Energy-saving air compressor running state monitoring system and method

Abstract

The invention relates to the technical field of state monitoring, in particular to an energy-saving air compressor running state monitoring system and method, and the system comprises a pressure slope analysis module, a deviation trend judgment module, an error fitting acquisition module, an error trend analysis module and an alarm triggering module. According to the method, a pressure range is calculated by dividing time slices, a slope ratio is generated, a dynamic threshold algorithm is combined to successively compare and capture an offset trend, frequency and current waveform data are synchronously acquired to match a timestamp to calculate a difference value, an error vector is generated in a standardized manner, and a periodic mean value is extracted to verify an incremental relationship to identify progressive anomaly. The cross validation slope ratio and error vector positioning wear or coordination abnormity, the dynamic threshold value is adjusted along with the mean value coefficient at the initial period to reduce fluctuation interference, the timestamp matching enhances the composite fault detection capability, and the range division operation quantifies the matching degree to provide an energy efficiency optimization basis. And a standardized error vector and mean value incremental mechanism improves the prediction precision and reduces false alarm and missing alarm.

Description

Technical Field

[0001] The present invention relates to the technical field of state monitoring, and in particular to a system and method for monitoring the operating state of an energy-saving air compressor. Background Art

[0002] The field of condition monitoring technology includes technical means for real-time or periodic monitoring of data such as the physical state, performance indicators, or abnormal performance of equipment during operation. The core content of this technical field includes data acquisition, processing, analysis, and diagnosis, which are used to identify possible signs of equipment failure, changing trends in operating performance, and changes in energy consumption levels. Condition monitoring systems are usually composed of a combination of sensor technology, control logic, and data communication methods. By observing multiple parameters such as temperature, pressure, vibration, current, and voltage, a comprehensive assessment of the equipment's operating status is achieved. The entire technical system emphasizes the continuous tracking and analysis of equipment status data, serving fault prevention, operation optimization, and maintenance management. It is widely used in industrial manufacturing, energy management, transportation, and other fields.

[0003] Among them, the energy-saving air compressor operation status monitoring system refers to a system device used to collect and monitor the key working status of the air compressor in the energy-saving operation mode. The technical matters covered by this patent subject mainly include operation data collection, data transmission and processing, energy consumption parameter determination and status alarm signal generation. It mainly installs a sensor device for detecting compressed air flow, power input and body temperature rise, transmits real-time data to the processing unit via the communication unit, and judges whether the current operating status deviates from the energy-saving target range based on the energy efficiency calculation model under preset conditions, and outputs a status diagnostic signal or alarm information. The system can also form a historical data curve based on the operating time and load frequency for trend analysis, thereby completing the full process monitoring of the operating status of the energy-saving air compressor.

[0004] Existing technologies use fixed thresholds to identify pressure anomalies, failing to dynamically adjust the threshold range based on initial equipment operation data. This can easily trigger false alarms or overlook early deviations in scenarios with frequent load changes. Air supply frequency and current parameter analysis are performed independently, lacking timestamp matching and amplitude difference standardization, making it difficult to detect faults associated with motor response delays or load mismatches. Trend prediction relies on comparing historical curves for a single parameter, lacking a multi-cycle mean increment verification mechanism, resulting in a lag in identifying progressive energy consumption anomalies. Pressure data processing relies on absolute value comparisons, failing to incorporate correlations between range and operating time, making it impossible to distinguish between normal fluctuations and abnormal deviations. Energy efficiency model input parameters primarily rely on static data and fail to integrate dynamic indicators such as pressure slope ratios, leading to biased attribution of energy consumption anomalies. Alarm triggering is based on single-point over-limit determination, lacking sequential verification of continuous time-slice data, making occasional interference prone to false alarms. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an energy-saving air compressor operating status monitoring system and method.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: an energy-saving air compressor operation status monitoring system includes: The pressure slope analysis module collects the compressed air outlet pressure sequence and the load end demand pressure sequence through the pressure sensor, divides the time slice to calculate the range between the two, divides the range by the time slice duration to generate the pressure slope ratio, and transmits it to the deviation trend determination module; The deviation trend determination module, based on the pressure slope ratio, calls a dynamic threshold generation algorithm to calculate the mean of the initial data of the current cycle and multiplies it by a coefficient to generate an adaptive threshold, performs a sequential comparison of the pressure slope ratio values ​​of three consecutive time slices, and generates a pressure deviation trend signal if all exceed the threshold, attaches a start time tag, and transmits it to the alarm trigger module; The error fitting acquisition module synchronously collects the gas supply frequency sequence and current waveform characteristic data, matches the timestamps of the frequency change points and the current peak and valley points, calculates the time difference and amplitude difference, and standardizes them to generate the fitting error vector, which is then passed to the error trend analysis module; The error trend analysis module extracts the fitting error vector of three consecutive cycles, calculates the mean value of each cycle and verifies the increasing relationship. If the increase exceeds the mean value of the previous two cycles, a fitting error trend signal is generated, the cycle number is attached and transmitted to the alarm trigger module.

[0007] As a further solution of the present invention, the pressure slope ratio is specifically the time slice duration ratio of the compressed air outlet extreme difference difference to the load end demand extreme difference difference, the pressure offset trend signal includes an adaptive threshold parameter and a trigger time label, the fitting error vector includes a time difference sequence and an amplitude difference sequence of frequency and current characteristic points, and the fitting error trend signal includes a cycle mean increment and a trigger cycle number.

[0008] As a further solution of the present invention, the pressure slope analysis module includes: The pressure sequence acquisition submodule collects the pressure value sequence output by the compressed air outlet pressure sensor, simultaneously obtains the load end demand pressure value sequence, and performs time axis alignment processing on the two sets of sequences to generate a dual pressure sequence; The time slice range calculation submodule divides the continuous time slices into fixed time slices based on the dual pressure sequence, extracts the maximum and minimum values ​​of the outlet pressure in each time slice, calculates the outlet pressure range, and simultaneously extracts the maximum and minimum values ​​of the load-side demand pressure, calculates the demand pressure range, and subtracts the absolute values ​​of the two ranges to generate the time slice range value; The slope ratio generation submodule calls the time slice extreme difference value, obtains the duration of the corresponding time slice, performs a division operation using the time slice extreme difference value as the numerator and the time slice duration as the denominator, rounds the calculation result to decimal places, and generates a pressure slope ratio.

[0009] As a further solution of the present invention, the deviation trend determination module includes: The threshold calculation submodule calls the pressure slope ratio, locates the values ​​corresponding to the first three consecutive time slices in the current cycle, adds the three values ​​and divides them by three to generate an arithmetic mean, calls a preset threshold coefficient, performs a multiplication operation on the arithmetic mean and the threshold coefficient to generate an adaptive threshold; The continuous comparison submodule obtains the pressure slope ratio of the subsequent three consecutive time slices based on the adaptive threshold, compares the value of each time slice with the threshold in turn, and marks it as true if the value is greater than the threshold, otherwise it is marked as false, generating a continuous over-threshold flag containing three Boolean values; The trend determination submodule calls the continuous exceeding threshold value flag to check whether the three Boolean values ​​are all true. If the conditions are met, the starting timestamp of the time slice corresponding to the first true value is extracted to generate a pressure deviation trend signal.

[0010] As a further solution of the present invention, the error fitting acquisition module includes: The synchronous acquisition submodule detects the sampling start time difference between the gas supply frequency sequence and the current waveform characteristic data, interpolates or truncates the timestamps of the two types of data, aligns the time axis using a linear interpolation algorithm, and merges them to generate a synchronous time series data set with a unified time stamp; The feature matching submodule extracts the moment when the difference between adjacent sampling points in the gas supply frequency sequence exceeds the set change threshold based on the synchronous time series data set as the frequency change point, extracts the local maximum and minimum values ​​in the current waveform as the peak and valley points, traverses the timestamps of the frequency change points and peak and valley points one by one, calculates the absolute time difference, and determines it as a valid feature point pair if the time difference is less than the matching fault tolerance window defined by the system, and generates a matching feature pair set; The error normalization submodule calls the time difference and amplitude difference in the matching feature pair set, arranges the time difference in ascending order and takes the median as the benchmark reference value, calculates the standard deviation of the amplitude difference dispersion, and uses the maximum-minimum normalization formula to map the two types of differences to the 0-1 interval, and merges them to generate a fitting error vector.

[0011] As a further solution of the present invention, the error trend analysis module includes: The cycle extraction submodule calls the fitting error vectors of three consecutive cycles, extracts vector data with continuous timestamps in each cycle in the order of cycle numbers, removes data points that overlap with adjacent cycles or have too large a time span, and generates a cycle error vector group; The mean calculation submodule traverses the vector set of each period in the period error vector group, takes the modulus length of all vectors in each period, accumulates and sums them, divides them by the total number of vectors in the period to obtain the single-period mean, arranges the mean results in the order of period numbers, and generates a period mean sequence; The trend verification submodule extracts the difference between the first period mean and the second period mean based on the period mean sequence as a benchmark increase, calculates the difference between the third period mean and the second period mean as the current increase, compares the ratio of the current increase to the benchmark increase, and updates a status flag and generates a trend verification flag if the current increase exceeds a set multiple threshold of the benchmark increase; The signal generation submodule is based on the state of the trend verification flag. If the flag is true, it extracts the numbers of the current three cycles, combines the total increase value of the third cycle mean and the first cycle mean, and generates a fitting error trend signal.

[0012] As a further embodiment of the present invention, the system further comprises: An alarm trigger module receives the pressure deviation trend signal and the fitting error trend signal, verifies the periodic overlap between the two, and if they overlap, calls a joint signal generation function to combine them into an energy-saving deviation alarm signal, and outputs an alarm data packet containing time, number of cycles, and intensity indicators; The energy-saving deviation alarm signal specifically includes a time overlap segment, a cycle superposition number, and a trend strength index.

[0013] As a further solution of the present invention, the alarm triggering module includes: The cycle verification submodule calls the cycle start time and end time of the pressure offset trend signal, and combines the cycle start time and end time of the fitting error trend signal to calculate the absolute duration of the intersection interval of the two signal time ranges, and performs a ratio operation on the intersection duration and the total duration of the shorter cycle of the two cycles. If the ratio exceeds the cycle overlap determination threshold set by the system, a cycle overlap flag is generated; The signal synthesis submodule is based on the periodic coincidence flag. If the flag is true, it calls the intensity parameter of the pressure offset trend signal and the amplification parameter of the fitting error trend signal, performs weighted summation according to the weight of the intensity parameter and the weight of the amplification parameter, and generates an energy-saving deviation alarm signal; The alarm generation submodule calls the current system timestamp, the associated cycle number and the strength value of the energy-saving deviation alarm signal, encapsulates data according to the triple structure of timestamp-cycle number-signal strength, and generates an energy-saving deviation alarm data packet.

[0014] A method for monitoring the operating status of an energy-saving air compressor is provided. The method is based on the above-mentioned energy-saving air compressor operating status monitoring system and comprises the following steps: S1: Obtain the compressed air outlet pressure sequence and the load end demand pressure sequence through the pressure sensor, divide the time slices and calculate the range value of the two sequences, and divide the range value by the corresponding time slice duration to generate the pressure slope ratio; S2: Based on the pressure slope ratio, a dynamic threshold generation algorithm is used to calculate the data mean of the initial stage of the current cycle, the mean is multiplied by a preset coefficient to generate an adaptive threshold, and the pressure slope ratio of three consecutive time slices is compared with the adaptive threshold to generate a pressure deviation trend signal and attach a time tag; S3: Synchronously collect the gas supply frequency sequence and current waveform characteristic data, match the timestamp difference between the frequency change point and the current peak and valley point, calculate the time difference and amplitude difference, perform normalization, and generate a fitting error vector; S4: extracting the fitting error vectors of three consecutive cycles, calculating the mean data in each cycle, verifying the increasing relationship between the means of adjacent cycles, and generating a fitting error trend signal when the increase exceeds the mean of the previous two cycles and attaching a cycle number; S5: Calling a joint signal generation function to verify the period overlap of the pressure offset trend signal and the fitting error trend signal, merging the signals to generate an energy-saving deviation alarm signal when they overlap, and outputting an alarm data packet containing a time tag, a period number, and an intensity index.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, the range between the compressed air outlet pressure and the load-side demand pressure is calculated by dividing the time slice, and the pressure slope ratio is generated by division operation. The continuous time slice data is compared successively in combination with the dynamic threshold generation algorithm to capture the pressure offset trend. The air supply frequency and current waveform characteristic data are collected synchronously, the time difference and amplitude difference are calculated by matching the timestamps, the fitting error vector is generated by standardization, the continuous cycle mean is extracted to verify the incremental relationship, and progressive anomalies are identified. The cross-validation mechanism of the pressure slope ratio and the fitting error vector locates mechanical wear or motor coordination anomalies. The dynamic threshold is adjusted based on the mean coefficient of the initial data of the cycle to reduce the interference of working condition fluctuations. Timestamp matching eliminates the defects of isolated analysis of single parameters and enhances the ability to detect complex faults. The range division operation quantifies the dynamic matching degree of pressure and load, providing data support for energy efficiency optimization. The standardized error vector and the mean increasing verification mechanism improve the trend prediction accuracy and reduce false alarms and omissions. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a system flow chart of the present invention; Figure 2 This is a flow chart of the pressure slope analysis module of the present invention; Figure 3 This is a flow chart of the deviation trend determination module of the present invention; Figure 4 This is a flow chart of the error fitting acquisition module of the present invention; Figure 5 This is a flow chart of the error trend analysis module of the present invention; Figure 6 This is a flow chart of the alarm trigger module of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0018] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0019] Example 1 See also Figure 1 The present invention provides a technical solution: an energy-saving air compressor operation status monitoring system includes: The pressure slope analysis module collects the compressed air outlet pressure sequence and the load end demand pressure sequence through the pressure sensor, divides the time slice to calculate the range between the two, divides the range by the time slice duration to generate the pressure slope ratio, and transmits it to the deviation trend determination module; The deviation trend determination module, based on the pressure slope ratio, calls the dynamic threshold generation algorithm to calculate the mean of the initial data of the current cycle and multiplies it by a coefficient to generate an adaptive threshold. It then compares the pressure slope ratios of three consecutive time slices one by one. If all of them exceed the threshold, a pressure deviation trend signal is generated, attached with a start time tag, and transmitted to the alarm trigger module. The error fitting acquisition module synchronously collects the gas supply frequency sequence and current waveform characteristic data, matches the timestamps of the frequency change points and the current peak and valley points, calculates the time difference and amplitude difference, and standardizes them to generate the fitting error vector, which is then passed to the error trend analysis module; The error trend analysis module extracts the fitting error vectors of three consecutive cycles, calculates the mean of each cycle and verifies the increasing relationship. If the increase exceeds the mean of the previous two cycles, a fitting error trend signal is generated, attached with the cycle number and passed to the alarm trigger module; The alarm trigger module receives the pressure deviation trend signal and the fitting error trend signal, verifies the periodic overlap between the two, and if they coincide, calls the joint signal generation function to merge them into an energy-saving deviation alarm signal, and outputs an alarm data packet containing time, number of cycles, and intensity indicators.

[0020] The pressure slope ratio is specifically the time slice duration ratio of the compressed air outlet extreme difference difference to the load end demand extreme difference difference. The pressure offset trend signal includes the adaptive threshold parameter and the trigger time label. The fitting error vector includes the time difference sequence and amplitude difference sequence of the frequency and current characteristic points. The fitting error trend signal includes the cycle mean increment and the trigger cycle number. The energy-saving deviation alarm signal is specifically the time overlap segment, the cycle superposition number, and the trend strength index.

[0021] See also Figure 2 , the pressure slope analysis module includes: The pressure sequence acquisition submodule collects the pressure value sequence output by the compressed air outlet pressure sensor, simultaneously obtains the load end demand pressure value sequence, and performs time axis alignment processing on the two sets of sequences to generate a dual pressure sequence; The pressure sequence acquisition submodule first connects to the pressure sensor installed at the outlet of the air compressor. The sensor model is MPX5700AP, with a range of 0-700kPa and an accuracy of ±1.5kPa. The data acquisition frequency is set to 10Hz, that is, the outlet pressure is recorded every 0.1 seconds. At the same time, the connected PLC (Programmable Logic Controller) is accessed through the industrial bus protocol (such as ModbusTCP), which monitors and records the target pressure required by each gas point (such as cylinder, nozzle) on the production line in real time. The recording frequency is also set to 10 Hz. The two sets of raw data sequences collected are first pre-processed by time stamp alignment to check the timestamp of the compressed air outlet pressure sequence. Timestamp of the load-side demand pressure sequence ,Due to the difference in the startup time of the acquisition device or network delay, there may be a slight time deviation, for example, at a certain moment, the collected The data timestamp is 14:25:30.105, and the corresponding The data timestamp is 14:25:30.115, the time difference is 10ms, and a maximum tolerance time deviation is set. is 20ms, if , then the timestamps are considered to match and the one with the closest timestamp will be and Data points are paired, if the deviation is greater than , then linear interpolation is used to calculate and insert an aligned data point based on the data points of adjacent timestamps. For example, if Corresponding Does not exist, but and Between, then ,This processing ensures that each outlet pressure data point has a demand pressure data point that ,corresponds precisely in time, forming a one-to-one corresponding ,double pressure sequence, as shown in Table 1, which contains the aligned continuous 1-second data (10 ,data points).

[0022] Table 1 Example of dual pressure sequence As shown in Table 1, this table shows part of the data of the dual pressure sequence generated after time axis alignment processing. Each row represents the outlet pressure and the corresponding demand pressure value at a time point. After processing, the dual pressure sequence with completely corresponding timestamps is generated. .

[0023] The time slice range calculation submodule is based on a dual pressure sequence and divides continuous time slices into fixed time slices. The maximum and minimum values ​​of the outlet pressure are extracted in each time slice to calculate the outlet pressure range. The maximum and minimum values ​​of the load-side demand pressure are simultaneously extracted to calculate the demand pressure range. The absolute values ​​of the two ranges are subtracted to generate the time slice range value. The time slice range calculation submodule calls the dual pressure sequence generated in the previous step , set a fixed time slice duration The duration is set according to the actual working conditions and pressure fluctuation characteristics, for example, = 5 seconds. Since the sampling frequency is 10 Hz, each time slice contains data points, the system divides the dual pressure sequence into consecutive time slices in sequence. For example, the first time slice contains timestamps from arrive The second time slice contains data from arrive Data, and so on, within each time slice, for example, examine the first time slice (time range [14:25:30.1, 14:25:35.0]), traverse all the outlet pressure data in the time slice , find the maximum value among them and minimum value , assuming that in this time slice, the maximum outlet pressure recorded is 658.2kPa and the minimum outlet pressure is 647.6kPa, then calculate the outlet pressure range of this time slice kPa, and at the same time, the load end pressure data corresponding to the time slice Perform the same operation to find the maximum value and minimum value , assuming that the maximum recorded demand pressure is 651.0kPa and the minimum demand pressure is 645.2kPa, calculate the demand pressure range kPa, then calculate the difference between the absolute values ​​of the two extremes to get the time slice extreme value of the time slice kPa, and repeat this process for all subsequent time slices. For example, the second time slice (time range [14:25:35.1,14:25:40.0]) is calculated as kPa, kPa, then kPa, arrange the time slice extreme difference values ​​calculated for each time slice in chronological order to generate a time slice extreme difference value sequence ,in is the total number of time slices.

[0024] The slope ratio generation submodule calls the time slice extreme value to obtain the duration of the corresponding time slice, performs a division operation using the time slice extreme value as the numerator and the time slice duration as the denominator, rounds the calculation result to decimal places, and generates the pressure slope ratio.

[0025] The slope ratio generation submodule calls the time slice extreme difference value sequence generated in the previous step , for each time slice extreme value in the sequence , get the actual duration of the corresponding time slice , the duration has been fixed to 5 seconds in the previous step, and the time slice extreme difference value As a numerator, the duration of the time slice As the denominator, perform the division operation to calculate the pressure slope ratio , taking the calculation of the first time slice as an example, the time slice extreme value kPa, duration of the time slice seconds, the corresponding pressure slope ratio , calculate the second time slice, kPa, seconds, then , perform decimal processing on the calculated ratio result, set the number of decimal places to 2, and use the rounding rule. For example, if the calculated result is 0.963kPa / s, it is rounded to 0.96kPa / s. If the result is 0.668kPa / s, it is rounded to 0.67kPa / s. In this example, 0.96 and 0.66 do not need to be processed. Arrange the calculated and processed results of all time slices in chronological order to generate the final pressure slope ratio sequence , for example, the first few items of the sequence are {0.96, 0.66, 0.75, 0.88, 1.05, 1.12, 1.30, 1.35, 1.40, …} (all units are kPa / s).

[0026] See also Figure 3 , the deviation trend determination module includes: The threshold calculation submodule calls the pressure slope ratio, locates the values ​​corresponding to the first three consecutive time slices in the current cycle, adds the three values ​​and divides them by three to generate the arithmetic mean, calls the preset threshold coefficient, multiplies the arithmetic mean by the threshold coefficient, and generates an adaptive threshold; The threshold calculation submodule calls the pressure slope ratio sequence generated in the previous step , locate the current analysis cycle (for example, if the cycle, you need to use the End of cycle and The pressure slope ratio values ​​corresponding to the first three consecutive time slices in the data at the beginning of the cycle, or the data within a sliding window, for example, the sequence is {…, 0.96, 0.66, 0.75, 0.88, 1.05, 1.12, 1.30, 1.35, 1.40, …}, we take the first three values , , , add these three values ​​and divide by three to calculate their arithmetic mean kPa / s, then call a preset threshold coefficient The setting of this coefficient is based on the normal fluctuation range and significant deviation in historical operating data. Statistical analysis of the pressure slope ratio and expert experience, for example, shows that when the short-term increase in the pressure slope ratio exceeds 25% of its recent average value, it often indicates a potential deviation, so the setting , the calculated arithmetic mean and threshold coefficient Performs a multiplication operation to generate an adaptive threshold for the current period or analysis window kPa / s, this adaptive threshold will be adjusted as the initial pressure fluctuation state changes.

[0027] The continuous comparison submodule obtains the pressure slope ratio of the next three consecutive time slices based on the adaptive threshold, and compares the value of each time slice with the threshold in turn. If the value is greater than the threshold, it is marked as true, otherwise it is marked as false, and a continuous over-threshold flag containing three Boolean values ​​is generated; The continuous comparison submodule is based on the adaptive threshold calculated in the previous step kPa / s, and obtain the pressure slope ratio values ​​of the three consecutive time slices immediately following the initial three time slices from the pressure slope ratio sequence, that is, , , , and compare the values ​​of these three time slices with the adaptive threshold Compare the size, the first value kPa / s, because , so the time slice is marked as false (False), the second value kPa / s, because , so the time slice is marked as true (True), the third value kPa / s, because , so the time slice is marked as true (True), and the results of the three comparisons are combined in sequence to generate a continuous over-threshold mark sequence containing three Boolean values If there are more time slices, the window slides forward one time slice, and the three initial values ​​are taken again to calculate the new average and adaptive threshold (or use fixed initial three values), and then the new subsequent three values ​​are compared. For example, if the window slides, the next comparison will be , , with a (possibly updated) adaptive threshold.

[0028] The trend determination submodule calls the continuous threshold value mark to check whether the three Boolean values ​​are all true. If the conditions are met, the starting timestamp of the time slice corresponding to the first true value is extracted to generate a pressure deviation trend signal.

[0029] The trend determination submodule calls the continuous threshold mark sequence generated in the previous step , checks whether all three Boolean values ​​contained in a sequence are true. In our first example, the sequence is ,Since the first value is False, the condition that all Boolean values ​​are true is not met. Therefore, it is determined that no definite pressure excursion trend has been formed at present, and no pressure excursion trend signal is generated. The system continues to monitor. Suppose that in a subsequent analysis window, the continuous exceeding threshold mark sequence obtained becomes , for example, the three pressure slope ratios compared are , , , and assuming that they are all greater than the currently calculated adaptive threshold (for example, if the average value of the initial window is low, or the threshold coefficient is set appropriately, this is possible), at this time, the condition that all three Boolean values ​​are true is met, and the system determines that the pressure deviation trend has been formed. Then, the starting timestamp of the time slice corresponding to the first true value in this sequence is extracted. In the example, the first truth value corresponds to , assuming that the starting timestamp of the 7th time slice is =14:25:50.1, the system extracts this timestamp and combines it with the trend confirmation information (for example, it can include the timestamp of the confirmation moment). =14:26:00.0, i.e. the end of the 9th time slice), generating a pressure deviation trend signal , which indicates the occurrence and starting time of the deviation trend.

[0030] See also Figure 4 , the error fitting acquisition module includes: The synchronous acquisition submodule detects the sampling start time difference between the gas supply frequency sequence and the current waveform characteristic data, interpolates or truncates the timestamps of the two types of data, aligns the time axis using a linear interpolation algorithm, and merges them to generate a synchronous time series data set with a unified time stamp; The synchronous acquisition submodule first connects to the inverter output to obtain the gas supply frequency sequence For example, through the analog output interface or digital communication interface, the acquisition frequency is set to 1Hz, that is, the frequency value is recorded once per second. At the same time, the three-phase current waveform data of the main motor of the air compressor is collected through the Hall effect sensor or current transformer. , set the acquisition frequency to 1kHz, that is, record the instantaneous value of the three-phase current every millisecond. Since the acquisition sources and acquisition frequencies of the two types of data are different, their original timestamps and There may be problems with starting time differences and misalignment. For example, frequency data may be recorded at 0ms per second, while current data may be recorded at any millisecond after the system is started, and the frequency is low. Time axis alignment processing is required to detect the first timestamp of the gas supply frequency sequence. and the first timestamp of the current waveform characteristic data , calculate the starting time difference ,like , then the length of the beginning of the cut-off current data sequence is If the data , then cut off the beginning of the frequency data sequence, or wait longer for alignment. In order to match the sampling rate, the frequency data is interpolated and padded. The linear interpolation algorithm is used to interpolate between two consecutive frequency sampling points. and Between, according to the timestamp of current data (in ), calculate the corresponding interpolation frequency , so that each current data sampling point There is a corresponding (possibly interpolated) frequency value. For current data, we usually focus on its characteristics rather than the original waveform, such as calculating its effective value (RMS) or extracting a specific frequency component. Suppose we calculate the effective value of the current in each 100ms window. , and interpolate the frequency to the corresponding time point , merge these two types of processed data to generate a unified time stamp Synchronous time series dataset , where the time resolution is 100ms.

[0031] Based on the synchronous time series data set, the feature matching submodule extracts the moment when the difference between adjacent sampling points in the gas supply frequency sequence exceeds the set change threshold as the frequency change point, extracts the local maximum and minimum values ​​in the current waveform as the peak and valley points, traverses the timestamps of the frequency change points and peak and valley points one by one, calculates the absolute time difference, and determines that if the time difference is less than the matching fault tolerance window defined by the system, it is determined to be a valid feature point pair, and generates a matching feature pair set; The feature matching submodule is based on the synchronized time series dataset generated in the previous step First, extract the frequency change point in the air supply frequency sequence and set a frequency change threshold The threshold is set according to the typical step size of frequency regulation during normal operation of the air compressor. For example, Hz, traverse the frequency sequence , calculate the frequency difference between adjacent sampling points , if the difference is greater than , then the timestamp Mark as a frequency change point , for example, if Hz, Hz, then , so is a frequency change point, then extract the current effective value sequence This can be achieved by comparing the values ​​of a point with its adjacent points before and after it. For example, if ,but Is a local maximum point (peak point) ,like ,but Is a local minimum point (valley point) , all the peak points and valley points found are collectively called current characteristic points Then, traverse the identified frequency change points one by one and current characteristic points , calculate the absolute difference between their timestamps , set a system-defined matching fault tolerance window ,The window size is set based on the empirical value of the system response delay, e.g. ms (0.15 seconds), if the calculated , then the frequency change point and the current characteristic point are considered to constitute a valid characteristic point pair, and the timestamps of the pair of points are recorded. And the corresponding frequency value and current value, for example, if a frequency change point is found at At this moment, the frequency becomes 45.9Hz, and A current peak point is detected at the moment, the current is 55.2A, then the time difference is , less than ,therefore Constitute a valid feature point pair, and all the found valid feature point pairs and their associated data (time difference , frequency value , current value , and the amplitude difference between the two can be defined as ,in Is the reference current before the frequency changes, or use directly itself as amplitude information) to generate a set of matching feature pairs .

[0032] The error normalization submodule calls the time difference and amplitude difference in the matching feature pair set, sorts the time differences in ascending order and takes the median as the benchmark reference value, calculates the standard deviation of the amplitude difference dispersion, and uses the maximum-minimum normalization formula to map the two types of differences to the 0-1 interval, and merges them to generate a fitting error vector.

[0033] The error normalization submodule calls the matching feature pair set generated in the previous step , the set contains The time difference of valid feature point pairs and amplitude difference (or amplitude) , first process the time difference , sort these time differences in ascending order, and then select the value in the middle position after sorting as the benchmark reference value ,if is an odd number, the median is A value, if is an even number, the median is Hedi For example, suppose there are 5 time difference values ​​{80ms, 100ms, 90ms, 110ms, 95ms}, and after sorting they are {80, 90, 95, 100, 110}ms, the median is ms, and then calculate the amplitude difference data The degree of dispersion, specifically calculate its standard deviation ,in is the average value of the amplitude difference. Next, the Min-Max Normalization formula is used to map the time difference and amplitude difference data to Within the interval, for the time difference , whose normalized value is ,in and are the minimum and maximum values ​​of all time differences, respectively. For example, using the time difference data above, , ,but , , , , , similarly, for the amplitude difference Normalize , the normalized time difference of each feature point pair Sum Amplitude Difference Merge into a two-dimensional vector , these vectors together constitute the set of fitting error vectors within the analysis period , this set as a whole is called the fitting error vector.

[0034] See also Figure 5 , the error trend analysis module includes: The cycle extraction submodule calls the fitting error vectors of three consecutive cycles, extracts the vector data with continuous timestamps in each cycle in the order of cycle numbers, removes data points that overlap with adjacent cycles or have too large time spans, and generates a cycle error vector group; The cycle extraction submodule calls three consecutive analysis cycles (e.g. marked as cycles ) The calculated fitting error vector set ,in Representative A set of fitting error vectors of periods, each vector Each is associated with a timestamp (e.g. the timestamp corresponding to the frequency change point), and is numbered by period. Extract all vector data with continuous timestamps in each period in the order of the timestamps. During the extraction process, it is necessary to check the validity of the data points and remove possible abnormal data points. For example, check whether the timestamp of a vector is too far away from the timestamps of the previous and next vectors, exceeding the preset maximum time interval. , or check whether the timestamp of a vector is too close to or even overlaps with the end time or start time of the adjacent cycle, and set a cycle boundary tolerance time , if the vector timestamp satisfy or , you may need to remove this vector to avoid periodic data confusion. Assuming the period The time range is [15:00:00,15:10:00], Yes [15:10:00,15:20:00], is [15:20:00,15:30:00], and seconds, if The timestamp of a vector is 15:09:58, because it is The end time is too close and may be removed, or if it is found The time stamps of two consecutive vectors differ by more than seconds, the data may be interrupted or abnormal and needs to be checked or eliminated. After screening and cleaning, a set of valid and time-continuous error vectors in each cycle is obtained. , these three sets are combined to generate the periodic error vector group .

[0035] The mean calculation submodule traverses the vector set of each cycle in the periodic error vector group, takes the modulus length of all vectors in each cycle, accumulates the sum, divides it by the total number of vectors in the cycle to obtain the single-cycle mean, arranges the mean results in the order of cycle numbers, and generates a periodic mean sequence; The mean calculation submodule traverses the periodic error vector group generated in the previous step The set of valid vectors for each cycle in (in ), for each cycle , traverse all valid error vectors it contains ,in ( It is a cycle The number of valid vectors in the Calculate its Magnitude, that is, the size of the vector, and the calculation formula is ,because and They are all normalized to the value of [0,1]. The modulus represents the degree of deviation of the feature point pair in the normalized error space. The modulus lengths of all vectors in the sum are accumulated and summed , then the sum Divide by the total number of valid vectors in that cycle , get the single cycle error mean of this cycle , for example, for the period , calculate , for the cycle , calculate , for the cycle , calculate , arrange the mean results of these three periods in the order of period numbers to generate a period mean sequence .

[0036] The trend verification submodule extracts the difference between the first and second period means based on the period mean sequence as the benchmark increase, calculates the difference between the third and second period means as the current increase, compares the ratio of the current increase to the benchmark increase, and updates the status flag and generates a trend verification flag if the current increase exceeds the set multiple threshold of the benchmark increase. The trend verification submodule is based on the periodic mean sequence generated in the previous step , first extract the first period mean and the second period mean , calculate the difference between them as the baseline increase in error growth , this value represents the average increase in error from the first cycle to the second cycle, then calculate the third cycle mean and the second period mean The difference is used as the current error increase , this value represents the average increase in error from the second cycle to the third cycle. Next, compare the current increase Compared with the benchmark increase The proportional relationship between the two factors is introduced, and a threshold value is set. This threshold is used to determine whether the error growth shows an accelerating trend. The setting basis is to conduct statistical analysis on a large amount of historical data. When the error growth rate exceeds a certain multiple, it is considered that the energy-saving state may deviate significantly. For example, according to experience, , indicating that the current increase needs to be more than twice the baseline increase to be considered a significant trend. Perform comparison: Calculate the ratio ,because , so the condition is met, the current increase exceeds the set multiple threshold of the baseline increase, at this time, update the internal status flag, such as setting a Boolean variable , generates a trend verification sign, which indicates that the fitting error shows an accelerating growth trend. If the ratio is less than or equal to ,but , no trend verification flag is generated.

[0037] The signal generation submodule verifies the status of the trend flag. If the flag is true, it extracts the numbers of the current three cycles, combines the total increase value of the third cycle mean and the first cycle mean, and generates a fitting error trend signal.

[0038] The signal generation submodule verifies the trend flag obtained in the previous step , checks if the flag is True. In our example, is set to True, the condition is met, so the module extracts the numbers of the three periods involved in the current analysis, namely , then calculate the total increase from the first cycle to the third cycle , which represents the overall growth of the fitting error over the three observed periods. and the calculated total increase value Merge to generate fitting error trend signal , the signal contains trend confirmation information, the period range involved and the total amplitude of error growth. If the trend verification flag is False, this signal is not generated or a signal indicating no significant trend is generated. .

[0039] See also Figure 6 , the alarm trigger module includes: The cycle verification submodule calls the cycle start time and end time of the pressure offset trend signal, and combines the cycle start time and end time of the fitting error trend signal to calculate the absolute duration of the intersection of the two signal time ranges. The intersection duration is then ratioed to the total duration of the shorter of the two cycles. If the ratio exceeds the cycle overlap determination threshold set by the system, a cycle overlap flag is generated. The periodic verification submodule calls the previously generated pressure deviation trend signal and fitting error trend signal , first from The time range in which the pressure deviation trend is detected is obtained from 14:25:50.1, the end time can be considered as the confirmation time 14:26:00.0, or it can be defined as the complete time period that includes those three consecutive exceeding threshold time slices, for example, from the 7th time slice to the 9th time slice, that is, [14:25:50.1, 14:26:00.0], its duration seconds, then from The period range involved in obtaining the fitting error trend is assumed to be The corresponding time ranges are [14:25:00,14:26:00], [14:26:00,14:27:00], [14:27:00,14:28:00], and the total time range involved in the error trend is , its duration Minutes = 180 seconds. Calculate the intersection interval of the two signal time ranges. The pressure trend time range is [14:25:50.1, 14:26:00.0], and the error trend time range is [14:25:00, 14:28:00]. Their intersection is [14:25:50.1, 14:26:00.0]. Calculate the absolute duration of the intersection interval. Seconds, next, find the total duration of the shorter of the two signal cycles (or valid time periods), where the duration of the pressure trend is Seconds, the total duration of the error trend Seconds, the shorter one is seconds, the intersection duration With shorter cycle length Perform ratio operation to calculate the degree of overlap , call the cycle coincidence judgment threshold set by the system , the threshold is set based on experience and is used to determine whether two independently detected trend signals are sufficiently correlated in time, for example, (i.e. 70%), because the calculated overlap Greater than , so it is determined that the cycles (or time periods) of the two signals overlap significantly, and a cycle overlap flag is generated .

[0040] The signal synthesis submodule is based on the periodic coincidence flag. If the flag is true, it calls the intensity parameter of the pressure offset trend signal and the amplification parameter of the fitting error trend signal, and performs weighted summation according to the weight of the intensity parameter and the weight of the amplification parameter to generate an energy-saving deviation alarm signal; The signal synthesis submodule is based on the periodic coincidence flag generated in the previous step Make a judgment, in our example , indicating that the two trend signals coincide in time, satisfying the conditions for signal synthesis. The module then calls the pressure offset trend signal The intensity parameter in and fitting error trend signal The amplification parameter in , it should be noted that in the original description The "intensity parameter" is not explicitly included. Here we assume that the parameter is calculated based on the degree or duration of the pressure slope ratio exceeding the threshold. For example, we can take the three consecutive thresholds Relative to the adaptive threshold The average excess , assuming ,but , and the error increase parameter is Next, call the preset intensity parameter weight and the weight of the increase parameter The two weights reflect the importance of the instability directly manifested by pressure and the efficiency reduction trend indirectly reflected by the frequency-current relationship when evaluating the overall energy-saving deviation risk. The weight setting is based on the analysis of historical fault data and energy-saving effects. For example, it is believed that direct pressure fluctuations can more immediately reflect the problem, so the setting , while the error accumulation trend reflects the long-term changes, setting ,make sure , perform weighted sum operation to calculate the intensity value of energy-saving deviation alarm signal , generating an energy-saving deviation alarm signal, whose intensity is .

[0041] The alarm generation submodule calls the current system timestamp, associated cycle number and the strength value of the energy-saving deviation alarm signal, encapsulates the data according to the triple structure of timestamp-cycle number-signal strength, and generates an energy-saving deviation alarm data packet.

[0042] The alarm generation submodule calls the current system’s precise timestamp, e.g. 2025-04-1415:30:05.123, and at the same time call the associated cycle number, that is, the cycle involved in the fitting error trend analysis , and call the intensity value of the energy-saving deviation alarm signal generated by the previous step signal synthesis submodule , according to the triple structure of timestamp-cycle number-signal strength Encapsulate these data and 2025-04-1415:30:05.123, ,as well as They are combined together to form a specific data record, for example, a JSON object {"timestamp":"2025-04-14T15:30:05.123Z","cycles":["C1","C2","C3"],"alarm_strength":0.3095}. This structured data record is the final energy-saving deviation alarm data packet, which can then be sent to the monitoring system, stored in the database, or trigger the corresponding alarm action.

[0043] A method for monitoring the operating status of an energy-saving air compressor is provided. The method is based on the above-mentioned energy-saving air compressor operating status monitoring system and includes the following steps: S1: Obtain the compressed air outlet pressure sequence and the load end demand pressure sequence through the pressure sensor, divide the time slices and calculate the range value of the two sequences, and divide the range value by the corresponding time slice duration to generate the pressure slope ratio; S2: Based on the pressure slope ratio, a dynamic threshold generation algorithm is used to calculate the data mean of the initial stage of the current cycle. This mean is multiplied by a preset coefficient to generate an adaptive threshold. The pressure slope ratio of three consecutive time slices is compared with the adaptive threshold to generate a pressure excursion trend signal and attach a time tag. S3: Synchronously collect the gas supply frequency sequence and current waveform characteristic data, match the timestamp difference between the frequency change point and the current peak and valley point, calculate the time difference and amplitude difference, perform normalization, and generate a fitting error vector; S4: Extract the fitting error vectors of three consecutive periods, calculate the mean data in each period, verify the increasing relationship between the means of adjacent periods, and generate a fitting error trend signal when the increase exceeds the mean of the previous two periods and add the period number; S5: Call the joint signal generation function to verify the period overlap of the pressure offset trend signal and the fitting error trend signal. When they overlap, the signals are merged to generate an energy-saving deviation alarm signal, and an alarm data packet containing a time tag, a period number, and an intensity index is output.

[0044] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. An energy-saving air compressor operating status monitoring system, characterized in that: The system comprises: The pressure slope analysis module collects the compressed air outlet pressure sequence and the load end demand pressure sequence through the pressure sensor, divides the time slice to calculate the range between the two, divides the range by the time slice duration to generate the pressure slope ratio, and transmits it to the deviation trend determination module; The deviation trend determination module, based on the pressure slope ratio, calls a dynamic threshold generation algorithm to calculate the mean of the initial data of the current cycle and multiplies it by a coefficient to generate an adaptive threshold, performs a sequential comparison of the pressure slope ratio values ​​of three consecutive time slices, and generates a pressure deviation trend signal if all exceed the threshold, attaches a start time tag, and transmits it to the alarm trigger module; The error fitting acquisition module synchronously collects the gas supply frequency sequence and current waveform characteristic data, matches the timestamps of the frequency change points and the current peak and valley points, calculates the time difference and amplitude difference, and standardizes them to generate the fitting error vector, which is then passed to the error trend analysis module; The error trend analysis module extracts the fitting error vector of three consecutive cycles, calculates the mean value of each cycle and verifies the increasing relationship. If the increase exceeds the mean value of the previous two cycles, a fitting error trend signal is generated, the cycle number is attached and transmitted to the alarm trigger module.

2. The energy-saving air compressor operation status monitoring system according to claim 1 is characterized in that: The pressure slope ratio is specifically the time slice duration ratio of the compressed air outlet extreme difference difference to the load end demand extreme difference difference. The pressure offset trend signal includes an adaptive threshold parameter and a trigger time label. The fitting error vector includes a time difference sequence and an amplitude difference sequence of frequency and current characteristic points. The fitting error trend signal includes a cycle mean increment and a trigger cycle number.

3. The energy-saving air compressor operation status monitoring system according to claim 2 is characterized in that: The pressure slope analysis module includes: The pressure sequence acquisition submodule collects the pressure value sequence output by the compressed air outlet pressure sensor, simultaneously obtains the load end demand pressure value sequence, and performs time axis alignment processing on the two sets of sequences to generate a dual pressure sequence; The time slice range calculation submodule divides the continuous time slices into fixed time slices based on the dual pressure sequence, extracts the maximum and minimum values ​​of the outlet pressure in each time slice, calculates the outlet pressure range, and simultaneously extracts the maximum and minimum values ​​of the load-side demand pressure, calculates the demand pressure range, and subtracts the absolute values ​​of the two ranges to generate the time slice range value; The slope ratio generation submodule calls the time slice extreme difference value, obtains the duration of the corresponding time slice, performs a division operation using the time slice extreme difference value as the numerator and the time slice duration as the denominator, rounds the calculation result to decimal places, and generates a pressure slope ratio.

4. The energy-saving air compressor operation status monitoring system according to claim 3 is characterized in that: The deviation trend determination module includes: The threshold calculation submodule calls the pressure slope ratio, locates the values ​​corresponding to the first three consecutive time slices in the current cycle, adds the three values ​​and divides them by three to generate an arithmetic mean, calls a preset threshold coefficient, performs a multiplication operation on the arithmetic mean and the threshold coefficient to generate an adaptive threshold; The continuous comparison submodule obtains the pressure slope ratio of the subsequent three consecutive time slices based on the adaptive threshold, compares the value of each time slice with the threshold in turn, and marks it as true if the value is greater than the threshold, otherwise it is marked as false, generating a continuous over-threshold flag containing three Boolean values; The trend determination submodule calls the continuous exceeding threshold value flag to check whether the three Boolean values ​​are all true. If the conditions are met, the starting timestamp of the time slice corresponding to the first true value is extracted to generate a pressure deviation trend signal.

5. The energy-saving air compressor operation status monitoring system according to claim 4 is characterized in that: The error fitting acquisition module includes: The synchronous acquisition submodule detects the sampling start time difference between the gas supply frequency sequence and the current waveform characteristic data, interpolates or truncates the timestamps of the two types of data, aligns the time axis using a linear interpolation algorithm, and merges them to generate a synchronous time series data set with a unified time stamp; The feature matching submodule extracts the moment when the difference between adjacent sampling points in the gas supply frequency sequence exceeds the set change threshold based on the synchronous time series data set as the frequency change point, extracts the local maximum and minimum values ​​in the current waveform as the peak and valley points, traverses the timestamps of the frequency change points and peak and valley points one by one, calculates the absolute time difference, and determines it as a valid feature point pair if the time difference is less than the matching fault tolerance window defined by the system, and generates a matching feature pair set; The error normalization submodule calls the time difference and amplitude difference in the matching feature pair set, arranges the time difference in ascending order and takes the median as the benchmark reference value, calculates the standard deviation of the amplitude difference dispersion, and uses the maximum-minimum normalization formula to map the two types of differences to the 0-1 interval, and merges them to generate a fitting error vector.

6. The energy-saving air compressor operation status monitoring system according to claim 5 is characterized in that: The error trend analysis module includes: The cycle extraction submodule calls the fitting error vectors of three consecutive cycles, extracts vector data with continuous timestamps in each cycle in the order of cycle numbers, removes data points that overlap with adjacent cycles or have too large a time span, and generates a cycle error vector group; The mean calculation submodule traverses the vector set of each period in the period error vector group, takes the modulus length of all vectors in each period, accumulates and sums them, divides them by the total number of vectors in the period to obtain the single-period mean, arranges the mean results in the order of period numbers, and generates a period mean sequence; The trend verification submodule extracts the difference between the first period mean and the second period mean based on the period mean sequence as a benchmark increase, calculates the difference between the third period mean and the second period mean as the current increase, compares the ratio of the current increase to the benchmark increase, and updates a status flag and generates a trend verification flag if the current increase exceeds a set multiple threshold of the benchmark increase; The signal generation submodule is based on the state of the trend verification flag. If the flag is true, it extracts the numbers of the current three cycles, combines the total increase value of the third cycle mean and the first cycle mean, and generates a fitting error trend signal.

7. The energy-saving air compressor operation status monitoring system according to claim 6 is characterized in that: The system further comprises: An alarm trigger module receives the pressure deviation trend signal and the fitting error trend signal, verifies the periodic overlap between the two, and if they overlap, calls a joint signal generation function to combine them into an energy-saving deviation alarm signal, and outputs an alarm data packet containing time, number of cycles, and intensity indicators; The energy-saving deviation alarm signal specifically includes a time overlap segment, a cycle superposition number, and a trend strength index.

8. The energy-saving air compressor operation status monitoring system according to claim 7 is characterized in that: The alarm triggering module includes: The cycle verification submodule calls the cycle start time and end time of the pressure offset trend signal, and combines the cycle start time and end time of the fitting error trend signal to calculate the absolute duration of the intersection interval of the two signal time ranges, and performs a ratio operation on the intersection duration and the total duration of the shorter cycle of the two cycles. If the ratio exceeds the cycle overlap determination threshold set by the system, a cycle overlap flag is generated; The signal synthesis submodule is based on the periodic coincidence flag. If the flag is true, it calls the intensity parameter of the pressure offset trend signal and the amplification parameter of the fitting error trend signal, performs weighted summation according to the weight of the intensity parameter and the weight of the amplification parameter, and generates an energy-saving deviation alarm signal; The alarm generation submodule calls the current system timestamp, the associated cycle number and the strength value of the energy-saving deviation alarm signal, encapsulates data according to the triple structure of timestamp-cycle number-signal strength, and generates an energy-saving deviation alarm data packet.

9. A method for monitoring the operating status of an energy-saving air compressor, characterized in that: The method is used to implement the energy-saving air compressor operating status monitoring system according to any one of claims 1 to 8, comprising the following steps: S1: Obtain the compressed air outlet pressure sequence and the load end demand pressure sequence through the pressure sensor, divide the time slices and calculate the range value of the two sequences, and divide the range value by the corresponding time slice duration to generate the pressure slope ratio; S2: Based on the pressure slope ratio, a dynamic threshold generation algorithm is used to calculate the data mean of the initial stage of the current cycle, the mean is multiplied by a preset coefficient to generate an adaptive threshold, and the pressure slope ratio of three consecutive time slices is compared with the adaptive threshold to generate a pressure deviation trend signal and attach a time tag; S3: Synchronously collect the gas supply frequency sequence and current waveform characteristic data, match the timestamp difference between the frequency change point and the current peak and valley point, calculate the time difference and amplitude difference, perform normalization, and generate a fitting error vector; S4: extracting the fitting error vectors of three consecutive cycles, calculating the mean data in each cycle, verifying the increasing relationship between the means of adjacent cycles, and generating a fitting error trend signal when the increase exceeds the mean of the previous two cycles and attaching a cycle number; S5: Calling a joint signal generation function to verify the period overlap of the pressure offset trend signal and the fitting error trend signal, merging the signals to generate an energy-saving deviation alarm signal when they overlap, and outputting an alarm data packet containing a time tag, a period number, and an intensity index.

Images