A reciprocating compressor air valve fault diagnosis method and related device

Abstract

The application discloses a reciprocating compressor air valve fault diagnosis method and related device, and relates to the field of reciprocating compressor fault diagnosis, and comprises the following steps: acquiring a plurality of source state signals, the plurality of source state signals comprising a first type of signal reflecting the mechanical impact state of the air valve and a second type of signal reflecting the thermodynamic state of the air valve; dividing the plurality of source state signals into a plurality of signal segments based on the operating cycle of the reciprocating compressor; extracting impact energy features and frequency domain features in a preset high-frequency frequency band for each signal segment; performing air valve leakage discrimination based on the energy relationship between the impact energy features and a preset reference baseline value and the frequency domain features, or performing valve piece fracture discrimination based on the energy relationship between the impact energy features and the preset reference baseline value and the temperature change features of the second type of signal. The application realizes early and accurate early warning of air valve faults and reliable differentiation of fault modes.

Description

Technical Field

[0001] This application relates to the field of reciprocating compressor fault diagnosis technology, and in particular to a method and related device for diagnosing valve faults in reciprocating compressors. Background Technology

[0002] During the operation of a reciprocating compressor, the valve is a critical and vulnerable component, and its failure mainly manifests in two forms: valve plate leakage and valve plate breakage. Currently, online monitoring of valve failures mainly relies on single-parameter monitoring methods such as vibration, temperature, or acoustic emission.

[0003] However, in the early stages of valve leakage, single vibration monitoring is easily overwhelmed by the weak high-frequency impact energy generated by the leak, which is then masked by the low-frequency background noise from the reciprocating motion of the compressor piston and the rotation of the crankshaft, resulting in a low signal-to-noise ratio and making early warning difficult. Single temperature monitoring is limited by the physical lag of heat conduction; it often takes several minutes or even longer for a significant temperature change to occur after a fault occurs, failing to meet the requirements for rapid response. While single acoustic emission monitoring is sensitive to high-frequency signals, it is easily interfered with by broadband noise generated by other moving parts such as bearings and piston rings, resulting in a high false alarm rate. Furthermore, existing single-parameter monitoring methods cannot accurately distinguish between valve leakage and valve plate breakage, which have different evolutionary patterns, leading to difficulties in maintenance decisions and a high risk of delays or unnecessary downtime due to misjudgment.

[0004] Therefore, there is an urgent need for a diagnostic method that can accurately distinguish fault modes and has a low false alarm rate. Summary of the Invention

[0005] To address the shortcomings of existing technologies, such as slow response in early warning of valve failures due to single-parameter monitoring, susceptibility to noise interference, and inability to distinguish between leakage and breakage fault modes, this application proposes a valve fault diagnosis method and related devices for reciprocating compressors to achieve accurate early warning of valve failures and reliable differentiation of fault modes. The specific solution is as follows:

[0006] The first aspect of this application provides a method for diagnosing valve faults in a reciprocating compressor, comprising:

[0007] Acquire multi-source state signals, which include a first type of signal reflecting the mechanical impact state of the valve and a second type of signal reflecting the thermodynamic state of the valve.

[0008] Based on the operating cycle of the reciprocating compressor, the multi-source status signal is divided into multiple signal segments;

[0009] For each signal segment, the impact energy characteristics within a preset high-frequency band are extracted, and the frequency domain characteristics are also extracted.

[0010] Based on the energy relationship between the impact energy characteristics and the preset reference baseline value, and the frequency domain characteristics, a valve leakage detection is performed; or, based on the energy relationship between the impact energy characteristics and the preset reference baseline value, and the temperature change characteristics of the second type of signal, a valve plate breakage detection is performed.

[0011] In one possible implementation, the step of performing valve leakage detection based on the energy relationship between the impact energy characteristics and a preset reference baseline value, and the frequency domain characteristics, includes:

[0012] When the impact energy characteristic is greater than a first preset multiple of the preset reference baseline value, and the frequency domain characteristic contains a preset fractional harmonic component, it is determined to be a valve leak.

[0013] The preset fractional harmonic components include either a half harmonic component or a one-third harmonic component.

[0014] In one possible implementation, the step of performing valve plate fracture detection based on the energy relationship between the impact energy characteristics and the preset reference baseline value, and the temperature change characteristics of the second type of signal, includes:

[0015] When the impact energy characteristic is less than a second preset ratio of the preset reference baseline value, and the temperature drop value indicated by the second type of signal exceeds a preset temperature drop threshold within a consecutive preset number of operating cycles, it is determined that the valve plate is broken.

[0016] In one possible implementation, the preset reference baseline value is determined based on the statistical average of the impact energy characteristics of the reciprocating compressor during a preset period of time prior to its healthy operation.

[0017] The extraction of impact energy characteristics within a preset high-frequency band includes:

[0018] The signal segments are analyzed using short-time Fourier transform to extract the impulse energy in the 2kHz to 8kHz frequency band as the impulse energy feature.

[0019] In one possible implementation, the first type of signal includes acceleration vibration signal and acoustic emission signal, and the second type of signal includes valve cover surface temperature signal;

[0020] The acquisition of multi-source state signals includes:

[0021] The acceleration vibration signal, the acoustic emission signal, and the valve cover surface temperature signal are collected synchronously at a preset sampling rate.

[0022] In one possible implementation, the multi-source state signal is divided into multiple signal segments based on the operating cycle of the reciprocating compressor, including:

[0023] The operating cycle is determined based on the crankshaft speed of the reciprocating compressor;

[0024] Using the operating cycle as a time window, the multi-source state signal is segmented to obtain the multiple signal segments.

[0025] One possible implementation also includes:

[0026] In response to a determination of valve leakage or valve plate breakage, a diagnostic result is generated that includes the fault type, time of occurrence, and severity.

[0027] The diagnostic results are sent to the monitoring system in real time and stored in the historical database.

[0028] A second aspect of this application provides a reciprocating compressor valve fault diagnosis device, comprising:

[0029] The signal acquisition module is configured to acquire multi-source state signals, which include a first type of signal reflecting the mechanical impact state of the valve and a second type of signal reflecting the thermodynamic state of the valve.

[0030] The cycle segmentation module is configured to divide the multi-source status signal into multiple signal segments based on the operating cycle of the reciprocating compressor;

[0031] The feature extraction module is configured to extract impact energy features within a preset high-frequency band and extract frequency domain features for each of the signal segments.

[0032] The fault detection module is configured to perform valve leakage detection based on the energy relationship between the impact energy characteristics and the preset reference baseline value, as well as the frequency domain characteristics; or, to perform valve plate breakage detection based on the energy relationship between the impact energy characteristics and the preset reference baseline value, as well as the temperature change characteristics of the second type of signal.

[0033] A third aspect of this application provides a computer program product including computer-readable instructions that, when executed on an electronic device, cause the electronic device to implement the reciprocating compressor valve fault diagnosis method described in the first aspect or any implementation thereof.

[0034] A fourth aspect of this application provides an electronic device, including at least one processor and a memory connected to the processor, wherein:

[0035] The memory is used to store computer programs;

[0036] The processor is used to execute the computer program so that the electronic device can implement the reciprocating compressor valve fault diagnosis method described in the first aspect or any implementation thereof.

[0037] The fifth aspect of this application provides a computer storage medium carrying one or more computer programs, which, when executed by an electronic device, enable the electronic device to perform a reciprocating compressor valve fault diagnosis method as described in the first aspect or any implementation thereof.

[0038] By employing the aforementioned technical solutions, this application constructs a multi-dimensional fault characterization system by integrating a first-type signal reflecting the mechanical impact state of the valve and a second-type signal reflecting the thermodynamic state of the valve. By segmenting the signal using the operating cycle and extracting the impact energy characteristics within a preset high-frequency band, interference from the low-frequency mechanical vibration background noise of the compressor body can be effectively avoided, significantly improving the signal-to-noise ratio of weak fault signals. Combining frequency domain characteristics for valve leakage detection utilizes the nonlinear dynamic characteristics induced by leaking airflow, achieving early and accurate identification of leakage faults. Simultaneously, valve plate fracture detection is performed based on the joint logic of impact energy attenuation and temperature change characteristics. Utilizing the dual physical phenomena of the disappearance of mechanical impact and abrupt change in thermodynamic state under fracture fault conditions, this effectively decouples the two fault modes of leakage and fracture, solving the problem that single-parameter monitoring cannot distinguish fault types. Furthermore, through the adaptive correction mechanism of the preset reference baseline value, the diagnostic model can be dynamically adjusted as the equipment operating status evolves, avoiding false alarms or missed alarms under different operating conditions due to fixed thresholds. This significantly improves the robustness and adaptability of the diagnostic method under complex operating conditions such as wide load and variable speed, providing reliable technical support for early warning and precise operation and maintenance of reciprocating compressor valve failures. Attached Figure Description

[0039] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0040] Figure 1 A flowchart of a method for diagnosing valve faults in a reciprocating compressor provided in this application;

[0041] Figure 2 This is a structural diagram of a reciprocating compressor valve fault diagnosis device provided in an embodiment of this application. Detailed Implementation

[0042] The embodiments of this application are described below with reference to the accompanying drawings. The terminology used in the implementation section of this application is for explaining specific embodiments only and is not intended to limit the scope of this application.

[0043] The embodiments of this application will now be described with reference to the accompanying drawings. Those skilled in the art will recognize that, with technological advancements and the emergence of new scenarios, the technical solutions provided in the embodiments of this application are equally applicable to similar technical problems.

[0044] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such terms are interchangeable where appropriate; this is merely a way of distinguishing objects with the same attributes in the embodiments of this application. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, so that a process, method, system, product, or apparatus that comprises a series of elements is not necessarily limited to those elements but may include other elements not explicitly listed or inherent to those processes, methods, products, or apparatuses.

[0045] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0046] This embodiment provides a reciprocating compressor valve fault diagnosis system. The reciprocating compressor valve fault diagnosis system includes: a reciprocating compressor and a fault diagnosis device communicatively connected to the reciprocating compressor.

[0047] A reciprocating compressor includes a crankcase, cylinder, piston, intake valve, exhaust valve, and cooling system.

[0048] The cylinder is the core working chamber of the compressor. Driven by the crankshaft, the piston reciprocates within the cylinder, sequentially completing the four working processes of intake, compression, exhaust, and expansion. The intake and exhaust valves are the most easily damaged components of the compressor, and their health directly affects the compressor's efficiency and operational safety.

[0049] In this embodiment, the operating speed range of the reciprocating compressor is 300 to 500 rpm, and the working medium can be a mixture of hydrogen and nitrogen, or natural gas, air or other process gases. This application does not limit this.

[0050] The fault diagnosis equipment includes sensor components, a data acquisition unit, and a processor. The sensor components are deployed at key monitoring locations on the reciprocating compressor and specifically include: an acceleration sensor mounted on the cylinder valve cover to collect broadband mechanical vibration signals reflecting valve opening and closing impacts and airflow turbulence; an acoustic emission sensor also mounted on the valve cover to capture material stress wave release and high-frequency leakage noise; a dynamic pressure sensor mounted on the intake and exhaust pipes to monitor pressure pulsations during the operating cycle; and a temperature sensor attached to the outer surface of the valve cover to sense thermodynamic state changes during valve operation. The data acquisition unit is connected to the signals of each sensor and is configured to synchronously acquire the output signals of the aforementioned sensors at a preset sampling rate, and transmit the digitized multi-source status signals to the processor. The processor stores a computer program implementing the reciprocating compressor valve fault diagnosis method of any one of Embodiments 1 to 8. When the program is executed by the processor, it can complete the entire diagnostic process from multi-source signal acquisition, periodic segmentation, feature extraction to dual-logic branch discrimination of valve leakage and valve plate breakage.

[0051] Furthermore, the fault diagnosis equipment also includes a monitoring system and a historical database. After completing fault identification, the processor sends structured diagnostic results, including fault type, occurrence time, and fault severity, to the monitoring system in real time to trigger audible and visual alarms and highlight fault details on the human-machine interface, enabling on-site operators to perceive the anomaly and take intervention measures immediately. Simultaneously, the processor stores the complete diagnostic record in the historical database to support subsequent fault tracing, trend analysis, and adaptive optimization of the diagnostic algorithm. The monitoring system can be an on-site DCS (Distributed Control System), a local monitoring screen, or a remote cloud monitoring platform; this application does not impose any restrictions on this.

[0052] It should be understood that although this embodiment describes a single reciprocating compressor equipped with a fault diagnosis device as an example, in other embodiments, the diagnostic system can also be expanded into a centralized monitoring network covering multiple compressor groups. For example, in a skid-mounted reciprocating compressor plant, the sensor components of multiple compressors are each connected to a local data acquisition unit and processor. Each processor aggregates the diagnostic results to the central monitoring system via industrial Ethernet, thereby achieving unified health management and fault early warning for the entire plant's compressor group. Furthermore, the processor and data acquisition unit can be integrated into the same edge computing box or deployed on different physical devices interconnected via a network, as long as the overall system can achieve the aforementioned diagnostic functions.

[0053] Example 1

[0054] To overcome the shortcomings of existing technologies in reciprocating compressor valve fault diagnosis, such as early warning lag, susceptibility to background noise interference, and inability to effectively distinguish between leakage and breakage fault modes, this embodiment provides a general basic process for reciprocating compressor valve fault diagnosis. This process, by constructing a multi-source heterogeneous signal fusion and dual-logic-branch discrimination architecture, fundamentally decouples the evolutionary characteristics of different fault types.

[0055] like Figure 1 The diagram shown is a flowchart of a reciprocating compressor valve fault diagnosis method provided in an embodiment of this application. The reciprocating compressor valve fault diagnosis method provided in this embodiment mainly includes the following steps S100 to S400.

[0056] Step S100: Obtain multi-source state signals, which include a first type of signal reflecting the mechanical impact state of the valve and a second type of signal reflecting the thermodynamic state of the valve.

[0057] Specifically, this step aims to establish a multi-dimensional perception capability of the valve's operating status.

[0058] The first type of signal mainly originates from the transient mechanical response of the valve assembly during the opening and closing process. This can be characterized by vibration signals collected by an accelerometer mounted on the valve cover, and stress wave signals collected by an acoustic emission sensor. This type of signal is extremely sensitive to high-frequency dynamic events such as valve plate impact and airflow turbulence, and can detect subtle mechanical anomalies in the early stages of a fault.

[0059] The second type of signal focuses on reflecting the thermodynamic equilibrium state during the operation of the valve, such as the temperature signal on the valve cover surface collected by a temperature sensor. Since valve leakage or breakage will lead to changes in compression efficiency and gas backflow, which in turn will cause changes in local heat exchange conditions, the thermodynamic signal, although relatively delayed in response, has extremely high stability and a cumulative indication of the severity of the fault.

[0060] By fusing these two types of signals with different physical mechanisms, this application can cross-verify the valve state from two orthogonal dimensions: dynamic transient and steady-state static, effectively avoiding the blind spots of a single signal source under certain operating conditions.

[0061] Step S200: Based on the operating cycle of the reciprocating compressor, the multi-source status signal is divided into multiple signal segments.

[0062] During operation, the internal pressure, flow rate, and vibration signals of a reciprocating compressor exhibit strong non-stationary periodic changes with the crankshaft angle. If time-domain statistical analysis is performed directly on long-term continuous signals, the superposition of signals at different phase points can lead to feature ambiguity or even mutual cancellation.

[0063] Therefore, this embodiment uses the compressor's operating cycle as a time scale to segment continuously acquired multi-source state signals into independent signal segments with complete physical meaning. Each signal segment corresponds to a complete intake → compression → exhaust → expansion process. This segmentation method based on physical cycles can transform non-stationary signals into a series of quasi-stationary signal samples, eliminating the interference of speed fluctuations and phase differences on subsequent feature extraction, and ensuring the consistency and comparability of feature calculations.

[0064] Step S300: For each of the signal segments, extract the impact energy characteristics within a preset high-frequency band and extract the frequency domain characteristics.

[0065] After obtaining the periodic signal segments, it is necessary to extract key indicators that can sensitively characterize faults. Considering that the reciprocating compressor body has a large amount of low-frequency strong background noise generated by the reciprocating motion of the piston and the rotation of the crankshaft, this noise often masks the weak early fault signals of the valves. Therefore, this embodiment specifically selects a preset high-frequency band as the extraction window for impact energy characteristics. The principle of setting this frequency band is to avoid the main mechanical structure resonance frequency and low-frequency motion components, and focus on the high-frequency components generated by valve plate impact and high-speed leakage airflow.

[0066] In addition to energy amplitude, frequency domain features are further extracted to capture changes in the signal's spectral structure. This is because valve malfunctions (especially leaks) often induce a coupling effect between airflow vortex shedding and valve plate flutter, resulting in specific nonlinear modulation components in the signal. Simple time-domain energy cannot fully characterize this complex dynamic behavior.

[0067] Step S400: Based on the energy relationship between the impact energy characteristics and the preset reference baseline value, and the frequency domain characteristics, perform valve leakage detection; or, based on the energy relationship between the impact energy characteristics and the preset reference baseline value, and the temperature change characteristics of the second type of signal, perform valve plate breakage detection.

[0068] Specifically, a dual logic branch architecture with either parallel or selective operation is adopted. For gas valve leakage, the essence is high-speed gas backflow caused by seal failure, which manifests in the signal as a significant increase in high-frequency impact energy characteristics (relative to the preset reference baseline value of the healthy state) and the emergence of nonlinear components in the spectrum. Therefore, a joint criterion of surge in impact energy characteristics and frequency domain anomalies is adopted.

[0069] As for valve plate fracture, its essence is the destruction of the mechanical structure, which leads to the disappearance of normal opening and closing impact, accompanied by the disruption of the thermal balance of the compression chamber. This is manifested in the signal as a sudden drop or even disappearance of high-frequency impact energy characteristics, as well as an abnormal drop in valve cover temperature. Therefore, a combined criterion of impact energy characteristic attenuation and temperature change characteristic decrease is adopted.

[0070] This application ingeniously utilizes the opposite characteristics of the two faults in the direction of energy change, such as the energy increase from valve leakage and the energy reduction from valve breakage, as well as the differentiated response of the auxiliary characteristic domain, to fundamentally achieve the decoupling of fault modes.

[0071] It should be understood that "or" in step S400 does not mean a mutually exclusive execution order. In actual systems, these two discrimination logics can run in parallel in real time, or they can be executed sequentially according to a preset priority, as long as they can cover the identification requirements of these two core fault modes.

[0072] This embodiment constructs a general fault diagnosis framework that is independent of specific sensor models or fixed threshold parameters through the above steps. This framework establishes a technical approach that includes multi-source state signal fusion, periodic signal segmentation processing, impact energy feature extraction within a preset high-frequency band, and a dual-logic-branch architecture.

[0073] Example 2

[0074] Building upon Example 1, this example further provides a specific valve leakage detection logic and feature extraction implementation method to capture subtle valve leakage faults and effectively suppress background noise interference in practical engineering applications. If only a general energy threshold is relied upon without specific frequency domain feature constraints, false alarms are easily generated under varying compressor operating conditions or strong vibration environments. Therefore, this example constructs a complete technical closed loop by refining the signal processing algorithm and combining it with multi-dimensional feature joint criteria.

[0075] Specifically, "based on the energy relationship between the impact energy characteristics and the preset reference baseline value, and the frequency domain characteristics, performing valve leakage judgment" includes: when the impact energy characteristics are greater than a first preset multiple of the preset reference baseline value, and the frequency domain characteristics contain a preset fractional harmonic component, it is determined that the valve is leaking.

[0076] The above discrimination logic employs a dual verification mechanism based on energy amplitude and spectral structure.

[0077] The first preset multiple is not arbitrarily selected, but is based on the confidence interval boundary obtained from a large amount of historical fault data. For example, in this embodiment, it can be selected as 3 times. This value can cover the energy growth range in the early stage of slight leakage and effectively avoid random fluctuation peaks during normal operation.

[0078] For example, the preset fractional harmonic components include half-harmonic components or one-third-harmonic components. From a nonlinear dynamics perspective, the normal opening and closing impact of a valve primarily excites the rotational frequency and its integer multiples of harmonics. However, when a valve leaks, the turbulent vortices generated by the high-speed airflow passing through the tiny gaps nonlinearly couple with the self-excited flutter of the valve plate. This complex fluid-structure interaction modulates fractional harmonics of the fundamental frequency in the frequency spectrum. These fractional harmonic components are unique to leakage faults; vibrations or noises from other mechanical components rarely produce such characteristics. Therefore, including them as a necessary condition can fundamentally eliminate interference from non-leakage factors.

[0079] In order to accurately obtain the above-mentioned impact energy characteristics, "extracting impact energy characteristics within a preset high-frequency band" includes: performing time-frequency analysis on the signal segments using short-time Fourier transform to extract the impact energy within the 2kHz to 8kHz frequency band as the impact energy characteristics.

[0080] The selection of the specific frequency band of 2kHz to 8kHz is based on the following clear reasoning: the mechanical vibrations of the reciprocating compressor, such as piston reciprocating motion, crankshaft rotation, and gear meshing, are mainly concentrated in the low-frequency range below 1kHz, while the high-frequency stress waves and turbulence noise caused by valve leakage are mainly distributed above 2kHz. By locking the analysis window to 2kHz-8kHz, it is equivalent to constructing a high-pass filter in the frequency domain, shielding most of the strong low-frequency background noise and significantly improving the signal-to-noise ratio of the target signal.

[0081] In the specific algorithm implementation, the parameter configuration of the Short-Time Fourier Transform (STFT) is crucial. In this embodiment, the preferred window length is 1024 points, and the overlap rate is 50%. This parameter combination is the engineering optimal solution between time resolution and frequency resolution: the 1024-point window length provides a frequency resolution of approximately 48.8Hz at a 50kHz sampling rate, which is sufficient to clearly distinguish adjacent fractional harmonic components; while the 50% overlap rate ensures that transient impact events occurring at the window edges are not missed during the sliding window process, ensuring the integrity and continuity of feature extraction.

[0082] Furthermore, the preset reference baseline value is determined based on the statistical average of the impact energy characteristics of the reciprocating compressor during a preset period of time before it reaches a healthy operating state.

[0083] The "preset duration" here is usually set to 72 hours of continuous healthy operation data during the initial stage of equipment commissioning or after major overhaul in actual deployment. Using statistical averages rather than single measurements or fixed empirical values ​​as a benchmark can effectively smooth out individual variations caused by sensor installation differences, changes in ambient temperature and humidity, and the break-in period of the compressor.

[0084] This dynamic benchmark generation method based on its own historical data enables the diagnostic system to have an adaptive capability of one file per machine, eliminating the need for tedious manual threshold calibration for each device, and also avoiding the problem of fixed threshold failure due to equipment aging.

[0085] This embodiment limits the leakage criterion to the high-frequency band of 2kHz-8kHz and combines it with fractional harmonic characteristics, achieving deep decoupling of fault characteristics and background noise at the physical mechanism level. This significantly improves the signal-to-noise ratio and identification accuracy of weak leakage faults. At the same time, the refined time-frequency analysis and dynamic baseline calculation strategy based on STFT solves the technical problems of early leakage signals being easily submerged and poor adaptability of fixed thresholds, providing a reliable and robust implementation path for early warning of gas valve leaks.

[0086] Example 3

[0087] In the dual-logic branch architecture constructed in Example 1, besides the detection of valve leakage, valve plate fracture is another extremely dangerous failure mode, and its evolution pattern is completely different from that of leakage. If only a single vibration monitoring method is relied upon, the signal often shows energy loss rather than abnormal enhancement because the mechanical impact disappears after valve plate fracture. This can easily be misjudged by the system as normal operation or background noise, leading to missed detections. If only temperature monitoring is relied upon, although the temperature drop caused by fracture can be captured, the response not only has significant thermal inertia hysteresis, but it is also difficult to effectively distinguish it from the temperature change of leakage under certain special operating conditions. Therefore, in order to fill this diagnostic blind spot and achieve accurate decoupling of failure modes, this embodiment provides a joint detection logic for valve plate fracture. It uses the attenuation of mechanical impact characteristics and the deterioration of thermodynamic state for dual verification, forming another core scheme independent of leakage detection.

[0088] Specifically, "based on the energy relationship between the impact energy characteristics and the preset reference baseline value, and the temperature change characteristics of the second type of signal, perform valve plate fracture detection" includes: based on the energy relationship between the impact energy characteristics and the preset reference baseline value, and the temperature change characteristics of the second type of signal, perform valve plate fracture detection.

[0089] From a deeper physical perspective, the design basis of this joint criterion lies in the unique inverse evolution characteristics of energy and temperature in valve plate fracture failure. First, regarding the sharp drop in impact energy characteristics, when the valve plate fractures, the previously regular rigid mechanical impact between the valve plate and the valve seat disappears or is significantly weakened. This directly leads to a precipitous drop in the first type of signal reflecting the mechanical impact state of the valve (such as the impact energy characteristics in the 2kHz-8kHz frequency band).

[0090] For example, in this embodiment, the second preset ratio can be set to 0.5, meaning that when the real-time impact energy characteristic is lower than 50% of the healthy baseline value, it is considered an abnormal absence of mechanical impact. This is in stark contrast to the physical characterization of the valve leakage in Embodiment 2, where the energy surge induced by high-speed airflow turbulence (i.e., the impact energy characteristic is greater than the first preset multiple of the preset reference baseline value). This opposition in the direction of energy change provides the most direct mechanical basis for this application to distinguish between valve leakage and valve plate breakage at the algorithm level, effectively avoiding the logical confusion of misjudging a breakage as a leak or a leak as a breakage.

[0091] Secondly, regarding the introduction of temperature change characteristics, valve plate breakage will lead to cylinder compression chamber seal failure, and high-pressure gas will flow back in large quantities during the expansion stroke, causing a sharp drop in compression efficiency and a significant reduction in heat generation, which in turn will cause the valve cover surface temperature to continuously decrease.

[0092] However, temperature signals, as thermodynamic parameters, are subject to hysteresis and are easily disturbed. Fluctuations in compressor cooling water flow, sudden changes in ambient temperature, and even poor sensor contact can all cause instantaneous abnormal temperature readings. If an alarm is triggered based solely on a single or short-term temperature drop, false alarms are highly likely. Therefore, this embodiment specifically introduces a timing constraint of "a continuously preset number of operating cycles." For example, the continuously preset number can be set to 2 or 3 complete operating cycles, and the preset temperature drop threshold can be set to 15°C. This means that only when the impact energy characteristic is confirmed to remain low, and the observed temperature drop trend maintains continuity for at least 2-3 cycles and exceeds a range of 15°C on the time axis, will a valve plate fracture be definitively diagnosed.

[0093] This timing consistency verification mechanism is equivalent to adding a low-pass filter to the thermodynamic criterion, which can effectively filter out transient thermal noise and occasional signal interference, ensuring the robustness and reliability of the diagnostic results.

[0094] It should be understood that the aforementioned 0.5 ratio, 15°C temperature drop threshold, and 2 cycles are merely preferred examples for explaining the principles of this application and are not intended to limit the application. In practical engineering applications, technicians can determine the most suitable second preset ratio, preset temperature drop threshold, and number of consecutive cycles based on factors such as the specific compressor model, medium characteristics, insulation conditions, and sensor accuracy, through experimental calibration or historical data analysis. For example, for small high-speed compressors with smaller heat capacity and faster temperature drop response, the number of consecutive cycles can be appropriately reduced; while for large low-speed compressors with larger thermal inertia, a longer timing confirmation window may be required.

[0095] This embodiment achieves accurate identification and reliable alarm for valve plate fracture faults by constructing a joint criterion of impact energy characteristic attenuation and continuous temperature drop. On the one hand, the sudden drop characteristic of impact energy characteristics compensates for the shortcomings of single temperature monitoring in terms of response lag and inability to qualitatively classify fault types; on the other hand, the continuous temperature drop characteristic solves the industry pain point that single vibration monitoring cannot detect faults due to signal loss after valve plate fracture. More importantly, this criterion is a mirror image of the leakage criterion in Embodiment 2 in the direction of energy change and is independent of each other in the auxiliary feature domain. Thus, within the same diagnostic framework, the two core fault modes of leakage and fracture are completely decoupled, significantly improving the comprehensiveness and accuracy of the reciprocating compressor valve fault diagnosis system.

[0096] Example 4

[0097] In the diagnostic methods described in Examples 1 to 3, the quality of acquiring multi-source state signals directly determines the accuracy of subsequent feature extraction and fault diagnosis. If only the functional attributes of the signals are defined without specific physical acquisition carriers and synchronization mechanisms, in practical engineering applications, improper sensor selection, installation position deviations, or timing misalignments of multi-channel data can easily lead to distortion of high-frequency impact characteristics or misalignment of cross-modal characteristics, thereby causing the aforementioned dual-logic discrimination to fail. Therefore, this embodiment further provides specific acquisition configurations and synchronization implementation methods for multi-source signals, providing reliable hardware entity support for the abstract signal definition.

[0098] Specifically, the first type of signal includes acceleration vibration signals and acoustic emission signals, while the second type of signal includes valve cover surface temperature signals.

[0099] This signal classification closely corresponds to the physical deployment of the sensors. Among them, the acceleration vibration signal is acquired by an acceleration sensor installed on the cylinder valve cover, which is used to capture broadband mechanical vibrations caused by valve opening and closing and airflow impact; the acoustic emission signal is acquired by an acoustic emission sensor also installed on the valve cover, which is specifically used to sense material stress wave release and high-frequency turbulent noise; and the valve cover surface temperature signal is acquired by a temperature sensor attached to the outer surface of the valve cover, which is used to monitor the thermal balance state during the operation of the valve.

[0100] It should be understood that although this embodiment preferably arranges both the acceleration sensor and the acoustic emission sensor on the valve cover, in other embodiments, as long as the mechanical impact state of the valve can be sensitively reflected, the installation position of the sensor can also be adjusted according to the structural characteristics of the compressor, such as arranging it on the cylinder head or on the intake and exhaust manifold near the valve. However, the valve cover position is usually the optimal choice for signal-to-noise ratio because it is closest to the fault source and is less affected by the noise of the main body transmission.

[0101] Regarding hardware selection, to ensure effective extraction of the impact energy characteristics within the 2kHz to 8kHz frequency band in Example 2, the preferred accelerometer is a piezoelectric accelerometer with a sensitivity of 100mV / g and a frequency response range covering 0.5Hz to 10kHz, ensuring complete coverage of the target analysis frequency band without aliasing. The preferred acoustic emission sensor is a resonant sensor with a center frequency of 100kHz and a bandwidth of 50kHz to 200kHz, which can effectively capture the ultrasonic signals generated by the leak while avoiding interference from most low-frequency mechanical friction noise. The temperature sensor can be a Pt100 platinum resistance thermometer or a K-type thermocouple, and its response time and measurement accuracy should meet the monitoring requirements for the gradual change trend of valve cover temperature.

[0102] These specific sensor parameters are merely preferred examples for explaining the principles of this application and are not intended to limit the application. Those skilled in the art can make equivalent substitutions based on the actual operating parameters and fault characteristic frequencies of the compressor.

[0103] Furthermore, acquiring multi-source state signals includes: synchronously acquiring the acceleration vibration signal, the acoustic emission signal, and the valve cover surface temperature signal at a preset sampling rate.

[0104] The following is an explanation of "synchronous acquisition".

[0105] In practice, the data acquisition unit has a unified high-precision clock source or hardware trigger bus. The signal conditioning circuits of the accelerometer, acoustic emission sensor and temperature sensor are all controlled by the same clock reference, thereby ensuring strict alignment of the three signals on the time axis.

[0106] For example, the preset sampling rate is preferably set to 50kHz in this embodiment. This value is determined based on the Nyquist sampling theorem and the effective bandwidth of the acoustic emission signal, which can ensure the distortion-free reproduction of high-frequency components and also take into account the real-time load capacity of the data processing system.

[0107] The necessity of synchronous acquisition is emphasized because the fault diagnosis of reciprocating compressor valves is highly dependent on the phase information of the signal within the operating cycle. As described in step S200 of Example 1, the signal needs to be segmented based on the operating cycle. If there is a timing deviation of milliseconds or even microseconds between the signals of each channel, the vibration peak, acoustic emission event and temperature reading within the same operating cycle cannot be accurately matched within the time window.

[0108] Specifically, when executing the joint discrimination logic in Embodiments 2 and 3, only when the surge in impact energy characteristics and the fractional harmonic characteristics, or the decay of impact energy characteristics and the drop in temperature, occur strictly within the same physical time period can they be confirmed as valid fault evidence. The phase difference introduced by asynchronous acquisition may be misread by the algorithm as missing or abnormal features, leading to missed or false alarms. Therefore, a hardware-level synchronous acquisition mechanism ensures precise alignment of multi-source heterogeneous signals in the time dimension, laying a solid data foundation for subsequent joint analysis of multi-dimensional features and accurate decoupling of fault modes.

[0109] This embodiment defines the sensor type, installation location, and synchronous acquisition mechanism, transforming the higher-level signal definition into an executable engineering solution. This not only ensures the complete acquisition of information in the 2kHz-8kHz key frequency band, but also fundamentally eliminates the potential interference of multi-channel data timing misalignment on the joint diagnostic logic, significantly improving the reliability and robustness of the diagnostic system in complex industrial environments.

[0110] Example 5

[0111] After establishing a general diagnostic framework and signal acquisition foundation in Examples 1 to 4, the time reference problem for signal segmentation must be addressed to ensure strict physical consistency in subsequent feature extraction and fault diagnosis. In actual industrial operation, reciprocating compressors are affected by power grid frequency fluctuations, load adjustments, or changes in process requirements, causing their crankshaft speed to be not absolutely constant but dynamically drifting within a certain range. If a fixed time length is used to mechanically segment continuous signals, the extracted signal segments may contain incomplete compression cycles or cross two adjacent cycle boundaries when the speed changes. This will cause the extracted impact energy features and frequency domain features to lose their clear physical correspondence, leading to misjudgments. Therefore, to enhance the feasibility and robustness of the solution under varying operating conditions, this example further provides a specific method for determining the operating cycle window.

[0112] Specifically, "dividing the multi-source status signal into multiple signal segments based on the operating cycle of the reciprocating compressor" includes the following steps A1 to A2.

[0113] Step A1: Determine the operating cycle based on the crankshaft speed of the reciprocating compressor.

[0114] Step A2: Using the running cycle as a time window, segment the multi-source state signal to obtain multiple signal segments.

[0115] The core of this process lies in establishing a real-time mapping relationship between signal processing time and the phase of mechanical motion.

[0116] Regarding the acquisition of crankshaft speed, this embodiment provides several implementation methods adapted to different field conditions. In one optional implementation, speed pulse signals can be directly acquired using a key-phase sensor or photoelectric encoder installed on the crankshaft end or flywheel. The instantaneous speed is calculated with high precision by measuring the time interval between adjacent pulses. This method has strong anti-interference capabilities and the highest accuracy. In another implementation, for older units or confined spaces without dedicated speed sensors, the crankshaft speed can also be estimated backward from the periodic components of the acquired acceleration vibration signals or dynamic pressure signals through autocorrelation analysis or envelope demodulation, thereby achieving adaptive segmentation under sensorless conditions. It should be understood that regardless of the speed acquisition method used, the purpose is to track the compressor's mechanical cycle time in real time.

[0117] After obtaining the real-time crankshaft speed, the current operating cycle is dynamically calculated based on the compressor's structural parameters (such as single-acting or double-acting compressors). For example, for a single-acting compressor, the operating cycle is equal to 60 divided by the current speed value; for a double-acting compressor, the alternating working characteristics of both sides of the cylinder need to be considered for corresponding calculations. Subsequently, the data processing unit uses this real-time calculated operating cycle as a sliding or fixed time window to precisely segment the multi-source state signal stream synchronously acquired in Example 4. This dynamic segmentation mechanism based on physical beats ensures that each generated signal segment completely and uniquely corresponds to a thermodynamic working cycle of intake → compression → exhaust → expansion, completely eliminating signal truncation errors or phase misalignments caused by speed fluctuations.

[0118] This embodiment introduces a dynamic window segmentation mechanism based on crankshaft speed, enabling the diagnostic method to possess excellent adaptability to operating conditions. Regardless of whether the compressor is in steady-state operation, variable speed regulation, or start-stop transition phase, the system can ensure the integrity and consistency of signal segments in the physical dimension. This provides a reliable data alignment basis for leakage and fracture discrimination based on periodic characteristics in subsequent embodiments 2 and 3, effectively avoiding feature distortion and diagnostic failure caused by speed changes.

[0119] Example 6

[0120] After constructing a complete algorithm chain from signal acquisition and feature extraction to fault diagnosis in Examples 1 to 5, in order to ensure that the diagnostic results can truly serve the operation and maintenance decisions in the industrial field, the interface problem between algorithm output and engineering application must be solved. If it only stays at the Boolean value judgment level without structured information encapsulation and standardized data flow mechanism, operators will not be able to quickly understand the specific nature and severity of the fault, nor will they be able to trace and analyze the historical health status of the equipment, thus greatly reducing the practical application value of the diagnostic system. Therefore, in order to improve the output end of the diagnostic method, this embodiment further provides a specific implementation method for diagnostic result output and archiving. Specifically, it also includes the following steps B1 to B2.

[0121] Step B1: In response to the determination that the valve is leaking or the valve plate is broken, generate a diagnostic result including the fault type, the time of occurrence, and the degree of fault.

[0122] The diagnostic results here are not simple alarm signals, but a multi-dimensional structured data packet.

[0123] Among them, the fault type field clearly distinguishes between valve leakage and valve plate breakage. This distinction directly echoes the completely different dual logic discrimination branches in the aforementioned Embodiment 2 and Embodiment 3.

[0124] Since leakage faults usually require planned shutdowns for maintenance, while fracture faults often mean the risk of emergency shutdowns, clear type identification can help maintenance personnel quickly match the correct emergency response plan and avoid misoperation or delays caused by ambiguous fault classification.

[0125] The occurrence time field accurately records the timestamp when the fault characteristics first meet the discrimination threshold, providing a time reference for subsequent fault tracing and correlation analysis.

[0126] The fault severity field is quantified and graded based on the multiple by which the impact energy characteristics exceed the preset reference baseline value (for leaks) or the magnitude by which the temperature drop value exceeds the preset temperature drop threshold (for fractures). For example, leaks can be divided into three levels: minor, moderate, and severe, so that maintenance personnel can intuitively perceive the evolution stage of the fault and thus formulate differentiated maintenance strategies.

[0127] Step B2: Send the diagnostic results to the monitoring system in real time and store the diagnostic results in the historical database.

[0128] In practical engineering implementation, this process is typically performed by edge computing devices or PLCs (Programmable Logic Controllers). After fault identification and diagnostic results are generated, the results are pushed to the on-site DCS (Distributed Control System) or local monitoring screen in milliseconds via standard communication protocols such as Industrial Ethernet, Modbus, or OPC UA (OPC Unified Architecture). This triggers audible and visual alarms and highlights fault details on the human-machine interface, ensuring that operators can perceive the anomaly and take intervention measures immediately.

[0129] At the same time, the system will write the complete diagnostic record, including the original feature values, discrimination logic parameters and the final conclusion, into the historical database.

[0130] The aforementioned dual-channel output mechanism, which combines real-time push notifications of diagnostic results with persistent storage of complete diagnostic records, ensures immediate access to fault information, preventing missed reports due to network latency or system refreshes, while also enabling long-term accumulation of health data throughout the device's lifecycle. The data stored in the historical database can be used not only for post-fault review and responsibility determination but also provides a valuable sample base for subsequent trend analysis, remaining life prediction, and adaptive optimization of diagnostic algorithms.

[0131] It should be understood that although this embodiment describes the method of sending data to the DCS system via a wired communication protocol, in other embodiments, wireless transmission such as 5G or Wi-Fi 6 can also be used to upload diagnostic results to a cloud platform or mobile terminal, or the results can be directly embedded into the compressor's local control logic to achieve automatic interlocking protection. Regardless of the transmission medium and receiving terminal used, as long as real-time distribution and persistent recording of diagnostic results can be achieved, they fall within the protection scope of this application.

[0132] Example 7

[0133] In Examples 2 and 4, although a preset reference baseline was established using statistical data from the initial stage of healthy operation, providing an initial benchmark for fault diagnosis, the background levels of vibration and acoustic emission signals of the equipment often drift slowly during the long-term actual operation of the reciprocating compressor due to natural wear of valve components, aging of seals, fine-tuning of process medium composition, and seasonal changes in temperature and humidity. If the fixed preset reference baseline value calibrated at the initial stage of commissioning is always used, the system is prone to frequent false alarms when the overall energy level of the equipment generally rises due to normal aging; conversely, if the background noise is reduced due to maintenance and adjustment but the baseline is not updated, weak fault signals may be missed due to falling below the threshold. To address the problem that a fixed baseline cannot adapt to the evolution of the equipment's state throughout its entire life cycle and to improve the robustness of the diagnostic scheme in long-term operation, this embodiment further provides a dynamic maintenance mechanism for the baseline value.

[0134] Specifically, after the reciprocating compressor completes the initial baseline calibration, the preset reference baseline value is adaptively corrected based on the impact energy characteristics extracted during subsequent operating cycles.

[0135] The core of this adaptive correction mechanism lies in constructing a closed-loop feedback system with state awareness capabilities. Its execution process includes two key stages: rigorous trigger condition verification and a smooth update strategy. First, regarding trigger conditions, the system does not update the baseline for all collected impact energy characteristics. Instead, it only allows new data to be included in the baseline calculation queue when the current operating cycle is confirmed to be in a healthy state. The confirmation logic for the healthy state relies on the reverse verification of the fault discrimination results in Examples 2 and 3. That is, only when the impact energy characteristics of the current cycle do not meet the criteria for valve leakage (greater than the first preset multiple and exhibiting fractional frequency doubling) or valve plate breakage (less than the second preset ratio and with continuous temperature drop exceeding the standard), and non-steady-state conditions such as start-up and shutdown transitions are excluded, will the data for that operating cycle be marked as a valid healthy sample. This fundamentally eliminates the risk of fault data or abnormal interference data contaminating the preset reference baseline value.

[0136] Secondly, regarding the specific correction algorithm, in order to avoid drastic disturbances to the baseline caused by single measurement fluctuations, while also being able to track the long-term trend changes of the equipment in a timely manner, this embodiment preferably adopts an exponentially weighted moving average algorithm or a recursive update strategy with a forgetting factor to iteratively calculate the preset reference baseline value.

[0137] For example, a new preset reference baseline value can be represented as a weighted sum of the old baseline value and the impact energy characteristics of the current healthy sample, where the weight of the new data is determined by a preset forgetting factor. The forgetting factor typically ranges from 0.01 to 0.1. A smaller forgetting factor means the system relies more on historically accumulated data, resulting in smoother baseline changes and stronger resistance to random disturbances, making it suitable for large, stable generating units. A larger forgetting factor, on the other hand, gives new data a greater influence, allowing the baseline to respond more quickly to sudden changes in state after operating condition switching or component replacement, making it suitable for generating units with frequent load fluctuations or those that have just undergone maintenance. Alternatively, as another optional implementation, a statistical average within a sliding time window can be used for updating, i.e., only the average impact energy characteristics of the most recent N healthy operating cycles are retained, automatically discarding outdated historical data. This method is more efficient in edge computing devices with limited computing resources. It should be understood that regardless of the specific mathematical model used, the essence is to progressively calibrate the initial understanding using subsequently accumulated healthy operating data, rather than simply replacing numerical values.

[0138] Example 8

[0139] like Figure 2 The diagram shown is a structural diagram of a reciprocating compressor valve fault diagnosis device provided in an embodiment of this application. The device includes:

[0140] The signal acquisition module 201 is configured to acquire multi-source state signals, which include a first type of signal reflecting the mechanical impact state of the valve and a second type of signal reflecting the thermodynamic state of the valve.

[0141] The cycle segmentation module 202 is configured to divide the multi-source status signal into multiple signal segments based on the operating cycle of the reciprocating compressor.

[0142] The feature extraction module 203 is configured to extract the impact energy features within a preset high-frequency band and extract frequency domain features for each of the signal segments.

[0143] The fault discrimination module 204 is configured to perform valve leakage discrimination based on the energy relationship between the impact energy characteristics and the preset reference baseline value, and the frequency domain characteristics; or, to perform valve plate breakage discrimination based on the energy relationship between the impact energy characteristics and the preset reference baseline value, and the temperature change characteristics of the second type of signal.

[0144] In an optional implementation, the fault detection module includes a leakage detection unit, configured as follows:

[0145] When the impact energy characteristic is greater than a first preset multiple of the preset reference baseline value, and the frequency domain characteristic contains a preset fractional harmonic component, it is determined to be a valve leak.

[0146] The preset fractional harmonic components include either a half harmonic component or a one-third harmonic component.

[0147] In an optional implementation, the fault detection module includes a fracture detection unit, configured as follows:

[0148] When the impact energy characteristic is less than a second preset ratio of the preset reference baseline value, and the temperature drop value indicated by the second type of signal exceeds a preset temperature drop threshold within a consecutive preset number of operating cycles, it is determined that the valve plate is broken.

[0149] In one optional implementation, the feature extraction module includes:

[0150] The baseline determination unit is configured to determine the preset reference baseline value based on the statistical average value of the impact energy characteristics of the reciprocating compressor during a preset period of time prior to its healthy operation.

[0151] The time-frequency analysis unit is configured to perform time-frequency analysis on the signal segments using short-time Fourier transform to extract the impulse energy in the 2kHz to 8kHz frequency band as the impulse energy feature.

[0152] In one optional implementation, the signal acquisition module includes:

[0153] The acceleration vibration signal acquisition unit is configured to acquire acceleration vibration signals;

[0154] The acoustic emission signal acquisition unit is configured to acquire acoustic emission signals;

[0155] The temperature signal acquisition unit is configured to acquire the temperature signal of the valve cover surface;

[0156] The first type of signal includes the acceleration vibration signal and the acoustic emission signal, the second type of signal includes the valve cover surface temperature signal, and each of the acquisition units is configured to synchronously acquire the corresponding signal at a preset sampling rate.

[0157] In an alternative implementation, the periodic segmentation module is configured as follows:

[0158] The operating cycle is determined based on the crankshaft speed of the reciprocating compressor;

[0159] Using the operating cycle as a time window, the multi-source state signal is segmented to obtain the multiple signal segments.

[0160] In an alternative implementation, the apparatus further includes:

[0161] The diagnostic result output module is configured to generate a diagnostic result containing the fault type, occurrence time, and fault severity in response to a determination of valve leakage or valve plate breakage, send the diagnostic result to the monitoring system in real time, and store the diagnostic result in a historical database.

[0162] In an alternative implementation, the apparatus further includes:

[0163] The baseline update module is configured to adaptively correct the preset reference baseline value based on the impact energy characteristics extracted during subsequent operating cycles after the reciprocating compressor completes the initial baseline calibration.

[0164] In one optional implementation, the diagnostic result output module includes:

[0165] The result sending unit is configured to send the diagnostic results to the monitoring system in real time.

[0166] The result storage unit is configured to store the diagnostic results in the historical database.

[0167] Example 9

[0168] Having detailed the method and modular device architecture for diagnosing reciprocating compressor valve faults in Examples 1 to 8, and considering the high diversity of physical hardware forms carrying the diagnostic logic in actual industrial applications, limiting the technical solution to specific functional module combinations may not fully cover different levels of computing platforms, from embedded edge terminals to cloud servers. To extend the scope of protection of this application to general-purpose computing hardware carriers and clarify the collaborative execution relationship between software programs and underlying hardware, this embodiment provides an electronic device.

[0169] Specifically, the electronic device includes at least one processor and a memory connected to the processor. The processor is the core of the electronic device's operation and control center, and its specific implementation is not limited to a particular type of chip. For example, the processor can be a general-purpose central processing unit (CPU), suitable for industrial control computers or server scenarios running operating systems and complex upper-layer applications; it can also be a digital signal processor (DSP) or a microcontroller (MCU), suitable for embedded monitoring terminals with high real-time requirements and limited resources; or it can be a field-programmable gate array (FPGA) or an application-specific integrated circuit (ASIC), suitable for edge computing boxes that need hardware acceleration to improve the efficiency of high-frequency signal processing. Regardless of the specific chip architecture used, as long as it has instruction execution capability and can interact with the memory, it can serve as the processor in this embodiment. Similarly, the memory is used to store computer programs, and its type includes, but is not limited to, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive, or solid-state drive. The computer program stored in the memory contains all the instruction codes or logical mapping tables for implementing the reciprocating compressor valve fault diagnosis method described in any one of Embodiments 1 to 8 of this application.

[0170] Furthermore, the processor executes the computer program to enable the electronic device to implement the reciprocating compressor valve fault diagnosis method as described in any one of Embodiments 1 to 8. This process embodies the technical essence of this application as a combination of hardware and software. When the electronic device is powered on or receives an external trigger signal, the processor reads and parses the computer program from the memory, converts it into a low-level electrical signal control sequence, and then drives the connected data acquisition interface, communication interface, and human-machine interaction unit to work together. For example, when executing step S100 in Embodiment 1 to acquire multi-source status signals, the processor controls the ADC (Analog-to-Digital Converter) chip to synchronously read the analog data from the accelerometer, acoustic emission sensor, and temperature sensor at a preset sampling rate by executing the corresponding driver program; when executing step S300 to extract impact energy features, the processor calls the FFT algorithm library or STFT transformation routine stored in the memory to perform mathematical operations on the signal segments in the buffer area; when executing step S400 to determine faults, the processor compares the feature value with the baseline value according to the program logic and generates corresponding control instructions or alarm messages according to the determination result.

[0171] It should be understood that although this embodiment describes a single device as an example, in other embodiments, the electronic device may also be a node in a distributed computing cluster, or a virtual computing entity composed of multiple physical devices interconnected by a network, as long as the whole can execute the above method flow.

[0172] Example 10

[0173] Having described the hardware architecture of the electronic device executing the diagnostic method in Embodiment 9 above, considering that in actual software product delivery, system upgrades, and embedded development processes, the program code carrying the diagnostic algorithm logic is often circulated and distributed as an independent commodity or technology carrier, it would be impossible to fully cover diverse software deployment forms, from firmware burning to cloud image downloads, if only the running electronic device is protected while ignoring its underlying software storage carrier. Therefore, this embodiment further provides a computer storage medium.

[0174] Specifically, the computer storage medium provided in this embodiment carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement the reciprocating compressor valve fault diagnosis method as described in any one of the foregoing embodiments 1 to 8.

[0175] Here, "carrier" refers to fixing or temporarily storing binary code representing the logical instructions of a diagnostic method on a medium in a physical or electromagnetic way, so that it has the ability to be read and restored into executable instructions by electronic devices.

[0176] It should be understood that the computer storage media described in this embodiment encompasses any form of tangible or non-transitory computer-readable media known in the art or that may emerge in the future. For example, in terms of non-volatile storage media, it may include read-only memory (ROM), programmable read-only memory (PROM), electrically erasable programmable read-only memory (EEPROM), flash memory, solid-state drive (SSD), optical disc (CD-ROM, DVD), disk (floppy disk, hard disk), and portable USB storage devices, etc. Such media are suitable for long-term storage, offline distribution, and firmware embedding of diagnostic programs. In terms of volatile storage media, it may include random access memory (RAM), dynamic random access memory (DRAM), etc. Such media are suitable for high-speed loading and temporary caching of diagnostic program code during the operation of electronic devices. Regardless of the specific physical form it adopts, as long as it can stably store the instruction set required to implement the diagnostic method of this application and trigger the corresponding diagnostic process when called by an electronic device, it falls within the protection scope of this application.

[0177] From the microscopic mechanism of functional implementation, after the aforementioned storage medium establishes a data connection with the electronic device (such as the processor and memory combination described in Example 9), the electronic device reads the computer program carried in the medium through a bus interface or communication protocol. This program code contains all the core algorithm steps, including multi-source state signal acquisition, signal segmentation based on the operating cycle, extraction of preset high-frequency band impact energy characteristics and frequency domain features, and dual-logic branch discrimination for valve leakage and valve plate breakage. When the processor executes these instructions, it transforms the abstract code logic into precise scheduling of underlying hardware resources. For example, it controls the data acquisition card to synchronously read sensor signals at a preset sampling rate, calls the mathematical operation unit to perform short-time Fourier transform, and compares the feature values ​​in the register with the baseline value, thereby reproducing the complete reciprocating compressor valve fault diagnosis process in the physical world. This closed-loop technology of "media storage - device reading - function reproduction" ensures the consistency and reliability of the diagnostic method when migrating between different hardware platforms.

[0178] Example 11

[0179] Following the description of the physical storage medium carrying the diagnostic program in Embodiment 10, and considering the rapid development of cloud computing, the Internet of Things, and the Industrial Internet, the delivery and usage models of software products are undergoing a profound transformation from traditional physical media distribution to networked and service-oriented approaches. In many practical application scenarios, the fault diagnosis algorithm for reciprocating compressor valves no longer relies on tangible carriers such as optical discs or USB drives for transmission, but is deployed on terminal devices in the form of pure data streams via network download, online updates, or API calls. To adapt to new software delivery models such as SaaS (Software as a Service), cloud-native technologies, and OTA (Over-the-Air) technology, this embodiment provides a computer program product.

[0180] Specifically, the computer program product provided in this embodiment includes computer-readable instructions that, when executed on an electronic device, cause the electronic device to implement the reciprocating compressor valve fault diagnosis method as described in any one of embodiments 1 to 8 above.

[0181] Unlike the computer storage medium described in Embodiment 10, which focuses on the static storage and physical carrying of data, the computer program product in this embodiment focuses more on the functional attributes of the instruction set itself and its dynamic execution process during runtime. Here, "computer-readable instructions" refers to a series of code sequences or logical opcodes that can be recognized, parsed, and executed by the processor of an electronic device. Specifically, these instructions can be source code, object code, bytecode, script files, or encapsulated API (Application Programming Interface) call packages. Regardless of the syntax format used or the protocol used for transmission, as long as the instructions, when loaded and run by an electronic device, can drive hardware resources to perform core diagnostic steps such as acquiring multi-source state signals, segmenting based on the operating cycle, extracting high-frequency impact energy characteristics and frequency domain characteristics, and performing dual-logic discrimination for leakage or fracture, they fall within the protection scope of this embodiment.

[0182] In practical applications, the delivery and deployment of this computer program product offer high flexibility. For example, in a typical cloud-edge collaborative scenario, the product can be hosted on a cloud server. When the compressor monitoring terminal at the edge initiates an update request, the cloud pushes a computer-readable instruction stream containing the latest diagnostic algorithms to the terminal device via HTTPS (HyperText Transfer Protocol Secure) or MQTT (Message Queuing Telemetry Transport) protocols. Upon receiving the request, the terminal device loads it into its memory and applies it immediately, enabling remote iterative optimization of the diagnostic strategy without replacing any hardware or storage media. As another example, in a SaaS service model, the product can be encapsulated as a standard RESTful API or SDK (Software Development Kit). Third-party system integrators can simply call the interface or integrate the SDK into their own monitoring platforms to equip them with professional reciprocating compressor valve fault diagnosis capabilities, without needing to concern themselves with the specific implementation details of the underlying algorithms. For example, for embedded diagnostic devices already deployed in the field, maintenance personnel can upload firmware upgrade packages containing the computer program product via a local maintenance port or wireless network, enabling the device to complete function enhancements or defect repairs without disassembling it. It should be understood that the delivery methods listed above are merely preferred examples for explaining the principles of this application and are not intended to limit the application. Any method of transmission, distribution, or authorization that enables computer-readable instructions to be operational on an electronic device falls within the scope of this embodiment.

[0183] From a microscopic perspective of technical implementation, when this computer program product is run by an electronic device, its contained computer-readable instructions actually constitute a precise set of hardware control logic. During the execution of these instructions, the processor sequentially activates underlying hardware resources such as the data acquisition interface, digital signal processing unit, memory read / write controller, and communication module according to a preset timing and data flow. For example, one piece of code in the instruction stream might be specifically used to configure the ADC chip to synchronously acquire acceleration and acoustic emission signals at a 50kHz sampling rate; another piece of code might call DSP library functions to perform short-time Fourier transforms on segments of the signal in the buffer to extract energy in the 2kHz-8kHz frequency band; and yet another piece of code might be responsible for comparing feature values ​​with baseline values ​​and generating alarm messages based on the results. It is this orderly execution of a series of instructions that transforms the abstract diagnostic method logic into observable device behavior and diagnostic results in the physical world. Therefore, the essence of protection in this embodiment is the set of instructions itself that endows general-purpose electronic devices with specific fault diagnosis capabilities, rather than merely the physical carrier storing this set of instructions.

[0184] Following the detailed explanations of the methods, device architecture, and various carrier forms for diagnosing reciprocating compressor valve faults in Examples 1 to 11, this example constructs a high-fidelity simulation test platform based on the process parameters of the reciprocating raw material gas compressor in a thousand-ton skid-mounted green ammonia synthesis unit, and conducts a system performance comparison experiment. This example, as an independent application verification step, fully reproduces the operational performance of the aforementioned technical solutions under typical fault scenarios, and demonstrates this through horizontal comparison with various traditional single-parameter monitoring methods.

[0185] Specifically, the simulation object of this verification experiment is a two-stage reciprocating compressor with a crankshaft speed set at 370 rpm, corresponding to an operating cycle of 0.162 s, an intake pressure of 1.6 MPa, an exhaust pressure of 8.2 MPa, and a medium of 75% hydrogen and 25% nitrogen. The cylinder diameter is 120 mm, and the piston stroke is 150 mm. Regarding sensor configuration, the scheme described in Example 4 was strictly followed: an accelerometer with a sensitivity of 100 mV / g and a frequency response of 0.5-10 kHz, and an acoustic emission sensor with a center frequency of 100 kHz and a bandwidth of 50-200 kHz were installed on the cylinder valve cover; a dynamic pressure sensor with a range of 0-15 MPa and a response frequency of 50 kHz was installed on the intake and exhaust pipes; and a Pt100 thermal resistor was attached to the valve cover surface for temperature monitoring. The data acquisition unit simultaneously acquired the above four signals at a sampling rate of 50 kHz and performed segmented processing based on the crankshaft speed determined in real time according to the method described in Example 5. The preset reference baseline value is based on the strategies described in Examples 2 and 7, taking the statistical average of the impact energy characteristics in the 2kHz-8kHz frequency band during the 72 hours prior to the compressor's healthy operation. On this basis, different degrees of valve leakage and valve plate breakage faults were simulated, and the diagnostic performance of the method in this application was comprehensively compared with three typical comparative methods.

[0186] First, to clarify the fundamental impact of hardware configuration differences on diagnostic capabilities, Table 1 shows a comparison of the sensor configuration and key algorithm parameters between the embodiments of this application and various comparative examples. As shown in Table 1, the embodiments of this application are equipped with an accelerometer, an acoustic emission sensor, a dynamic pressure sensor, and a temperature sensor, and use the 2-8kHz frequency band for impact energy characteristic analysis, combined with 1 / 2 and 1 / 3 harmonic criteria. In contrast, Comparative Example 1 only uses an accelerometer, lacking acoustic emission and dynamic pressure information; Comparative Example 2 relies solely on a temperature sensor, completely lacking the ability to perceive the mechanical impact state; Comparative Example 3, although using an acoustic emission sensor, does not integrate acceleration vibration signals, and the analysis frequency band is limited to the 100-200kHz ultrasonic band, without employing subharmonic criteria. This configuration difference directly determines the dimension of physical information that each method can acquire: This application, through the complementarity of multi-source heterogeneous signals, utilizes the sensitivity of acceleration signals to low- and mid-frequency mechanical vibrations, leverages the advantage of acoustic emission signals in capturing high-frequency stress waves, and combines the thermodynamic steady-state indication function of temperature signals, thereby constructing a comprehensive fault characterization system. The comparative examples, due to their single signal source, inevitably have information blind spots. For example, comparative example 1 cannot detect the weak ultrasonic turbulence at the initial stage of leakage, comparative example 2 cannot capture transient mechanical impact events, and comparative example 3 is susceptible to ultrasonic noise interference from other components due to the lack of auxiliary verification with vibration signals. This hardware-level difference is the physical root of all subsequent performance gaps.

[0187] Table 1 Comparison of sensor configurations and parameters between the embodiment and the comparative example

[0188]

[0189] Secondly, regarding the most common and difficult-to-detect minor leak scenario in valve failures, Table 2 compares in detail the diagnostic performance of Embodiment 1 of this application and various comparative examples under a minor leak (equivalent diameter 0.5 mm) in the intake valve. Data shows that, with the fault injection time at 30 seconds, the embodiment of this application issues an accurate alarm at 35 seconds, with an alarm delay of only 5 seconds. Furthermore, the diagnostic accuracy reaches 100% in the initial leak stage (first 10 seconds) and the steady-state phase, with zero false alarms during healthy operation. In contrast, the comparative examples show that Comparative Example 1 (single vibration) suffers from extremely low signal-to-noise ratio because the high-frequency impact energy characteristics generated in the early stage of leakage are overwhelmed by the strong low-frequency mechanical vibration of the compressor body. The alarm is only triggered 45 seconds after the fault injection, when the leakage intensifies and the energy amplitude barely exceeds the threshold. The delay is as high as 45 seconds, which is 40 seconds slower than the present application. Moreover, the accuracy in the early stage of leakage is almost zero. Comparative Example 2 (single temperature) is limited by the huge thermal inertia caused by the heat capacity of the valve cover metal. It takes a long time for the temperature to rise by 15°C. The alarm delay is as long as 180 seconds, which is nearly 3 minutes slower than the present application. It completely loses the significance of early warning. Although Comparative Example 3 (single acoustic emission) is theoretically sensitive to leakage, it has 3 false alarms during the healthy operation period due to the lack of collaborative filtering of acceleration signals and the constraint of fractional harmonic characteristics. Even with slight leakage, it still takes 35 seconds to accumulate enough energy to trigger the threshold, which is 30 seconds slower than the present application. The core reason why this application can achieve a 5-second rapid warning with zero false alarms is that the 2kHz-8kHz frequency band selection described in Example 2 effectively avoids strong background noise below 1kHz, significantly improving the signal-to-noise ratio. Simultaneously, the 1 / 2 and 1 / 3 harmonic frequencies, as unique nonlinear fingerprints generated by the coupling of leaking airflow vortex shedding and valve flutter, fundamentally eliminate interference from non-fault factors, enabling the system to lock onto faults through spectral structure anomalies before the energy amplitude significantly exceeds the limit. This dual verification mechanism of "energy + spectrum" is a technical level that single-parameter methods cannot achieve.

[0190] Table 2. Comparison of diagnostic performance between the examples and comparative examples (minor leakage faults)

[0191]

[0192] Furthermore, regarding the highly hazardous yet subtly characterized valve plate fracture fault, Table 3 compares the diagnostic performance of Embodiment 3 of this application with that of various comparative examples. The results show that after the valve plate fracture fault was detected at 70 seconds, this application accurately issued an emergency alarm at 78 seconds, with a delay of only 8 seconds, and a fault identification accuracy of 100%, without any misjudgments. In contrast, in Comparative Example 1 (single vibration), the mechanical impact disappeared after the valve plate fracture, and the signal manifested as energy loss rather than abnormal enhancement. The system misjudged this as normal operation or background noise fluctuation, and no alarm was triggered, resulting in a fatal missed detection. In Comparative Example 2 (single temperature), although a temperature drop exceeding 15°C was detected and an alarm was triggered at 78 seconds, the method could not distinguish between fracture and leakage because leakage faults can also cause temperature changes (although the direction is usually opposite, it can be confused under complex operating conditions), leading to uncertainty in maintenance decisions. Comparative Example 3 (single acoustic emission) also failed to identify fracture due to the disappearance of the signal. This application's ability to accurately identify fracture without confusion with other faults is attributed to the joint criterion of impact energy characteristic attenuation and continuous temperature drop constructed in Embodiment 3. On the one hand, by utilizing the mechanical characteristic of the impact energy suddenly dropping to less than 0.5 times the preset reference baseline value, the instant the valve plate impact disappeared was accurately captured. On the other hand, by utilizing the thermodynamic characteristic of a temperature drop exceeding 15°C for more than two consecutive cycles, the continuous thermal equilibrium disruption caused by the failure of the compression chamber seal was confirmed. The strict alignment of these two conditions on the time axis and the logical "AND" relationship completely decoupled the two failure modes of fracture and leakage, filling the gap in the industry's field of online diagnosis of valve plate fracture.

[0193] Table 3. Comparison of diagnostic performance between the examples and comparative examples (valve plate fracture failure)

[0194]

[0195] Finally, to verify the adaptability and robustness of the method in this application under different fault severity levels, Table 4 shows a comparison of the diagnostic performance of the method described in Example 1 for different leakage equivalent diameters (0.3 mm, 0.5 mm, 0.8 mm, and 1.2 mm). The data shows that as the leakage equivalent diameter increases from 0.3 mm to 1.2 mm, the time required for the impact energy characteristic to reach three times the preset reference baseline value gradually decreases from 12 seconds to 1.5 seconds, the time for the subharmonic characteristic to appear also decreases from 15 seconds to 1.5 seconds, and the corresponding alarm delay is optimized from 15 seconds to 1.5 seconds, while the diagnostic accuracy remains 100% across all test levels. This gradient change fully demonstrates the high correlation between the feature indicators extracted in this application and the physical evolution of the fault: the more severe the leakage, the greater the intensity of the high-speed airflow turbulence, the stronger the induced high-frequency impact energy characteristics and nonlinear modulation effect, and the faster the characteristics exceed the threshold. More importantly, even with a leakage as weak as 0.3 mm, the system can still reliably detect it within 15 seconds. This further confirms the outstanding contribution of 2kHz-8kHz frequency band analysis and fractional harmonic feature extraction to improving the signal-to-noise ratio. If traditional full-band analysis or methods without spectral constraints were used, such weak signals would have been submerged in the noise floor long ago.

[0196] Table 4. Comparison of diagnostic performance under different leakage levels (Method of Example 1)

[0197]

[0198] By constructing a high-fidelity simulation environment and conducting multi-dimensional comparative experiments, quantitative data were used to intuitively demonstrate the significant advancements of this application in alarm delay, diagnostic accuracy, and fault mode differentiation capabilities. Specifically, compared to single vibration monitoring, this application advances the alarm time by 40 seconds under slight leakage conditions; by 175 seconds compared to single temperature monitoring; and by 30 seconds compared to single acoustic emission monitoring, while eliminating false alarms. In valve plate fracture diagnosis, it completely solves the industry challenges of the inability to identify faults from single vibrations and the inability to characterize faults from single temperatures, achieving accurate alarms within 8 seconds and mode decoupling. These data not only strongly support the inventive claims of this application but also provide a reliable performance benchmark for industrial field deployment, demonstrating that this technical solution can bring substantial safety and economic benefits to early warning and precise maintenance of reciprocating compressor valve faults.

[0199] It should also be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. In addition, in the device embodiment drawings provided in this application, the connection relationship between modules indicates that they have a communication connection, which can be implemented as one or more communication buses or signal lines.

[0200] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware, or it can be implemented by special-purpose hardware including application-specific integrated circuits, special-purpose CPUs, special-purpose memory, special-purpose components, etc. Generally, any function performed by a computer program can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be diverse, such as analog circuits, digital circuits, or special-purpose circuits. However, for this application, software program implementation is more often the preferred implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium, such as a computer floppy disk, USB flash drive, mobile hard disk, ROM, RAM, magnetic disk, or optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, training equipment, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0201] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product.

[0202] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, training device, or data center to another website, computer, training device, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a training device or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)).

Claims

1. A method for diagnosing valve faults in a reciprocating compressor, characterized in that, include: Acquire multi-source state signals, which include a first type of signal reflecting the mechanical impact state of the valve and a second type of signal reflecting the thermodynamic state of the valve. Based on the operating cycle of the reciprocating compressor, the multi-source status signal is divided into multiple signal segments; For each signal segment, the impact energy characteristics within a preset high-frequency band are extracted, and the frequency domain characteristics are also extracted. Based on the energy relationship between the impact energy characteristics and the preset reference baseline value, and the frequency domain characteristics, a valve leakage detection is performed; or, based on the energy relationship between the impact energy characteristics and the preset reference baseline value, and the temperature change characteristics of the second type of signal, a valve plate breakage detection is performed.

2. The method for diagnosing valve faults in a reciprocating compressor according to claim 1, characterized in that, The process of determining valve leakage based on the energy relationship between the impact energy characteristics and the preset reference baseline value, and the frequency domain characteristics, includes: When the impact energy characteristic is greater than a first preset multiple of the preset reference baseline value, and the frequency domain characteristic contains a preset fractional harmonic component, it is determined to be a valve leak. The preset fractional harmonic components include either a half harmonic component or a one-third harmonic component.

3. The method for diagnosing valve faults in a reciprocating compressor according to claim 1, characterized in that, The valve plate fracture detection based on the energy relationship between the impact energy characteristics and the preset reference baseline value, and the temperature change characteristics of the second type of signal, includes: When the impact energy characteristic is less than a second preset ratio of the preset reference baseline value, and the temperature drop value indicated by the second type of signal exceeds a preset temperature drop threshold within a consecutive preset number of operating cycles, it is determined that the valve plate is broken.

4. The method for diagnosing valve faults in a reciprocating compressor according to claim 1, characterized in that, The preset reference baseline value is determined based on the statistical average value of the impact energy characteristics of the reciprocating compressor during a preset period of time before it is in a healthy operating state; The extraction of impact energy characteristics within a preset high-frequency band includes: The signal segments are analyzed using short-time Fourier transform to extract the impulse energy in the 2kHz to 8kHz frequency band as the impulse energy feature.

5. The method for diagnosing valve faults in a reciprocating compressor according to claim 1, characterized in that, The first type of signal includes acceleration vibration signal and acoustic emission signal, and the second type of signal includes valve cover surface temperature signal; The acquisition of multi-source state signals includes: The acceleration vibration signal, the acoustic emission signal, and the valve cover surface temperature signal are collected synchronously at a preset sampling rate.

6. The method for diagnosing valve faults in a reciprocating compressor according to claim 1, characterized in that, Based on the operating cycle of the reciprocating compressor, the multi-source status signal is divided into multiple signal segments, including: The operating cycle is determined based on the crankshaft speed of the reciprocating compressor; Using the operating cycle as a time window, the multi-source state signal is segmented to obtain the multiple signal segments.

7. The method for diagnosing valve faults in a reciprocating compressor according to claim 1, characterized in that, Also includes: In response to a determination of valve leakage or valve plate breakage, a diagnostic result is generated that includes the fault type, time of occurrence, and severity. The diagnostic results are sent to the monitoring system in real time and stored in the historical database.

8. A fault diagnosis device for a reciprocating compressor valve, characterized in that, include: The signal acquisition module is configured to acquire multi-source state signals, which include a first type of signal reflecting the mechanical impact state of the valve and a second type of signal reflecting the thermodynamic state of the valve. The cycle segmentation module is configured to divide the multi-source status signal into multiple signal segments based on the operating cycle of the reciprocating compressor; The feature extraction module is configured to extract impact energy features within a preset high-frequency band and extract frequency domain features for each of the signal segments. The fault detection module is configured to perform valve leakage detection based on the energy relationship between the impact energy characteristics and the preset reference baseline value, as well as the frequency domain characteristics; or, to perform valve plate breakage detection based on the energy relationship between the impact energy characteristics and the preset reference baseline value, as well as the temperature change characteristics of the second type of signal.

9. An electronic device, characterized in that, It includes at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program so that the electronic device can implement the reciprocating compressor valve fault diagnosis method as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that, The storage medium carries one or more computer programs that, when executed by an electronic device, enable the electronic device to implement the reciprocating compressor valve fault diagnosis method as described in any one of claims 1 to 7.

11. A computer program product, characterized in that, It includes computer-readable instructions that, when executed on an electronic device, cause the electronic device to implement the reciprocating compressor valve fault diagnosis method as described in any one of claims 1 to 7.

Images