Large reciprocating compressor continuous stepless regulation system and method based on intelligent control

Abstract

The invention provides a large reciprocating compressor continuous stepless regulation system and method based on intelligent control, and the method comprises the steps: building a continuous stepless regulation model according to a load change trend and process fluctuation characteristics, and determining an initial parameter set of a regulation instruction through a preset load range regulation threshold value and a dynamic response speed requirement; according to the non-equilibrium stress adjustment scheme, dynamically adjusting operation parameters of a compressor execution mechanism, monitoring the abrasion state of key components through a real-time feedback control technology, and determining operation data of the components after stress equilibrium; and the operation data of the parts after stress balance are obtained, the operation parameters of the compressor are subjected to secondary calibration through an intelligent control system in combination with an energy efficiency improvement target, and the optimized optimal pressure ratio state parameters are obtained.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of information technology, and in particular to a large reciprocating compressor continuous stepless regulation system and method based on intelligent control BACKGROUND

[0002] Problem background: Compressor technology, as a core field in industrial production, is directly related to energy efficiency, equipment safety and production continuity, and its importance is self-evident. With energy saving and consumption reduction becoming the key direction of national policy and enterprise development, the gas regulation technology of large reciprocating compressors has attracted much attention due to its energy consumption performance and operational stability. Traditional gas regulation methods, such as bypass return, speed regulation and variable clearance chamber regulation, while meeting the process requirements to some extent, have generally been flawed: bypass return results in 20% to 30% of the compression power being wasted, speed regulation is economically inefficient and slow to respond, and variable clearance chamber regulation is less reliable due to its complex structure. These limitations make it difficult for existing solutions to meet the requirements of modern industry for efficient, accurate and reliable operation, and energy efficiency and equipment wear and tear problems are increasingly prominent. In this context, the core challenges facing large reciprocating compressors are concentrated on the balance between the continuity of gas regulation, response speed and equipment life. Specifically, existing technologies cannot achieve continuous stepless regulation of load, resulting in the operation of the compressor deviating from the optimal compression ratio state, limiting energy utilization; at the same time, the dynamic response speed in the regulation process is slow, making it difficult to respond to rapidly changing process requirements; in addition, the one-sided stress problem caused by long-term high-load operation or improper regulation methods further exacerbates the wear and tear of key components, increasing safety hazards. These technical factors have not been effectively addressed, resulting in the unique problems of high energy consumption, inadequate control and increased operational risks. Therefore, how to achieve continuous stepless regulation of gas volume within a load range of 30% to 100%, while improving the dynamic response capability of milliseconds and reducing the non-uniform stress of key components, has become a key problem in improving the energy efficiency and operational reliability of compressors. The solution to this problem not only needs to break through the limitations of traditional regulation methods, but also needs to seek innovation in intelligent control and hardware design to meet the urgent needs of industrial production for energy saving, safety and efficiency. SUMMARY

[0003] The present application provides a large reciprocating compressor continuous stepless regulation system and method based on intelligent control, mainly comprising: Obtain real-time operation data of the compressor, collect load range regulation parameters and process demand change data through sensors, and use data preprocessing technology to filter and normalize the collected signals to obtain load change trend and process fluctuation characteristics; According to the load change trend and the process fluctuation characteristics, a continuous stepless regulation model is constructed, and through preset load range regulation threshold and dynamic response speed requirement, initial parameter set of the regulation instruction is determined; For the initial parameter set, an intelligent control algorithm is used to optimize the continuous stepless regulation model in real time, target pressure ratio values are obtained from the optimal pressure ratio state database, and it is judged whether the regulation instruction meets the millisecond level response condition or not; If the regulation instruction meets the millisecond level response condition, an execution signal is generated through a hardware design innovation module, stress distribution data of key components are obtained from a pre-established hardware response mapping table, and a non-uniform stress adjustment scheme is obtained; According to the non-uniform stress adjustment scheme, the operation parameters of the compressor actuator are dynamically adjusted, the wear state of the key components is monitored through real-time feedback control technology, and the component operation data after stress balance is determined; The component operation data after stress balance is obtained, the energy efficiency improvement target is combined, the operation parameters of the compressor are calibrated again through the intelligent control system, and the optimal pressure ratio state parameters after optimization are obtained; For the optimal pressure ratio state parameters after optimization, a dynamic response speed evaluation algorithm is used to extract response delay indicators from real-time data of process demand changes, and it is judged whether the regulation process meets the standard of strong operation reliability or not; If the regulation process meets the standard of strong operation reliability, the control logic of the actuator is updated through the hardware design innovation module, the life prediction value is extracted from the key component wear monitoring data, and the stability parameters of the long-term operation of the compressor are obtained; According to the matching degree of the stability parameters and the process demand changes, the intelligent control algorithm is used to iteratively update the continuous stepless regulation model, the energy efficiency improvement effect is extracted from the real-time operation data, and the final regulation optimization scheme is determined.

[0004] The technical scheme provided by the embodiment of the application can include the following beneficial effects: The application discloses a large reciprocating compressor continuous stepless regulation system and method based on intelligent control, aiming at the problems of high energy consumption, slow response and low reliability of traditional regulation modes, and proposes a solution based on intelligent control and hardware innovation. The application realizes accurate regulation in the load range of 30% to 100% by collecting compressor operation data and process demand changes in real time, building a continuous stepless regulation model, and optimizing regulation parameters by using intelligent algorithms. The hardware design module is innovatively introduced to generate millisecond-level response execution signals, and through the non-uniform stress adjustment scheme, the wear of key components is reduced. The application also integrates energy efficiency optimization, life prediction and other functions, and continuously improves the operation efficiency and reliability of the compressor by iteratively updating the regulation model. The method significantly improves the energy utilization rate of the compressor, enhances the adaptability to rapidly changing process demands, and prolongs the service life of the equipment, providing strong support for energy saving and consumption reduction and safe and stable operation of industrial production. BRIEF DESCRIPTION OF DRAWINGS

[0005] Fig. 1 A flowchart of a large reciprocating compressor continuous stepless regulation system and method based on intelligent control of the application.

[0006] Fig. 2 A schematic diagram of a large reciprocating compressor continuous stepless regulation system and method based on intelligent control of the application.

[0007] Fig. 3 Another schematic diagram of a large reciprocating compressor continuous stepless regulation system and method based on intelligent control of the application. DETAILED DESCRIPTION

[0008] In order to further understand the content of the application, the application will be described in detail in combination with the drawings and examples. It can be understood that the specific examples described herein are only used to explain the related application, and not to limit the application. In addition, it should be noted that only the parts related to the application are shown in the drawings for convenience of description.

[0009] As Figs. 1-3 , the large reciprocating compressor continuous stepless regulation system and method based on intelligent control of the embodiment can specifically include: S101, acquiring real-time operation data of the compressor, collecting load range regulation parameters and process demand change data through sensors, and adopting data preprocessing technology to filter and normalize the collected signals to obtain load change trend and process fluctuation characteristics.

[0010] Real-time signal data is acquired from the compressor operating environment by sensors to obtain an original data set. The original data set is denoised by signal filtering technology to obtain filtered data. The filtered data is standardized by a normalization processing method to obtain standardized data. A load change feature is extracted from the standardized data. If the load change feature exceeds a preset threshold range, it is marked as abnormal fluctuation to obtain the load change feature. According to the load change feature, the fluctuation situation related to the process requirement is analyzed. If the fluctuation frequency or amplitude exceeds the process requirement range, it is recorded as a process deviation to determine the process fluctuation feature. Through joint analysis of the process fluctuation feature and the load change feature, a support vector machine algorithm is used to classify and process the data to determine the potential operation risk category. The operation risk category is compared and analyzed in combination with historical operation data to determine the final risk assessment result.

[0011] For example, in the data processing and risk assessment of the compressor operating environment, the acquisition of signal data is the first step. Sensors collect vibration, temperature, and pressure signals of the compressor in real time to form an original data set. Suppose in a certain operation, the frequency range of the vibration signal is between 10 and 500 Hz, but it is mixed with environmental noise interference such as mechanical impact or wind noise.

[0012] In one possible implementation, for the denoising of the original data set, a low-pass filtering technique can be used to filter out high-frequency noise above 300 Hz and retain low-frequency signals related to compressor operation. After filtering, the vibration signal waveform is smoother, and noise interference is significantly reduced, providing a more reliable data basis for subsequent analysis. The beneficial effect of this is to improve data quality and avoid noise interference in feature extraction.

[0013] For example, the standardization of filtered data can be achieved by normalizing the amplitude of the vibration signal to the range of 0 to 1. Suppose the amplitude of the original vibration signal is between -5 and 5, and after normalization, the data can be mapped to the standard interval, facilitating subsequent feature extraction and model training. The beneficial effect is to eliminate dimensional differences and ensure the comparability of different types of data.

[0014] In one possible implementation, when extracting the load change feature from the standardized data, attention can be paid to the peak value change rate of the vibration signal. Suppose the peak value change rate is within 0.1 under normal operation, but reaches 0.3 in a certain detection, exceeding the preset threshold of 0.2, it is marked as abnormal fluctuation. This feature reflects the risk of compressor overload or component wear and helps to provide early warning.

[0015] For example, when analyzing the process fluctuation characteristics, if the process requirements require the vibration frequency to be stable at about 50 Hz, and the amplitude fluctuation is not more than 0.05, and the actual detection frequency drifts to 60 Hz, and the amplitude fluctuation reaches 0.1, it is recorded as a process deviation. The benefit of this analysis is to find the deviation of the running state from the production requirements in time, and to ensure product quality.

[0016] In one possible implementation, through joint analysis of load change characteristics and process fluctuation characteristics, support vector machine algorithm is used for data classification. Assuming that the data is divided into three categories: normal, slight risk and serious risk, combined with the high and low of the characteristic value, the model can judge that the current running belongs to the slight risk category. The advantage is that the classification accuracy is high, and potential problems can be effectively identified.

[0017] For example, for the evaluation of the running risk category, combined with the comparison and analysis of historical running data, it is assumed that similar characteristics in historical data have caused bearing overheating, and the current risk evaluation result may point to bearing fault hidden danger. Through this comparison, the risk assessment is more accurate, guiding the maintenance decision and reducing the downtime loss.

[0018] In one possible implementation, the comprehensive application of the above method not only improves the monitoring ability of the compressor running state, but also realizes the whole chain management from data to risk prediction through feature extraction and classification analysis. Its beneficial effect is to reduce the equipment failure rate, prolong the service life, optimize the process flow, and ensure the stability of production.

[0019] S102, according to the load change trend and the process fluctuation characteristics, a continuous stepless regulation model is constructed, and through the preset load range regulation threshold and dynamic response speed requirement, the initial parameter set of the regulation instruction is determined.

[0020] The load change data and process fluctuation data are acquired from a real-time monitoring system, and a fluctuation feature data set is constructed. For the fluctuation feature data set, a pre-established continuous stepless regulation model is used for feature mapping to determine the correlation parameters of load change and process fluctuation, and an input feature set is formed. If the load change in the input feature set exceeds the preset load range threshold, a regulation threshold detection module is triggered, historical data is compared and analyzed to determine whether the dynamic response condition is met, and a response trigger signal is obtained. According to the response trigger signal, a regulation instruction generation module is called, and based on the preset speed demand and response characteristics, an initial parameter set is calculated to obtain a preliminary configuration scheme of the regulation instruction. The support vector machine algorithm is used to optimize and adjust the parameter set in the preliminary configuration scheme, and the real-time load range data and dynamic response demand are combined to determine the final regulation instruction parameter set. If the final regulation instruction parameter set meets the preset response characteristic requirement, it is transmitted to the execution system for load regulation, and the fluctuation data in the regulation process is recorded to obtain execution feedback information. Through the execution feedback information, the input feature set of the continuous stepless regulation model is updated, the model parameter correction is performed for the load change and process fluctuation characteristics, and an updated regulation model is obtained.

[0021] For example, in the business field of compressor operation monitoring, acquiring load change data and process fluctuation data from a real-time monitoring system is a core link. For load change data, the pressure and power signals during compressor operation can be collected in real time through sensors. Assuming that the pressure value reaches 120% of the rated value at a certain time, it exceeds the preset threshold range, and the system will record it as an abnormal signal. Process fluctuation data may exhibit periodic changes in temperature or flow rate, such as a temperature fluctuation of more than 5 degrees Celsius within 10 minutes, which is also marked as a potential deviation. Through these data, a fluctuation feature data set is constructed to lay the foundation for subsequent analysis.

[0022] For example, for the feature mapping of the fluctuation feature data set and the continuous stepless regulation model, it can be understood as quantitatively processing the correlation between load change and process fluctuation. Assuming that when the load increases by 10%, the process temperature fluctuation frequency increases by 15%, the system will calculate the correlation parameters between the two through the model to form an input feature set. This mapping helps identify key influencing factors and provides data support for regulation.

[0023] For example, when triggering the regulation threshold detection module, if the load change exceeds the preset range, such as the power value exceeding 110% of the rated value, the system will compare and analyze the past 24 hours of historical data to determine whether dynamic response is needed. If similar situations in the historical data have caused equipment overheating, a response trigger signal will be generated. This approach can timely capture potential risks.

[0024] For example, during the call process of the adjustment instruction generation module, the system calculates an initial parameter set according to the speed requirement and response characteristics. Assuming that the speed requirement is to adjust the power by 5% per minute and the response characteristics require that the temperature fluctuation be controlled within 2 degrees Celsius, the system will preliminarily configure an adjustment scheme. This scheme lays the foundation for subsequent optimization.

[0025] For example, when optimizing the preliminarily configured scheme using a support vector machine algorithm, the system adjusts the parameter set in combination with real-time load range data, such as the current load being 90% of the rated value, and dynamic response requirements, to ensure that the final adjustment instruction parameter set meets the requirements. This optimization can improve the adjustment accuracy.

[0026] For example, after transmitting the final adjustment instruction parameter set to the execution system, the system records the fluctuation data during the adjustment process. Assuming that the temperature fluctuation after adjustment is reduced to within 1.5 degrees Celsius, the execution feedback information shows that the adjustment effect is good. This feedback helps to verify the effectiveness of the instruction.

[0027] For example, when updating the continuous stepless adjustment model through execution feedback information, the system corrects the model parameters according to the latest load changes and process fluctuation characteristics. For example, if the feedback shows that a certain parameter causes the response speed to be too slow, the model will adjust the related weights to improve the response efficiency of the next adjustment. This updating mechanism can continuously optimize system performance.

[0028] S103, for the initial parameter set, an intelligent control algorithm is used to optimize the continuous stepless adjustment model in real time, and the target pressure ratio value is obtained from the best pressure ratio state database to determine whether the adjustment instruction meets the millisecond-level response condition.

[0029] The running data of the current system is obtained from the preset initial parameter set, and the intelligent control method is used to preliminarily process the running data to obtain the basic parameter characteristics. For the basic parameter characteristics, a support vector machine algorithm is used to optimize the continuous stepless adjustment model in real time to determine the dynamic adjustment range of the model. According to the dynamic adjustment range, the corresponding target pressure ratio value is extracted from the preset pressure ratio state database to determine whether the target pressure ratio value matches the current running state. If the target pressure ratio value does not match the current running state, an adjustment instruction is generated through the intelligent control method to obtain adjusted running parameters. For the adjusted running parameters, it is detected whether the adjustment instruction meets the millisecond-level response requirement to obtain a response time evaluation result. If the response time evaluation result exceeds the preset threshold, the adjustment instruction is optimized again to determine the final control scheme. Through the final control scheme, the running state of the continuous stepless adjustment model is updated to complete the real-time optimization process.

[0030] For example, in the context of industrial load regulation, obtaining the current system's operational data from a pre-set initial parameter set can be understood as extracting key indicators such as power, current, or pressure from real-time monitoring devices. Assuming that a certain chemical production line has a power data of 500 kilowatts and a pressure value of 3.5 bars, these data serve as the basis for input, which is preliminarily processed through intelligent control methods to extract basic parameter features such as fluctuation frequency or trend. Based on these features, the intelligent control method may perform smoothing processing on the data to filter out short-term noise interference, ensuring the accuracy of subsequent analysis.

[0031] For example, when using a support vector machine algorithm to optimize the continuous stepless regulation model based on the basic parameter features, the model can be trained using historical data to identify potential correlations between load changes and process parameters. Assuming that in a certain optimization, the model identifies that the power fluctuation has a high correlation with the pressure change, and the dynamic adjustment range is set to within ±10% of the power. The determination of this range helps to accurately generate subsequent regulation instructions, avoiding excessive adjustment range that may cause system instability.

[0032] For example, when extracting the target pressure ratio value from the pressure ratio state database, assuming that the ideal pressure ratio value under the current operating state is 1.2, and the value matched in the database is 1.5, which is obviously not matched. At this time, the intelligent control method will generate a regulation instruction, which may be through adjusting the valve opening or motor speed to gradually approach the pressure ratio value of 1.2. The generation logic of this regulation instruction is based on the combination of real-time data and historical experience, ensuring that the system operates in the best state.

[0033] For example, when detecting whether the regulation instruction meets the millisecond-level response requirement, assuming that the response time of a certain regulation instruction is 120 milliseconds, and the preset threshold is 100 milliseconds, which is out of range. At this time, secondary optimization is needed, which may be through simplifying the instruction execution path or priority adjustment, and finally the response time is shortened to 90 milliseconds. This optimization process is crucial to ensure the rapid response of the system under high load, especially in chemical production, which can effectively avoid process deviations caused by delays.

[0034] For example, when updating the running state of the continuous stepless regulation model, the actual effect data after each regulation, such as power stabilizing at 505 kilowatts and pressure maintaining at 3.6 bars, is re-input into the model for parameter correction. This continuous updating mechanism enables the model to adapt to different operating conditions, gradually improving the regulation accuracy and stability. Through the above-mentioned multiple implementation manners, from data acquisition to model optimization, to instruction generation and response detection, a complete closed-loop control system is formed, which helps to improve the reliability and efficiency of system operation.

[0035] S104, if the adjustment instruction meets the millisecond response condition, generate an execution signal through the hardware design innovation module, obtain the stress distribution data of the key components from the pre-established hardware response mapping table, and obtain the non-uniform stress adjustment scheme.

[0036] The hardware innovation module detects the adjustment instruction in real time. If the detection result meets the preset response time condition, a signal generation process is triggered to obtain corresponding execution signal data. According to the execution signal data, the stress distribution information related to the key components is extracted from the preset response mapping table, and the initial state of the stress distribution information is determined. The stress distribution information is analyzed by using a data processing tool. If non-uniform force distribution characteristics are found, a preliminary adjustment strategy framework is generated. The hardware innovation module optimizes the preliminary adjustment strategy framework to obtain specific adjustment parameters for non-uniform force distribution, and obtains a refined adjustment strategy. According to the refined adjustment strategy, a stress adjustment instruction for the key components is generated, and it is judged whether the stress adjustment instruction conforms to the execution range of the hardware module. If the stress adjustment instruction conforms to the execution range, the stress adjustment instruction is issued through the hardware innovation module, execution feedback data is obtained, and the adjustment effect is determined. The stress distribution feedback data is analyzed again. If the feedback data still has non-uniform force characteristics, the hardware innovation module iteratively optimizes the adjustment strategy to obtain a final stress balance scheme.

[0037] For example, in the process of real-time detection of the adjustment instruction by the hardware innovation module, the embedded sensor and high-performance processor can be used to accurately monitor the response time of the instruction. Assuming that in a continuous stepless adjustment system, the preset response time condition is 5 milliseconds, the hardware module detects the time difference between the instruction issuing and the feedback by real-time collection, and finds that the actual response time is 4.8 milliseconds, which meets the condition and triggers the signal generation process immediately. This method can ensure the stability of the system in the high-frequency adjustment scene.

[0038] For example, for the link of extracting stress distribution information from response mapping table according to execution signal data, it can be understood that through the pre-constructed database, the signal data is one-to-one corresponding to the stress state of the key components. Assuming that the system detects that the stress peak of an execution signal corresponding to a component is 200 Newton, and the mapping table shows that the normal stress range of the component is 150 to 180 Newton, indicating that there is a non-uniform characteristic. This method is convenient for quickly locating the problem component and providing data support for subsequent adjustment.

[0039] For example, when analyzing the force distribution information with data processing tools, the force data can be presented in the form of a heat map through visualization tools, intuitively showing the non-equilibrium area. Assuming that the heat map shows that the left side of a key component is subjected to a force of 220 Newtons, and the right side is only subjected to a force of 130 Newtons, the system will generate a preliminary adjustment strategy framework based on this difference, prioritizing load reduction on the left side. This intuitive analysis helps to quickly develop targeted strategies.

[0040] For example, in the specific implementation of the hardware innovation module optimizing the preliminary adjustment strategy framework, the non-equilibrium force distribution can be parameterized adjusted through built-in intelligent algorithms. Assuming that the left side load reduction target is to reduce to 180 Newtons, the system will calculate the corresponding adjustment parameters, such as reducing the drive power by 20%, thereby forming a refined adjustment strategy. This optimization method can improve the accuracy of adjustment.

[0041] For example, when generating force adjustment instructions for key components and determining whether they meet the execution range, the performance boundary database of the hardware module can be used for verification. Assuming that the adjustment instruction requires a 25% reduction in drive power, and the maximum adjustment range of the hardware module is 30%, the instruction meets the execution range and can be issued. This verification mechanism ensures the executability of the instruction.

[0042] For example, in the secondary analysis of feedback data on the force distribution after execution, the effect can be evaluated by comparing the force data before and after adjustment. Assuming that after adjustment, the left side force is reduced to 185 Newtons, and the right side is increased to 145 Newtons, there is still a slight non-equilibrium, the system will iteratively optimize through the hardware innovation module, fine-tune the power parameter to reduce by 22%, and finally achieve force balance. This iterative method can continuously improve the system state.

[0043] For example, in the process of forming the final force balance scheme, historical data and real-time feedback can be combined to form a dynamic adjustment model. Assuming that the system records the force change trend of the past 10 adjustments, combined with the current feedback data, a long-term applicable balance scheme is generated. This method can improve the adaptability and stability of the system.

[0044] S105, according to the non-equilibrium force adjustment scheme, dynamically adjusting the operating parameters of the compressor actuator, and monitoring the wear state of the key components through real-time feedback control technology to determine the component operating data after force balance.

[0045] The running parameter data of the compressor actuator is acquired through real-time feedback technology, and the deviation distribution of the running parameter is obtained by analyzing the non-equilibrium stress condition. According to the deviation distribution, the running parameter is dynamically optimized by using an adjustment strategy to generate an optimized parameter configuration scheme, and the correction direction of the non-equilibrium stress is determined. If the correction direction deviates from the preset stress equilibrium target, the running parameter of the compressor actuator is fine-tuned through control technology to obtain adjusted running data. Through monitoring technology, real-time data acquisition is performed on the wear state of the key components, and the wear state of the key components is judged to be in a stress equilibrium state in combination with the adjusted running data. If the wear state data of the key components exceeds the preset threshold, the running parameter is recalculated by a dynamic optimization method to obtain a new parameter adjustment scheme. According to the new parameter adjustment scheme, the running state of the compressor actuator is continuously tracked by using real-time feedback technology to determine whether the final component data meets the stress equilibrium requirement. Through continuous analysis of the component data, the long-term wear state of the key components is predicted by using a support vector machine algorithm to obtain a potential risk distribution, and an optimization closed loop of the running parameter is completed.

[0046] For example, in the application of real-time feedback technology of compressor running parameters, the vibration frequency, temperature change and pressure value of the actuator can be collected in real time by a sensor device. Assuming that the vibration frequency of a certain compressor reaches 50 times per second during operation, which exceeds the normal range of 30 times per second, it indicates that there may be a non-equilibrium stress problem. For this data, the technician can find that the stress deviation of a specific bearing part exceeds 20% by analyzing the vibration distribution diagram, so as to determine the direction that needs to be adjusted. This way can quickly locate the problem source and provide data support for subsequent optimization.

[0047] For example, in the analysis of non-equilibrium stress condition and the determination of deviation distribution, the current parameters can be compared with the historical running data. Assuming that the pressure value of a certain key component is 5.0 megapascals under normal conditions, and the current monitoring value is 6.2 megapascals, the deviation exceeds 20%, which can be judged as non-equilibrium stress. By drawing a deviation distribution diagram, the technician can directly see the specific area of non-uniform stress, such as the left bearing pressure being too high and the right side being too low. This analysis helps to clarify the adjustment priority and ensure that the subsequent strategy is targeted.

[0048] For example, in the dynamic optimization of running parameters and the generation of configuration scheme, the speed and load distribution of the compressor can be adjusted according to the deviation distribution. Assuming that the speed is reduced from 3000 revolutions per minute to 2800 revolutions per minute, and the load is adjusted from 60% to 55%, the pressure of the left bearing can be effectively reduced. This optimization scheme can gradually approach the stress equilibrium target without affecting the overall operation efficiency, which helps to prolong the service life of the equipment.

[0049] For example, in the fine-tuning process of correcting the deviation of the target pressure, the intake volume of the compressor can be slightly adjusted through control technology. Assuming the target pressure is 5.0 MPa, and the current pressure is 5.3 MPa, the pressure value can be reduced by reducing the intake volume by 5%. The adjusted operating data shows that the pressure has dropped to 5.1 MPa, close to the target value. This fine-tuning method can avoid the instability caused by large adjustments and ensure smooth transition of the equipment.

[0050] For example, in real-time monitoring of the wear state of key components, the bearing temperature can be detected using infrared thermal imaging technology. Assuming the normal temperature is 40°C, and the current monitoring value is 48°C, which exceeds the threshold by 5°C, indicating that the wear is intensifying. By combining the adjusted operating data, it is determined whether the equilibrium state is reached, if not, further optimization is needed. This monitoring method can timely discover potential problems and reduce the risk of sudden failure.

[0051] For example, in the recalculation of operating parameters and the generation of new adjustment schemes, the lubricating oil supply can be adjusted according to the wear data. Assuming the original supply is 2 liters per hour, and the wear is intensifying, it can be increased to 2.5 liters per hour to reduce friction. After the implementation of the new scheme, the wear rate has decreased from 0.2 mm per month to 0.1 mm per month, with significant effect. This dynamic optimization can effectively respond to changes in equipment state.

[0052] For example, in the prediction of long-term wear state using support vector machine algorithm, based on historical data and current operating parameters, the wear trend of key components in the next three months can be predicted. Assuming the prediction result shows that the wear of a certain bearing may reach 0.5 mm, exceeding the safety threshold of 0.3 mm, a maintenance plan can be developed in advance. This prediction method helps to prevent problems from occurring and ensures long-term stable operation of the equipment.

[0053] S106, obtain the operating data of the components after force balance, combine the energy efficiency improvement target, and through the intelligent control system, the operating parameters of the compressor are calibrated again to obtain the optimized best pressure ratio state parameters.

[0054] Real-time data is obtained from the operation of compressor components through a sensor network to obtain operation state information of the components under different working conditions. According to the operation state information, an analysis is performed for a force balance state, a mechanical model established in advance is used to calculate force distribution of each key point to determine a preliminary evaluation result of force balance. If the force balance evaluation result deviates from a preset threshold value, an adjustment instruction is generated by an intelligent control system to dynamically calibrate the compressor parameters, and whether the adjusted force distribution meets the balance requirement is judged. According to the adjusted force distribution data, in combination with an energy efficiency improvement target, a support vector machine algorithm is used to optimize and calculate the compressor operation parameters to obtain a candidate parameter combination of the best compression ratio state. Through simulation operation test on the candidate parameter combination, energy efficiency data under a simulation environment is obtained to judge whether a preset energy efficiency standard is reached. If the simulation test result does not reach the preset standard, the compressor parameters are readjusted, and data feedback of the analysis state is combined to perform cyclic calculation until the best compression ratio state parameters meeting the energy efficiency improvement target are determined.

[0055] For example, in the process of obtaining compressor component operation data through a sensor network, a specific industrial scenario can be imagined in which multi-point pressure sensors and vibration sensors are deployed to collect data of compressor key components such as bearings and blades under different rotating speeds in real time. It is assumed that under a certain working condition, the sensor detects that the vibration frequency of the bearing part abnormally increases, and the data indicates that there may be a local force imbalance. In this case, the system compares the collected vibration value with historical normal operation data to preliminarily judge whether the force distribution deviates from the standard range.

[0056] For example, for the force balance state analysis and mechanical model calculation, a mechanical model constructed in advance can be used to simulate the stress distribution of each key point inside the compressor. It is assumed that the model shows that the force value of a blade area reaches 1200 Newton, while the average force value of other areas is only 800 Newton, and there is obviously an imbalance. Based on this, the system will mark this area as a key object of attention and generate a preliminary evaluation report to provide a basis for subsequent adjustment.

[0057] For example, when the intelligent control system generates an adjustment instruction and dynamically calibrates the parameters, it can be imagined that the system automatically reduces the operating speed of the related components in this area from 3000 revolutions per minute to 2500 revolutions per minute according to the evaluation result, and increases the cooling liquid flow to reduce the accumulation of thermal stress. After adjustment, the real-time feedback data of the sensor indicates that the force distribution tends to be balanced, and the stress value of the blade area decreases to 850 Newton, close to the average level of other areas. This way can effectively protect the components and prolong the service life.

[0058] For example, when optimizing the pressure ratio state in combination with the energy efficiency target and support vector machine algorithm, the system may predict three candidate parameter combinations based on historical operation data and current working conditions. Assuming the first combination has a pressure ratio of 1.5, simulation shows that energy efficiency is improved by 8%; the second pressure ratio is 1.8, which improves by 10%; the third pressure ratio is 2.0, but the efficiency is only improved by 7%. Through comparison, the system selects the combination with a pressure ratio of 1.8 as the optimal solution. This optimization method helps to maximize energy utilization while ensuring force balance.

[0059] For example, in the simulation running test and energy efficiency data judgment link, the above optimal parameter combination can be tested through a virtual simulation platform. Assuming that the test environment simulates high temperature and high load working conditions, the energy efficiency data reaches 105% of the preset standard, indicating that the parameter combination still performs stably under extreme conditions. If it does not meet the standard, the system will adjust the speed or pressure ratio parameters and cycle test until the requirements are met. This repeated verification method can ensure the reliability of parameter adjustment.

[0060] For example, in the cycle calculation to determine the optimal pressure ratio state parameters, multiple data feedback and parameter fine-tuning can be used to gradually approach the target value. Assuming that the efficiency is only improved by 5% after initial adjustment, the system will further optimize the cooling system parameters based on the temperature and pressure data feedback from the sensors, and finally improve the efficiency to 12%. This closed-loop adjustment mechanism can significantly improve the overall operation stability of the compressor and provide reliable protection for industrial production.

[0061] S107, for the optimized optimal pressure ratio state parameters, a dynamic response speed evaluation algorithm is used to extract response delay indicators from real-time data of process demand changes to determine whether the adjustment process meets the standard of strong operation reliability.

[0062] The change data is obtained from real-time data related to process requirements to obtain a preliminary data set. According to the preliminary data set, the change data is screened using a preset threshold to extract key information related to response delay, and an initial value of the delay indicator is determined. For the initial value of the delay indicator, a support vector machine algorithm is applied to feature analysis of dynamic response and speed evaluation to determine whether the response delay meets a preset standard range. If the response delay exceeds the preset standard range, the state parameters of the adjustment process are compared in real time to obtain deviation data related to the optimal pressure ratio to obtain adjustment basis. According to the adjustment basis, parameter correction is performed for the relationship between the adjustment process and the operation reliability, the corrected state parameters are extracted, and a new operation configuration is determined. Through the new operation configuration, the change data of the process requirement is continuously monitored to obtain a real-time extracted response delay update value to determine whether the standard requirement is met. If the response delay update value still does not meet the standard requirement, the parameter correction and operation configuration update process are repeatedly performed to obtain a final reliability evaluation result.

[0063] For example, in the field of compressor operation parameter optimization, obtaining change data from real-time data related to process requirements is a key link. Assuming that the focus is on the pressure fluctuation data of the compressor under different loads, the change values of the inlet pressure and the outlet pressure are collected in real time through sensors to form a preliminary data set. For example, under a certain working condition, the inlet pressure fluctuation range is between 2.5 and 3.0 bar, and the outlet pressure fluctuation is between 8.0 and 8.5 bar. These data will be classified as a preliminary set for subsequent analysis.

[0064] Specifically, when the preliminary data set is screened using a preset threshold, data with a pressure fluctuation amplitude exceeding 0.3 bar can be set as an abnormal point, and key information related to response delay is extracted. For example, in the above data, if the outlet pressure jumps from 8.0 bar to 8.5 bar in a short time, exceeding the threshold range, it is recorded as delay-related information, and the initial value of the delay indicator is preliminarily determined as 0.2 seconds. This indicator reflects the response speed of the system to pressure changes.

[0065] For example, when the support vector machine algorithm is applied for feature analysis, the correlation between dynamic response speed and pressure change can be taken as the analysis object. Assuming that through algorithm analysis, it is found that when the response delay is 0.2 seconds, the stability of the system under high load decreases, exceeding the preset standard range of 0.1 to 0.15 seconds. At this time, the state parameters of the adjustment process need to be further compared to obtain deviation data related to the optimal pressure ratio. For example, the current pressure ratio is 3.2, and the target pressure ratio is 3.0, with a deviation of 0.2, which is used as the adjustment basis.

[0066] Specifically, when adjusting the parameters for the relationship between the regulation process and operational reliability, the pressure ratio can be gradually adjusted to 3.1 by reducing the intake flow, and the corrected state parameters are extracted to form a new operating configuration. Assuming that the response delay update value is 0.16 seconds after adjustment and still does not meet the standard range, the parameter correction process is continued until the delay value is reduced to 0.13 seconds, meeting the requirements. This process ensures the reliability of the system in dynamic changes.

[0067] For example, when continuously monitoring process demand change data, the response delay value can be updated based on real-time collected pressure and flow data. Assuming that the monitoring finds that the delay value stabilizes at 0.13 seconds, indicating that the new operating configuration effectively improves the response capability of the system. This way not only optimizes the parameter configuration, but also ensures the stable operation of the equipment under complex working conditions.

[0068] Specifically, the final reliability evaluation result can be obtained by comprehensively analyzing the data after multiple cycle adjustments to determine whether the system can maintain the delay index within the standard range under different loads. For example, in a 24-hour continuous operation test, the delay value is maintained between 0.12 and 0.14 seconds, indicating that the system reliability has been significantly improved. This method provides data support for subsequent optimization and lays a foundation for long-term stable operation of the equipment.

[0069] S108, if the regulation process meets the standard of high operational reliability, update the control logic of the actuator through the hardware design innovation module, extract the life prediction value from the key component wear monitoring data, and obtain the stability parameters of the compressor for long-term operation.

[0070] Obtain the state information of the key components from the wear monitoring data, use the pre-established feature extraction model to obtain the wear feature value of the key components. According to the wear feature value combined with historical operation data, apply the support vector machine model to judge the life prediction trend of the key components. If the life prediction trend is lower than the preset threshold, adjust the control logic through the hardware innovation module to obtain the optimized actuator parameters. According to the optimized actuator parameters, real-time update the regulation process of the compressor operation, and determine the stability index of the compressor for long-term operation. If the stability index does not meet the preset standard, extract the deviation data from the regulation process to obtain the deviation correction value. Adjust the output strategy of the hardware innovation module through the deviation correction value to determine whether the compressor operation has returned to a high reliability state. If the high reliability state has been restored, continuously collect wear monitoring data to determine the subsequent life prediction trend of the key components.

[0071] For example, when obtaining key component state information from wear monitoring data, the vibration frequency and temperature data of key components such as bearings or blades inside the compressor can be collected in real time through sensors. Assuming that the vibration frequency of a certain bearing is normally 5 Hz, and the monitoring data suddenly shows 8 Hz, it may indicate early signs of wear. Based on this, combined with the pre-established feature extraction model, the vibration frequency, temperature change rate, etc. can be converted into specific wear characteristic values, laying the foundation for subsequent analysis.

[0072] For example, when applying a support vector machine model to determine the life prediction trend, trend analysis can be performed based on historical operation data and current wear characteristic values. Assuming that historical data shows that the bearing usually has a remaining life of 500 hours when the vibration frequency reaches 9 Hz, and the current data is 8 Hz, it can be predicted that the remaining life is about 600 hours. If the preset threshold is 700 hours, intervention measures need to be taken. This method intuitively reflects the component state through data comparison, facilitating timely decision-making.

[0073] For example, for the case where the life prediction trend is below the threshold, the control logic can be adjusted through the hardware innovation module to optimize the actuator parameters. Assuming that the lubrication frequency of the bearing is increased from 1 time per hour to 2 times per hour after adjustment, and the operating load is reduced by 10%, the wear rate can be effectively slowed down. This adjustment logic can extend the service life of the component and reduce the risk of sudden failure.

[0074] For example, when updating the compressor operation adjustment process in real time, based on the optimized parameters, the stability indicators can be monitored. Assuming that the vibration amplitude of the compressor is reduced from 0.5 mm to 0.3 mm after adjustment, but the preset standard is 0.2 mm, further correction is needed. This continuous monitoring and parameter adjustment helps to ensure long-term stable operation of the equipment.

[0075] For example, when extracting deviation data from the adjustment process and obtaining correction values, the reasons for the vibration amplitude exceeding the standard can be analyzed. Assuming that the excessive standard is found to be due to insufficient lubrication, the correction value can be set to increase the amount of lubricant by 20%. Through deviation correction, the ideal operating state is gradually approached, and the reliability of the equipment is improved.

[0076] For example, after adjusting the output strategy of the hardware innovation module, it can be determined whether the compressor has returned to a high reliability state by continuously monitoring the vibration and temperature data. Assuming that the vibration amplitude stabilizes at 0.2 mm and the temperature remains within the normal range after adjustment, it can be considered that the high reliability state has been restored. This strategy adjustment can effectively respond to dynamically changing operating environments.

[0077] For example, when continuously collecting wear monitoring data to determine the trend of subsequent life prediction, the data can be updated every 24 hours, and the bearing life is re-evaluated in combination with the aforementioned model. Assuming that the predicted life is increased from 600 hours to 750 hours, it indicates that the intervention measures are effective. This continuous monitoring mechanism provides a reliable basis for equipment maintenance, ensuring the continuity and safety of operation.

[0078] S109, according to the matching degree of stability parameters and process demand changes, the intelligent control algorithm is used to iteratively update the continuous stepless adjustment model, the energy efficiency improvement effect is extracted from the real-time running data, and the final adjustment optimization scheme is determined.

[0079] The process demand data and stability parameter data are obtained from the system, the correlation between the two is calculated using a pre-established matching degree evaluation model, and a preliminary matching analysis result is obtained. According to the preliminary matching analysis result, the intelligent control algorithm is used to adjust the parameters of the continuous stepless adjustment model, and the part with a matching degree lower than the preset threshold is optimized to determine the update direction of the adjustment model. The change trend of real-time data and running data is obtained, and the support vector machine algorithm is used to classify the abnormal points that exceed the preset threshold in the change trend to determine the energy efficiency influencing factors. According to the classification result of the energy efficiency influencing factors, the key improvement indicators are extracted, and the data of the indicators with a priority higher than the preset threshold is deeply mined to obtain the efficiency improvement space distribution. According to the efficiency improvement space distribution, the parameter weights in the adjustment model are adjusted, and if the deviation between the model output after the weight adjustment and the real-time data exceeds the preset range, the iterative optimization is performed to determine the final adjustment parameter combination. The running effect data of the adjustment parameter combination is obtained, the time series analysis is performed on the fluctuation in the effect data to determine whether there is a stability hidden danger, and the optimized adjustment scheme is obtained.

[0080] For example, when obtaining process demand data and stability parameter data from the system, real-time collection of process demand data such as pressure and temperature during compressor operation can be performed, and correlation analysis can be performed in combination with historical stability parameter data such as vibration frequency and running time. Assuming that the pressure fluctuation range in the process demand data is 2-5 bar, and the stability parameter shows that the vibration frequency continuously increases in a certain period, the correlation degree can be quantified by a pre-established matching degree evaluation model, and the preliminary analysis result may show that the matching degree is 75%, which is lower than the preset threshold of 80%, and further optimization is required.

[0081] For example, the intelligent control algorithm adjusts the parameters of the continuous stepless adjustment model.

[0082] It can be understood that this adjustment is often based on real-time feedback data. Assuming that the part with a low matching degree is concentrated in a high-load running scenario, the intelligent algorithm will preferentially adjust the response speed parameter in the adjustment model, increasing it from the default 0.5 times per second to 1 time per second, to adapt to pressure fluctuations more quickly. This adjustment direction can better match the process demand and stability requirements.

[0083] For example, when obtaining real-time data and running data trends, if it is found that the energy consumption data abnormally increases in a certain period of time, exceeding the preset threshold of 10%, the support vector machine algorithm can be used for classification processing, and the influencing factors are classified into categories such as equipment aging and external environment changes. Assuming that the classification result shows that the proportion of equipment aging is 60%, the subsequent analysis will focus on this, and key improvement indicators such as bearing wear rate will be extracted, and parts with priority higher than the preset threshold of 50% will be subjected to data deep mining, and it is concluded that the efficiency improvement space is mainly concentrated in the optimization of bearing maintenance period.

[0084] For example, when adjusting the weight of the adjustment model parameter for the efficiency improvement space distribution, if it is found that the bearing maintenance period has a weight of 40% on the overall energy consumption, it can be increased to 60% to more accurately allocate resources. If the deviation between the output of the adjusted model and the real-time data still exceeds the preset range of 5%, other parameters such as lubrication frequency will be gradually adjusted through iterative optimization, and the final adjustment parameter combination will be determined. This way can effectively improve the efficiency of the compressor.

[0085] For example, when analyzing the running effect data of the adjustment parameter combination, the fluctuation in the effect data can be analyzed by time series analysis to determine whether there is a stability risk. Assuming that after a certain parameter combination is run, the vibration frequency fluctuation amplitude is reduced from ±0.2 to ±0.1, but there are still periodic abnormal peak values, the adjustment scheme needs to be further optimized, which may be to eliminate the risk by increasing the frequency of preventive maintenance. This analysis helps to ensure the stability of long-term operation.

[0086] For example, for data deep mining of key improvement indicators, the bearing temperature variation trend can be extracted from historical running data, combined with wear rate data, and it is analyzed that the wear rate increases by about 2% for every 5 degrees Celsius increase in temperature. This mining result provides data support for subsequent optimization, ensuring that the adjustment scheme is more targeted.

[0087] For example, when optimizing the adjustment scheme, the key component life prediction trend mentioned in the historical dialogue can be combined to focus on the parts with severe wear and tear, and a differentiated maintenance strategy can be developed. This strategy can effectively extend the service life of the equipment and improve the overall operation reliability.

[0088] The above embodiment is only one of the preferred embodiments of the present application, and should not be used to limit the protection scope of the present application, but any modification or polishing without substantial meaning made in the main design idea and spirit of the present application, and the technical problems solved are still consistent with the present application, and should be included in the protection scope of the present application.

Claims

1. A continuous stepless adjustment system and method for a large reciprocating compressor based on intelligent control, characterized in that: The following steps are involved: (1) Collect the real-time operating data of the compressor through sensors, filter and normalize the data, and extract the load change trend and process fluctuation characteristics; (2) Based on the load change trend and process fluctuation characteristics, a continuous stepless regulation model is constructed, and the initial parameter set of the regulation instruction is generated in combination with the preset load range regulation threshold and dynamic response speed requirements; (3) Using intelligent control algorithms to optimize the initial parameter set, matching the target pressure ratio value from the optimal pressure ratio state database, and determining whether the adjustment command meets the millisecond-level response condition; (4) If the millisecond-level response condition is met, an execution signal is generated, the force distribution data of key components is obtained according to the hardware response mapping table, and an unbalanced force adjustment plan is generated; (5) Dynamically adjust the operating parameters of the actuator according to the unbalanced force adjustment scheme, and monitor the wear status of key components through real-time feedback to obtain the component operation data after force balance; (6) Combined with the goal of improving energy efficiency, the compressor operating parameters are recalibrated to obtain the optimized optimal pressure ratio state parameters; (7) Evaluate the dynamic response speed based on the real-time data of process demand changes and determine whether the adjustment process meets the operational reliability standards; (8) If the standard is met, the control logic of the actuator is updated, the life span is predicted based on the wear monitoring data of key components, and the stability parameters of the long-term operation of the compressor are obtained; (9) According to the matching degree between the stability parameters and the changes in process requirements, the continuous stepless adjustment model is iteratively updated, and the energy efficiency improvement effect is extracted to determine the final adjustment optimization plan.

2. The method according to claim 1, characterized in that Extracting load change trends and process fluctuation characteristics in step (1) specifically includes: De-noising and normalizing the collected real-time signal data to obtain standardized data; Extract load variation characteristics from standardized data and mark them as abnormal fluctuations if they exceed the preset threshold; Analyze fluctuations related to process requirements, record deviations beyond the process requirements, and identify process fluctuation characteristics; Combining the load variation characteristics and process fluctuation characteristics, the support vector machine algorithm is used to classify potential operation risks.

3. The method according to claim 1, characterized in that The initial parameter set for generating the adjustment instruction in step (2) specifically includes: Construct an input feature set based on load variation data and process fluctuation data; If the load change exceeds the preset threshold, the adjustment instruction generation module is triggered to calculate the initial parameter set based on the dynamic response requirements; The support vector machine algorithm is used to optimize the parameter set, and the feedback information of the adjustment process is recorded to update the model.

4. The method according to claim 1, wherein The step (3) of determining whether the adjustment instruction satisfies the millisecond-level response condition specifically includes: Extract operating data from the initial parameter set and use intelligent control algorithms to optimize the regulation model; Match the target pressure ratio value from the optimal pressure ratio state database, and generate an adjustment instruction if there is a mismatch; Detect the response time of the adjustment instruction and perform secondary optimization if it exceeds the threshold.

5. The method according to claim 1, wherein The generation of the unbalanced force adjustment scheme described in step (4) specifically includes: Extracting force distribution data of key components from a hardware response mapping table according to the execution signal; Analyze the force distribution data and generate adjustment strategies if there are any non-equilibrium characteristics; Through hardware design innovation module optimization and adjustment strategy, the final force balance solution is generated.

6. The method according to claim 1, characterized in that The step (5) of dynamically adjusting the operating parameters and monitoring the wear status specifically includes: Optimize the operating parameters of the actuator according to the unbalanced force adjustment plan; Monitor the wear status of key components in real time and recalculate operating parameters if the threshold is exceeded; Predicting the long-term wear risk of key components using support vector machine algorithm.

7. The method according to claim 1, characterized in that The secondary calibration operating parameters in step (6) specifically include: Optimize compressor operating parameters by combining load balance data and energy efficiency targets; The energy efficiency improvement effect is verified through simulation tests until the preset standards are met.

8. The method according to claim 1, characterized in that The evaluation of dynamic response speed described in step (7) specifically includes: Extract response delay indicators from process demand change data; If the delay is outside the preset range, correct the tuning parameters and re-evaluate.

9. The method according to claim 1, characterized in that The step (8) of predicting the life span and obtaining the stability parameters specifically includes: Extracting life prediction values ​​for key components from wear monitoring data; If the predicted lifetime value is lower than the threshold, the control logic is adjusted and the stability parameters are updated.

10. The method according to claim 1, characterized in that The iterative update of the adjustment model in step (9) specifically includes: Evaluate how well stability parameters match changes in process requirements; Optimize parameters with low matching degree to extract energy efficiency improvement effects; Through time series analysis, we can identify potential stability risks and determine the final adjustment plan.

Images