Hydraulic machine intelligent optimization control method based on big data analysis

Abstract

The invention discloses a hydraulic machine intelligent optimization control method based on big data analysis, and relates to the technical field of hydraulic machine intelligent control, and the method comprises the following steps: carrying out the real-time frequency domain analysis of pressure, flow and energy consumption signals during the operation of a hydraulic machine, constructing a load spectrum analysis module, extracting main frequency fluctuation characteristics, and building a corresponding spectrum template; and performing sliding cross matching on the control threshold change curve of the energy-saving mode switching and the frequency spectrum template, identifying a periodic coupling risk section, and labeling and recording the periodic coupling risk section. According to the invention, through frequency domain analysis and a self-adaptive lag mechanism, frequency resonance type oscillation in energy-saving control is effectively suppressed, and the operation stability and the response reliability of the system are improved; meanwhile, a strategy self-learning mechanism based on resonance labels and energy efficiency feedback is introduced, parameters and control strategies are dynamically optimized and judged, the adaptability and long-term robustness of the system to complex working conditions are enhanced, and intelligentization and high efficiency of energy-saving control of the hydraulic equipment are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent control of hydraulic presses, and in particular to an intelligent optimization control method for hydraulic presses based on big data analysis. Background Art

[0002] Intelligent optimization control of hydraulic presses based on big data analysis refers to the use of multi-source data (such as pressure, flow, oil temperature, load response, energy consumption, etc.) collected from the operation of the hydraulic press to build an operation feature database, and use big data mining technology to perform pattern recognition, trend analysis and anomaly detection on historical and real-time data, extract the coupling relationship between the key performance indicators of the hydraulic system and the control variables, and further combine machine learning algorithms to generate adaptive control strategies to achieve dynamic optimization and adjustment of parameters such as pressure output, energy-saving mode, and response speed during the execution of the hydraulic press, thereby improving the equipment's operating efficiency, stability and intelligence level, reducing energy consumption and failure rate, and promoting the evolution of hydraulic equipment towards digitalization and intelligence.

[0003] The existing technology has the following deficiencies: Existing energy-saving control systems based on power factor usually rely on real-time collection of voltage, current and load data during the operation of the hydraulic press, and dynamically adjust the start and stop thresholds of the energy-saving mode after calculating the corresponding power factor. However, in certain complex working conditions with frequent dynamic load fluctuations, if the instantaneous fluctuation frequency of the system power factor is periodically coupled with the judgment cycle or threshold adjustment mechanism of the energy-saving controller, it is easy to cause the system to frequently judge the activation and exit status of the energy-saving mode within the boundary interval, thereby causing control logic oscillation. This "frequency resonance oscillation phenomenon" will cause the energy-saving module to start and stop repeatedly in a short period of time, causing core components such as hydraulic pumps, proportional valves, and motors to be in a non-steady-state alternating operation state, which will not only cause system response delays and actuator fatigue, but may also cause drastic fluctuations in oil temperature, controller logic dead loops, and premature aging of key components. In severe cases, it may even cause mechanical shock or seal failure, threatening system safety and reliability.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent optimization control method for hydraulic presses based on big data analysis. Through frequency domain analysis and adaptive hysteresis mechanism, the frequency resonance oscillation in energy-saving control is effectively suppressed, and the system operation stability and response reliability are improved; at the same time, a strategy self-learning mechanism based on resonance tags and energy efficiency feedback is introduced to dynamically optimize judgment parameters and control strategies, enhance the system's adaptability to complex working conditions and long-term robustness, and realize the intelligence and efficiency of energy-saving control of hydraulic equipment to solve the problems in the above-mentioned background technology.

[0006] In order to achieve the above object, the present application provides the following technical scheme: a hydraulic machine intelligent optimization control method based on big data analysis, comprising the following steps: S1, real-time frequency domain analysis is performed on the pressure, flow and energy consumption signals in the operation of the hydraulic machine, a load spectrum analysis module is constructed, the main frequency fluctuation characteristics are extracted, and a corresponding spectrum template is established; S2, the control threshold value change curve of the energy-saving mode switching is matched with the spectrum template through sliding cross, the periodic coupling risk section is identified and labeled; S3, an adaptive lag adjustment function is introduced to the marked high-risk section, time delay is applied to destroy the synchronization relationship between the switching rhythm and the load main frequency; S4, a two-way confirmation criterion is constructed based on the moving average and multi-scale trend analysis, and the energy-saving mode switching is triggered only when the trend continuously crosses the set interval and meets the positive and negative judgment conditions; S5, after the energy-saving mode is enabled, the minimum residence time limit is applied, the state retention counter is used to prevent short-term exit, and the control continuity is enhanced; S6, a strategy self-learning and evolution mechanism is constructed by using the historical resonance event label and energy efficiency feedback, the judgment parameters and threshold weights are dynamically optimized, and adaptive regulation and robustness improvement are realized.

[0007] Preferably, step S1 comprises: The pressure, flow and energy consumption signals in the operation of the hydraulic machine are collected and uniformly time-stamped; The collected signals are windowed and then fast Fourier transform is performed to obtain frequency domain data; The frequency domain data is feature-extracted and denoised to generate a spectrum feature vector; The spectrum feature vector is matched and analyzed with the historical spectrum template to establish a dynamic load fluctuation spectrum template.

[0008] Preferably, step S2 comprises: A dynamic change curve of the energy-saving mode switching control threshold value is constructed and a smoothing function model is generated; The load fluctuation spectrum template under the current operating condition is extracted and a main frequency trend curve is generated; The dynamic change curve and the main frequency trend curve are matched section by section by using a sliding window and a weighted cross-correlation coefficient algorithm, and the time section with a matching score exceeding a threshold value is identified as a periodic coupling risk section; The periodic coupling risk section is labeled and the label number, start and end time, matching score, control threshold value period parameter and spectrum feature are recorded for calling in the subsequent control process.

[0009] Preferably, step S3 comprises: determine whether the current control period is in a high coupling risk section, and start a hysteresis control module; set a corresponding delay time window according to the coupling risk level, and dynamically adjust the delay time within a limited range; introduce a delay control link in the original energy-saving switching judgment logic, and only execute the energy-saving switching instruction after the delay time ends and the judgment condition is still met; record and feed back each hysteresis execution condition and control effect to the strategy optimization module for updating the hysteresis parameters to realize dynamic regulation and control.

[0010] Preferably, step S4 comprises: collect energy efficiency related signals during the operation of the hydraulic machine, and extract short period and long period trend data through a sliding window; determine whether the energy efficiency trend develops continuously in the energy-saving entering or exiting direction within a set time range and meets the condition of crossing a preset threshold interval; establish a positive confirmation criterion and a reverse confirmation criterion respectively, the positive confirmation criterion is used to determine whether the energy-saving mode entering condition is met, and the reverse confirmation criterion is used to determine whether the energy-saving mode exiting condition is met; only when the positive confirmation criterion and the reverse confirmation criterion both meet the set requirements, execute the energy-saving mode switching instruction.

[0011] Preferably, step S5 comprises: start an energy-saving state retention counter when the energy-saving mode is enabled, and record the energy-saving starting time; set a corresponding minimum residence time according to the current operating condition of the hydraulic machine, and lock the state; mask the energy-saving exit trigger logic within the minimum residence time, and buffer the monitored exit judgment signals; unlock the state after the minimum residence time ends, and determine whether to execute the energy-saving mode exit instruction according to the buffered signals.

[0012] Preferably, step S6 comprises: record the resonance label event and the corresponding energy efficiency feedback result in the energy-saving control process, and establish a historical database; analyze the correlation characteristics between the control parameters and the energy efficiency performance, extract the optimal configuration combination of the judgment parameters and the weight coefficients; dynamically adjust the judgment parameters and the weight coefficients according to the historical effects, and limit the parameter change amplitude to prevent strategy shock; establish a strategy version control mechanism, generate version records for each strategy adjustment, and maintain the strategy evolution path in the system operation.

[0013] In the above technical solution, the technical effects and advantages provided by the present application are: By introducing frequency domain analysis and periodic coupling identification mechanisms, the present invention can extract the main frequency fluctuation characteristics during the operation of the hydraulic press in real time, and perform sliding matching with the energy-saving control threshold change rhythm, so as to identify in advance the periodic resonance sections that may cause logical oscillations. By applying an adaptive hysteresis adjustment function to the high-coupling risk section and constructing a state maintenance mechanism with a minimum residence time, the synchronous resonance relationship between the control judgment and the load response is effectively destroyed, and the repeated start-stop phenomenon triggered by transient signal disturbances is blocked. This mechanism significantly improves the robustness of the energy-saving switching judgment, allowing core components such as hydraulic pumps, motors, and proportional valves to operate in a stable state, avoiding problems such as actuator fatigue, oil temperature fluctuations, and controller dead loops caused by frequent non-steady-state switching, and comprehensively improving the control stability and response reliability of the system under complex working conditions.

[0014] The present invention constructs a strategy self-learning and evolution mechanism based on resonance tags and energy efficiency feedback, so that the control system can continuously accumulate decision-making history and working condition feedback data during long-term operation, automatically identify the causal relationship between various control criteria and system responses, and dynamically optimize and adjust the judgment parameters and threshold weight coefficients. Through strategy version management and parameter evolution mechanism, the system can continuously optimize its own control strategy and gradually realize the intelligent transition from rule-driven to data-driven. This solution not only improves the adaptive ability of the energy-saving control system under different working conditions, but also significantly enhances its robustness to nonlinear disturbances, unknown working conditions and signal noise, ensuring the long-term effective operation of the energy-saving control logic in a variety of scenarios, and ultimately achieving a coordinated improvement in the energy efficiency, lifespan and intelligence level of hydraulic equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction to the drawings required for use in the embodiments will be given below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0016] Figure 1 This is a flow chart of the method for intelligent optimization control of a hydraulic press based on big data analysis of the present invention. DETAILED DESCRIPTION

[0017] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of this disclosure will be thorough and complete and will fully convey the concepts of the example embodiments to those skilled in the art.

[0018] The present invention provides Figure 1The hydraulic machine intelligent optimization control method based on big data analysis shown comprises the following steps: The pressure, flow and energy consumption signals collected during the operation of the hydraulic machine are analyzed in real time in the frequency domain, a dynamic load spectrum analysis module is constructed, the main frequency fluctuation characteristics under the current operating condition are extracted, and a load fluctuation spectrum template corresponding to the operating condition is established; In order to realize high-precision modeling and energy-saving control optimization of the operating characteristics of the hydraulic machine under complex conditions, a dynamic load spectrum analysis method based on big data analysis is proposed to construct a load fluctuation spectrum template to support periodic coupling identification and control strategy adjustment in the subsequent energy-saving mode switching process, which comprises the following steps: Firstly, a real-time acquisition mechanism of hydraulic machine operation data is established. A plurality of sensor nodes are arranged in the main hydraulic circuit of the hydraulic machine, including pressure sensors, flow sensors and energy consumption measuring modules, which are used to acquire working cavity pressure signals, hydraulic oil flow signals and system power consumption signals in real time. All sensor output signals are transmitted to the edge processing unit for data collection after being processed by the front-end signal conditioning module (including filtering, amplification, A / D conversion, etc.). The sampling frequency is set to be not less than 500Hz according to the dynamic response requirement of the hydraulic system, so as to ensure sufficient frequency domain resolution. In order to improve the time sequence consistency and analysis accuracy of the signals, a unified time stamp mechanism is used to synchronize and calibrate various types of sensor data, and a continuous and high-precision multivariate time sequence data set is constructed.

[0019] Secondly, the collected pressure, flow and energy consumption signal sequences are processed by multi-channel real-time frequency domain conversion. The edge processing unit is integrated with a Fast Fourier Transform (FFT) module, which performs real-time spectrum analysis on the windowed time sequence signals. In this process, a windowing technique (such as Hanning window or Blackman window) is used to preprocess each data window to reduce the influence of frequency spectrum leakage; and a sliding window mechanism is used to continuously segment and analyze the data stream to obtain the spectrum sequence evolving with time. For the frequency domain results in each window, the main frequency, peak frequency, energy concentration degree and other key frequency domain characteristic parameters are extracted, and the analysis results are stored in the spectrum database in real time for subsequent pattern recognition and modeling analysis. In order to improve the availability and noise reduction capability of the frequency domain data, the empirical mode decomposition (EMD) or wavelet threshold filtering method is used to suppress the high-frequency disturbance components after the spectrum is generated, so as to ensure the accuracy and stability of the main frequency signal.

[0020] Third, a multivariate spectrum fusion model is constructed to comprehensively characterize the fluctuation characteristics of the hydraulic system under current operating conditions. Principal Component Analysis (PCA) or an autoencoder neural network is used to perform feature compression and information fusion on the frequency response results of the three dimensions of pressure spectrum, flow spectrum, and energy consumption spectrum, extracting representative spectrum vectors. These spectrum vectors are used to describe the dynamic load response characteristics of the hydraulic system under the current operating conditions and are compared with stored spectrum templates for typical operating conditions. To enhance the template representation capability, a segmented spectrum template system is established for different operating stages (such as fast forward, pressurization, pressure maintenance, and return) based on data from multiple rounds of operating condition experiments, enabling dynamic modeling of complex multi-stage operating conditions. Each spectrum template records parameters such as the dominant frequency value, average frequency fluctuation range, and frequency energy distribution for that stage to support subsequent periodic coupling identification and strategy matching.

[0021] Finally, the current spectrum vector is matched and analyzed with the existing spectrum template to construct a dynamic load fluctuation spectrum template. The matching process uses a sliding time window cross-correlation analysis method to compare the spectrum similarity between the current operating data and the historical template, identify whether the current operation is within a certain known operating condition spectrum range, and establish a matching index in the template library. When it is detected that the current operating condition is unknown or deviates from the known template, the spectrum template update process is automatically triggered, and the current spectrum feature vector is recorded in the database and marked as a "new operating condition" for the system learning module to subsequently update the template library. The entire spectrum template system is maintained in the form of a tree structure or graph neural network, which supports rapid retrieval and evolutionary adjustment, and provides basic data support for the identification of resonance sensitive areas during subsequent energy-saving mode switching.

[0022] Through the above steps, dynamic frequency domain modeling of the hydraulic press under any operating conditions can be achieved, and a refined fluctuation template can be constructed based on the spectrum data, effectively supporting intelligent optimization functions such as periodic resonance identification and adaptive adjustment in subsequent energy-saving control logic, significantly improving the system's adaptability to complex working conditions and the robustness of the control strategy.

[0023] Perform sliding cross-matching between the control threshold change curve used for energy-saving mode switching and the load fluctuation spectrum template to identify potential periodic coupling risk segments and label the identification results for subsequent control process calls. To realize accurate identification and intervention of the risk of periodic coupling that may cause control logic oscillation in the energy-saving mode switching process, a periodic coupling sensitive zone identification and labeling method based on sliding cross matching is proposed. By coupling and matching the control threshold value change curve with the load fluctuation frequency spectrum template under the current operating condition of the hydraulic system, the sensitive time period of the system with resonance trend is accurately identified, and the sensitive section is digitally labeled, providing prior input basis for the subsequent adaptive control mechanism. Specifically, the following steps are included: First, a time-domain expression model of the energy-saving control threshold value change curve is constructed. The energy-saving controller generates start-stop threshold judgment rules during operation according to power factor, energy efficiency curve and other parameters. The judgment rule can be abstracted as a judgment curve with periodic dynamic change characteristics. In this invention, the curve is sampled at equal intervals to construct a set of continuous time series, and a high-resolution smooth function curve is generated by interpolation to complete the processing, which is used as the standard input model of the control judgment curve. Further, the curve can be reconstructed based on the historical decision-making behavior of the controller, so as to truly restore the threshold fluctuation rhythm of the control system during operation.

[0024] Secondly, the load fluctuation frequency spectrum template corresponding to the current operating condition is obtained. The template is derived from the frequency domain analysis results generated in the previous step, and contains the frequency principal component features extracted from the pressure signal, flow signal and energy consumption signal of the current hydraulic machine after fast Fourier transform processing. Each frequency spectrum template includes main frequency position, bandwidth distribution, energy density and other core parameters, forming a set of time sequence template set of frequency response. In this embodiment, the spectrum main axis fitting algorithm is used to convert the multivariate frequency distribution into a one-dimensional main frequency trend curve, so that it has structural comparability with the threshold value change curve for synchronous matching.

[0025] Third, perform sliding cross matching operation to identify the resonance trend section. The control threshold value change curve constructed is used as the target reference function, and the frequency spectrum template main frequency trend curve is used as the comparison object. In the set time window, the sliding window is gradually pushed forward, and the weighted cross-correlation coefficient (Weighted Cross-Correlation Coefficient) algorithm is used to match and calculate the two. The algorithm considers the amplitude fitting degree and phase synchronization degree, and scores the curve segments in the sliding window on the time axis, so as to obtain the coupling degree measurement results at different time points. When the matching score in the sliding window exceeds the set coupling risk judgment threshold, it is identified as a periodic coupling sensitive section, and the corresponding time interval and related frequency parameters are recorded.

[0026] Finally, the identified coupling risk section is labeled and managed by digital tags. In the data structure, a unique identifier is assigned to each identified sensitive section, and a tag table is established to record information such as tag number, start time, end time, matching score, control threshold period characteristics, spectral main frequency characteristics, etc. All tag information is stored in the tag database in real time and called during controller operation. In the energy-saving control logic judgment process, if the current time step falls into a section that has been marked as high coupling risk, the corresponding intervention mechanism (such as adaptive lag, two-way confirmation, etc.) is automatically triggered to avoid the control logic entering the oscillation state. The tag database has an update and removal mechanism that can dynamically adjust the validity of the tags based on the running period and resonance elimination state to ensure safety while maintaining response efficiency.

[0027] Through the above steps, the pre-identification and explicit labeling of the periodic resonance trend that may occur in the energy-saving control process can be achieved, improving the system's self-awareness and the forward-looking nature of intelligent control in dynamic complex working conditions, and providing key support for the intelligent and robust evolution of the energy-saving control strategy of the hydraulic press.

[0028] For the labeled high coupling risk section, an adaptive lag adjustment function is applied to introduce an adjustable time lag window to the original threshold decision logic to break the synchronization relationship between the threshold switching rhythm and the main frequency fluctuation rhythm, reducing the probability of resonance triggering; To solve the problem of control oscillation caused by periodic resonance due to the change in energy-saving decision threshold and the fluctuation rhythm of the load main frequency in the energy-saving control process of the hydraulic press, an adaptive lag adjustment method for periodic coupling risk sections is proposed. This method introduces a time delay control mechanism into the original energy-saving switching logic to break the synchronization relationship between the control rhythm and the system main frequency, thereby suppressing the repeated switching and control instability caused by frequency coupling. The method specifically includes the following steps: First, identify the target time section that needs to apply lag adjustment. The system has completed the sliding cross-matching analysis of the energy-saving mode switching control threshold change curve and the hydraulic press running load fluctuation frequency spectrum template in the early running process, and has labeled the identified periodic coupling risk section. In this step, the system dispatching control module retrieves in real time whether the current time section is within the high coupling risk section that has been marked, and judges whether to trigger the lag adjustment logic based on the start and end time, coupling degree level, frequency information, etc. in the tag record. Only when the current control period falls completely within the high-risk section, the adaptive lag control module is started to ensure that the delay mechanism does not interfere with the response efficiency under normal working conditions.

[0029] Secondly, the time lag parameters are set dynamically according to the risk level. In order to achieve flexible adjustment of the hysteresis mechanism, the system introduces a configurable delay time window, the size of which is dynamically adjusted according to the identification score results of the current high coupling risk section. Specifically, the system evaluates the strength of the current coupling based on the matching score, main frequency duration and spectrum energy distribution characteristics calculated in the early stage, and selects the corresponding hysteresis level within the set range. The higher the hysteresis level, the more likely the system is to be in an unstable state that is prone to resonance. Correspondingly, the hysteresis time window should be larger to fully widen the overlapping interval between the control judgment time and the load fluctuation period. The time window is finely controlled in milliseconds and has upper and lower limit protection values ​​to prevent excessive hysteresis from affecting the control response.

[0030] Third, a hysteresis control link is inserted into the energy-saving switching judgment logic. In the original energy-saving control logic, when the power factor or energy efficiency signal meets the judgment conditions for entering or exiting the energy-saving mode, a switching instruction will be issued immediately. After the introduction of the adaptive hysteresis mechanism, the system first sets a "waiting time" for the current switching judgment condition, that is, delaying the preset hysteresis time window based on the current time point. Only when the delay expires, the system re-judges whether the original switching condition is still valid. If the condition is still met after the end of the delay window, the control logic is allowed to continue to execute the activation or exit of the energy-saving mode. This is equivalent to constructing a dynamic buffer zone, which effectively filters out those short-period phase synchronization oscillation triggers in the load main frequency fluctuations, thereby reducing false triggering of switching.

[0031] Finally, the real-time recording and subsequent adjustment of the hysteresis control effect are performed. In order to improve the self-optimization capability of the control system, the hysteresis control module will record the duration of the hysteresis execution, the final switching decision state, the system energy efficiency change curve, and whether switching jitter occurs after each delay processing, and upload the data to the strategy optimization module. In the subsequent operation cycle, the strategy optimization module will update and adjust the hysteresis time parameters corresponding to different risk levels based on the performance of the historical hysteresis effect, thereby realizing the dynamic evolution of the delay control mechanism and enhancing the adaptability and stability of the control strategy. In addition, the system also supports the operation cycle locking of the hysteresis mechanism, that is, once the hysteresis control is triggered in a high-risk section, the hysteresis strategy will be maintained throughout the validity period of the entire risk section to avoid frequent switching of hysteresis parameters causing control inconsistency.

[0032] Through the above steps, dynamic timing disturbances can be effectively introduced during the energy-saving mode switching process, breaking the synchronization relationship between the energy-saving control logic and the main frequency response of the hydraulic system, significantly reducing control oscillations, frequent equipment start-up and shutdown, response delays and other problems caused by frequency resonance, and improving the control stability and system robustness of the hydraulic press under complex working conditions. It is an intelligent control method with obvious innovation and engineering practical value.

[0033] Based on the sliding average algorithm and multi-scale trend analysis method, a two-way confirmation criterion is constructed to determine whether the energy efficiency trend continuously crosses the threshold interval within the set time range. The energy-saving mode switching instruction is only allowed to take effect when both the positive and negative criteria are met. In order to enhance the judgment stability during the switching process of the hydraulic press's energy-saving mode and avoid the phenomenon of frequent start and stop of the energy-saving mode caused by control misjudgment due to short-term data fluctuations or local signal anomalies, a two-way confirmation criterion construction method based on the sliding average algorithm and multi-scale trend analysis is proposed. This method comprehensively evaluates the trend of energy efficiency data at different time scales and performs continuity verification in both forward and reverse directions before the energy-saving mode is enabled or exited. The energy-saving control logic is only allowed to take effect if both bidirectional conditions are met. This technical path effectively enhances the control strategy's ability to resist random disturbances and improves the decision-making robustness of energy-saving control instructions. It specifically includes the following four steps: First, a real-time trend extraction mechanism for energy efficiency data was established. The system continuously collects energy efficiency-related signals, such as real-time power consumption data, energy utilization indicators, and power factor change curves, during the operation of the hydraulic press at a fixed sampling frequency. A sliding window processing module was then implemented within the control system. The sliding window performs local smoothing on the collected raw data, extracting the local average trend value and eliminating transient fluctuations and spike interference to form a smooth curve representing the direction of energy efficiency change. The control system also constructs multiple sliding windows of different time scales in parallel. For example, a short-period window is used to capture rapidly changing trends, while a long-period window is used to reflect global trend evolution, providing a data foundation for subsequent multi-scale analysis.

[0034] Secondly, a multi-scale trend direction determination analysis is carried out. In each control determination cycle, the system compares the energy efficiency trend data extracted from the short-cycle window and the long-cycle window respectively to determine whether it continues to develop in the direction of entering or exiting the energy-saving mode within the set time range. For example, before the energy-saving mode is determined, the system requires that the energy efficiency index shows a downward trend in both the short-cycle and long-cycle windows, and continues to be lower than the energy efficiency start-up threshold throughout the set period to meet the trend consistency standard. In this process, the system uses a trend direction identification algorithm to analyze the slope sign and change amplitude of the data to clarify whether the trend meets the three basic requirements of "continuous, unidirectional, and across thresholds." If the above conditions are not met on any time scale, it is considered that the trend is unclear and the energy-saving control logic is not triggered.

[0035] Third, a two-way confirmation criterion structure is constructed. Based on the energy efficiency trend analysis module, the system establishes two sets of judgment logics: positive confirmation criteria and reverse confirmation criteria. Forward confirmation is used to determine whether the system currently meets the conditions for entering energy-saving mode; reverse confirmation is used to determine whether the system meets the conditions for exiting energy-saving mode. Before the energy-saving mode is started, the positive trend is required to stably cross the entry threshold interval within the specified time period, while the reverse trend remains not meeting the exit threshold conditions; before the exit judgment, the reverse trend is required to stably cross the exit threshold interval within the time period, while the positive trend cannot meet the entry criteria. The system only issues energy-saving control instructions when the two sets of judgments do not conflict and the confirmation direction is clear. The construction of two-way criteria avoids misjudgments caused by single trend judgments and reduces the probability of switching jitter under boundary conditions.

[0036] Finally, the execution of the judgment result triggers the energy-saving instruction and records the effect feedback. After the two-way confirmation judgment conditions are met at the same time, the system starts the energy-saving mode switching process, including the adjustment of the hydraulic pump energy-saving parameters, the change of the motor working state and the switching of the control logic state. At the same time, the system records the parameters such as the judgment result, the time taken for the judgment process, the success rate of trend judgment and the energy efficiency improvement, and forms a historical track of the control judgment in the operation database. If the trend misjudgment or invalid switching occurs repeatedly in the future, the strategy learning module will automatically adjust the sliding window length, judgment time range or trend weight coefficient to further improve the accuracy of trend analysis. Through the judgment feedback learning mechanism, the system can continuously correct the control strategy and improve the effectiveness and stability of energy-saving mode switching.

[0037] Through the above four steps, the two-way confirmation judgment mechanism can effectively constrain the triggering conditions of energy-saving control instructions, so that the hydraulic system will only execute switching commands when the energy efficiency change trend is clear, the direction is clear, and there is sufficient time, reducing the risk of false operation caused by transient disturbances and boundary fluctuations, thereby significantly improving the system's intelligent control quality and operational reliability under complex dynamic working conditions.

[0038] After the energy-saving mode is enabled, an energy-saving state holding counter is set and a minimum dwell time limit is imposed on the state so that the energy-saving mode cannot be exited due to system feedback fluctuations during the dwell time, thereby enhancing the stability of the control logic; To effectively avoid the control logic shock problem caused by frequent enablement and exit of the energy-saving mode due to short-time feedback changes under dynamic load fluctuation conditions, an energy-saving state retention method based on a minimum residence time mechanism is proposed. The method introduces a state retention counter to impose a minimum residence time limit once the energy-saving mode is entered, thereby shielding temporary disturbances or reverse trends within the determination period and ensuring stable operation of the energy-saving mode. The method not only suppresses the response delay and mechanical impact caused by frequent switching of the system, but also helps to improve the robustness and consistency of the overall energy efficiency control strategy. The implementation includes the following four steps: First, start the energy-saving state retention counter when the energy-saving mode is successfully enabled. The system enters the energy-saving mode after meeting the energy-saving switching criterion, and immediately activates the energy-saving state retention counter at this time and records the current system time as the start time of the energy-saving mode. The counter is driven by a high-precision clock built into the control system, with millisecond-level resolution, ensuring accurate control of the residence time. At the same time, the system marks the current energy-saving state as "locked", i.e. the state cannot be interrupted by any feedback changes or trend analysis results before the end of the minimum residence time window.

[0039] Second, set the minimum residence time parameter and adaptively associate it with the operating condition characteristics. The length of the residence time is a system preset parameter, which is determined by factors such as the typical response time of the hydraulic machine, the energy-saving mode switching delay, and the control frequency. In this implementation, the residence time is not a fixed value, but a dynamic parameter that matches the current operating condition. For example, when the system operates under high load conditions, the impact of energy-saving switching is greater, and the residence time is appropriately extended to ensure state stability; while in light load or medium speed conditions, the residence time is relatively short to improve system regulation flexibility. This dynamic parameter adjustment mechanism is realized through the operating condition recognition module, enabling the retention counter to have environmental perception capability, thereby improving the intelligent adaptability of the control strategy.

[0040] Third, suppress the energy-saving exit trigger logic within the residence time window. During the energy-saving state retention period, although the system continues to perform power factor detection, load trend analysis, and frequency spectrum feature recognition, the controller sets a shield flag for the trigger logic to exit the energy-saving mode. During this period, even if the system feedback data temporarily breaks through the energy-saving exit threshold, it will not trigger the energy-saving exit command. The control system buffers and records the relevant exit signals during the state retention period, and then uniformly judges whether to trigger state transition after the retention window ends. Through this delayed response mechanism, the system can effectively filter false exit judgments caused by short-time disturbances or transient fluctuations, ensuring that the energy-saving state is functionally persistent and strategically robust.

[0041] Finally, the state lock is released and an exit decision is made after the minimum residence time ends. When the energy-saving state retention counter reaches the preset residence time, the system automatically changes the energy-saving state flag from "locked" to "switchable", restoring the response ability to the energy-saving exit decision logic. At this time, if the feedback data still meets the energy-saving exit conditions, such as a significant decrease in power factor, a sharp change in load, or a drop in energy efficiency indicators, the system can normally issue an energy-saving exit instruction; otherwise, the energy-saving state is maintained and the next residence time judgment cycle is entered. This mechanism realizes a closed-loop control mode of "minimum duration + on-demand exit", ensuring the flexibility of control response while enhancing the system's immunity to oscillatory behavior.

[0042] Through the above steps, the energy-saving residence control strategy based on the state retention counter proposed in this embodiment can effectively improve the stability of the energy-saving state in the intelligent optimization control system of the hydraulic machine, avoid the adverse effects of frequent start-stop on control accuracy, equipment life, and energy efficiency performance, and demonstrate good technical practicality and engineering feasibility, suitable for various energy-saving control application scenarios that require high switching stability.

[0043] According to the resonance label events and the feedback results of the corresponding energy efficiency indicators recorded in the energy-saving control process, a strategy self-learning and evolution mechanism is constructed to dynamically adjust the decision parameters and threshold weight coefficients, realizing adaptive optimization and robustness improvement of the control strategy; To solve the problem of insufficient strategy rigidity and environmental adaptability of the energy-saving control system in long-term operation, a control strategy self-learning and evolution mechanism based on historical resonance events and energy efficiency feedback analysis is proposed. This mechanism systematically analyzes the resonance label events, decision behavior records, and energy efficiency performance data accumulated in the energy-saving control process, dynamically optimizes and adjusts the decision parameters and threshold weight coefficients in the control strategy, thereby enhancing the adaptability and robustness of the control system to uncertain working conditions. This mechanism not only has online learning characteristics, but also supports long-term evolution of the strategy, especially suitable for hydraulic machine systems operating in high-frequency dynamic load environments. It specifically includes the following four steps: Firstly, the historical label and energy efficiency feedback database in the energy-saving control process is established. During the continuous operation of the hydraulic machine energy-saving control system, the controller records the judgment basis, system state data, threshold matching results, spectrum resonance identification situation and final execution behavior before and after each energy-saving mode switching, and labels the periodic coupling risk section with "resonance label". At the same time, the system collects the energy efficiency index feedback results during the energy-saving mode execution, including average energy consumption, system response time, oil temperature rise trend, frequent switching times and other key performance parameters, forms a one-to-one corresponding data structure between label events and energy efficiency results, and stores it in the strategy learning database in real time, providing real, continuous and quantifiable sample sources for subsequent strategy analysis.

[0044] Secondly, the correlation characteristics between strategy decision behavior and execution effect are extracted. The control strategy learning module periodically calls the sample data in the historical database, analyzes the control behavior corresponding to each resonance label event and the subsequent system feedback results, extracts the key decision parameters that can significantly affect the energy-saving control effect, such as power factor change rate threshold, frequency spectrum similarity threshold, hysteresis window width, bidirectional confirmation criterion span, state retention counter duration, etc., and statistics the correlation between them and the energy efficiency improvement degree under different working conditions. In addition, the system also analyzes the decision reliability and misjudgment probability of the above parameters under different label levels (mild coupling, moderate coupling and severe coupling), thereby constructing a mapping relationship model between "control parameters-energy efficiency effect", which is used to guide the strategy optimization direction.

[0045] Third, the strategy self-learning and parameter evolution update are executed. Based on the above model, the control system uses the strategy evolution mechanism to dynamically update the key decision parameters and the weight coefficients of each criterion item. This process can be iteratively optimized by the adjustment algorithm based on sample scores. In the parameter adjustment period, the system selects the best parameter configuration in the historical samples as the current strategy reference template. At the same time, to avoid overfitting and strategy shock, the system sets boundary limits for the change range of parameter adjustment, and only when the multiple cycles perform well does it confirm to take effect. In addition, the system supports comparing and evaluating the running results before and after the latest round of strategy adjustment, and continues to strengthen the positive adjustment direction or reverse rollback adjustment path according to the evaluation results, gradually realizing the convergence and stability of the control strategy.

[0046] Finally, a long-term strategy evolution mechanism is constructed and strategy version control is maintained. In order to keep the traceability and manageability of system strategy adjustment, the controller introduces a strategy version control module, which records the version number, modification time, use period, main optimization target and effect score of each adjusted parameter combination and criterion weight configuration. At the same time, the system supports the rollback operation of strategy version, which can quickly recover to the previous stable version when the new parameter configuration causes the degradation of strategy. In addition, the controller can automatically enable the "exploratory strategy" version when it detects that the system enters a new unknown working condition (the spectrum feature deviates significantly from the historical template), which can explore a more suitable strategy for the current working condition by expanding the judgment range and adjusting the strategy sensitivity, and the process is included in the evolution database to support the rapid response of similar scenarios in the future.

[0047] Through the above four steps, the strategy self-learning and evolution mechanism can continuously optimize the control logic parameters based on the resonance events and control feedback data during the long-term operation of the hydraulic machine, improve the immunity of the system to nonlinear load disturbance, and improve the stability of the energy-saving control effect, which is an important technical path to improve the intelligent level and robust control ability of the hydraulic system, and has significant innovation and engineering application value.

[0048] By introducing the frequency domain analysis and cycle coupling identification mechanism, the present application can extract the main frequency fluctuation characteristics in the operation process of the hydraulic machine in real time, and perform sliding matching with the energy-saving control threshold change rhythm, to identify the cycle resonance section that may cause logic shock in advance. By applying an adaptive hysteresis adjustment function to the high coupling risk section and constructing a state retention mechanism with minimum residence time, the synchronous resonance relationship between control judgment and load response is effectively destroyed, and the repeated start-stop phenomenon triggered by transient signal disturbance is blocked. This mechanism significantly improves the judgment robustness of energy-saving switching, so that the core components such as hydraulic pump, motor and proportional valve can run in a stable state, avoiding problems such as actuator fatigue, oil temperature fluctuation and controller dead loop caused by non-steady-state frequent switching, and comprehensively improving the control stability and response reliability of the system under complex working conditions.

[0049] The present invention constructs a strategy self-learning and evolution mechanism based on resonance tags and energy efficiency feedback, so that the control system can continuously accumulate decision-making history and working condition feedback data during long-term operation, automatically identify the causal relationship between various control criteria and system responses, and dynamically optimize and adjust the judgment parameters and threshold weight coefficients. Through strategy version management and parameter evolution mechanism, the system can continuously optimize its own control strategy and gradually realize the intelligent transition from rule-driven to data-driven. This solution not only improves the adaptive ability of the energy-saving control system under different working conditions, but also significantly enhances its robustness to nonlinear disturbances, unknown working conditions and signal noise, ensuring the long-term effective operation of the energy-saving control logic in a variety of scenarios, and ultimately achieving a coordinated improvement in the energy efficiency, lifespan and intelligence level of hydraulic equipment.

[0050] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0051] The above description is merely illustrative of certain exemplary embodiments of the present invention. It goes without saying that those skilled in the art will be able to modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims.

[0052] It should be noted that, in this document, if there are relational terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device that includes the element.

[0053] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0054] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0055] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0056] 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 these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0057] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0058] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0059] The above description is merely illustrative of certain exemplary embodiments of the present invention. It goes without saying that those skilled in the art will be able to modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims.

Claims

1. The intelligent optimization control method of hydraulic press based on big data analysis is characterized by: The following steps are involved: S1, performs real-time frequency domain analysis on the pressure, flow, and energy consumption signals during the operation of the hydraulic press, builds a load spectrum analysis module, extracts the main frequency fluctuation characteristics and establishes the corresponding spectrum template; S2: Perform sliding cross-matching between the control threshold change curve of energy-saving mode switching and the spectrum template to identify the periodic coupling risk section and record it with a label; S3, introduces an adaptive hysteresis adjustment function to the marked high-risk sections, imposes a time delay to destroy the synchronization relationship between the switching rhythm and the load main frequency; S4, a two-way confirmation criterion is constructed based on sliding average and multi-scale trend analysis, which triggers the energy-saving mode switch only when the trend continuously crosses the set interval and satisfies the positive and negative judgment conditions; S5, imposes a minimum dwell time limit after energy-saving mode is enabled, prevents short exits through a state retention counter, and enhances control continuity; S6 uses historical resonance event labels and energy efficiency feedback to build a strategy self-learning and evolution mechanism to dynamically optimize judgment parameters and threshold weights.

2. The intelligent optimization control method for hydraulic press based on big data analysis according to claim 1 is characterized in that: Step S1 includes: Collect pressure, flow and energy consumption signals during the operation of the hydraulic press and perform unified time stamp calibration; After windowing the acquired signal, perform fast Fourier transform to obtain frequency domain data; Perform feature extraction and denoising on frequency domain data to generate spectrum feature vectors; The spectrum feature vector is matched and analyzed with the historical spectrum template to establish a dynamic load fluctuation spectrum template.

3. The intelligent optimization control method for hydraulic press based on big data analysis according to claim 1 is characterized in that: Step S2 includes: Construct a dynamic change curve of the energy-saving mode switching control threshold and generate a smooth function model; Extract the load fluctuation spectrum template under the current operating conditions and generate the main frequency trend curve; The sliding window and weighted cross-correlation coefficient algorithm are used to match the dynamic change curve and the main frequency trend curve segment by segment, and the time segment where the matching score exceeds the threshold is identified as the periodic coupling risk segment; Label the periodic coupling risk section and record the label number, start and end time, matching score, control threshold period parameters and spectrum characteristics for subsequent call in the control process.

4. The intelligent optimization control method for hydraulic press based on big data analysis according to claim 1 is characterized in that: Step S3 includes: Determine whether the current control cycle is in a high coupling risk section and start the hysteresis control module; The corresponding delay time window is set according to the coupling risk level, and the delay time is dynamically adjusted within a limited range; A delay control link is introduced into the original energy-saving switching decision logic, and the energy-saving switching instruction is executed only when the delay time ends and the decision condition is still met; The execution status and control effect of each hysteresis are recorded and fed back to the strategy optimization module to update the hysteresis parameters to achieve dynamic regulation.

5. The intelligent optimization control method for hydraulic press based on big data analysis according to claim 1 is characterized in that: Step S4 includes: Collect energy efficiency-related signals during the operation of the hydraulic press and extract short-term and long-term trend data through a sliding window; Determine whether the energy efficiency trend continues to move toward energy saving entry or exit within a set time range and meets the conditions of crossing the preset threshold range; Establish positive confirmation criteria and reverse confirmation criteria respectively. The positive confirmation criteria is used to determine whether the energy-saving mode entry conditions are met, and the reverse confirmation criteria is used to determine whether the energy-saving mode exit conditions are met. The energy-saving mode switching instruction is executed only when both the positive confirmation criterion and the reverse confirmation criterion meet the set requirements.

6. The intelligent optimization control method for hydraulic press based on big data analysis according to claim 1 is characterized in that: Step S5 includes: When the energy saving mode is enabled, the energy saving state holding counter is started and the energy saving start time is recorded; Set the corresponding minimum dwell time according to the current operating conditions of the hydraulic press to keep the state locked; Shield the energy-saving exit trigger logic within the minimum residence time and cache the monitored exit judgment signal; After the minimum residence time ends, the state lock is released, and it is determined whether to execute the energy-saving mode exit instruction according to the cache signal.

7. The intelligent optimization control method for hydraulic press based on big data analysis according to claim 1 is characterized in that: Step S6 includes: Record the resonance tag events and corresponding energy efficiency feedback results during the energy-saving control process, and establish a historical database; Analyze the correlation characteristics between control parameters and energy efficiency performance, and extract the optimal configuration combination of judgment parameters and weight coefficients; Dynamically adjust the judgment parameters and weight coefficients based on historical results, and limit the range of parameter changes to prevent strategy fluctuations; Establish a policy version control mechanism, generate version records for each policy adjustment, and maintain the policy evolution path during system operation.

Images