Energy-saving control system and method of energy-saving generator set

Abstract

The invention relates to the technical field of generator set energy-saving control, and discloses an energy-saving control system and method of an energy-saving generator set. The system comprises a vibration monitoring module, a load demand prediction module, an energy efficiency optimization decision module, an abnormal state recognition module and a dynamic regulation and control execution module. The vibration monitoring module collects mechanical vibration intensity in real time; the load demand prediction module generates a load demand prediction curve by using a dynamic load distribution algorithm; the energy efficiency optimization decision module determines an operation mode switching strategy through a multi-objective optimization algorithm; the abnormal state recognition module analyzes vibration frequency domain characteristics to evaluate equipment health; and the dynamic regulation and control execution module adjusts fuel supply and power output through a closed-loop control algorithm according to the strategy and the evaluation index. The method can accurately predict the load, optimize the operation mode, monitor the equipment health, effectively improve the energy utilization efficiency of the generator set, guarantee the stable operation of the equipment, reduce the cost, and adapt to various working conditions and new energy development requirements.

Description

Technical Field

[0001] The present invention relates to the technical field of energy-saving control of generator sets, and in particular to an energy-saving control system and method for energy-saving generator sets. Background Art

[0002] Generator sets, as key equipment in the energy supply system, are widely used in various fields such as industrial production, power supply, and emergency power generation. With the increasingly tense global energy situation and the continuous improvement of energy conservation and emission reduction requirements, energy-saving control of generator sets has become a major issue that needs to be addressed urgently.

[0003] Traditional generator sets generally have low energy efficiency during operation. For one thing, their operating modes are often relatively fixed, making it difficult to flexibly adjust to dynamic changes in actual load. For example, in industrial production scenarios, the power load of production equipment changes frequently as the production process progresses. If generator sets continue to operate in high-power mode during low-load phases, a large amount of energy is wasted simply to maintain operation, resulting in inefficient energy utilization and increased production costs. Furthermore, the lack of accurate load forecasting methods makes it impossible for generator sets to plan energy supply in advance. During peak and trough periods of power supply, the inability to accurately predict load fluctuations can lead to power shortages due to insufficient power generation, or energy waste due to excessive power generation, seriously impacting the stability and economic viability of power supply.

[0004] Traditional generator sets also have numerous shortcomings when it comes to equipment health management. Existing monitoring methods mostly focus solely on equipment operating parameters, paying insufficient attention to potential fault signals such as mechanical vibration. Mechanical vibration is a key indicator of generator set health. If problems such as loose components or wear occur within the equipment, vibration intensity and frequency will change. However, traditional monitoring systems are unable to detect these subtle changes in a timely manner, failing to provide early warning and address faults. This not only worsens equipment failures, increasing repair costs and downtime, but can also lead to safety incidents, posing a threat to human life and property.

[0005] Furthermore, the various control links in the generator set's control process lack effective coordination. Fuel supply systems, power output regulation systems, and other systems often operate independently, failing to coordinate unified control based on the overall operating status of the equipment. This decentralized control approach makes it difficult for the generator set to achieve optimal performance under complex operating conditions, hindering its full energy-saving potential.

[0006] With the rapid development of new energy sources such as solar energy and wind energy, the power supply pattern has become increasingly diversified. Generator sets need to operate better in coordination with new energy sources to cope with the changes in the energy structure. The traditional energy-saving control technology of generator sets is difficult to adapt to this new development trend, and there is an urgent need for a more advanced and intelligent energy-saving control system and method to meet the requirements of efficient energy utilization, stable operation of equipment, and coordinated development with new energy sources. Summary of the Invention

[0007] The purpose of the present invention is to provide an energy-saving control system and method for an energy-saving generator set to solve the problems raised in the above background technology.

[0008] To achieve the above purpose, the present invention provides the following technical solution: An energy-saving control system for an energy-saving generator set, the system includes:

[0009] Vibration monitoring module: used to collect the mechanical vibration intensity of the generator set in real time;

[0010] Load demand prediction module: According to historical operation data and real-time load changes, generate a load demand prediction curve for the future period through a dynamic load distribution algorithm;

[0011] Energy efficiency optimization decision module: Based on the load demand prediction curve, use a multi-objective optimization algorithm to determine the operation mode switching strategy of the generator set, and the strategy includes power output gradient, energy utilization rate threshold, and equipment start-stop priority;

[0012] Abnormal state identification module: Perform frequency domain feature analysis on the mechanical vibration intensity, and generate an equipment health status evaluation index in combination with preset vibration reference parameters;

[0013] Dynamic regulation execution module: According to the operation mode switching strategy and health status evaluation index, adjust the fuel supply rate and power output through a closed-loop control algorithm to generate a real-time control instruction sequence.

[0014] Preferably, the implementation steps of the dynamic load distribution algorithm include:

[0015] Perform moving average statistics on historical operation data according to a time window, and extract the periodic characteristics of load changes;

[0016] Based on the fuzzy logic algorithm, process real-time load fluctuation data to generate a load fluctuation tolerance interval;

[0017] Overlay and fuse the periodic characteristics with the fluctuation tolerance interval to construct a load demand prediction curve for the future period;

[0018] Dynamically divide the multi-level power output gears of the generator set according to the prediction curve.

[0019] Preferably, the method for setting parameters of the multi-objective optimization algorithm includes:

[0020] Define the trade-off coefficient between power output and fuel consumption according to the energy utilization rate threshold, traverse different operation mode combinations based on the genetic algorithm, and screen candidate strategies that meet the preset energy efficiency constraints;

[0021] Conduct a stability simulation test on the candidate strategies, and select the operation mode switching strategy with the highest comprehensive score.

[0022] Preferably, the execution steps of the frequency domain feature analysis include:

[0023] Perform wavelet transform decomposition on the mechanical vibration intensity data, and extract the energy distribution characteristics of different frequency bands;

[0024] Identify the abnormal vibration frequency band through the peak detection algorithm, and calculate its energy proportion and duration;

[0025] Associate the energy proportion and duration with the preset vibration reference parameters to generate an evaluation index for the equipment health status.

[0026] Preferably, the regulation logic of the closed-loop control algorithm includes:

[0027] Generate an initial adjustment amount of the fuel supply rate according to the operation mode switching strategy;

[0028] Dynamically correct the initial adjustment amount based on the health status evaluation index to generate an optimized fuel supply instruction;

[0029] Perform error compensation on the power output through a proportional-integral-derivative controller to generate a final control instruction sequence.

[0030] Preferably, the method for constructing the rule base of the fuzzy logic algorithm includes: defining the fuzzy input variables of the load fluctuation and the corresponding membership functions, generating the fuzzy inference rules for the load tolerance interval according to the expert experience base, and converting the inference result into the boundary values of the specific fluctuation tolerance interval through the defuzzification algorithm.

[0031] Preferably, the execution steps of the stability simulation test include:

[0032] Construct a multi-physics field coupling simulation model of the generator set to simulate the mechanical stress distribution under different operation modes;

[0033] Statistically analyze the overrun ratio and duration of the stress concentration area, and calculate the stability score;

[0034] Eliminate the candidate strategies with the risk of equipment overload according to the scoring results.

[0035] Preferably, the implementation steps of the peak detection algorithm include:

[0036] Perform sliding window mean calculation on the decomposed frequency band energy distribution to generate a dynamic energy baseline, detect the target frequency band exceeding the baseline threshold, and record its energy amplitude and occurrence timestamp;

[0037] Calculate the comprehensive index of abnormal vibration according to the occurrence frequency and amplitude weight of the target frequency band.

[0038] Preferably, the parameter adaptive method of the proportional-integral-derivative controller includes:

[0039] Train an error change rate prediction model based on the historical data of the power output error, and dynamically adjust the gain coefficients of the proportional, integral, and differential terms based on the prediction results.

[0040] Preferably, the present invention further includes an energy-saving control method for an energy-saving generator set, and the method includes the following steps:

[0041] S1: Real-time collect the mechanical vibration intensity of the generator set through the vibration monitoring module;

[0042] S2: Use the load demand prediction module to generate a load demand prediction curve for the future period through a dynamic load distribution algorithm according to the historical operation data and the real-time load change;

[0043] S3: With the help of the energy efficiency optimization decision module, based on the load demand prediction curve, determine the operation mode switching strategy of the generator set by using a multi-objective optimization algorithm, and the strategy includes the power output gradient, the energy utilization rate threshold, and the equipment start-stop priority;

[0044] S4: Use the abnormal state recognition module to perform frequency domain feature analysis on the mechanical vibration intensity, and generate an equipment health state evaluation index in combination with the preset vibration reference parameters;

[0045] S5: Rely on the dynamic regulation execution module to adjust the fuel supply rate and power output through a closed-loop control algorithm according to the operation mode switching strategy and the health state evaluation index, and generate a real-time control instruction sequence.

[0046] Compared with the prior art, the beneficial effects of the present invention are:

[0047] The energy-saving control system and method of the energy-saving generator set provided by the present invention show significant advantages in aspects such as improving energy utilization efficiency, equipment health management, and system collaborative control. In terms of energy utilization efficiency, the system, through the load demand prediction module, based on historical operation data and real-time load changes, uses a dynamic load distribution algorithm to generate an accurate load demand prediction curve. This curve provides a reliable basis for adjusting the operation mode of the generator set, enabling the unit to make preparations for energy allocation in advance. When it is predicted that the load is about to increase, the fuel supply can be increased in advance to ensure that the power output can keep up with the demand in a timely manner, avoiding production interruptions or power shortages caused by insufficient power. The energy efficiency optimization decision module, based on the load demand prediction curve, uses a multi-objective optimization algorithm to determine the operation mode switching strategy. This strategy covers key elements such as power output gradient, energy utilization rate threshold, and equipment start-stop priority, and can find the best combination of operation modes under different working conditions, effectively improving the energy utilization rate. During low-load periods, the system can reduce the power output gradient and reduce unnecessary energy consumption; during high-load periods, by optimizing the equipment start-stop sequence, it ensures that the unit operates in the most efficient way, thus realizing the rational use of energy and reducing operating costs.

[0048] In terms of equipment health management, the abnormal state recognition module conducts frequency-domain feature analysis on the mechanical vibration intensity of the generator set. Through wavelet transform decomposition and peak detection algorithms, it can accurately extract the energy distribution characteristics of different frequency bands, identify abnormal vibration frequency bands, and calculate their energy proportion and duration. Combining with the equipment health status evaluation index generated by preset vibration reference parameters, it can timely detect potential fault hazards of the equipment. Once abnormal vibration is detected, the system can arrange maintenance in advance to avoid the further deterioration of equipment failures, reduce maintenance costs and downtime, and improve the reliability and service life of the equipment.

[0049] In terms of system collaborative control, the dynamic regulation execution module adjusts the fuel supply rate and power output through a closed-loop control algorithm according to the operation mode switching strategy and health status evaluation index. This algorithm can not only generate an initial adjustment amount of the fuel supply rate according to the operation mode, but also dynamically correct this adjustment amount based on the health status evaluation index to ensure that the fuel supply matches the actual needs of the equipment. At the same time, the proportional-integral-derivative controller compensates for the power output error, realizing the collaborative work of each link such as the fuel supply system and the power output regulation system, enabling the generator set to operate stably and efficiently under various complex working conditions.

[0050] In addition, the energy-saving control system and method of the present invention have good adaptability. Whether in the scenario of frequent load changes in industrial production or in the case of obvious peak-valley differences in power supply, it can play its advantages of energy saving and stable operation. With the continuous development of new energy, this system can also provide technical support for the coordinated operation of generator sets and new energy, contributing to the optimization of the energy structure and sustainable development. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is the working principle diagram of the energy-saving control system of the energy-saving generator set of the present invention;

[0052] Figure 2 is the working flow chart of the dynamic load distribution algorithm;

[0053] Figure 3 is the flow chart of parameter setting and strategy screening of the multi-objective optimization algorithm;

[0054] Figure 4 is the flow chart of generating equipment health status evaluation indicators through frequency domain feature analysis. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0055] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0056] Please refer to Figures 1 - 4 , the present invention relates to an energy-saving control system and method for an energy-saving generator set, and its specific implementation will be elaborated in detail below.

[0057] During the operation of the generator set, the vibration monitoring module uses professional vibration sensors to collect the mechanical vibration intensity of the generator set in real time. These vibration sensors are reasonably installed at key parts of the generator set, such as the engine cylinder block, bearing seat, etc., to ensure that vibration information reflecting the operation state of the generator set can be accurately obtained. The collected vibration data is transmitted to the subsequent processing unit in the form of electrical signals, providing basic data for equipment health status evaluation.

[0058] The load demand prediction module collects the historical operation data of the generator set, which covers power output, load change conditions, etc. at different time periods. At the same time, it also obtains the current load change information in real time. On this basis, a load demand prediction curve for the future period is generated through a dynamic load distribution algorithm. This algorithm comprehensively considers the load change law in historical data and the real-time load fluctuation situation to improve the prediction accuracy.

[0059] Based on the curve generated by the load demand prediction module, the energy efficiency optimization decision-making module uses a multi-objective optimization algorithm to determine the operation mode switching strategy of the generator set. This strategy includes key elements such as power output gradient, energy utilization rate threshold, and equipment start-stop priority. The multi-objective optimization algorithm analyzes different combinations of operation modes and, on the premise of meeting the preset energy efficiency constraints, searches for the optimal operation mode switching plan to achieve efficient energy utilization and stable operation of the generator set.

[0060] After receiving the mechanical vibration intensity data collected by the vibration monitoring module, the abnormal state identification module conducts frequency-domain feature analysis on it. By converting the mechanical vibration intensity data from the time domain to the frequency domain and combining the preset vibration reference parameters, equipment health status evaluation indicators are generated. These evaluation indicators can intuitively reflect the health status of the generator set and provide a basis for subsequent regulation.

[0061] Based on the operation mode switching strategy determined by the energy efficiency optimization decision-making module and the health status evaluation indicators generated by the abnormal state identification module, the dynamic regulation execution module uses a closed-loop control algorithm to adjust the fuel supply rate and power output. In this way, a real-time control instruction sequence is generated to achieve precise control of the generator set and ensure its efficient and stable operation under different working conditions.

[0062] Next, the technical solution of the present invention will be further described in detail through specific embodiments.

[0063] Embodiment 1:

[0064] Regarding the specific implementation of the dynamic load distribution algorithm. The historical operation data is statistically analyzed by sliding average according to a time window. Assuming the time window length is , within each time window, the average value of the load data is calculated. Let the historical load data sequence be , , then the sliding average within the th time window is calculated by the formula: , where represents the time window length, represents the serial number of the time window, is the time index of the load data, and is the load data at the

[0065] Processing real-time load fluctuation data based on a fuzzy logic algorithm to generate a load fluctuation tolerance interval. First, define the fuzzy input variables of the load fluctuation and the corresponding membership functions. Let the load fluctuation amount be , which is fuzzified into fuzzy sets such as "negative large", "negative small", "zero", "positive small", "positive large", etc. The corresponding membership functions can use triangular functions, trapezoidal functions, etc. Generate fuzzy inference rules for the load tolerance interval according to the expert experience base. For example, if the load fluctuation is "positive small", the load tolerance interval is appropriately relaxed; if the load fluctuation is "negative large", the load tolerance interval is correspondingly tightened. Through defuzzification algorithms such as the centroid method, the maximum membership degree method, etc., convert the inference result into the boundary value of the specific fluctuation tolerance interval.

[0066] Superimpose and fuse the periodic characteristics with the fluctuation tolerance interval to construct a load demand prediction curve for the future period. Combine the extracted periodic characteristics with the load fluctuation tolerance interval in chronological order, fully considering the periodic changes and real-time fluctuations of the load, so as to construct a load demand prediction curve for the future period that better conforms to the actual situation.

[0067] Dynamically divide the multi-level power output gears of the generator set according to the prediction curve. For example, when the predicted load is low, divide the power output gears of the generator set into lower gears; when the predicted load is high, correspondingly increase the range of the power output gears. Through this dynamic division, the power output of the generator set can better match the load demand and improve the energy utilization efficiency.

[0068] Example 2:

[0069] Implementation of the parameter setting method of the multi-objective optimization algorithm. Define the trade-off coefficient between power output and fuel consumption according to the energy utilization rate threshold. Let the energy utilization rate threshold be , the power output be , the fuel consumption be , and the trade-off coefficient can be determined by the formula , where represents the energy utilization rate threshold, is the power output, is fuel consumption. By traversing different combinations of operating modes based on the genetic algorithm, the genetic algorithm simulates the selection, crossover, and mutation operations in the process of biological evolution. Encode different combinations of operating modes to form an initial population. In each iteration, evaluate the individuals in the population according to the fitness function, and the fitness function can be designed based on multiple objectives such as energy utilization rate and power output stability. Retain the individuals with higher fitness through the selection operation, generate new individuals through the crossover and mutation operations, and continuously evolve the population. Screen candidate strategies that meet the preset energy efficiency constraints, and the preset energy efficiency constraints can be that the energy utilization rate is not lower than a certain specific value, such as , only when the energy utilization rate of the combination of operating modes meets , will it be used as a candidate strategy.

[0070] Conduct a stability simulation test on the candidate strategy. Build a multi-physics coupling simulation model of the generator set, which considers the interaction of multiple physical fields such as the mechanical, thermal, and electrical fields of the generator set. Simulate the mechanical stress distribution under different operating modes in the model, and calculate the stress conditions of each component of the generator set through methods such as finite element analysis. Statistically analyze the overrun ratio and duration of the stress concentration area. Assume that the stress value in the stress concentration area is , the stress threshold is , and the overrun ratio The calculation formula of is , where represents the number of areas where the stress exceeds the threshold, represents the total number of areas counted. The duration is determined by recording the time period when the stress exceeds the threshold. Calculate the stability score based on these data. For example, the weighted sum method can be used to assign different weights and to the overrun ratio and duration respectively. The stability score , where is the duration of the stress concentration area exceeding the limit, , are the weight coefficients. Select the operating mode switching strategy with the highest comprehensive score to ensure the generator set operates stably while achieving high efficiency and energy conservation.

[0071] Example 3:

[0072] During the operation of the generator set, the vibration monitoring module continuously collects mechanical vibration intensity data. After the frequency domain feature analysis process is started, the first step is to perform wavelet transform decomposition on the mechanical vibration intensity data. Wavelet transform is a powerful time-frequency analysis tool that can dissect the signal at different time and frequency scales. Assume that the collected mechanical vibration intensity signal is , select an appropriate wavelet basis function (such as Daubechies wavelet) to perform wavelet transform on it, and decompose into multiple sub-signals of different frequency bands , where represents different frequency band ranges, and each frequency band corresponds to the vibration characteristics of different components or operating states of the generator set.

[0073] After the decomposition is completed, the energy distribution characteristics of different frequency bands need to be extracted. For each sub-signal , according to the basic principle of energy calculation, its energy is calculated by the formula . In actual calculation, since the collected data is discrete, numerical integration methods will be used to approximate the calculation. For example, using the trapezoidal integration method, assuming the discrete data points are , , is the sampling interval, then . In this way, the energy value of each frequency band can be accurately obtained, providing data support for subsequent analysis.

[0074] Next, identify the abnormal vibration frequency band through the peak detection algorithm. First, calculate the moving window mean of the decomposed frequency band energy distribution to generate a dynamic energy baseline. Set the length of the moving window to , in the th window, the average value of the frequency band energy is calculated according to the formula , where determines the size of the window, represents the window number, is the energy value of the frequency band at time . The window slides point by point on the energy distribution data, and the dynamic energy baseline at each position is calculated. Set the baseline threshold to , when the frequency band energy value exceeds this threshold, it is identified as the target frequency band, and record the energy amplitude of this target frequency band and the timestamp when it appears. In order to more comprehensively evaluate the abnormal vibration situation, calculate the comprehensive index of abnormal vibration according to the occurrence frequency of the target frequency band and the amplitude weight . The calculation formula is . The amplitude weight is preset according to actual experience and equipment characteristics, and the frequency

[0075] Finally, associate the energy proportion and the duration with preset vibration reference parameters to generate an evaluation index for the equipment health status. Calculate the proportion of the energy in the abnormal vibration frequency band to the total energy , and the formula is , where is the energy in the abnormal vibration frequency band, represents the total energy of all frequency bands. At the same time, record the duration of the abnormal vibration frequency band . Compare the energy proportion and the duration with preset vibration reference parameters, such as the energy proportion threshold and the duration threshold . If and , it indicates that there is a greater risk of equipment failure, and the generated evaluation index for the equipment health status will output corresponding warning information; if and partially exceed the threshold or neither exceeds the threshold, the health status of the equipment will be classified according to the specific situation, providing a basis for equipment maintenance and operation adjustment.

[0076] Example 4:

[0077] When the generator set is operating, the closed-loop control algorithm starts to take effect. First, generate an initial adjustment amount for the fuel supply rate according to the operation mode switching strategy. Assume that the operation mode switching strategy determines the target value of the power output , and the actual power output of the current generator set is , and the difference between the two reflects the deviation of the power output. Combine the pre-established power-fuel supply relationship model of the generator set to calculate the initial adjustment amount of the fuel supply rate. For example, a common power-fuel supply relationship model can be expressed as , where and are model parameters obtained through experiments or theoretical analysis according to the characteristics of the generator set, is the power output, is the fuel supply rate. Transform this model to get . Through this formula, the amount by which the fuel supply rate needs to be adjusted to make the power output reach the target value can be initially determined.

[0078] However, relying solely on the initial adjustment amount is not precise enough, and the initial adjustment amount needs to be dynamically corrected based on the evaluation index of the health status. When the evaluation index of the equipment health status given by the abnormal status identification module shows that there are certain abnormalities in the equipment, such as abnormal mechanical vibration, this may affect the combustion efficiency and energy conversion efficiency of the generator set. At this time, introduce a correction coefficient Correct the initial adjustment amount. The correction coefficient is determined according to the specific situation of the health status evaluation index, usually by establishing a mapping relationship. For example, when the degree of vibration abnormality is greater, it indicates that the response of the equipment to fuel supply may be more unstable, the value is smaller to avoid over-adjusting the fuel supply and causing the equipment to operate more unstably. The optimized fuel supply instruction can better fit the current actual operating condition of the equipment and ensure the accuracy of fuel supply.

[0079] Finally, the power output is compensated for errors through a proportional-integral-derivative (PID) controller to generate the final control instruction sequence. The PID controller uses the power output error as the input. Its output is calculated by the formula where is the proportional coefficient, which adjusts the control output proportionally according to the magnitude of the current error to enable the system to respond quickly to error changes; is the integral coefficient, and the integral term is used to accumulate past errors and eliminate the steady-state error of the system; is the derivative coefficient, reflects the rate of change of the error, and the derivative term can adjust the control output in advance according to the trend of error change to enhance the stability of the system. In actual operation, by continuously adjusting , and values, the power output can quickly and stably track the target value, and finally generate a series of precise control instructions to achieve precise control of the fuel supply and power output of the generator set and ensure the efficient and stable operation of the generator set.

[0080] Example 5:

[0081] When constructing the rule base of the fuzzy logic algorithm, first define the fuzzy input variables of load fluctuation and the corresponding membership functions. Taking the load fluctuation amount as the key fuzzy input variable, it is fuzzified and divided into multiple fuzzy sets such as "negative large", "negative small", "zero", "positive small", "positive large", etc. For the "negative large" fuzzy set, a trapezoidal membership function is used to describe its membership degree. Let the parameters of this trapezoidal membership function be , , , , then the expression of the membership function is:

[0082]

[0083] In this formula, Indicates the load fluctuation amount Degree of membership belonging to the "negative large" fuzzy set 、 、 、 Are parameters that determine the trapezoidal shape and Define the transition interval from completely belonging to the "negative large" fuzzy set to not completely belonging and Then specify the range from not belonging to the "negative large" fuzzy set to completely not belonging. In a similar way, membership functions for other fuzzy sets such as "negative small", "zero", "positive small", "positive large" are defined respectively, and these functions together constitute a fuzzy description system for the load fluctuation amount

[0084] Next, fuzzy inference rules for the load tolerance interval are generated based on the expert experience database. The expert experience database is established based on long-term observation and analysis of the operating data of the generator set and practical operation experience. For example, Rule 1: If the load fluctuation is "positive small" and the current load is close to the upper limit , then the load tolerance interval is appropriately tightened. This is because when the load is already close to the upper limit and there is still a small positive fluctuation, in order to ensure the safe and stable operation of the equipment, it is necessary to strictly control the load change range. Rule 2: If the load fluctuation is "negative large" and the current load is close to the lower limit , then the load tolerance interval is appropriately relaxed. Because when the load is close to the lower limit and there is a large negative fluctuation, appropriately relaxing the tolerance interval can avoid unnecessary frequent adjustments. These rules are presented in the form of "if... then...", closely linking the load fluctuation situation, the current load state with the adjustment strategy of the load tolerance interval

[0085] Finally, the inference result is converted into the boundary values of the specific fluctuation tolerance interval through the defuzzification algorithm. Here, the centroid method is used for defuzzification. Let the fuzzy set of the load tolerance interval obtained from fuzzy inference be , then the calculation formula for the boundary values of the specific fluctuation tolerance interval is . In actual calculation, since the fuzzy set is discrete, the integral operation will be converted into a summation operation for approximate calculation. Through this defuzzification operation, the fuzzy result obtained from fuzzy inference is converted into specific numerical values, and these values are the key parameters for subsequent load demand prediction and power output gear division, enabling the system to make more reasonable decisions based on the actual situation

[0086] Example 6:

[0087] The proportional-integral-derivative (PID) controller plays a crucial role in the control of generator sets. The adaptive adjustment of its parameters can significantly improve the control effect. First, a prediction model for the rate of change of error is trained based on the historical data of power output error. During the operation of the generator set, the historical data of power output error is continuously recorded For example, at time points , , , , the collected error data is , , , . The neural network algorithm in machine learning is used to train the prediction model for the rate of change of error. A neural network with a suitable structure is constructed, such as a multi-layer perceptron (MLP) containing an input layer, a hidden layer, and an output layer. The historical data of power output error is used as the input of the neural network. After non-linear transformation by the hidden layer, the predicted rate of change of error is output . During the training process, the weights and biases of the neural network are continuously adjusted through a large amount of historical data, so that the predicted rate of change of error is as close as possible to the actual value. Common training methods include the backpropagation algorithm, etc.

[0088] Based on the prediction results, the gain coefficients of the proportional, integral, and derivative terms are dynamically adjusted. When the predicted rate of change of error is large, it means that the power output error changes rapidly, and the system needs to respond quickly to correct the error. At this time, the proportional coefficient is appropriately increased to make the controller more sensitive to the current error and accelerate the adjustment speed of the system; at the same time, the integral coefficient is decreased because when the error changes rapidly, the rapid accumulation of the integral term may cause system overshoot, and reducing can avoid this situation. On the contrary, when the predicted rate of change of error is small, it indicates that the system error changes slowly. At this time, the proportional coefficient can be appropriately reduced to prevent the system from over-adjusting; the integral coefficient is increased to strengthen the cumulative effect of past errors and improve the steady-state accuracy of the system, making the power output closer to the target value. For the gain coefficient of the derivative term, it is adjusted according to the change trend of the rate of change of error. If the rate of change of error gradually decreases, it indicates that the system has a tendency to stabilize, and is appropriately reduced to avoid excessive interference of the derivative term on the stable process; if the rate of change of error has an increasing trend, then is appropriately increased, enhance the damping effect of the system and suppress the further expansion of errors. By this method of dynamically adjusting the parameters of the PID controller according to the prediction results, the controller can better adapt to the operating requirements of the generator set under different working conditions, effectively improve the control effect, and ensure the stable and efficient operation of the generator set.

[0089] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0090] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An energy-saving control system for an energy-saving generator set, characterized in that, Including: Vibration monitoring module: used to collect the mechanical vibration intensity of the generator set in real time; Load demand prediction module: according to historical operation data and real-time load changes, generate a load demand prediction curve for the future period through a dynamic load distribution algorithm; Energy efficiency optimization decision module: based on the load demand prediction curve, use a multi-objective optimization algorithm to determine the operation mode switching strategy of the generator set, and the strategy includes power output gradient, energy utilization rate threshold, and equipment start-stop priority; Abnormal state identification module: perform frequency-domain feature analysis on the mechanical vibration intensity, and generate an equipment health status evaluation index in combination with preset vibration reference parameters; Dynamic regulation execution module: according to the operation mode switching strategy and health status evaluation index, adjust the fuel supply rate and power output through a closed-loop control algorithm, and generate a real-time control instruction sequence.

2. The energy-saving control system of the energy-saving generator set according to claim 1, characterized in that, The implementation steps of the dynamic load distribution algorithm include: Perform moving average statistics on historical operation data according to a time window, and extract the periodic characteristics of load changes; Process real-time load fluctuation data based on a fuzzy logic algorithm to generate a load fluctuation tolerance interval; Overlay and fuse the periodic characteristics and the fluctuation tolerance interval to construct a load demand prediction curve for the future period; Dynamically divide the multi-level power output gears of the generator set according to the prediction curve.

3. The energy-saving control system of the energy-saving generator set according to claim 1, characterized in that, The parameter setting method of the multi-objective optimization algorithm includes: Define the trade-off coefficient between power output and fuel consumption according to the energy utilization rate threshold, traverse different operation mode combinations based on a genetic algorithm, and screen candidate strategies that meet the preset energy efficiency constraints; Perform stability simulation tests on the candidate strategies, and select the operation mode switching strategy with the highest comprehensive score.

4. The energy-saving control system of the energy-saving generator set according to claim 1, characterized in that, The execution steps of the frequency-domain feature analysis include: Perform wavelet transform decomposition on the mechanical vibration intensity data, and extract the energy distribution characteristics of different frequency bands; Identify abnormal vibration frequency bands through a peak detection algorithm, and calculate their energy proportion and duration; Associate the energy proportion and duration with preset vibration reference parameters to generate an equipment health status evaluation index.

5. The energy-saving control system of the energy-saving generator set according to claim 1, characterized in that The regulation logic of the closed-loop control algorithm includes: Generate an initial adjustment amount of the fuel supply rate according to the operation mode switching strategy; Dynamically correct the initial adjustment amount based on the health status evaluation index to generate an optimized fuel supply instruction; Perform error compensation on the power output through a proportional-integral-derivative controller to generate a final control instruction sequence.

6. The energy-saving control system of the energy-saving generator set according to claim 2, characterized in that, The method for constructing the rule base of the fuzzy logic algorithm includes: defining fuzzy input variables of load fluctuations and corresponding membership functions, generating fuzzy inference rules for the load tolerance interval according to an expert experience library, and converting the inference result into specific fluctuation tolerance interval boundary values through a defuzzification algorithm.

7. The energy-saving control system of the energy-saving generator set according to claim 3, characterized in that The execution steps of the stability simulation test include: Construct a multi-physical field coupling simulation model of the generator set to simulate the mechanical stress distribution under different operation modes; Statistical overrun ratio and duration of stress concentration areas, and calculate the stability score; Eliminate candidate strategies with equipment overload risks according to the scoring results.

8. The energy-saving control system of the energy-saving generator set according to claim 4, characterized in that, The implementation steps of the peak detection algorithm include: Perform sliding window mean calculation on the decomposed frequency band energy distribution to generate a dynamic energy baseline, detect the target frequency band exceeding the baseline threshold, and record its energy amplitude and the occurrence timestamp. Calculate the comprehensive index of abnormal vibration according to the occurrence frequency and amplitude weight of the target frequency band.

9. The energy-saving control system of the energy-saving generator set according to claim 5, characterized in that, The parameter adaptive method of the proportional-integral-derivative controller includes: Train the error change rate prediction model based on the historical data of the power output error, and dynamically adjust the gain coefficients of the proportional, integral, and derivative terms based on the prediction results.

10. A method for energy-saving control of an energy-saving generator set, characterized in that, It includes the following steps: S1: Real-time collect the mechanical vibration intensity of the generator set through the vibration monitoring module. S2: Use the load demand prediction module to generate a load demand prediction curve for the future period through the dynamic load distribution algorithm according to the historical operation data and the real-time load change. S3: With the help of the energy efficiency optimization decision module, based on the load demand prediction curve, use the multi-objective optimization algorithm to determine the operation mode switching strategy of the generator set, and the strategy includes the power output gradient, the energy utilization rate threshold, and the equipment start-stop priority. S4: Use the abnormal state recognition module to perform frequency domain feature analysis on the mechanical vibration intensity, and generate an equipment health state evaluation index in combination with the preset vibration reference parameters. S5: Rely on the dynamic regulation execution module to adjust the fuel supply rate and power output through the closed-loop control algorithm according to the operation mode switching strategy and the health state evaluation index, and generate a real-time control instruction sequence.

Images