Ship generator set power load control method and system

Abstract

The invention relates to the technical field of ship power systems, in particular to a ship generator set power load control method and system, and the method comprises the following steps: collecting a ship total power load signal; the total power load signal is processed, basic load information and instantaneous disturbance information are obtained, the basic load information is obtained by smoothing the total power load signal, and the instantaneous disturbance information is obtained through the difference value between the total power load signal and the basic load information; on the basis of the basic load information, starting and stopping decisions of the generator set are made, and on-line power fine adjustment of the generator set is carried out on the basis of the instantaneous disturbance information; through the technical scheme, the basic load and the instantaneous disturbance in the total power load of the ship can be effectively distinguished, and different control strategies are adopted in a targeted manner, so that the problem of frequent start and stop or low-efficiency operation caused by complex load characteristics in a traditional method is avoided.

Description

Technical Field

[0001] This invention relates to the field of marine power systems, and more particularly to a method and system for controlling the power load of a marine generator set. Background Technology

[0002] In modern ship electrical systems, generator sets are the core power source, and the effectiveness of their power load control directly affects the ship's navigation safety and operational efficiency. Traditional power management systems typically rely on preset load thresholds to control the start-up and shutdown of generator sets and load distribution. However, in actual navigation, especially when the ship carries a large number of intermittent high-power loads (such as refrigeration units for refrigerated containers), the composition of the ship's total electrical load becomes complex, exhibiting a combination of a stable base load and numerous randomly occurring load spikes that are short in duration but have very high peak values.

[0003] Existing technologies exhibit significant flaws in their control logic when dealing with such complex load characteristics. If the system strictly adheres to the total electrical load threshold, it may misinterpret transient load spikes as sustained increases, leading to frequent generator starts and stops, fuel waste, and accelerated wear on critical equipment such as starter motors and circuit breakers. Conversely, manually starting the generator for extended periods to avoid frequent starts and stops results in prolonged low-load operation. Under low load, diesel generators experience incomplete combustion, producing substantial carbon deposits, accelerating wear, and deteriorating engine oil, causing irreversible damage and significantly increasing maintenance costs. Therefore, existing marine generator load control methods struggle to effectively distinguish between the fundamental and transient components of the ship's total electrical load, failing to strike a balance between preventing frequent generator starts and stops and ensuring healthy and efficient operation.

[0004] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a method and system for controlling the electrical load of a ship generator set.

[0006] In a first aspect, the present invention provides a method for controlling the electrical load of a ship generator set, the method comprising the following steps: Collect the ship's total electrical load signal; The total power load signal is processed to obtain basic load information and instantaneous disturbance information. The basic load information is obtained by smoothing the total power load signal, and the instantaneous disturbance information is obtained by the difference between the total power load signal and the basic load information. Based on the basic load information, start-up and shutdown decisions for the generator set are made, and based on the instantaneous disturbance information, online power fine-tuning of the generator set is performed.

[0007] Through this technical solution, this application can effectively distinguish between the basic load and instantaneous disturbance in the total electrical load of a ship, and adopt different control strategies accordingly, thereby avoiding the problems of frequent start-stop or inefficient operation caused by the complex load characteristics in traditional methods, and realizing the healthy and efficient operation of the generator set.

[0008] The core innovation of this application lies in the introduction of a refined decomposition mechanism for the ship's total electrical load signal. By smoothing the total electrical load signal, this method can accurately separate the "basic load information" reflecting the ship's long-term, stable power demand trend. At the same time, through difference calculation, it can accurately extract the "instantaneous disturbance information" reflecting short-term, rapid changes. This dual information extraction mechanism is the key difference between this application and existing technologies.

[0009] Secondly, a ship generator set power load control system is provided, the system comprising: The acquisition module is used to acquire the ship's total electrical load signal; The processing module is used to process the total power load signal to obtain basic load information and instantaneous disturbance information. The basic load information is obtained by smoothing the total power load signal, and the instantaneous disturbance information is obtained by the difference between the total power load signal and the basic load information. The start / stop decision module is used to make start / stop decisions for the generator set based on the basic load information. The power fine-tuning module is used to perform online power fine-tuning of the generator set based on the instantaneous disturbance information.

[0010] Compared with the prior art, the present invention has the following beneficial effects: By collecting the ship's total electrical load signal and processing it to separate the basic load information and instantaneous disturbance information, a refined identification of the ship's electrical load is achieved. Specifically, this method obtains the basic load information by smoothing the total electrical load signal and calculates the difference between the total electrical load signal and the basic load information to obtain the instantaneous disturbance information. This effectively distinguishes between the stable and continuous basic load in the ship's electrical system and the randomly occurring, short-duration but high-peak instantaneous load spikes.

[0011] Based on this accurate load decomposition, this application enables targeted generator start-up and shutdown decisions and online power fine-tuning of generator sets. For basic load information, the system makes generator start-up and shutdown decisions, avoiding the frequent start-ups and shutdowns caused by misinterpreting short-term load spikes as continuous increases, as is common in traditional methods. This significantly reduces fuel consumption and wear on critical equipment such as starter motors and circuit breakers. For instantaneous disturbance information, the system performs online power fine-tuning of generator sets, ensuring that generator sets can adjust their power output promptly and accurately when dealing with rapidly changing loads. This avoids grid instability caused by insufficient power and prevents multiple generator sets from operating at low loads for extended periods to cope with instantaneous spikes. This effectively solves problems such as incomplete combustion, carbon buildup, accelerated wear, and oil deterioration caused by diesel generators operating under low loads, significantly reducing maintenance costs. Attached Figure Description

[0012] Figure 1 This is a flowchart of the method of the present invention.

[0013] Figure 2 This is a schematic diagram of the system structure of the present invention.

[0014] In the diagram: 201, data acquisition module; 202, processing module; 203, start / stop decision module; 204, power fine-tuning module. Detailed Implementation

[0015] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0016] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0017] like Figure 1 The method for controlling the electrical load of a ship generator set, as shown, includes the following steps: S101. Collect the ship's total electrical load signal; This step is fundamental to the entire control method. Specifically, it can be achieved by installing high-precision power sensors at the ship's main switchboard or main load circuits. For example, Hall effect current and voltage sensors can be used to convert the acquired analog signals into digital signals via an analog-to-digital converter (ADC) and transmit them to the central processing unit at a preset sampling frequency (e.g., 100 times per second). Alternatively, digitized total power load data can be directly obtained through smart meters or power monitoring units. This data typically includes parameters such as instantaneous power, current, and voltage, and is transmitted via industrial Ethernet or serial communication interfaces.

[0018] S102. Process the total power load signal to obtain basic load information and instantaneous disturbance information. The basic load information is obtained by smoothing the total power load signal, and the instantaneous disturbance information is obtained by the difference between the total power load signal and the basic load information. Specifically, to obtain "baseline load information," the total power load signal can be smoothed. For example, a simple moving average filter can be used to take the average of the total power load signal over a past period (e.g., 5 minutes) as the current baseline load information. Another approach is to use an exponentially weighted moving average (EWMA) algorithm, assigning higher weights to recent data to more sensitively reflect the slow changing trend of the baseline load. After obtaining the baseline load information, "instantaneous disturbance information" can be obtained by the difference between the total power load signal and the baseline load information. For example, in each sampling period, the current total power load signal is subtracted from the currently calculated baseline load information; the difference is considered the instantaneous disturbance information. As an alternative, a high-pass filter can also be used to process the total power load signal, directly filtering out low-frequency baseline load components to obtain high-frequency instantaneous disturbance information.

[0019] S103. Based on the basic load information, make start-up and shutdown decisions for the generator set, and based on the instantaneous disturbance information, perform online power fine-tuning of the generator set.

[0020] The power load control method for ship generator sets disclosed in this application aims to optimize the operating efficiency and stability of the ship's power system. This method is typically deployed within the ship's power management system or energy management system, implemented through a computer system integrated into the ship's central control room or a dedicated embedded controller.

[0021] Specifically, the "total ship power load signal" refers to the overall power consumption data of the ship, collected in real time by sensors such as current transformers and voltage transformers in the ship's electrical system. It reflects the total power demand of all electrical equipment on board. This signal is typically collected in analog or digital form and periodically transmitted to the processing unit. "Basic load information" refers to the load component extracted from the total ship power load signal, reflecting the ship's long-term, stable power demand. It represents the ship's basic power consumption under normal navigation or berthing conditions, such as continuous loads like lighting, navigation equipment, and living facilities. "Instantaneous disturbance information" refers to the load component remaining after deducting the basic load information from the total ship power load signal, reflecting short-term, rapidly changing loads. These disturbances are usually caused by the instantaneous connection or disconnection of large intermittent loads (such as cranes, pumps, refrigeration compressors, propeller pitch, etc.), characterized by rapid changes, short duration, but high peak values. "Smoothing processing" is a data processing technique designed to eliminate high-frequency noise and instantaneous fluctuations in the signal, extracting the long-term trend or low-frequency components. In power load control, smoothing processing can be used to separate basic load information from the original total power load signal. "Generator start-stop decision-making" refers to the process of deciding whether to start standby generators or shut down online generators based on changes in the ship's power demand. Reasonable start-stop decisions can avoid frequent generator starts and stops, reducing fuel consumption and equipment wear. "Online generator power fine-tuning" refers to making small, rapid adjustments to the power output of currently operating generators to respond to instantaneous load changes and maintain the stability of grid frequency and voltage.

[0022] This method first acquires real-time overall power demand data of the ship by "collecting the ship's total electrical load signal." These raw signals contain slow changes in the base load and rapid fluctuations in transient disturbances. Subsequently, the collected signals undergo crucial "processing." Through "smoothing" technology, high-frequency transient fluctuations in the total electrical load signal are filtered out, thereby accurately extracting the "base load information" that reflects the ship's long-term, stable power demand trend. Simultaneously, by performing a difference operation between the raw total electrical load signal and the extracted base load information, "transient disturbance information" can be accurately separated. This information characterizes short-term, rapid load spikes or drops in the ship's power system caused by the start-up and shutdown of large equipment.

[0023] Therefore, this method decomposes the complex load control problem into two relatively independent but interrelated sub-problems. On the one hand, it "makes generator start-up and shutdown decisions based on basic load information." This means that the start-up or shutdown of generator sets is no longer subject to instantaneous, short-term load spikes, but rather judged based on the ship's long-term, stable power demand trends. For example, only when the basic load information continuously rises and reaches a certain level will it be determined that a standby generator set needs to be started, thus effectively avoiding frequent start-ups and shutdowns caused by instantaneous load fluctuations, significantly reducing fuel consumption and equipment wear. On the other hand, it "performs online generator set power fine-tuning based on instantaneous disturbance information." This means that for rapid, short-term load changes, the system can quickly identify and only make small power adjustments to the currently operating generator sets to quickly respond to load changes and maintain the stability of the grid frequency and voltage. This fine-tuning mechanism avoids unnecessary generator set start-ups and shutdowns while ensuring the continuity and quality of power supply.

[0024] In summary, by decomposing the total power load signal into basic load information and instantaneous disturbance information, and then performing start-stop decisions and power fine-tuning separately, this method achieves refined and intelligent control of ship generator sets. Basic load information is used for macroscopic resource allocation, ensuring that generator sets operate within a reasonable load range and avoiding inefficiency and excessive wear; instantaneous disturbance information is used for microscopic dynamic response, ensuring grid stability and responding to sudden load shocks. This collaborative control mechanism enables generator sets to operate more healthily and efficiently, significantly improving the overall performance and reliability of the ship's power system.

[0025] As one embodiment of the present invention, before making the start-up and shutdown decision of the generator set based on the basic load information, the method further includes the following steps: When the basic load information continues to rise and exceeds the preset startup threshold, a load nature judgment is performed; Load characteristics determination includes: Send a power increase command to the currently operating generator set, so that the operating generator set increases its power output by a preset amount within a preset time; Monitor frequency changes in the ship's electrical grid during periods of increased generator power; Based on the frequency changes, determine whether the increase in the base load information is caused by the actual increase in the base load; If the increase is determined to be caused by a genuine increase in base load, the standby generator set will be started. If the increase is determined to be caused by a non-genuine increase in base load, the standby generator set will not be started, and the power output of the generator set in operation will be reduced.

[0026] Specifically, when the baseline load information is detected to be continuously rising and its value exceeds the preset start-up threshold, the system will not immediately start the backup generator set, but will trigger a load nature judgment process. Here, "continuous rise" can be understood as the baseline load information showing a stable growth trend within a certain time window, rather than a short-term fluctuation; "preset start-up threshold" is a load level pre-set based on factors such as the design capacity of the ship's power system, the rated power of the generator set, and safety margin. When the baseline load reaches or exceeds this threshold, it usually means that it is necessary to consider increasing the power generation capacity.

[0027] During the load characteristic assessment process, a power increase command is first sent to the currently operating generator sets. This command aims to cause the operating generator sets to increase their power output by a preset amount within a predetermined time period, such as several seconds to tens of seconds. The purpose of this is to test the ship's electrical grid response to current load changes by actively increasing power generation. For example, the preset amount could be 5% to 10% of the rated power of the currently operating generator sets, and the preset time could be 5 seconds.

[0028] During periods of increased generator power, the frequency changes of the ship's power grid are continuously monitored. Grid frequency is a key indicator of the balance between load and generation in a power system. When load increases but generation is insufficient, the frequency decreases; conversely, when generation is excessive, the frequency increases. Monitoring frequency changes indirectly reflects the actual demand of the current base load. Subsequently, based on the monitored frequency changes, it is determined whether the increase in base load information is caused by a genuine increase in base load. Specifically, if the online generators increase power output, but the frequency of the ship's power grid continues to decrease or only slightly increases, this usually indicates that the ship's actual base load is indeed continuously increasing, and the existing generation capacity is insufficient to meet the demand. Conversely, if after the online generators increase power output, the frequency of the ship's power grid quickly recovers and stabilizes within the normal range, or even increases, it may indicate that the previous increase in base load information was not caused by a genuine, continuous load increase, but rather by instantaneous load fluctuations, measurement noise, or temporary adjustments within the system.

[0029] This application's solution effectively addresses the misjudgment problem that may arise from relying solely on smoothed baseline load information for start-up and shutdown decisions by introducing a proactive load characteristic judgment mechanism. When the baseline load information reaches the start-up threshold, the system no longer blindly starts standby generators. Instead, it verifies the authenticity of the load increase by sending power increase commands to online generators and monitoring the resulting grid frequency response. If the frequency response indicates that the existing generating capacity is still insufficient to maintain grid stability, then it confirms a genuine load increase and starts the standby generators. Conversely, if the frequency response shows that the existing generating capacity is sufficient, unnecessary startup is avoided. This "trial and error" verification method makes generator start-up and shutdown decisions more accurate and reliable.

[0030] In one specific implementation, suppose that after the ship's total electrical load signal has been smoothed, its baseline load information continues to rise and at a certain moment exceeds a preset start-up threshold, for example, from 500kW to 650kW, while the start-up threshold is set at 600kW. At this time, the system will not immediately start the standby generator set. Instead, the system will send a power increase command to the two generator sets currently operating online, for example, each generator set will increase its power output by 50kW within 5 seconds, for a total increase of 100kW.

[0031] During this period, the system will continuously monitor the frequency of the ship's power grid. Scenario 1: If, after the online generator sets increase their power output by 100kW, the frequency of the ship's power grid continues to drop from 59.8Hz to 59.5Hz, or only slightly rebounds before dropping again, this indicates that even with increased power generation, the actual load demand still exceeds the supply, meaning the increase in base load is real. In this case, the system will determine that the increase is caused by a real increase in base load and immediately start a standby generator set, connecting it to the grid to meet the continuously increasing power demand.

[0032] Scenario 2: If the frequency of the ship's power grid rapidly rises from 59.8Hz to 60.0Hz and remains stable, or even rises slightly, after the online generator sets increase their power output by 100kW, this indicates that the increased power from the online generator sets is sufficient to cope with the current increase in base load, or that the previous load increase was not a sustained, genuine increase. In this case, the system will determine that the increase is not due to a genuine increase in base load, will not activate the standby generator sets, and will restore the power output of the online generator sets to the level before the increase, thus avoiding unnecessary waste of resources.

[0033] As one embodiment of the present invention, the step of processing the total power load signal to obtain basic load information and instantaneous disturbance information includes: Continuously monitor the fluctuation range of the total power load signal within a preset reference time window; When the fluctuation amplitude exceeds the preset threshold, the smoothing time window is extended; when the fluctuation amplitude is below the preset threshold, the smoothing time window is shortened. The total power load signal is smoothed based on a smoothing time window to obtain basic load information. The difference between the latest acquired total power load signal and the basic load information is calculated to obtain the instantaneous disturbance information.

[0034] Specifically, continuous monitoring refers to the system acquiring the total power load signal in real time or at a preset sampling frequency during operation, and calculating its fluctuation amplitude within a preset reference time window. This fluctuation amplitude can be understood as the difference between the maximum and minimum values, standard deviation, mean absolute deviation, and other statistical quantities of the signal within a specific time period. Its purpose is to quantify the activity or instability of the current load signal. The preset reference time window is an initially set time length used to preliminarily assess the short-term dynamic characteristics of the load signal. When the monitored fluctuation amplitude exceeds a preset threshold, it indicates that the current load signal has large instantaneous changes or noise. In this case, to more effectively filter out these instantaneous disturbances and obtain more stable basic load information, the smoothing processing time window will be extended. Conversely, when the fluctuation amplitude is below the preset threshold, it indicates that the load signal is relatively stable. In this case, to improve the response speed to changes in the basic load, the smoothing processing time window will be shortened. This dynamic adjustment mechanism ensures the adaptability of the smoothing processing, allowing it to be optimized according to the dynamic characteristics of the actual load. In practical applications, smoothing processing can be implemented using various algorithms, such as moving average filtering, exponential smoothing filtering, and Kalman filtering. The selected smoothing algorithm operates based on a dynamically adjusted time window, processing the original total power load signal to effectively separate the stable component representing the ship's long-term or trend-based power demand, i.e., the base load information. Therefore, by subtracting the smoothed base load information from the most recently acquired total power load signal, instantaneous disturbance information can be accurately extracted. This instantaneous disturbance information reflects rapid, short-term, or sudden load changes superimposed on the base load, such as the instantaneous start-up or shutdown of large equipment, or the starting current surge of motors.

[0035] This application's solution overcomes the limitations of traditional fixed smoothing windows in handling variable load characteristics by introducing continuous monitoring of the total power load signal fluctuation amplitude and dynamically adjusting the smoothing processing time window based on the monitoring results. Specifically, when load fluctuations are severe, extending the smoothing time window can more effectively suppress instantaneous noise and short-term disturbances, making the extracted baseline load information more stable and accurate. This avoids misjudging baseline load changes due to instantaneous fluctuations, thus providing a more reliable basis for generator start-up and shutdown decisions. Conversely, when load fluctuations are small, shortening the smoothing time window can improve the system's response speed to slow changes in baseline load, reduce information lag, and allow baseline load information to reflect actual power demand trends more promptly. Therefore, by adaptively adjusting the smoothing intensity, it ensures that baseline load information can effectively filter out noise and reflect real load changes in a timely manner, while allowing instantaneous disturbance information to more purely reflect true transient events, providing more accurate input for subsequent power fine-tuning.

[0036] As one embodiment of the present invention, the steps for fine-tuning the power of an online generator set based on instantaneous disturbance information include: Dynamic characteristic analysis is performed on instantaneous disturbance information to identify load components with different dynamic response characteristics in the instantaneous disturbance information; Based on the identified load components with different dynamic response characteristics, differentiated power fine-tuning commands are determined; The power of the online generator set is fine-tuned according to the differentiated power fine-tuning instructions.

[0037] Specifically, dynamic characteristic analysis of instantaneous disturbance information refers to in-depth analysis of multiple dimensions of instantaneous disturbance signals, such as waveform, rate of change, duration, and frequency components, to reveal their inherent dynamic response characteristics. Its purpose is to distinguish instantaneous disturbances caused by different types of loads, such as those caused by large motor startups, lighting load switching, short-term impulsive loads, or power grid failures. Through this analysis, load components with different dynamic response characteristics within the instantaneous disturbance information can be identified, such as rapidly rising loads, pulsed loads, and continuously fluctuating loads. Based on the identified load components with different dynamic response characteristics, differentiated power fine-tuning commands are determined. This can be understood as generating customized power adjustment strategies that match the dynamic characteristics of each identified load component type. For example, a faster and larger power response may be needed for rapidly rising loads; a smoother and more continuous power adjustment may be needed for continuously fluctuating loads; and a short-term impulsive load may require a momentary high power output to maintain system stability, followed by a rapid decline. These differentiated instructions are designed to ensure that the generator set's response can accurately adapt to load changes, avoiding over- or under-adjustment.

[0038] In practical applications, power fine-tuning is performed on online generator sets based on differentiated power fine-tuning commands. Specifically, this refers to applying the aforementioned customized power adjustment strategies to the currently operating generator sets. For example, the fuel supply, excitation current, or governor setpoint of the generator set can be adjusted according to the commands to achieve precise power output adjustment. This fine-tuning is dynamic and real-time, designed to quickly respond to instantaneous load changes and maintain frequency and voltage stability of the ship's power grid.

[0039] This application's solution overcomes the limitations of traditional methods in handling complex and variable instantaneous loads by introducing a step of dynamic characteristic analysis of instantaneous disturbance information. It is precisely this detailed dynamic characteristic analysis of instantaneous disturbance information that enables the system to identify load components with different dynamic response characteristics. This identification capability is key to achieving precise control because it allows the system to distinguish the source and nature of instantaneous disturbances, rather than treating all instantaneous disturbances indiscriminately. Based on this, the system can determine and issue differentiated power fine-tuning commands for each identified load component. The generation of these differentiated commands ensures that the generator set's power output adjustment is highly matched to the actual load changes, avoiding unnecessary overshoot or lag. Ultimately, by executing these differentiated power fine-tuning commands, the power output of the online generator set is precisely adjusted, effectively responding to the complex changes in instantaneous load.

[0040] As one embodiment of the present invention, the step of performing dynamic characteristic analysis on instantaneous disturbance information to identify load components with different dynamic response characteristics in the instantaneous disturbance information includes: Identify transient change events in transient disturbance information; For each identified instantaneous change event, its multidimensional transient feature parameters within a preset time window are extracted. The multidimensional transient feature parameters include the power change rate, power rise slope, power fall slope, and power change duration. Based on multidimensional transient characteristic parameters, instantaneous change events are classified to identify load components with different dynamic response characteristics in instantaneous disturbance information.

[0041] Specifically, identifying transient events involves real-time monitoring and analysis of transient disturbance information to detect moments or intervals where power values ​​change significantly and rapidly. These changes may manifest as sudden increases, decreases, or drastic fluctuations in power, typically corresponding to the connection, disconnection, or change of operating status of large loads in the ship's electrical network. For example, the occurrence of transient events can be determined by setting power change thresholds and duration thresholds. When transient disturbance information exceeds the preset power change threshold within a short period of time, it is identified as a transient event.

[0042] Furthermore, for each identified transient event, its multidimensional transient feature parameters within a preset time window need to be extracted. The preset time window can be flexibly set according to the actual application scenario and load characteristics; for example, it can be set to a fixed period before and after the transient event, or dynamically adjusted according to the event's duration. The multidimensional transient feature parameters aim to comprehensively characterize the dynamic characteristics of the transient event, specifically including the power change rate, power ramp-up slope, power descent slope, and power change duration. The power change rate can be understood as the magnitude of power change per unit time, reflecting the severity of load changes; the power ramp-up slope and power descent slope describe the trend and speed of power increase and decrease, respectively, and are important for distinguishing between impulsive and gradual loads; the power change duration represents the length of time the transient event lasts from start to finish, helping to determine the nature and scale of the load change. These parameters can be extracted using mathematical methods such as differencing, regression analysis, or curve fitting of the transient disturbance information.

[0043] Therefore, based on multidimensional transient feature parameters, transient events can be classified to identify load components with different dynamic response characteristics within transient disturbance information. The purpose of classification is to group transient events with similar transient characteristics into one category, thereby identifying different load types, such as large motor startup, pump load switching, and heating equipment activation. Each load type typically corresponds to specific dynamic response characteristics; for example, impulsive loads may exhibit a high power change rate and short duration, while gradual loads may exhibit a lower power change rate and longer duration. Classification methods can employ machine learning algorithms, such as Support Vector Machines (SVM), decision trees, clustering algorithms (such as K-means), or rule-based expert systems, by comparing extracted feature parameters with preset feature ranges for various load components. In this way, various load components and their dynamic characteristics contained in transient disturbance information can be accurately identified.

[0044] This application's solution, through detailed dynamic characteristic analysis of instantaneous disturbance information, deconstructs the instantaneous disturbance component of the complex total power load signal into a series of discrete instantaneous change events. By extracting multidimensional transient characteristic parameters for each instantaneous change event, such as power change rate, power ramp-up slope, power drop slope, and power change duration, the dynamic behavior of these events can be quantified and described from multiple dimensions. It is precisely because these multidimensional characteristic parameters can comprehensively reflect the unique transient responses exhibited by different load types when they are connected or disconnected that subsequent classification steps can effectively categorize these instantaneous change events into load components with different dynamic response characteristics. This refined identification mechanism lays the foundation for subsequently determining differentiated power fine-tuning commands for different load components, thereby ensuring that the generator set can respond accurately and efficiently to various instantaneous load changes.

[0045] As one embodiment of the present invention, the step of classifying transiently changing events based on multidimensional transient characteristic parameters to identify load components with different dynamic response characteristics in transient disturbance information includes: Compare the various parameters in the multidimensional transient characteristic parameters with the pre-determined characteristic ranges of various load components; Based on the comparison results, the type of load component to which the instantaneous change event belongs is determined, so as to identify load components with different dynamic response characteristics in the instantaneous disturbance information.

[0046] Specifically, the multidimensional transient characteristic parameters include the power change rate, power ramp-up slope, power descent slope, and power change duration. The characteristic ranges of various load components refer to the typical value ranges or patterns of these multidimensional transient characteristic parameters, predetermined through historical data analysis or experimental testing, for various typical loads that may occur in a ship's electrical system (e.g., large motor startup, heating equipment activation, pump switching, etc.). For example, when a large motor starts, its power change rate and power ramp-up slope are usually very large, but the power change duration is relatively short; while when a heater turns on, the power change rate may be smaller, but the power change duration may be longer. The comparison process can employ various matching algorithms, such as rule-based logical judgment, fuzzy matching, or machine learning classifiers, to evaluate the degree of agreement between the multidimensional transient characteristic parameters extracted from the current transient event and the preset characteristic ranges of various load components.

[0047] Determining the load component type of a transient event based on comparison results means that when the multidimensional transient characteristic parameters of a transient event highly match the characteristic range of a preset load component, the event is classified into that load component type. For example, if the power change rate and power ramp-up slope of a transient event both fall within the characteristic range of "large motor starting," and the power change duration conforms to its typical value, then the event is identified as a large motor starting load. Through this feature-matching-based classification mechanism, the complex fluctuations in transient disturbance information can be decomposed into identifiable load components with specific dynamic response characteristics.

[0048] The proposed solution systematically compares the multidimensional transient characteristic parameters of instantaneous events with pre-established characteristic ranges for various load components, enabling accurate identification of different load components in instantaneous disturbance information. Specifically, each type of load generates a unique transient response in the ship's total electrical load signal upon connection or disconnection. These responses can be quantified using multidimensional transient characteristic parameters such as power change rate, power ramp-up slope, power descent slope, and power change duration. By pre-defining typical ranges for these characteristic parameters under different load types, the system can match the characteristic parameters of real-time monitored instantaneous events with these preset ranges. When the matching degree reaches a certain level, it can be determined which type of load caused the instantaneous event. This feature-matching-based classification mechanism allows the system to accurately separate load components with different dynamic response characteristics from complex instantaneous disturbances, providing a precise basis for subsequent differentiated power fine-tuning.

[0049] As one embodiment of the present invention, when the identified load component does not match the preset load component type, the step of determining differentiated power fine-tuning commands based on the identified load components with different dynamic response characteristics includes: The loaded component was identified as an unclassified loaded component. For unclassified load components, power fine-tuning commands are generated based on the power change rate and duration of the unclassified load components. During the execution of power fine-tuning commands, monitor the frequency and voltage of the ship's electrical network; Based on the monitoring results of frequency and voltage, adjust the amplitude or duration of the power fine-tuning command.

[0050] Specifically, after classifying transient events using multidimensional transient characteristic parameters, if an event is found not to belong to any preset known load component type, it is identified as an unclassified load component. For such unclassified load components, the system does not simply ignore them or adopt a general strategy, but rather generates a preliminary power fine-tuning command based on their most basic dynamic characteristic parameters: the power change rate and the power change duration. The power change rate reflects the severity of the load change, while the power change duration reflects the persistence of the load change. Based on these two parameters, the required power compensation amount and duration can be preliminarily estimated. During the execution of this preliminarily generated power fine-tuning command, the system monitors the frequency and voltage of the ship's power grid in real time. Frequency and voltage are key indicators for measuring grid stability, and their fluctuations directly reflect the balance between generator power output and load demand. In practical applications, monitoring can be performed by dedicated sensors or a power monitoring system, and the real-time data can be fed back to the control module. Furthermore, based on the monitoring results of frequency and voltage, the preliminarily generated power fine-tuning command can be dynamically adjusted. For example, if a large deviation in frequency or voltage is detected, or if the deviation recovers too slowly, it indicates that the initial command may not be sufficient to effectively handle the unclassified load. In this case, the amplitude of the power fine-tuning command can be increased or its duration extended accordingly. Conversely, if the frequency and voltage recover quickly and stably, it may indicate that the initial command was too aggressive. The amplitude or duration can be appropriately reduced to avoid overcompensation.

[0051] This application's solution effectively overcomes the limitations of traditional classification methods by introducing an identification and adaptive processing mechanism for unclassified load components. When encountering instantaneous load changes that cannot be accurately classified, the system no longer passively waits or adopts suboptimal strategies. Instead, it generates an initial response based on the basic dynamic characteristics of the load and dynamically and in a closed-loop manner adjusts power fine-tuning commands by monitoring the frequency and voltage of the power grid in real time. This adaptive adjustment capability enables the generator set's power output to more accurately match actual load demands, ensuring the stable operation of the ship's power grid even in the face of unknown or abnormal load shocks.

[0052] As one embodiment of the present invention, the step of adjusting the amplitude or duration of the power fine-tuning command based on the monitoring results of frequency and voltage includes: Based on the monitoring results of frequency and voltage, the deviation between frequency and voltage and the preset stable value, as well as the instantaneous rate of change of the deviation, are obtained; The adjustment amount of the power fine-tuning command is determined based on the deviation and the instantaneous rate of change of the deviation. The adjustment amount is positively correlated with the instantaneous rate of change of the deviation. Apply the adjustment amount to the magnitude or duration of the power fine-tuning command.

[0053] Specifically, the "preset stable value" refers to the ideal target values ​​that the frequency and voltage should maintain under normal operating conditions of the ship's power grid. For example, the frequency can be set to 50Hz or 60Hz, and the voltage can be set to the rated voltage value. "Deviation" refers to the difference between the currently monitored frequency or voltage value and its respective preset stable value, reflecting the degree to which the power grid's operating state deviates from the ideal state. "Instantaneous rate of change" refers to the speed at which the frequency or voltage changes within a very short time. This can be obtained, for example, by differentiating continuously sampled data or calculating using a digital filter, and its purpose is to capture the dynamic trend and intensity of change of power grid parameters. The "adjustment amount of the power fine-tuning command" is used to correct the value of the original power fine-tuning command to make it more accurately adapt to the current power grid state. "The adjustment amount is positively correlated with the instantaneous rate of change of the deviation" means that the faster the frequency or voltage deviates from the stable value, the larger the determined adjustment amount, and vice versa. This positive correlation ensures that the system can respond more proactively to rapidly changing power grid conditions. In practical applications, this positive correlation can be achieved through linear functions, piecewise functions, or more complex nonlinear functions. For example, the adjustment amount can be set as K* (the instantaneous rate of change of the deviation), where K is a proportionality coefficient, which can be optimized according to system characteristics and control requirements.

[0054] This application's solution overcomes the limitations of adjustments based solely on static deviations by incorporating consideration of the "instantaneous rate of change" of frequency and voltage deviations. When the frequency or voltage of the ship's power grid fluctuates, it not only focuses on the degree of deviation from the stable value (deviation), but more importantly, it monitors the trend and speed of this deviation in real time (instantaneous rate of change). This perception of dynamic trends allows the system to predict the future direction of the power grid. For example, when the frequency is not only below the preset stable value but also rapidly decreasing, it means that the grid load may be increasing sharply, requiring stronger power output. In this case, since the adjustment amount is positively correlated with the instantaneous rate of change of the deviation, the system generates a larger power fine-tuning command, prompting the online generators to increase power output more quickly to rapidly offset the impact of load changes. This proactive adjustment mechanism enables the system to respond to dynamic changes in the power grid more promptly and accurately, avoiding grid instability that may result from delayed response.

[0055] As one embodiment of the present invention, the process of determining the characteristic range of various loading components includes: Continuously monitor the matching degree between the multidimensional transient feature parameters of classified instantaneous change events and the preset feature range; When the matching degree is consistently below the preset threshold, or when classified transient change events are frequently identified as unclassified load components, an update to the feature range is triggered. After the update is triggered, collect multidimensional transient feature parameters of recently classified transient change events; The feature range is adjusted based on the collected multidimensional transient feature parameters.

[0056] Specifically, "characteristic ranges of various load components" refer to the power change characteristic boundaries used to distinguish different types of transient events (e.g., large motor startup, pump load switching, heating equipment activation, etc.). These characteristic ranges are typically defined by multi-dimensional transient characteristic parameters such as power change rate, power ramp-up slope, power drop slope, and power change duration. "Matching degree" can be understood as the degree of agreement between the actual multi-dimensional transient characteristic parameters of a classified transient event and the corresponding preset characteristic range. This can be measured, for example, by calculating the distance or similarity between feature vectors. A high matching degree indicates that the actual parameters fall within the preset range; conversely, a low matching degree indicates a lower matching degree. "Preset threshold" is a configurable parameter used to define whether the matching degree has reached a level requiring an update. "Continuous monitoring" means that the system will evaluate each successfully classified transient event in real time or periodically, comparing its multi-dimensional transient characteristic parameters with the currently used preset characteristic range. "Classified transient events" refer to events that have been successfully identified and categorized by the system into a specific load component type. The conditions for "triggering an update to the feature range" are twofold: first, when the matching degree consistently falls below a preset threshold, it indicates that the existing feature range may no longer accurately reflect the actual load characteristics; second, when classified transient events are frequently identified as unclassified load components, this suggests that the existing classification system has blind spots or deficiencies and needs to be expanded or adjusted. "Frequent identification" can be determined by statistically analyzing the proportion or number of unclassified events within a certain time window. After triggering the update, the system will "collect multidimensional transient feature parameters of recently classified transient events." Here, "recent" can be a sliding time window, such as data from the last few hours, days, or hundreds of events, to ensure that the collected data reflects the latest load characteristics. "Collection" can involve storing this data in a temporary database or buffer for subsequent analysis. Finally, "adjust the feature range based on the collected multidimensional transient feature parameters." Adjustment methods may include, but are not limited to: recalculating the feature boundaries of various load components based on statistical methods (such as mean, standard deviation, cluster analysis, etc.); or retraining the classification model using machine learning algorithms (such as support vector machines, neural networks, etc.) to update its internal feature discrimination logic, thereby reflecting an adjustment of the feature range. The purpose of the adjustment is to enable the new feature range to more accurately encompass the feature distribution of the current actual various load components, thereby improving the accuracy and robustness of classification.

[0057] This application's solution effectively addresses the problem of preset feature ranges becoming ineffective due to changes in load characteristics in traditional methods by introducing a dynamic update mechanism for the feature ranges of various load components. Specifically, by continuously monitoring the matching degree between the actual multidimensional transient feature parameters of classified transient events and the existing preset feature ranges, the system can perceive the "health status" of the classification model in real time. When the matching degree continuously declines, indicating a deviation between the existing feature ranges and actual load behavior, or when the system frequently misidentifies events that should belong to known categories as unclassified load components, both situations serve as clear signals indicating that the current feature ranges are no longer applicable. Consequently, the system is triggered to perform a feature range update operation. During the update process, the system collects the latest feature data of recently occurring and successfully classified transient events, which represent the true dynamic characteristics of the current ship's electrical load. Based on this latest and representative data, the original feature ranges are recalculated or adjusted, thereby ensuring that the classification model can always accurately identify and distinguish load components with different dynamic response characteristics. This adaptive update mechanism enables the entire electrical load control system to better adapt to the long-term changes and evolution of the ship's electrical load, maintaining high-precision load classification capabilities.

[0058] like Figure 2 The diagram illustrates a marine generator set power load control system, which includes: Acquisition module 201 is used to acquire the ship's total electrical load signal; The processing module 202 is used to process the total power load signal to obtain basic load information and instantaneous disturbance information. The basic load information is obtained by smoothing the total power load signal, and the instantaneous disturbance information is obtained by the difference between the total power load signal and the basic load information. The start / stop decision module 203 is used to make start / stop decisions for the generator set based on the basic load information. The power fine-tuning module 204 is used to perform online power fine-tuning of the generator set based on instantaneous disturbance information.

[0059] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. A method for controlling the electrical load of a ship's generator set, characterized in that, The method includes the following steps: Collect the ship's total electrical load signal; The total power load signal is processed to obtain basic load information and instantaneous disturbance information. The basic load information is obtained by smoothing the total power load signal, and the instantaneous disturbance information is obtained by the difference between the total power load signal and the basic load information. Based on the basic load information, start-up and shutdown decisions for the generator set are made, and based on the instantaneous disturbance information, online power fine-tuning of the generator set is performed.

2. The method for controlling the electrical load of a ship generator set according to claim 1, characterized in that, Before making generator start-up and shutdown decisions based on the aforementioned basic load information, the method further includes the following steps: When the basic load information continues to rise and exceeds the preset start-up threshold, a load nature judgment is performed; The load characteristics determination includes: Send a power increase command to the currently online generator set, so that the online generator set increases its power output by a preset amount within a preset time; Monitor the frequency changes of the ship's power grid during periods of increased generator power; Based on the frequency change, determine whether the increase in the base load information is caused by the actual increase in base load; If the increase is determined to be caused by a genuine increase in base load, the standby generator set will be started. If the increase is determined to be caused by a non-genuine increase in base load, the standby generator set will not be started, and the power output of the online generator set will be reduced.

3. The method for controlling the electrical load of a ship generator set according to claim 1, characterized in that, The step of processing the total power load signal to obtain basic load information and instantaneous disturbance information includes: Continuously monitor the fluctuation range of the total power load signal within a preset reference time window; When the fluctuation amplitude exceeds a preset threshold, the smoothing time window is extended; when the fluctuation amplitude is below the preset threshold, the smoothing time window is shortened. The total power load signal is smoothed based on the time window of the smoothing process to obtain the basic load information. The difference between the latest acquired total power load signal and the basic load information is calculated to obtain the instantaneous disturbance information.

4. The method for controlling the electrical load of a ship generator set according to claim 1, characterized in that, The step of fine-tuning the power of the generator set based on the instantaneous disturbance information includes: Dynamic characteristic analysis is performed on the instantaneous disturbance information to identify load components with different dynamic response characteristics in the instantaneous disturbance information; Based on the identified load components with different dynamic response characteristics, differentiated power fine-tuning commands are determined; The power of the online generator set is fine-tuned according to the differentiated power fine-tuning command.

5. The method for controlling the electrical load of a ship generator set according to claim 4, characterized in that, The step of performing dynamic characteristic analysis on the instantaneous disturbance information to identify load components with different dynamic response characteristics in the instantaneous disturbance information includes: Identify instantaneous change events in the instantaneous disturbance information; For each identified instantaneous change event, its multidimensional transient feature parameters within a preset time window are extracted. The multidimensional transient feature parameters include the power change rate, power rise slope, power fall slope, and power change duration. Based on the multidimensional transient characteristic parameters, the instantaneous change events are classified to identify load components with different dynamic response characteristics in the instantaneous disturbance information.

6. The method for controlling the electrical load of a ship generator set according to claim 5, characterized in that, The step of classifying the transient events based on the multidimensional transient feature parameters to identify load components with different dynamic response characteristics in the transient disturbance information includes: Compare the various parameters in the multidimensional transient characteristic parameters with the predetermined characteristic ranges of various load components; Based on the comparison results, the type of load component to which the instantaneous change event belongs is determined, so as to identify load components with different dynamic response characteristics in the instantaneous disturbance information.

7. The method for controlling the electrical load of a ship generator set according to claim 5, characterized in that, When the identified load component does not match the preset load component type, the steps for determining differentiated power fine-tuning commands based on the identified load components with different dynamic response characteristics include: The loaded component was identified as an unclassified loaded component; For the unclassified load components, a power fine-tuning command is generated based on the power change rate and power change duration of the unclassified load components; During the execution of the power fine-tuning command, the frequency and voltage of the ship's electrical network are monitored; Based on the monitoring results of the frequency and voltage, adjust the amplitude or duration of the power fine-tuning command.

8. The method for controlling the electrical load of a ship generator set according to claim 7, characterized in that, The step of adjusting the amplitude or duration of the power fine-tuning command based on the monitoring results of the frequency and voltage includes: Based on the monitoring results of the frequency and voltage, the deviation between the frequency and voltage and the preset stable value, as well as the instantaneous rate of change of the deviation, are obtained; Based on the deviation and the instantaneous rate of change of the deviation, the adjustment amount of the power fine-tuning command is determined, and the adjustment amount is positively correlated with the instantaneous rate of change of the deviation; The adjustment amount is applied to the magnitude or duration of the power fine-tuning command.

9. A method for controlling the electrical load of a ship generator set according to claim 6, characterized in that, The process of determining the characteristic ranges of the various loading components includes: Continuously monitor the matching degree between the multidimensional transient feature parameters of classified instantaneous change events and the preset feature range; When the matching degree is consistently lower than a preset threshold, or when the classified instantaneous change event is frequently identified as an unclassified load component, an update to the feature range is triggered. After the update is triggered, collect the multidimensional transient feature parameters of the recently classified transient change events; The feature range is adjusted based on the collected multidimensional transient feature parameters.

10. A power load control system for a ship generator set, characterized in that, The system includes: The acquisition module is used to acquire the ship's total electrical load signal; The processing module is used to process the total power load signal to obtain basic load information and instantaneous disturbance information. The basic load information is obtained by smoothing the total power load signal, and the instantaneous disturbance information is obtained by the difference between the total power load signal and the basic load information. The start / stop decision module is used to make start / stop decisions for the generator set based on the basic load information. The power fine-tuning module is used to perform online power fine-tuning of the generator set based on the instantaneous disturbance information.

Images