Power Input Power Optimization Method and System Based on Intelligent Data Analysis

Abstract

The present invention relates to the field of intelligent control technology, and discloses a power input power optimization method and system based on intelligent data analysis, including: collecting multimodal data, constructing a power optimization model according to the multimodal data, and obtaining an input power set value; using a fuzzy PID controller to dynamically adjust the input power according to the input power set value; obtaining a load prediction value through a load prediction algorithm, and adjusting the input power set value according to the load prediction value to achieve feedforward power regulation. Compared with the traditional scheme that only relies on single feedback, the present invention can better balance system stability, energy storage life protection and overall economic benefits, and truly realizes the comprehensive goals of peak shaving and valley filling, reducing operation costs, and improving the utilization efficiency of energy storage devices, and has significant popularization value and good application prospects for various microgrid, power management and industrial energy application scenarios.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent control, and specifically to a power input power optimization method and system based on intelligent data analysis. Background Art

[0002] With the rapid development of power electronics technology and intelligent control technology, emergency power supply (EPS) systems are increasingly widely used in fields such as construction, industry, and healthcare. Traditional EPS systems mainly rely on fixed threshold control or simple feedback regulation mechanisms to achieve the switching and power distribution between the main power supply and the emergency power supply. In recent years, with the popularization of the Internet of Things (IoT) and big data technology, intelligent power management systems have gradually become a research hotspot. In the prior art, some studies have attempted to optimize the power distribution efficiency of the power input by real-time monitoring of the grid status and load demand, combined with PID control algorithms or fuzzy control theories. In addition, the maturity of bidirectional inverter technology has also provided technical support for energy recovery and reuse, enabling the energy efficiency of EPS systems to be improved under non-emergency conditions. However, although the prior art has made certain progress in power input power optimization, there are still many limitations and it is difficult to meet the requirements of efficient and stable operation in complex application scenarios.

[0003] Traditional power optimization methods mostly rely on a single data source and lack the comprehensive utilization of multi-modal data (such as ambient temperature, battery status, load type, etc.), resulting in insufficient accuracy and adaptability of the optimization model. Secondly, existing control strategies often have difficulty achieving fast and accurate power regulation when facing non-linear and time-varying load characteristics, and are prone to overshoot or oscillation phenomena, affecting system stability. In addition, the prior art generally lacks the ability to predict the load change trend and cannot achieve feed-forward power regulation, resulting in system response lag and reduced energy efficiency. These deficiencies limit the performance improvement of EPS systems in complex application scenarios, and there is an urgent need for a more intelligent and comprehensive power input power optimization method. Summary of the Invention

[0004] In view of the above problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is that the prior art has obvious deficiencies in the utilization of multi-modal data, dynamic power regulation, load prediction, and energy recovery, and it is difficult to meet the requirements of efficient and stable power input power optimization.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: A power input power optimization method based on intelligent data analysis, including:

[0007] Collect multi-modal data, construct a power optimization model according to the multi-modal data, and obtain the input power set value;

[0008] Adopt a fuzzy PID controller to dynamically adjust the input power according to the input power set value;

[0009] Obtain the load prediction value through the load prediction algorithm, and adjust the input power set value according to the load prediction value to achieve feed-forward power regulation.

[0010] As a preferred solution of the power input power optimization method based on intelligent data analysis according to the present invention, wherein: the acquisition of multi-modal data includes collecting real-time grid status data and load data, and introducing environmental data and battery status as auxiliary optimization parameters;

[0011] The real-time grid status data and load data include grid voltage, current, load prediction data, power cost data, and the power exchange value between the load and the grid;

[0012] The auxiliary optimization parameters include the internal temperature of the battery, the equivalent life loss of the battery charge and discharge cycle, the battery data and the battery charge and discharge current data.

[0013] As a preferred solution of the power input power optimization method based on intelligent data analysis according to the present invention, wherein: the construction of the power optimization model according to the multi-modal data includes obtaining the load prediction data and the power cost data according to the multi-modal data, and constructing a multi-objective optimization function, expressed as:

[0014] ;

[0015] Among them, Indicates taking the minimum value; Indicates the input power set value at time ; , , Indicates the weight coefficient; Indicates the operating cost; Indicates the life decay cost of the energy storage device; Indicates the grid impact index;

[0016] The operating cost Is expressed as the weighted sum of the power costs obtained from the grid side within the scheduling period:

[0017] ;

[0018] Among them, Indicates the start time of the scheduling period; Indicates the length of the optimization rolling time domain; Indicates the power exchanged with the grid side; Indicates the time step; Represents the cost of power exchange with the grid side; Time series index;

[0019] Quantify the equivalent life loss of charge and discharge cycles through an empirical model to obtain the life attenuation cost of the energy storage device , expressed as:

[0020] ;

[0021] Wherein, Represents the cost consumption function of charge and discharge operations at the current moment on the battery life; Represents the moment State of charge; Represents Change; Represents the battery charge and discharge current; Represents the internal temperature of the battery;

[0022] The grid impact index Is represented by the power exchange value between the load on the grid side and the grid:

[0023] ;

[0024] Wherein, Represents the moment And the power exchange value between the load on the grid side and the grid; Represents the moment And the power exchange value between the load on the grid side and the grid.

[0025] As a preferred scheme of the power input power optimization method based on intelligent data analysis described in the present invention, wherein: the empirical model includes expressing the average life loss caused by each unit of charge and discharge under the basic working condition with the basic attenuation rate , obtaining auxiliary optimization parameters through multi-modal data, and calculating the life attenuation cost in combination with the Change amount of the energy storage within each time step , expressed as:

[0026] ;

[0027] Wherein, Represents the basic attenuation rate; Represents the Change amount within the time step; Represents the temperature correction factor; Represents the rate correction factor;

[0028] The temperature correction factor Describes the impact of temperature exceeding the basic working condition on battery aging, expressed as:

[0029] ;

[0030] Among them, represents the sensitivity of temperature to the attenuation rate; represents the reference temperature under the basic working condition;

[0031] The magnification correction factor reflects the influence of the difference between the actual charge-discharge current and the basic working condition current on attenuation, and is expressed as:

[0032] ;

[0033] Among them, represents the sensitivity of the charge-discharge current magnification to the attenuation rate; represents the actual charge-discharge current; represents the reference current under the basic working condition; By introducing the temperature correction factor and the magnification correction factor, the battery life attenuation is evaluated in real time in each scheduling cycle; When the temperature or magnification deviation is too large, the life attenuation cost is adaptively increased , guiding the scheduling plan to balance power distribution and life protection.

[0034] As a preferred solution of the power input power optimization method based on intelligent data analysis described in the present invention, wherein: the fuzzy PID controller includes collecting basic input information, and the basic input information includes the input power set value , the actual input power and the energy storage life attenuation index , and the energy storage life attenuation index is calculated and obtained according to the life attenuation cost ;

[0035] Calculate the error between the input power set value and the actual input power and the error change amount ;

[0036] The , and are fuzzified to construct membership functions and linguistic values;

[0037] A fuzzy rule base is established. Based on the two-dimensional rules of error and error change, the is used as a branch judgment to correct the judgment result of the two-dimensional rules of error and error change. The result of the fuzzy rule is converted into a specific value by the centroid method, and the PID parameters are updated according to the specific value:

[0038] ;

[0039] Among them, represents the final control quantity; represents the proportional gain; represents the integral gain; represents the sum of the collected and accumulated error change amounts over time; represents the derivative gain; converts the control quantity into an input power adjustment command.

[0040] As a preferred solution of the power input power optimization method based on intelligent data analysis according to the present invention, wherein: the adjusting the input power setting value according to the load prediction value includes, according to multi-modal data, predicting the output load curve within a future rolling window through time series ;

[0041] Obtain the input power setting value , use to compare with the load prediction value, obtain the load deviation, generate a feedforward correction amount according to the load deviation and energy storage constraints, and pre-adjust the input power setting value, expressed as:

[0042] ;

[0043] wherein, represents the corrected power setting value; represents the feedforward correction amount.

[0044] As a preferred solution of the power input power optimization method based on intelligent data analysis according to the present invention, wherein: the feedforward power regulation includes, sending the input power setting sequence after feedforward correction to a fuzzy PID controller, and the fuzzy PID continues to finely adjust according to the actual error on a short time scale to form a double guarantee of feedforward and feedback;

[0045] When the deviation between the actual load and the predicted load exceeds the allowable range, the actual input power is adjusted through the feedback control of the fuzzy PID to quickly eliminate the deviation caused by the error, and the feedforward control is updated again when the next prediction period arrives.

[0046] A power input power optimization system based on intelligent data analysis adopting any of the methods of the present invention, wherein: a collection module, which collects multi-modal data, constructs a power optimization model according to the multi-modal data, and obtains the input power setting value;

[0047] A control module, which adopts a fuzzy PID controller and dynamically adjusts the input power according to the input power setting value;

[0048] An adjustment module, which obtains the load prediction value through a load prediction algorithm, adjusts the input power setting value according to the load prediction value, and realizes feedforward power regulation.

[0049] A computer device, comprising: a memory and a processor; the memory stores a computer program, including: when the processor executes the computer program, the steps of the method described in any one of the present inventions are implemented.

[0050] A computer-readable storage medium, on which a computer program is stored, including: when the computer program is executed by a processor, the steps of the method described in any one of the present inventions are implemented.

[0051] Advantages of the present invention: By introducing multimodal data analysis, load prediction, and energy storage life attenuation models on the basis of conventional power scheduling, the method of the present invention can not only dynamically obtain the optimal input power setting value, but also achieve refined adjustment of the actual input power by combining feedforward and feedback. The fuzzy PID controller makes a high-speed response to the real-time error, effectively avoiding power fluctuations caused by sudden changes in the load; at the same time, through the energy management and feedforward scheduling of the bidirectional inverter, electric energy can be reserved in advance before the load peak, or the excess energy can be fed back in time when the load drops suddenly, greatly reducing the instantaneous impact on the power grid and the energy storage system. Compared with the traditional scheme that only relies on single feedback, the present invention can better balance system stability, energy storage life protection, and overall economic benefits, truly achieving the comprehensive goals of peak shaving and valley filling, reducing operating costs, and improving the utilization efficiency of energy storage devices, and has significant promotion value and good application prospects for various microgrid, power management, and industrial energy application scenarios. Description of the Drawings

[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for description in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0053] Figure 1 It is the overall flowchart of the power input power optimization method based on intelligent data analysis provided by an embodiment of the present invention. Detailed Embodiments

[0054] To make the above objects, features, and advantages of the present invention more obvious and understandable, the specific embodiments of the present invention will be described in detail below with reference to the drawings of the specification. Obviously, the described embodiments are some embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0055] Example 1, refer to Figure 1, which is an embodiment of the present invention, provides a power input power optimization method based on intelligent data analysis, including:

[0056] S1: Collect multi-modal data, construct a power optimization model according to the multi-modal data, and obtain the input power setting value.

[0057] Furthermore, the collection of multi-modal data includes collecting real-time grid status data and load data, and introducing environmental data and battery status as auxiliary optimization parameters. Specifically, multi-type data collection devices are respectively deployed at the system end and the load end to collect real-time grid status data and load data. The real-time grid status data and load data include basic electrical information such as grid voltage, current, harmonic content, and load power demand. The battery status includes internal temperature of the battery, battery data, battery charge and discharge current data, and internal status such as operation health. At the same time, environmental data such as temperature, humidity, and light intensity can also be obtained by using external environmental sensors to provide support for the subsequent adaptive adjustment of the model in different climates or places. To improve the integrity and accuracy of the data, a data fusion module can be set in the main control platform to perform preliminary cleaning and formatting processing on multi-modal data from electrical monitoring, energy storage management, and environmental perception, etc., to form a unified feature vector that can be called by the subsequent power optimization model.

[0058] Even further, the core goal of constructing the power optimization model is to output a decision-making scheme that can achieve the optimal input power setting value within a certain time domain according to the multi-modal data and system operation requirements. For this purpose, the problem can be abstracted into an optimization model with multi-objective functions and containing several physical and operation constraints, and prediction can be used to improve the forward-looking and real-time performance of the model. The construction of the power optimization model according to the multi-modal data includes obtaining load prediction data and power cost data according to the multi-modal data, and constructing a multi-objective optimization function, which is expressed as:

[0059] ;

[0060] Among them, means taking the minimum value; means the moment of the input power setting value; , , means the weight coefficient; means the operating cost; means the life attenuation cost of the energy storage device; means the grid impact index.

[0061] The operating cost is expressed as the weighted sum of the power costs obtained from the grid side within the scheduling period:

[0062] ;

[0063] Among them, represents the start time of the scheduling period; represents the length of the optimized rolling time domain; represents the power of the power exchange with the grid side; represents the time step; represents the cost of the power exchange with the grid side; Time series index. It should be noted that if considering selling energy to the grid (two-way trading), a negative revenue item needs to be added to the above model, or the power purchase and power sale are separately modeled as power flows in positive and negative directions and the corresponding costs / revenues.

[0064] The life of energy storage devices (such as lithium batteries) is generally related to factors such as their depth of discharge (DOD), charge and discharge rate (C-rate), and temperature. In practical applications, it is not necessary to calculate the life loss of energy storage devices very precisely. Therefore, it can be simplified, and the equivalent life loss of charge and discharge cycles is quantified through an empirical model to obtain the life attenuation cost of the energy storage device , expressed as:

[0065] ;

[0066] Among them, represents the cost function of the current charge and discharge operation on the battery life; represents the time state of charge; represents change; represents the battery charge and discharge current; represents the internal temperature of the battery.

[0067] The empirical model includes using the basic attenuation rate to represent the average life loss caused by each unit of charge and discharge under the basic working conditions (normal temperature, normal charge and discharge rate), obtaining auxiliary optimization parameters through multi-modal data, and calculating the life attenuation cost in combination with the change amount of the energy storage within each time step , expressed as:

[0068] ;

[0069] Among them, represents the basic attenuation rate; represents the change amount within the time step; represents the temperature correction factor; represents the rate correction factor.

[0070] The temperature correction factor Describes the impact of temperature outside the basic operating conditions on battery aging. The chemical and physical properties of the battery or energy storage system are quite sensitive to temperature: too high or too low temperature will accelerate processes such as increased internal resistance and attenuation of active substances, resulting in shortened lifespan. Therefore, a correction factor is used to represent the impact of "the deviation of the current temperature from the standard temperature (or optimal temperature)" on the lifespan decay rate. The temperature correction factor is expressed as:

[0071] ;

[0072] Wherein, represents the sensitivity of temperature to the decay rate; represents the reference temperature under basic operating conditions. Through experiments, it is found that the exponential form is closer to the data law of many actual life acceleration tests (Arrhenius-type relationship). However, considering that the calculation of life loss in the present invention is for adjusting the input power and does not require precise life loss values, a simpler linear form is used, which is convenient for rapid calculation in real-time control and can also be approximately applicable in cases where the temperature change range is not large.

[0073] Similar to temperature, there is also data basis for the impact of discharge rate on attenuation (accelerated life tests, manufacturer test curves, etc.). Battery aging is closely related to the discharge rate (C-rate): during high-rate charge and discharge, side reactions, heat generation, and material stress inside the battery are more severe. In order to penalize the additional aging caused by excessive current during operation, a correction factor is usually defined relative to the standard rate (such as 1C). The rate correction factor reflects the impact of the difference between the actual charge and discharge current and the current under basic operating conditions on attenuation, and is expressed as:

[0074] ;

[0075] Wherein, represents the sensitivity of the charge and discharge current rate to the decay rate; represents the actual charge and discharge current; represents the reference current under basic operating conditions; By introducing the temperature correction factor and the rate correction factor, the battery life decay is evaluated in real-time for each scheduling cycle; when the temperature or rate deviation is too large, the life decay cost is adaptively increased , guiding the scheduling plan to balance power distribution and life protection.

[0076] By introducing temperature / rate correction, the system will evaluate the battery life decay in real-time for each scheduling cycle; if it is found that the temperature is too high or the rate is too large, the decay cost weight will be adaptively increased, guiding the scheduling plan to better balance power distribution and life protection.

[0077] The grid impact index is represented by the power exchange value between the load on the grid side and the grid:

[0078] ;

[0079] wherein, represents the moment and the power exchange value between the load on the grid side and the grid; represents the moment and the power exchange value between the load on the grid side and the grid. The larger the grid impact index, the more severe the power peak or fluctuation on the grid; by introducing this index into the objective function or using it as a constraint, the instantaneous impact on the grid can be reduced.

[0080] It should be noted that the power optimization model also needs to be limited by constraint conditions. When solving the optimal input power setting value, it can meet the actual engineering and system safety requirements, and avoid pursuing the optimal solution only from a mathematical perspective while ignoring physical or operating limits. The constraint conditions include but are not limited to power balance constraints, energy storage state evolution constraints, charge and discharge power limits, and grid-side power limits.

[0081] S2: Adopt a fuzzy PID controller to dynamically adjust the input power according to the input power setting value.

[0082] Furthermore, take the optimal input power setting value output by the multi-objective optimization model as the target input of the fuzzy PID. Through fuzzy inference and PID parameter self-adaptation, the real-time adjustment of the actual input power is realized, ensuring both good tracking performance and strong robustness and adaptability during the system operation.

[0083] The fuzzy PID controller includes collecting basic input information, and the basic input information includes the input power setting value , the actual input power and the energy storage life attenuation index , and the energy storage life attenuation index is calculated based on the life attenuation cost .

[0084] Specifically, the life attenuation index is, in fuzzy control, expressed by an intuitive quantization index such as 10 to represent the current energy storage's tolerance to fast charge and discharge, the larger it is, the higher the pressure or aging risk of the battery / energy storage, and more cautious high-power operations are required; the smaller it is, the better the battery health condition, and the power can be adjusted more actively.

[0085] Map the life attenuation cost linearly to Domain of definition, when the attenuation cost is very high, it indicates that the battery is experiencing high-rate / high-temperature / deep-discharge conditions currently or in the recent period, and the health stress accumulates or surges substantially; through mapping, let is higher, enabling the controller to schedule carefully and reduce intense charge and discharge, thereby achieving life protection. The mapping formula is implemented in the form of linear + saturation value, expressed as:

[0086] ;

[0087] Among them, represents the average value of the attenuation cost within the control period; represents the low threshold of the attenuation cost; represents the high threshold of the attenuation cost. When the attenuation cost is lower than , it is considered that the life stress can be almost ignored; when the attenuation cost exceeds , it is considered to be very high, and intense operations should be minimized.

[0088] Calculate the error between the input power set value and the actual input power and the error change .

[0089] Fuzzify , and , construct membership functions and linguistic values, and use 7 levels or 5 levels such as NB (Negative Big), NM (Negative Medium), NS (Negative Small), Z (Zero), PS (Positive Small), PM (Positive Medium), PB (Positive Big), and represent them with appropriate membership functions (such as trigonometric functions or trapezoidal functions) respectively. represents the current health stress of the energy storage - the higher the value, the greater the risk of battery aging or the higher the temperature, etc. Give a reasonable range and divide it into three fuzzy subsets: L (Low), M (Medium), and H (High).

[0090] Establish a fuzzy rule base. Based on the two-dimensional rules of error and error change, use as the branch judgment to correct the judgment results of the two-dimensional rules of error and error change, convert the results of the fuzzy rules into specific values through the centroid method, and update the PID parameters according to the specific values:

[0091] ;

[0092] Among them, represents the final control quantity; represents the proportional gain; represents the integral gain; represents collecting and accumulating the total error change over time; represents the differential gain; converts the control quantity into an input power adjustment command.

[0093] S3: Obtain the load prediction value through the load prediction algorithm, adjust the input power set value according to the load prediction value, and achieve feed-forward power regulation.

[0094] Furthermore, the adjusting the input power set value according to the load prediction value includes, according to multi-modal data, predicting the output load curve within the future rolling window through time series inside.

[0095] Obtain the input power set value , use to compare with the load prediction value, obtain the load deviation, generate a feed-forward correction amount according to the load deviation and energy storage constraints, and pre-adjust the input power set value, expressed as:

[0096] ;

[0097] wherein, represents the corrected power set value; represents the feed-forward correction amount.

[0098] It should be noted that when adjusting, the energy storage system range, the power upper limit / current upper limit of the bidirectional inverter, and the grid safety limit need to be considered to avoid violating physical and safety constraints. If the energy storage is already close to full SoC, it is impossible to absorb excess energy in advance; on the contrary when it is too low, it is also difficult to provide sufficient power before the load peak, and it is necessary to charge in advance or cooperate with the grid for coordinated scheduling.

[0099] Furthermore, adjust the charge and discharge strategy of the bidirectional inverter according to the adjusted power set value. The charge and discharge strategy includes that when it is predicted that there will be a load peak and the current energy storage is relatively insufficient, it is possible to charge the energy storage in advance during the low-load period or the low electricity price period; within this period, if the grid status permits or the electricity price is low, increase the power purchase and store it in the battery to prepare for discharging during the subsequent peak period.

[0100] When entering the load peak, the bidirectional inverter discharges (releases the energy storage energy) to assist in meeting the peak load, thereby reducing the instantaneous demand peak from the grid side; reducing the large impact on the grid and also saving the power purchase cost during the peak electricity price period (if there is a peak-valley difference in the electricity price mechanism).

[0101] When the load prediction shows an impending decrease or the system enters a period of extremely low power consumption, the discharge of the energy storage can be reduced and the remaining energy can be retained; if the system allows two-way trading and the electricity price is high, the excess energy can also be fed back to the grid to obtain benefits.

[0102] Furthermore, the feedforward power regulation includes sending the input power setting sequence after feedforward correction to the fuzzy PID controller, and the fuzzy PID continues to finely adjust according to the actual error on a short time scale, forming a double guarantee of feedforward and feedback.

[0103] When the deviation between the actual load and the predicted load exceeds the allowable range, the actual input power is adjusted through the feedback control of the fuzzy PID to quickly eliminate the deviation caused by the error, and the feedforward control is updated again when the next prediction cycle arrives.

[0104] Through the process design of prediction → feedforward scheduling → bi-directional inverter energy management → real-time execution → rolling update, the present invention can efficiently utilize the energy storage system and intelligently allocate the grid power in scenarios with strong load fluctuations or large peak-valley differences. While ensuring the load demand, it greatly alleviates the grid impact, reduces the operating cost, and prolongs the energy storage life. Compared with simple PID or passive post-control, this method is more forward-looking and flexible, and is an important technical means to achieve the "feedforward power regulation" goal of the present invention.

[0105] Embodiment 2, in an exemplary embodiment, a power input power optimization system based on intelligent data analysis is further provided, including a collection module, a data management module, a power optimization model, a control module (fuzzy PID controller), an adjustment module (feedforward adjustment unit), and an execution module.

[0106] Among them, the collection module obtains multi-modal data on the grid side, load side, and energy storage side in real time through various sensors, metering devices, and interfaces, including voltage, current, load power, ambient temperature, humidity, and battery SoC, etc. Historical operating conditions and fault information can also be collected for subsequent model training and parameter correction.

[0107] The collection module uploads the collected raw data to the data management module for preliminary calibration and storage; and provides the latest available data to the load prediction algorithm of the adjustment module.

[0108] The data management module cleans, converts the format, and extracts features from the multi-modal data uploaded by the collection module, providing high-quality and uniformly formatted inputs for the power optimization model or the load prediction algorithm. It can include a database or cache for parallel management of historical data and real-time data.

[0109] After receiving the data from the collection module, the data management module outputs it to the power optimization model and the prediction unit of the adjustment module to ensure data consistency and traceability.

[0110] The power optimization model is based on the fused data provided by the data management module, comprehensively considers multiple objective factors such as energy cost, energy storage life, power fluctuation, etc., constructs and solves the power optimization problem, and obtains the input power setting value of the current or next cycle.

[0111] The power optimization model sends the optimization results (input power set value) to the control module and the regulation module, providing a target benchmark for subsequent feedback and feedforward regulation.

[0112] The control module (fuzzy PID controller) uses the input power setting value given by the power optimization model as a reference and performs real-time feedback control in seconds or shorter cycles. The current error is calculated through fuzzy PID, and the actual input power is dynamically adjusted to ensure rapid tracking of the target value and stable power output.

[0113] The control module (fuzzy PID controller) obtains the set value from the power optimization model or the regulation module, and outputs instructions to the inverter or energy storage management system in the "execution module" to achieve closed-loop regulation of the actual power.

[0114] The regulation module (feedforward regulation unit) includes a load prediction algorithm, which predicts subsequent load change trends by analyzing historical and real-time data. If it is predicted that the load is about to change significantly, the original set value of the power optimization model is corrected or updated to achieve feedforward power regulation, and the bidirectional inverter is used to store or release energy at appropriate times to further smooth the system operation.

[0115] The regulation module (feedforward regulation unit) interacts with the data management module to obtain the latest multimodal data, and shares or updates the set value with the power optimization model; and outputs the corrected input power reference value to the control module.

[0116] The execution module ultimately executes the control and regulation instructions, including hardware units such as inverters, energy storage management systems, and grid interaction interfaces; it completes actual charging and discharging or power conversion according to the power instructions issued by the control module or regulation module.

[0117] The execution module receives the adjustment commands output by the control module and feeds back the execution results (such as current power, SoC status, etc.) to the acquisition module and data management module in real time.

[0118] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs.

[0119] The logic and / or steps represented in the flowchart or described in other ways herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch instructions from the instruction execution system, apparatus, or device and execute the instructions), or in combination with these instruction execution systems, apparatus, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in combination with an instruction execution system, apparatus, or device.

[0120] More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection part with one or more wirings (electronic device), a portable computer disk cartridge (magnetic device), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber device, and portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, then editing, interpreting, or otherwise processing it as appropriate, and then storing it in a computer memory.

[0121] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0122] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A power input power optimization method based on intelligent data analysis, characterized in that: include: Collect multimodal data, build a power optimization model based on the multimodal data, and obtain the input power setting value; The constructing of the power optimization model according to the multimodal data includes acquiring load prediction data and power cost data according to the multimodal data, and constructing a multi-objective optimization function, which is expressed as: ； in, Indicates taking the minimum value; Indicates time Input power setting value; , , represents the weight coefficient; Indicates the operating cost; Represents the life attenuation cost of the energy storage device; Indicates the power grid impact index; The running cost It is expressed as the weighted sum of the cost of obtaining electricity from the grid during the dispatch period: ； in, Indicates the start time of the scheduling cycle; Indicates the length of the optimized rolling time domain; Indicates the power exchanged with the grid side; represents the time step; Represents the cost of exchanging electric energy with the grid; The equivalent life loss of the charge and discharge cycle is quantified through an empirical model to obtain the life attenuation cost of the energy storage device , expressed as: ； in, Represents the cost function of battery life consumed by the current charging and discharging operation; Indicates time The state of charge; express changes; Indicates the battery charging and discharging current; Indicates the internal temperature of the battery; The grid impact index The power exchange value between the load on the grid side and the grid is expressed as: ； in, Indicates time The power exchange value between the grid-side load and the grid; Indicates time The power exchange value between the grid-side load and the grid; The fuzzy PID controller is used to dynamically adjust the input power according to the input power setting value; The load prediction value is obtained through the load prediction algorithm, and the input power setting value is adjusted according to the load prediction value to realize feedforward power regulation.

2. The power input power optimization method based on intelligent data analysis according to claim 1, characterized in that: The collecting of multimodal data includes collecting real-time grid status data and load data, and introducing environmental data and battery status as auxiliary optimization parameters; The real-time grid status data and load data include grid voltage, current, load forecast data, power cost data, and power exchange value between load and grid; The auxiliary optimization parameters include the internal temperature of the battery, the equivalent life loss of the battery charge and discharge cycle, the battery Data and battery charge and discharge current data.

3. The power input power optimization method based on intelligent data analysis according to claim 2, characterized in that: The empirical model includes the following steps: Indicates that the auxiliary optimization parameters are obtained through multimodal data, combined with the energy storage in each time step Calculate the life-time decay cost by changing the amount , expressed as: ； in, represents the basic decay rate; Indicates the time step Amount of change; represents the temperature correction factor; Indicates the rate correction factor; The temperature correction factor Describe the impact of temperature exceeding the basic operating conditions on battery aging, expressed as: ； in, Indicates the sensitivity of temperature to the decay rate; Indicates the reference temperature under basic working conditions; Ratio correction factor The effect of the difference between the actual charge and discharge current and the basic operating current on attenuation is expressed as: ； in, Indicates the sensitivity of the charge and discharge current rate to the decay rate; Indicates the actual charge and discharge current; Indicates the reference current under basic working conditions; by introducing temperature correction factor and rate correction factor, the battery life attenuation is evaluated in real time in each scheduling cycle; when the temperature or rate deviation is too large, the life attenuation cost is adaptively increased , guiding the dispatch plan to balance power allocation and life protection.

4. The power input power optimization method based on intelligent data analysis according to claim 3, characterized in that: The fuzzy PID controller includes collecting basic input information, wherein the basic input information includes an input power setting value , Actual input power and energy storage life attenuation index The energy storage life decay index is based on the life decay cost Calculate acquisition; Calculate the error between the input power setting value and the actual input power and error variation ; Will , and Perform fuzzification, construct membership functions and language values; Establish a fuzzy rule base, based on the two-dimensional rules of error and error change As a branch judgment, the judgment results of the two-dimensional rules of error and error change are corrected, and the results of the fuzzy rules are converted into specific values ​​through the centroid method, and the PID parameters are updated according to the specific values: ； in, Indicates the final control amount; represents proportional gain; represents the integral gain; It means collecting and accumulating the sum of error changes over time; Represents the differential gain; the control quantity Converts to input power adjustment commands.

5. The power input power optimization method based on intelligent data analysis according to claim 4, characterized in that: The step of adjusting the input power setting value according to the load prediction value includes: outputting the input power setting value in the future rolling window through time series prediction according to the multimodal data. Predicted load curve within; Get the input power setting value ,use Compare with the load prediction value to obtain the load deviation. According to the load deviation and energy storage constraint, a feedforward correction is generated to pre-adjust the input power setting value, which is expressed as: ； in, Indicates the corrected power setting value; Represents the feedforward correction amount.

6. The power input power optimization method based on intelligent data analysis according to claim 5, characterized in that: The feedforward power regulation includes sending the feedforward corrected input power setting sequence to the fuzzy PID controller, and the fuzzy PID continues to make fine adjustments according to the actual error in a short time scale, forming a double guarantee of feedforward and feedback; When the deviation between the actual load and the predicted load exceeds the allowable range, the actual input power is adjusted through fuzzy PID feedback control to quickly eliminate the deviation caused by the error, and the feedforward control is updated again when the next prediction cycle arrives.

7. A power supply input power optimization system based on intelligent data analysis, applied to a power supply input power optimization method based on intelligent data analysis as claimed in any one of claims 1 to 6, characterized in that: include, An acquisition module collects multimodal data, builds a power optimization model based on the multimodal data, and obtains an input power setting value; The control module uses a fuzzy PID controller to dynamically adjust the input power according to the input power set value; The regulation module obtains the load prediction value through the load prediction algorithm, adjusts the input power setting value according to the load prediction value, and realizes feedforward power regulation.

8. A computer device comprising: Memory and processor; The memory stores a computer program, characterized in that: when the processor executes the computer program, the steps of the power input power optimization method based on intelligent data analysis as described in any one of claims 1-6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the power supply input power optimization method based on intelligent data analysis as described in any one of claims 1 to 6 are implemented.

Images