Wind-solar complementary cold chain energy storage energy scheduling method and system

Abstract

This invention belongs to the field of cold chain scheduling technology, specifically relating to a wind-solar hybrid cold chain energy storage scheduling method and system. The method includes: collecting operating data from each compressor head in a multi-parallel unit; establishing an energy efficiency model between load rate and energy efficiency ratio based on Gaussian process regression; monitoring the start-stop state changes and duration of each compressor head, and calculating mechanical fatigue penalty values; calculating the marginal energy efficiency gradient of each compressor head according to the energy efficiency model; constructing a global objective function by comprehensively predicting operating power consumption, mechanical fatigue penalty values, and insufficient cooling capacity penalty; and using the marginal energy efficiency gradient to heuristically guide the solution process of the global objective function, obtaining the start-stop state combinations and load rate setpoints for each compressor head. This invention can effectively balance system energy consumption and equipment mechanical losses while meeting cooling load requirements, improving solution efficiency and scheduling accuracy.

Description

Technical Field

[0001] This invention relates to the field of cold chain scheduling technology. More specifically, this invention relates to a method and system for scheduling energy in a wind-solar hybrid cold chain energy storage system. Background Technology

[0002] With the rapid development of the cold chain logistics industry, the energy consumption of cold storage facilities and refrigerated trucks is increasing. In order to reduce operating costs and carbon emissions, wind-solar hybrid power generation systems and energy storage technologies are being widely used in cold chain energy supply.

[0003] However, existing control strategies for multiple parallel refrigeration units in cold chain refrigeration systems have significant limitations when dealing with wind-solar hybrid power supply scenarios. Cold chain refrigeration systems typically consist of multiple parallel compressor units, and the energy efficiency characteristics of each compressor head differ nonlinearly under different operating conditions. Traditional control methods are usually based on simple temperature thresholds or proportional-integral-derivative (PID) algorithms, starting the compressor heads in a fixed proportion or sequence. This coarse control logic ignores the significant nonlinear characteristics of the compressor's energy efficiency ratio under different load rates, meaning that efficiency is usually highest only in specific load ranges. Blindly allocating load proportionally often leads to multiple units operating simultaneously in low-load, inefficient ranges, resulting in huge energy waste. In addition, existing scheduling strategies lack unified coordination for the start-up and shutdown of fixed-frequency units and the regulation of variable-frequency units, making it difficult to achieve optimal energy efficiency at the system level.

[0004] Because wind and solar energy are intermittent and fluctuating, they can easily cause frequent start-stop of industrial frequency compressor units. The mechanical wear at the moment of compressor start-up is far greater than that during normal operation. Frequent start-stop will cause serious mechanical fatigue and shorten the equipment life, and will also cause unnecessary fluctuations in storage temperature.

[0005] Therefore, there is an urgent need for a wind-solar hybrid cold chain energy storage energy dispatching method and system that can balance nonlinear energy efficiency, source-side fluctuations and mechanical lifespan. Summary of the Invention

[0006] To address the technical problems of existing technologies in wind-solar hybrid cold chain energy storage scenarios, which struggle to balance the nonlinear energy efficiency characteristics of generating units, the random fluctuations in wind and solar power, and the lifespan of mechanical equipment, leading to energy waste and severe mechanical wear, this invention provides solutions in the following aspects.

[0007] In a first aspect, the present invention provides a wind-solar hybrid cold chain energy storage energy dispatch method, comprising:

[0008] The system collects operating data from each compressor head in a multi-parallel unit, calculates the load rate and energy efficiency ratio of each compressor head, and establishes an energy efficiency model between load rate and energy efficiency ratio based on Gaussian process regression. It monitors the start-stop state changes and current state duration of each compressor head, and calculates the corresponding mechanical fatigue penalty value. Based on the energy efficiency function of each compressor head in the energy efficiency model and the derivative of the energy efficiency function with respect to the load rate, it calculates the marginal energy efficiency gradient of each compressor head. Based on the predicted operating power consumption, mechanical fatigue penalty value, and insufficient cooling capacity penalty determined based on the total cooling load demand and predicted total output cooling capacity of each compressor head calculated using the energy efficiency model, a global objective function is constructed. The marginal energy efficiency gradient is used to heuristically guide the solution process of the global objective function, obtaining the start-stop state combination and load rate setpoint of each compressor head.

[0009] This invention collects operating data from multiple parallel units and uses Gaussian process regression to establish an energy efficiency model between load rate and energy efficiency ratio, which can accurately fit the nonlinear energy efficiency characteristics of the compressor head under complex operating conditions. Simultaneously, it monitors start-up and shutdown status changes and introduces mechanical fatigue penalty values ​​to measure the impact of frequent start-ups and shutdowns on equipment wear, thereby effectively suppressing unnecessary operations during scheduling to extend equipment lifespan. Furthermore, it calculates the marginal energy efficiency gradient based on the energy efficiency model and constructs a global objective function by combining predicted operating power consumption, mechanical fatigue, and insufficient cooling capacity penalties. The marginal energy efficiency gradient provides heuristic guidance for solving the objective function, which not only improves the convergence speed and solution accuracy of the optimization algorithm but also effectively balances the contradictions between system energy consumption, equipment protection, and cooling requirements, ensuring the economical, efficient, and reliable operation of the cold chain energy storage system.

[0010] Preferably, the step of collecting operating data of each compressor head in the multi-parallel unit and calculating the load rate and energy efficiency ratio of each compressor head includes: collecting the operating parameters of each compressor head in real time through a sensor network deployed on the refrigeration unit, the operating parameters including real-time input power, cooling capacity, suction pressure, discharge pressure, condensing temperature and evaporating temperature; determining the rated cooling capacity of each compressor head under the current operating conditions based on the collected condensing temperature and evaporating temperature; using the ratio of the collected cooling capacity to the rated cooling capacity under the current operating conditions as the load rate, and the ratio of the collected cooling capacity to the input power as the energy efficiency ratio.

[0011] Preferably, the step of establishing an energy efficiency model between load rate and energy efficiency ratio includes: using historical data sequences as a training set, and using Gaussian process regression to fit the mapping relationship between load rate and energy efficiency ratio of each compressor head, as an energy efficiency model between load rate and energy efficiency ratio.

[0012] Preferably, the mechanical fatigue penalty value satisfies the expression: ;in, Indicates the first The mechanical fatigue penalty value of the compressor head at the current moment; Indicates the fatigue loss coefficient; Indicates the first The compressor head at the current moment The start / stop status, with a value of 0 or 1; Indicates the first The compressor head at the previous moment Start-stop status; Indicates the first The duration for which the compressor head maintains its current state; This represents the safe cooling time constant of the compressor head; This represents an exponential function with the natural constant e as its base.

[0013] This invention constructs a mechanical fatigue penalty value that includes start-stop state changes and state durations, and introduces an exponential decay term with a base of the natural constant to correlate with the safe cooling time. This can measure the risk of mechanical wear and tear on equipment caused by frequent start-stop operations, thereby suppressing unnecessary short-term frequent switching during scheduling, effectively protecting the compressor head and extending the service life of the equipment.

[0014] Preferably, the marginal energy efficiency gradient satisfies the expression: ;in, Indicates the first The marginal energy efficiency gradient of the compressor head; Indicates the first The compressor head is at a load rate Energy efficiency function at time; Indicates the first The load rate of the compressor head; This represents the derivative of the energy efficiency function with respect to the load factor.

[0015] This invention obtains the marginal energy efficiency gradient by calculating the derivative of the energy efficiency function with respect to the load rate. It can measure the potential contribution of each compressor head to the overall energy efficiency when increasing the load at the current load point, thereby prioritizing the adjustment of compressor heads with high potential for energy efficiency improvement during scheduling, providing clear guidance for achieving refined load allocation and energy efficiency optimization.

[0016] Preferably, constructing the global objective function includes: ;in, Represents the global generalized equivalent power cost; Indicates the total number of compressor heads; Indicates the first compressor head at load rate Predicted operating power consumption; Indicates the first Mechanical fatigue penalty value of the compressor head; Indicates the system's scheduling and control cycle; This indicates the penalty weight for insufficient cooling capacity; This indicates the total cooling load demand of the cold storage at the current moment; This indicates the predicted total cooling output under the current compressor head configuration; This represents the maximum value function.

[0017] This invention comprehensively considers predicted operating power consumption, mechanical fatigue penalty, and insufficient cooling capacity penalty to construct a global objective function. It unifies the three mutually constraining objectives of system energy saving, equipment protection, and meeting refrigeration needs into a generalized cost model, ensuring that the scheduling strategy can minimize the overall cost of operating energy consumption and mechanical losses while ensuring that the cold storage temperature meets the standard.

[0018] Preferably, the predicted operating power consumption satisfies the expression: The predicted total output cooling capacity satisfies the expression: ;in, This is the rated cooling capacity under the current operating conditions; Indicates the first The load rate of the compressor head; For the first The compressor head is at a load rate The energy efficiency function at that time.

[0019] Preferably, the heuristic guidance for solving the global objective function using marginal energy efficiency gradients includes: using a discrete differential evolution algorithm to solve for the minimum value of the global objective function; mapping the marginal energy efficiency gradients of each compressor head to normalized selection probabilities during the solution process; and using a modified mutation operator that introduces a heuristic drift term when generating the mutation vector of the population. ;in, Indicates the first The first generation of the population The 1st mutant individual Dimensional component values; , as well as Indicates from the first The third generation of three different individuals randomly selected from the population Dimensional component values; Indicates the scaling factor; This represents the heuristic guidance strength coefficient; For the first The probability of selecting a compressor head; This represents the average probability threshold under unbiased conditions.

[0020] This invention introduces a heuristic guidance mechanism based on marginal energy efficiency gradients into the discrete differential evolutionary algorithm. By mapping gradient information to selection probabilities and constructing a modified mutation operator with heuristic drift terms, it can guide the search direction using gradient information while preserving population diversity. This effectively overcomes the problems of traditional evolutionary algorithms being prone to getting trapped in local optima and having slow convergence speed, and improves the efficiency and accuracy of solving the global optimal scheduling strategy.

[0021] Preferably, mapping the marginal energy efficiency gradient of each compressor head to a normalized selection probability includes: normalizing the marginal energy efficiency gradient of each compressor head using a softmax function to obtain the selection probability of each compressor head.

[0022] Secondly, the present invention provides a wind-solar hybrid cold chain energy storage energy dispatching system, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned wind-solar hybrid cold chain energy storage energy dispatching method is implemented.

[0023] By adopting the above technical solution, a computer program is generated from the above-mentioned wind-solar hybrid cold chain energy storage scheduling method and stored in a memory so that it can be loaded and executed by a processor. Terminal equipment can then be made based on the memory and processor for convenient use.

[0024] The beneficial effects of this invention are as follows: By collecting operating data from multiple parallel units and constructing a nonlinear energy efficiency model using Gaussian process regression, this invention can accurately identify the high-efficiency operating range of the compressor head, avoiding the inefficient operation problem caused by traditional proportional allocation; by constructing a mechanical fatigue penalty function to convert start-up and shutdown losses into virtual operating costs, this invention establishes an anti-vibration mechanism for random fluctuations in wind and solar power, effectively suppressing unnecessary frequent start-ups and shutdowns and extending equipment lifespan; this invention utilizes the marginal energy efficiency gradient to reflect the instantaneous efficiency enhancement capability of the compressor head, and uses this as a heuristic guiding factor to correct the search direction of the discrete difference evolution algorithm, guiding the system load to preferentially allocate to compressor heads with high efficiency enhancement potential, ensuring that the unit always operates on the ridge line of the synthetic energy efficiency curve; this invention constructs a global objective function that includes operating power consumption, mechanical fatigue, and insufficient cooling capacity penalties, seeking the minimum comprehensive operating cost under the premise of strictly meeting cooling requirements, solving the problem of mismatch between random fluctuations on the source side and stable demand on the load side, and realizing efficient scheduling of wind-solar complementary cold chain energy storage. Attached Figure Description

[0025] Figure 1 This is a schematic flowchart illustrating a wind-solar hybrid cold chain energy storage energy dispatching method according to the present invention;

[0026] Figure 2 A graph showing the actual cooling capacity versus the load demand;

[0027] Figure 3 This is a Gantt chart for multi-machine collaborative operation in this invention;

[0028] Figure 4 A bar chart comparing the number of unit start-ups and shutdowns;

[0029] Figure 5 This is a comparison chart of total operating costs. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0032] This invention discloses a wind-solar hybrid cold chain energy storage energy dispatching method, referring to... Figure 1 This includes steps S1-S4:

[0033] S1. Collect the operating data of each compressor head in the multi-parallel unit, calculate the load rate and energy efficiency ratio of each compressor head, and establish an energy efficiency model between load rate and energy efficiency ratio based on Gaussian process regression.

[0034] It should be noted that the operating energy efficiency of refrigeration units in cold chain industrial parks depends not only on the load rate but also significantly on environmental conditions. For example, under the same load rate, the energy efficiency ratio differs greatly between high-temperature, high-pressure conditions in summer and low-temperature, low-pressure conditions in winter. Directly using raw data for fitting would result in poor model generalization ability, failing to accurately reflect the true characteristics of the unit. Furthermore, traditional control logic typically allocates load proportionally, ignoring the significant nonlinear characteristics of the compressor head's energy efficiency ratio under different load rates; that is, efficiency is usually highest within a specific load range, exhibiting a single-peak characteristic. Therefore, this invention first calculates normalized load rates and real-time energy efficiency based on multi-dimensional operating parameters to eliminate the impact of operating condition fluctuations. Then, it constructs a digital energy efficiency model for each compressor head, accurately identifying the high-efficiency operating range of each compressor head, providing a physical model basis for subsequent precise scheduling.

[0035] Specifically, a sensor network deployed on the refrigeration unit collects the operating parameters of each compressor head in real time. These parameters include real-time input power, cooling capacity, suction pressure, discharge pressure, condensing temperature, and evaporating temperature. Based on the collected condensing and evaporating temperatures, the rated cooling capacity of each compressor head under the current operating conditions is determined by consulting performance curves provided by the compressor manufacturer or by using refrigeration thermodynamics equations. The ratio of the collected real-time cooling capacity to the rated cooling capacity under the current operating conditions is taken as the load factor, denoted as . The ratio of the collected real-time cooling capacity to the real-time input power is taken as the energy efficiency ratio, denoted as . Using the processed historical data sequence As a training set, Gaussian process regression was used to fit the mapping relationship between the load rate and energy efficiency ratio of each compressor head. The compressor head is at a load rate The energy efficiency function at that time satisfies the following expression:

[0036]

[0037] in, Indicates the first The compressor head is at a load rate Energy efficiency ratio at that time; Indicates the first The load rate of the compressor head, the range of values ​​is: ; This indicates a Gaussian process regression model, which can capture nonlinear trends in data and eliminate noise interference; The first one obtained through training The hyperparameters of the compressor head model. When the load rate... When within a specific range, the energy efficiency ratio It reaches its peak value and exhibits the characteristics of a unimodal convex function.

[0038] S2. Monitor the start-stop status changes of each compressor head and the duration of the current status, and calculate the corresponding mechanical fatigue penalty value.

[0039] It should be noted that, due to the random fluctuations in the output power of wind-solar hybrid power generation systems, maintaining power balance can easily lead to frequent start-stops of the fixed-frequency compressor head. The mechanical wear during compressor start-up is far greater than during normal operation, resulting in a shortened equipment lifespan. Therefore, this invention converts mechanical wear into virtual equivalent energy consumption and constructs an anti-vibration mechanism through a mechanical fatigue penalty function. When the compressor head switches frequently, a large penalty value is generated to suppress this behavior, thereby protecting equipment lifespan while participating in the scheduling game.

[0040] Specifically, the first The mechanical fatigue penalty function of the compressor head satisfies the following expression:

[0041]

[0042] in, Indicates the first The mechanical fatigue penalty value of the compressor head at the current moment; It represents the fatigue loss coefficient, which represents the electrical energy value lost during the startup of a compressor; Indicates the first The compressor head at the current moment The start / stop status, with a value of 0 or 1; Indicates the first The compressor head at the previous moment Start-stop status; This indicates a state switching indicator. The value is 1 if and only if the compressor head state changes, otherwise it is 0. Indicates the first The duration for which the compressor head maintains its current state; This represents the safe cooling time constant of the compressor head; This represents an exponential function with base e, used to describe... Perform negative correlation mapping.

[0043] If the duration of maintaining the current state The smaller the value, that is, when the compressor head has just started moving, the smaller the exponent. The closer it gets to 1, the greater the mechanical fatigue penalty value. The larger the value, the more the system tends to maintain the compressor head state unchanged; when the duration of maintaining the current state... The larger the exponent, meaning the compressor head has been running stably for a long time, the greater the exponent. The closer it gets to 0, the lower the mechanical fatigue penalty value becomes. The smaller the value, the more the system allows the compressor head to switch states.

[0044] The fatigue loss coefficient is obtained by calculating the equipment loss value per start-stop cycle based on the average number of start-stop cycles and the unit price of the equipment provided by the compressor manufacturer, and then converting it into an equivalent electrical energy value. The empirical value is 2 kWh. The safe cooling time constant is set according to the minimum downtime or minimum running time specified in the compressor technical manual, for example, 300 seconds. In other embodiments, the implementer can set the fatigue loss coefficient and the safe cooling time constant according to the actual implementation situation.

[0045] S3. Calculate the marginal energy efficiency gradient of each compressor head based on the energy efficiency function of each compressor head in the energy efficiency model and the derivative of the energy efficiency function with respect to the load rate.

[0046] It should be noted that since the total available power of the system varies with the output of wind and solar power, the system needs to dynamically allocate the load increment among multiple compressor heads. Therefore, this invention utilizes the marginal utility principle of optimization theory and uses the marginal energy efficiency gradient to reflect the instantaneous efficiency enhancement capability of the compressor head. During scheduling, the algorithm prioritizes allocating the increased load to the compressor head with the largest marginal energy efficiency gradient, or prioritizes reducing the load of the compressor head with the smallest marginal energy efficiency gradient, thereby ensuring that all compressor heads, as a whole, always work on the ridge line of the composite energy efficiency curve, achieving energy efficiency optimization at the system level.

[0047] Specifically, calculate the marginal energy efficiency gradient for each compressor head that is ready to start:

[0048]

[0049] in, Indicates the first The marginal energy efficiency gradient of the compressor head; Indicates the first The compressor head is at a load rate Energy efficiency function at time; Indicates the first The load rate of the compressor head; This represents the derivative of the energy efficiency function with respect to the load factor. When the marginal energy efficiency gradient... The larger the value, the more efficient the operation becomes for each additional unit of input power. The more cooling capacity a compressor head can produce, the greater its efficiency potential; when the marginal efficiency gradient... The smaller the value, the more it indicates that for every unit increase in input power, the [number]th [unit] [unit] [is needed]. The less cooling capacity a compressor head can produce, the less revenue it may generate, and it may even enter an inefficient zone.

[0050] S4. Based on the predicted operating power consumption and mechanical fatigue penalty value of each compressor head calculated based on the energy efficiency model, as well as the insufficient cooling penalty determined based on the total cooling load demand and the predicted total output cooling capacity, a global objective function is constructed. The marginal energy efficiency gradient is used to heuristically guide the solution process of the global objective function to obtain the start-stop state combination and load rate setpoint of each compressor head.

[0051] It should be noted that multi-compressor scheduling is essentially a process of seeking the optimal solution under multiple constraints. If the goal is simply to minimize power consumption, the algorithm may tend to reduce the number of compressors in operation, resulting in insufficient cooling output to meet actual needs, which in turn leads to increased storage temperature and affects the quality of goods. On the other hand, if the goal is simply to meet cooling demand without considering cost, the compressors may operate in inefficiently or experience frequent start-stop cycles. Therefore, this invention constructs a global objective function that converts mechanical fatigue losses into equivalent power losses within the current scheduling cycle, finding a balance between meeting cooling requirements, saving energy, and protecting the machine.

[0052] Specifically, construct the global objective function for the current moment:

[0053]

[0054] in, This represents the global generalized equivalent power cost, and the goal is to minimize it. Indicates the total number of compressor heads; Indicates the first compressor head at load rate Predicted operating power consumption: ,in This is the rated cooling capacity under current operating conditions. For the first The compressor head is at a load rate Energy efficiency function at time; Indicates the first Mechanical fatigue penalty value of the compressor head; Indicates the system's scheduling and control cycle; This indicates the penalty weight for insufficient cooling capacity; This indicates the total cooling load demand of the cold storage at the current moment; This indicates the predicted total cooling output under the current compressor head configuration: ; This represents the maximum value function.

[0055] When predicting the total output cooling capacity Less than the total cooling load demand At that time, the difference If it is a positive number, then... A positive penalty value is generated, and the penalty weight is increased due to insufficient cooling. Setting it to a maximum value leads to a global generalized equivalent power cost. A dramatic increase; when predicting total output cooling capacity Greater than or equal to the total cooling load demand hour, The value is 0, at which point the global generalized equivalent power cost is 0. It mainly depends on the predicted operating power consumption and the equivalent mechanical fatigue power.

[0056] The scheduling control cycle is set according to the actual instruction refresh frequency of the control system, in hours, for example, 0.167 hours. This invention uses the scheduling control cycle to convert the mechanical fatigue penalty value from energy to power. The insufficient cooling capacity penalty weight should be much greater than the sum of the operating power consumption item and the mechanical fatigue item, ensuring that meeting the cooling demand is the first priority. In this embodiment, the insufficient cooling capacity penalty weight is set to 1000. In other embodiments, implementers can set the scheduling control cycle and the insufficient cooling capacity penalty weight according to the actual implementation situation.

[0057] It should be noted that the global objective function It concerns the load rate of all compressor heads. The function, for a fixed-frequency compressor head, has a load rate The value constraint is a discrete set. These correspond to the shutdown and full-load states, respectively; for the variable frequency compressor head, its load rate The value of is constrained to be a continuous interval This corresponds to shutdown, partial load to full load, and so on. Therefore, the process of finding the minimum of the global objective function is essentially finding the optimal load rate vector in the multidimensional solution space. The process.

[0058] The minimum value of the global objective function is solved using the Discrete Differential Evolutionary Algorithm (DDE). In the DDE, the population represents a set of candidate scheduling schemes, and each individual represents a specific scheduling scheme. The dimension of the individual corresponds to the number of compressor heads. , No. The value of the dimensional component is the first dimensional component. The candidate load rate of each compressor head. To improve optimization efficiency, this invention introduces a heuristic guidance mechanism based on marginal energy efficiency gradient: First, the marginal energy efficiency gradient of each compressor head is calculated using the Softmax function. Mapped to normalized selection probabilities: Subsequently, in generating the first When calculating the mutation vector of the population, a modified mutation operator that introduces a heuristic drift term is used:

[0059]

[0060] in, Indicates the first The first generation of the population The 1st mutant individual Dimensional component values; , , Indicates from the first The third generation of three different individuals randomly selected from the population Dimensional component value, that is, the first dimensional component value in the previous generation population. Candidate load rate of the compressor head; This represents the scaling factor, used to control the magnitude of random disturbances; This represents the heuristic guidance strength coefficient, used to adjust the weight of the influence of prior physical knowledge on the direction of variation. For the first The probability of selecting a compressor head; The average probability threshold under unbiased conditions is represented by the value of . .

[0061] In this modified mutation operator, For the heuristic drift term, when the first... Marginal energy efficiency gradient of a compressor head A larger probability leads to its selection Above the average probability threshold When the drift term is positive, the mutated load rate tends to increase, meaning the algorithm tends to increase the load on the high-efficiency compressor head during evolution; conversely, it tends to decrease the load. Scaling factor This determines the algorithm's ability to explore the solution space. A value that is too large may cause the algorithm to fail to converge, while a value that is too small may lead to getting trapped in local optima. It is typically set between 0.5 and 0.9; in this embodiment, it is set to 0.6. Heuristic guidance strength coefficient. This determines the degree to which prior physical knowledge influences the algorithm's search direction. An excessively large value may cause the algorithm to degenerate into a greedy search and lose diversity, while an excessively small value will not be able to exert the acceleration effect of gradient guidance. It is usually set between 0.1 and 0.5. In this embodiment, it is set to 0.3. In other embodiments, implementers can set the scaling factor and heuristic guidance strength coefficient according to the actual implementation situation.

[0062] Through iterative solution using the discrete differential evolution algorithm described above, the optimal load factor vector that minimizes the global generalized equivalent power cost is finally obtained. This vector is then converted into a scheduling execution signal and transmitted to the refrigeration system execution end, driving each compressor head to perform corresponding start-stop operations or frequency changes. This achieves energy scheduling of the wind-solar hybrid cold chain system while meeting the cooling demand.

[0063] For example, Figure 2The graph shows the actual cooling capacity versus load demand. It can be seen that even under conditions of drastic load fluctuations, the red solid line still closely follows the gray dashed line, without any obvious insufficient or excessive cooling. This indicates that the insufficient cooling penalty term introduced in the global objective function plays a crucial role, ensuring that the system always prioritizes meeting cooling needs while pursuing energy saving and lifespan protection.

[0064] Figure 3 This is a Gantt chart for multi-machine collaborative operation in this invention. Figure 3 The horizontal axis represents time, and the vertical axis represents different compressor head numbers. The orange blocks indicate that the fixed-frequency compressor head is in the on state, and the blue bars represent the load rate changes of the variable-frequency compressor head; the darker the color, the higher the load rate. From... Figure 3 It can be clearly seen that during load fluctuations, the fixed-frequency compressor head is in a long-term, continuous state of operation, avoiding frequent start-stop cycles; while the variable-frequency compressor head fills the load gap by flexibly adjusting the load rate, undertaking the main task of fluctuation regulation.

[0065] Figure 4 A bar chart comparing the number of unit start-ups and shutdowns. Figure 4 The total number of unit actions under the same operating conditions was compared between the traditional greedy strategy and the optimized strategy of this invention. The traditional greedy strategy had a total of 94 actions, while the optimized strategy of this invention significantly reduced the total number of actions to 28. The method of this invention reduced the start-stop switching actions of the unit by about 70.2%, which demonstrates the excellent performance of the mechanical fatigue penalty value in this invention in suppressing unnecessary actions and reducing mechanical wear.

[0066] Figure 5 The comparison chart shows that although the power consumption of the optimization strategy of this invention is basically the same as that of the traditional strategy, the proportion of mechanical fatigue loss is significantly reduced due to the large reduction in the number of actions, thus making the generalized total cost of the system significantly better than that of the traditional strategy.

[0067] This invention also discloses a wind-solar hybrid cold chain energy storage energy dispatching system, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a wind-solar hybrid cold chain energy storage energy dispatching method according to the present invention is implemented.

[0068] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

Claims

1. A wind-solar hybrid cold chain energy storage energy dispatching method, characterized in that, include: The operating data of each compressor head in the multi-parallel unit is collected, the load rate and energy efficiency ratio of each compressor head are calculated, and an energy efficiency model between load rate and energy efficiency ratio is established based on Gaussian process regression. Monitor the start-stop status changes of each compressor head and the duration of the current status, and calculate the corresponding mechanical fatigue penalty value; Based on the energy efficiency function of each compressor head in the energy efficiency model and the derivative of the energy efficiency function with respect to the load rate, calculate the marginal energy efficiency gradient of each compressor head; A global objective function is constructed based on the predicted operating power consumption and mechanical fatigue penalty value of each compressor head calculated based on the energy efficiency model, as well as the insufficient cooling penalty determined based on the total cooling load demand and the predicted total output cooling capacity. The marginal energy efficiency gradient is used to heuristically guide the solution process of the global objective function, thereby obtaining the start-stop state combination and load rate setpoint of each compressor head; The mechanical fatigue penalty value satisfies the expression: ; in, Indicates the first The mechanical fatigue penalty value of the compressor head at the current moment; Indicates the fatigue loss coefficient; Indicates the first The compressor head at the current moment The start / stop status, with a value of 0 or 1; Indicates the first The compressor head at the previous moment Start-stop status; Indicates the first The duration for which the compressor head maintains its current state; This represents the safe cooling time constant of the compressor head; This represents an exponential function with the natural constant e as its base. Construct the global objective function, including: ;in, Represents the global generalized equivalent power cost; Indicates the total number of compressor heads; Indicates the first compressor head at load rate Predicted operating power consumption; Indicates the first Mechanical fatigue penalty value of the compressor head; Indicates the system's scheduling and control cycle; This indicates the penalty weight for insufficient cooling capacity; This indicates the total cooling load demand of the cold storage at the current moment; This indicates the predicted total cooling output under the current compressor head configuration; This represents the maximum value function.

2. The energy dispatching method for wind-solar hybrid cold chain energy storage according to claim 1, characterized in that, The process of collecting operating data from each compressor head in a multi-parallel unit and calculating the load rate and energy efficiency ratio of each compressor head includes: The operating parameters of each compressor head are collected in real time by a sensor network deployed on the refrigeration unit. The operating parameters include real-time input power, cooling capacity, suction pressure, discharge pressure, condensing temperature, and evaporating temperature. Based on the collected condensing temperature and evaporating temperature, the rated cooling capacity of each compressor head under the current operating conditions is determined. The ratio of the collected cooling capacity to the rated cooling capacity under the current operating conditions is used as the load rate, and the ratio of the collected cooling capacity to the input power is used as the energy efficiency ratio.

3. The energy dispatching method for wind-solar hybrid cold chain energy storage according to claim 1, characterized in that, The establishment of the energy efficiency model between load rate and energy efficiency ratio includes: Using historical data sequences as the training set, Gaussian process regression was used to fit the mapping relationship between the load rate and energy efficiency ratio of each compressor head, which served as the energy efficiency model between the load rate and energy efficiency ratio.

4. The wind-solar hybrid cold chain energy storage energy dispatching method according to claim 1, characterized in that, The marginal energy efficiency gradient satisfies the expression: ； in, Indicates the first The marginal energy efficiency gradient of the compressor head; Indicates the first The compressor head is at a load rate Energy efficiency function at time; Indicates the first The load rate of the compressor head; This represents the derivative of the energy efficiency function with respect to the load factor.

5. The energy dispatching method for wind-solar hybrid cold chain energy storage according to claim 1, characterized in that, The predicted operating power consumption satisfies the expression: ; The predicted total output cooling capacity satisfies the expression: ; in, This is the rated cooling capacity under the current operating conditions; Indicates the first The load rate of the compressor head; For the first The compressor head is at a load rate The energy efficiency function at that time.

6. The energy dispatching method for wind-solar hybrid cold chain energy storage according to claim 1, characterized in that, The heuristic guidance for solving the global objective function using marginal energy efficiency gradients includes: using a discrete differential evolution algorithm to find the minimum value of the global objective function; mapping the marginal energy efficiency gradients of each compressor head to normalized selection probabilities during the solution process; and using a modified mutation operator that introduces a heuristic drift term when generating the mutation vector of the population. ;in, Indicates the first The first generation of the population The first mutant individual Dimensional component values; , as well as Indicates from the first The third generation of three different individuals randomly selected from the population Dimensional component values; Indicates the scaling factor; This represents the heuristic guidance strength coefficient; For the first The probability of selecting a compressor head; This represents the average probability threshold under unbiased conditions.

7. The energy dispatching method for wind-solar hybrid cold chain energy storage according to claim 6, characterized in that, The step of mapping the marginal energy efficiency gradient of each compressor head to a normalized selection probability includes: normalizing the marginal energy efficiency gradient of each compressor head using a softmax function to obtain the selection probability of each compressor head.

8. A wind-solar hybrid cold chain energy storage energy dispatch system, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement a wind-solar hybrid cold chain energy storage energy dispatching method according to any one of claims 1-7.

Images