A port yard resource optimization method and device, a storage medium and an electronic device

Abstract

The present application provides a kind of port yard resource optimization method, device, storage medium and electronic equipment.The method comprises: obtaining the operation parameter of port yard;The operation parameter is input into the energy consumption prediction model trained, and the energy consumption prediction value of future preset time period is output by energy consumption prediction model;The energy consumption prediction value is input into lighting effect evaluation model, and the score of lighting effect is obtained by lighting effect evaluation model, obtains lighting evaluation score;Whether lighting evaluation score is greater than preset score threshold is judged, and according to the result of judging, the light modulation coefficient of port yard is calculated by light modulation coefficient calculation model, so that the first light modulation coefficient of ship area is greater than first preset value, and the second light modulation coefficient of storage area is greater than second preset value.According to the operation parameter and its prediction result obtained in real time, lighting intensity can be dynamically adjusted to adapt to the change of port operation.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and more specifically, to a method, apparatus, storage medium, and electronic device for optimizing port yard resources. Background Technology

[0002] Currently, port yard lighting energy consumption accounts for 25%-35% of the total energy consumption of the port. It is a key infrastructure for maintaining 24-hour operation of the yard. Calculating port yard lighting energy consumption and further controlling energy consumption can reduce port costs.

[0003] Traditional port yard lighting systems employ either constant illuminance or timed switching control. Constant illuminance control ensures a consistent light intensity regardless of weather conditions or time periods, based on a preset lighting brightness level. Timed switching control automatically turns the lighting system on or off according to a set schedule.

[0004] However, this type of control method may lead to energy waste, making it difficult to dynamically adjust the lighting intensity according to actual usage needs, and easily causing the lights to be turned on too early or too late. Summary of the Invention

[0005] To address the aforementioned problems, the present invention aims to provide a method, apparatus, storage medium, and electronic device for optimizing port yard resources, capable of calculating port yard lighting energy consumption.

[0006] Firstly, this application provides a method for optimizing port yard resources, including: Obtain the operational parameters of the port yard; the operational parameters include energy consumption, illuminance, number of ships arriving at the port, and cargo density; The operation parameters are input into the trained energy consumption prediction model, and the energy consumption prediction model outputs the predicted energy consumption value for a future preset time period. The predicted energy consumption value is input into the lighting effect evaluation model, and the lighting effect is scored by the lighting effect evaluation model to obtain a lighting evaluation score; The system determines whether the lighting evaluation score is greater than a preset score threshold. Based on the determination result, it calculates the dimming coefficient of the port yard using a dimming coefficient calculation model, so that the first dimming coefficient of the ship area is greater than the first preset value, and the second dimming coefficient of the storage area is greater than the second preset value. The dimming coefficient is used to provide the control module for dimming control.

[0007] In some alternative implementations, the method further includes, before inputting the operation parameters into the trained energy consumption prediction model: Calculate the third-order difference of multiple consecutive operation parameter measurement points ;in, For the first The third-order difference of a series of consecutive points starting from a measurement point, where the measurement point refers to the energy consumption value or illuminance value; Determine the outlier detection threshold Calculate the standard deviation of the residuals ;make ,like The measurement point is then determined to be an outlier. Linear interpolation is used to complete outliers, and the completed data is then standardized using Z-score.

[0008] In some optional implementations, the training process of the energy consumption prediction model includes: The root mean square error (RMSE) is defined as the evaluation index for prediction accuracy. ; The formula for mean absolute percentage error is: ; in, It is the root mean square error value. It is the mean absolute percentage error value. This is the energy consumption value. The energy consumption prediction value is the output of the energy consumption prediction model, and n is the number of samples. The pre-collected data was divided into training, validation, and test sets in a 7:2:1 ratio. The energy consumption prediction model was trained using the pre-collected data. Train the loss function for the model; Set a termination condition for model training: when the validation set has been tested for 5 consecutive rounds... ≤0.35 and Training is complete when the value is ≤0.02.

[0009] In some optional implementations, the lighting effect evaluation model calculates the lighting evaluation score using a fuzzy comprehensive evaluation method, including: Construct an evaluation index set U={u1,u2,u3,u4,u5}; where u1 is the average illuminance of the work area, u2 is the illuminance compliance rate, u3 is the illuminance uniformity, u4 is the glare value, and u5 is the energy consumption per unit area. The weight vector W = (w1, w2, w3, w4, w5) for each indicator is determined using the analytic hierarchy process (AHP), satisfying the following conditions: Among them, the weight ratio of the three criteria of lighting safety, lighting comfort and energy efficiency is 4:3:3; The membership function of the index weight vector is quantified using a linear membership function. The membership function of the average illuminance x in the work area is as follows: ; The membership function of the illuminance uniformity y (unit: lx) of the work area is: ; The membership functions of indicators u2, u4, and u5 are obtained by linearly mapping the actual measured values ​​to the preset threshold intervals. According to the formula Calculate the lighting evaluation score f; where wi is the weight of the i-th indicator and ui is its corresponding membership value.

[0010] In some optional implementations, determining whether the lighting evaluation score is greater than a preset score threshold, and calculating the dimming coefficient of the port yard based on the determination result using a dimming coefficient calculation model, includes: The average lighting energy consumption over a preset period of time is used as the baseline energy consumption. ; according to Calculate the difference in energy consumption prediction ,in, Output energy consumption prediction values ​​for the energy consumption prediction model; If the lighting evaluation score If the score is greater than or equal to the preset score threshold, then it will be processed according to... Calculate the dimming factor k; If the lighting evaluation score If the illuminance is less than the preset score threshold, the average illuminance of the port yard will first be adjusted to be no less than [a certain value]. Adjust the illuminance uniformity to not less than 0.7, and then proceed according to... Calculate the dimming factor k.

[0011] In some optional implementations, the method further includes: Regularly collect the actual illuminance values ​​of the port yard operation area and calculate the deviation between the actual illuminance values ​​and the commanded illuminance values ​​determined by the dimming coefficient; If the deviation is greater than the preset deviation threshold, the dimming coefficient adjustment amount is calculated based on the dimming coefficient, and the dimming coefficient parameter is adjusted based on the dimming coefficient adjustment amount.

[0012] In some optional implementations, the calculation of the deviation between the actual illuminance value and the commanded illuminance value determined by the dimming coefficient includes: pass Calculate the percentage of deviation; where, It's a percentage of deviation. This represents the actual illuminance value in the work area. The commanded illuminance value determined by the dimming factor; The calculation of the dimming coefficient adjustment amount based on the dimming coefficient includes: pass Calculate the dimming factor adjustment amount, where, It is the dimming factor adjustment amount. This is the dimming factor.

[0013] Secondly, this application also provides a port yard resource optimization device, comprising: The parameter acquisition module is used to acquire the operational parameters of the port yard; the operational parameters include energy consumption value, illuminance value, number of ships arriving at the port, and cargo density. The energy consumption prediction module is used to input the operation parameters into the trained energy consumption prediction model and output the energy consumption prediction value for a future preset time period through the energy consumption prediction model. The lighting evaluation module is used to input the energy consumption prediction value into the lighting effect evaluation model, and to score the lighting effect through the lighting effect evaluation model to obtain a lighting evaluation score; The dimming coefficient determination module is used to determine whether the lighting evaluation score is greater than a preset score threshold. Based on the determination result, the dimming coefficient of the port yard is calculated through the dimming coefficient calculation model so that the first dimming coefficient of the ship area is greater than the first preset value and the second dimming coefficient of the storage area is greater than the second preset value. The dimming coefficient is used to provide the control module for dimming control.

[0014] Thirdly, embodiments of the present invention also provide a computer storage medium storing computer-executable instructions for use in any of the port yard resource optimization methods described above.

[0015] Fourthly, embodiments of the present invention also provide an electronic device, comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed, enable the at least one processor to perform any of the port yard resource optimization methods described above.

[0016] In the solution provided by the first aspect of this invention, by acquiring the operational parameters of the port yard and using an energy consumption prediction model to predict energy consumption, future energy consumption needs can be effectively assessed. This allows port management personnel to optimize the lighting system and reduce unnecessary energy consumption without affecting operational efficiency. By combining the predicted energy consumption values ​​with a lighting effect evaluation model, it can be ensured that the yard maintains an appropriate lighting level under different operational conditions. Good lighting effects not only improve the comfort of the working environment but also significantly enhance operational safety and reduce the risk of accidents caused by insufficient light. Based on the real-time acquired operational parameters and their prediction results, the lighting intensity can be dynamically adjusted to adapt to changes in port operations.

[0017] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart of a port yard resource optimization method provided by an embodiment of the present invention is shown; Figure 2 A flowchart of another port yard resource optimization method provided by an embodiment of the present invention is shown; Figure 3 This diagram illustrates the structure of a port yard resource optimization device provided in an embodiment of the present invention. Figure 4 A schematic diagram of the structure of an electronic device for performing a port yard resource optimization method, provided in an embodiment of the present invention, is shown. Detailed Implementation

[0020] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0021] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0022] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0023] Currently, port yard lighting energy consumption accounts for 25%-35% of the total energy consumption of the port, and is a key infrastructure for maintaining 24-hour operation of the yard. However, the existing port yard lighting system is still constrained by two major issues: First, the energy consumption control method is crude. The existing port yard lighting system mostly adopts constant illuminance or timed switching control, without integrating dynamic parameters such as ship arrival and cargo density, and without considering the characteristics of the port's open-air environment, which cannot support precise energy consumption control. Second, the evaluation of lighting effect is lacking. The energy-saving strategy of the existing port yard lighting system only aims to reduce energy consumption, without a safety constraint mechanism. If the illuminance and illuminance uniformity in the port operation area do not meet the safety requirements, it is easy to cause operational accidents.

[0024] See Figure 1 One embodiment provides a method for optimizing port yard resources, including: Step 101: Obtain the operating parameters of the port yard; the operating parameters include energy consumption, illuminance, number of ships arriving at the port, and cargo density.

[0025] For example, the resource optimization method can be a lighting optimization method.

[0026] For example, the energy consumption value can be the amount of electricity consumed per unit time.

[0027] For example, the illuminance value can be a physical quantity of light intensity per unit area.

[0028] For example, storage density can be the number of storage units placed per unit area.

[0029] Step 102: Input the operation parameters into the trained energy consumption prediction model, and output the energy consumption prediction value for the future preset time period through the energy consumption prediction model.

[0030] For example, the energy consumption prediction model is an LSTM model (Long Short-Term Memory network).

[0031] Step 103: Input the predicted energy consumption value into the lighting effect evaluation model, score the lighting effect through the lighting effect evaluation model, and obtain the lighting evaluation score.

[0032] For example, the lighting effect evaluation model scores the lighting results, and the lighting evaluation scores can be used to evaluate the illuminance under the predicted energy consumption values.

[0033] Step 104: Determine whether the lighting evaluation score is greater than the preset score threshold. Based on the determination result, calculate the dimming coefficient of the port yard through the dimming coefficient calculation model, so that the first dimming coefficient of the ship area is greater than the first preset value and the second dimming coefficient of the storage area is greater than the second preset value. The dimming coefficient is used to provide the control module for dimming control.

[0034] For example, the dimming coefficient calculation model adjusts the first dimming coefficient and the second dimming coefficient based on the judgment result, and the control module can control the illuminance of the port yard according to the first dimming coefficient and the second dimming coefficient.

[0035] The aforementioned port yard resource optimization method effectively assesses future energy demands by acquiring port yard operational parameters and using energy consumption prediction models. This allows port managers to optimize lighting systems and reduce unnecessary energy consumption without impacting operational efficiency. By combining energy consumption predictions with lighting effect evaluation models, appropriate lighting levels can be maintained in the yard under different operational conditions. Good lighting not only improves the comfort of the working environment but also significantly enhances operational safety, reducing the risk of accidents caused by insufficient light. Based on real-time acquired operational parameters and their prediction results, lighting intensity can be dynamically adjusted to adapt to changes in port operations.

[0036] In some alternative implementations, the method further includes calculating the third-order difference of multiple consecutive operation parameter measurement points before inputting the operation parameters into the trained energy consumption prediction model. ;in, For the first The third-order difference of multiple consecutive points starting from a measurement point, where the measurement point refers to energy consumption or illuminance value; determining the threshold for outlier detection. Calculate the standard deviation of the residuals ;make ,like If the measurement point is determined to be an outlier, linear interpolation is used to complete the outlier, and the completed data is then standardized using Z-score.

[0037] For example, the B-spline curve method is used to detect outliers, and the third-order difference of four consecutive measurement points is calculated using the following formula: in, For the first The third difference of four consecutive points starting from the first measurement point. The first Energy consumption at any given moment or measured illuminance value.

[0038] For example, determining an outlier detection threshold First, calculate the standard deviation of the residuals using the formula. The formula is in, For the target parameter residual, The actual measured value of the target parameter. For the target parameter smoothing value, To measure the total number of target parameters; then take ,like It is then determined to be an outlier.

[0039] For example, linear interpolation is used to complete outliers, as shown in the formula: in, This is an outlier. These are the normal measurements immediately before and after the outlier; outliers at the beginning and end are padded with the mean of the three adjacent normal data.

[0040] For example, the completed data is subjected to Z-score standardization to eliminate dimensional differences. Z-score standardization, also known as standard deviation standardization, is a method to transform data into a standard normal distribution. This method adjusts the mean of the data to 0 and the standard deviation to 1, allowing data with different characteristics to be compared on the same scale.

[0041] In some optional implementations, the training process of the energy consumption prediction model includes: defining a prediction accuracy evaluation index, with the root mean square error formula as follows: ; The formula for mean absolute percentage error is: ; in, It is the root mean square error value. It is the mean absolute percentage error value. This is the energy consumption value, in kW·h. This represents the predicted energy consumption value output by the energy consumption prediction model, in kW·h, where n is the number of samples.

[0042] The pre-collected data was divided into training, validation, and test sets in a 7:2:1 ratio. The energy consumption prediction model was trained using the pre-collected data, with the root mean square error as the benchmark. Use the loss function to train the model.

[0043] Set a termination condition for model training: when the root mean square error of the validation set reaches 5 consecutive rounds. ≤0.35 and mean absolute percentage error Training is complete when the value is ≤0.02.

[0044] For example, the LSTM model verifies the prediction accuracy through root mean square error and mean absolute percentage error, ensuring that the prediction results can support precise energy consumption control.

[0045] In some optional implementations, the lighting effect evaluation model calculates the lighting evaluation score using the fuzzy comprehensive evaluation method, specifically including: constructing an evaluation index set U={u1,u2,u3,u4,u5}; where u1 is the average illuminance of the work area, u2 is the illuminance compliance rate, u3 is the illuminance uniformity, u4 is the glare value, and u5 is the energy consumption per unit area; and using the analytic hierarchy process (AHP) to determine the weight vector W=(w1,w2,w3,w4,w5) for each index, satisfying... The weighting ratio for the three criteria—lighting safety, lighting comfort, and energy efficiency—is 4:3:3. A linear membership function is used to quantify the index weight vector. The membership function for the average illuminance x in the work area is: ; The membership function of the illuminance uniformity y (unit: lx) of the work area is: ; The membership functions of indicators u2, u4, and u5 are obtained by linearly mapping the actual measured values ​​to the preset threshold intervals. According to the formula Calculate the lighting evaluation score f; where wi is the weight of the i-th indicator and ui is its corresponding membership value.

[0046] For example, the lighting effect evaluation method uses the analytic hierarchy process (AHP) to determine the weights of three-dimensional indicators, ensuring that the weight allocation conforms to the principle of safety priority, and constructs a three-dimensional indicator system of "safety-comfort-energy saving". The criterion layer includes lighting safety, lighting comfort, and energy saving, while the indicator layer includes average illuminance of the work area, illuminance compliance rate, illuminance uniformity, glare value, and energy consumption per unit area; including the following steps: Step A1: Construct an evaluation index system: Let the evaluation index set be U={u1,u2,u3,u4,u5}, which correspond to: u1: average illuminance of the work area (unit: lx), u2: illuminance compliance rate (%), u3: illuminance uniformity (dimensionless), u4: glare value (dimensionless), u5: energy consumption per unit area (W / ㎡).

[0047] Step A2: Determine the indicator weights: Using the analytic hierarchy process (AHP), multiple port lighting experts were invited to score the importance of the indicators, constructing a judgment matrix and calculating the weight vector W=(w1,w2,w3,w4,w5), which satisfies... An example weight allocation is: w1=0.22, w2=0.18, w3=0.18, w4=0.12, w5=0.30.

[0048] Step A3: Design the membership function: A linear membership function is used to map the measured values ​​of each indicator to membership degrees in the interval of 0 and 1: Average illuminance in the work area x: ; Lighting uniformity y in the work area: ; The membership functions of the other indicators u2, u4, and u5 are obtained by linearly mapping the actual measured values ​​to the preset threshold interval, ensuring that the functions are continuous and reflect the quality of the indicators.

[0049] Step A4: Calculate the overall evaluation score: The formula for calculating the lighting evaluation score f is: ; Among them, u i It is its corresponding membership value.

[0050] Example: If the measured values ​​are u1=0.6, u2=1.0, u3=0.5, u4=1.0, u5=0.75, then: f=0.22×0.6+0.18×1.0+0.18×0.5+0.12×1.0+0.30×0.75=0.747. This score is used to determine whether the lighting effect meets the safety and energy-saving requirements.

[0051] In some optional implementations, determining whether the lighting evaluation score is greater than a preset score threshold, and calculating the dimming coefficient of the port yard based on the determination result using a dimming coefficient calculation model, includes using the average lighting energy consumption over a preset time period as the baseline energy consumption. ;according to Calculate the difference in energy consumption prediction ,in, Output the predicted energy consumption value for the energy consumption prediction model; if the lighting evaluation score If the score is greater than or equal to the preset score threshold, then it will be processed according to... Calculate the dimming factor k; if the lighting evaluation score If the illuminance is less than the preset score threshold, the average illuminance of the port yard will first be adjusted to be no less than [a certain value]. Adjust the illuminance uniformity to not less than 0.7, and then proceed according to... Calculate the dimming factor k.

[0052] For example, determining the baseline energy consumption Take the average lighting energy consumption of the same period over the past 30 days, for example, the average energy consumption from 19:00 to 20:00 every Monday.

[0053] For example, calculate the energy consumption prediction difference. The formula is ,in, The energy consumption prediction model outputs the predicted energy consumption value for the next hour (unit: kW·h). Baseline energy consumption (unit: kW·h).

[0054] For example, determining the lighting effect evaluation score The preset score threshold can be 0.8. The dimming coefficient can be calculated directly using the formula. The formula is ,like First, adjust the average illuminance of the work area to no less than [amount missing]. 1. Adjust the uniformity of illumination in the work area to not less than 0.7, and then recalculate the dimming coefficient k using the above formula; For example, after determining the dimming coefficient k, a zone dimming command can be generated based on the dimming coefficient k to adjust the dimming coefficient of the ship's berthing operation area. The dimming coefficient of the cargo storage area shall not be less than 0.85. Not less than 0.75.

[0055] In some optional implementations, the method further includes: periodically collecting the actual illuminance value of the port yard operation area, calculating the deviation between the actual illuminance value and the commanded illuminance value determined by the dimming coefficient; if the deviation is greater than a preset deviation threshold, calculating the dimming coefficient adjustment amount based on the dimming coefficient, and adjusting the dimming coefficient parameters based on the dimming coefficient adjustment amount.

[0056] For example, after determining the dimming coefficient adjustment amount, the dimming coefficient is adjusted based on the dimming coefficient adjustment amount. If the dimming coefficient is found to be too small, the dimming coefficient adjustment amount is added to the initial dimming coefficient value to obtain the corrected dimming coefficient.

[0057] In some optional implementations, calculating the deviation between the actual illuminance value and the commanded illuminance value determined by the dimming factor includes: through... Calculate the percentage of deviation; where, It's a percentage of deviation. This represents the actual illuminance value in the work area. The commanded illuminance value determined by the dimming factor; calculating the dimming factor adjustment amount based on the dimming factor, including: through... Calculate the dimming factor adjustment amount, where, It is the dimming factor adjustment amount. This is the dimming factor.

[0058] For example, closed-loop feedback adjustment maintains stable lighting effects while also providing fault tolerance.

[0059] For example, the actual illuminance of the work area is collected every 5 minutes. (unit: The percentage of illuminance deviation is calculated using a formula. The formula is in, Actual illuminance in the work area (unit: ); For example, Greater than a preset deviation threshold, for example if The dimming coefficient adjustment amount is calculated using a formula. The formula is ,in, This is the current dimming factor; use "+" when the actual illuminance value is lower than the target value and "-" when it is higher than the target value, and ensure this after adjustment. .

[0060] For example, if the lighting controller fails to respond to commands three times consecutively or Then maintain the illuminance of the work area as follows: At the same time, it outputs an alarm signal to the port monitoring center.

[0061] For example, the LSTM energy consumption prediction model determines the input feature dimension through the correlation integration method to ensure that the features cover the dynamic laws of port operations. The specific implementation steps include: Step B1: Define the related integral function, the formula is: ; in, For the integral of association, For the embedding dimension, The total length of the energy consumption time series. Distance threshold For time intervals, The number of phase points after phase space reconstruction and For phase point With phase point Euclidean distance, It is a step function and (when hour), (when hour).

[0062] Step B2: Traversal Calculate the correlation integral under different parameter combinations. .

[0063] Step B3: Construct the statistic S(m,r,t) and determine the optimal time interval and embedding dimension.

[0064] The statistic S(m,r,t) is used to measure the structural stability of a time series under different embedding dimensions m and time intervals t. Its calculation formula is as follows: ; in: C s (m,N / t,r,t) represents the correlation integral of the sth subsequence under the embedding dimension m and the distance threshold r after dividing a time series of length N into t subsequences; C s (1,N / t,r,t) represents the correlation integral of the corresponding subsequence when the embedding dimension is 1; r is taken as twice the standard deviation of the time series, i.e., r = 2σ.

[0065] By iterating through parameter combinations of m=2-8 and t=1-6, the values ​​of S(m,r,t) under different (m,t) parameters are calculated. When S(m,r,t) first approaches 0, it indicates that the time series has the optimal deterministic structure under these parameters, and the corresponding t is the optimal time interval.

[0066] In one embodiment of the present invention, when m=4 and t=4, S(4,r,4)=0.008 is calculated, which is close to 0 for the first time, so the optimal time interval t=4 is determined.

[0067] Furthermore, according to the embedded window width formula τ w =(m−1)t, take τ w =12, substituting t=4, we can solve for the embedding dimension m=4.

[0068] Finally, considering the dynamic characteristics of port operations, the input features of the LSTM energy consumption prediction model were determined to be 9-dimensional, including: historical energy consumption from t−1 to t−6 hours (6 dimensions in total); ambient illuminance at time t (1 dimension); number of ships arriving at time t (1 dimension); and cargo density at time t (1 dimension).

[0069] For example, the LSTM energy consumption prediction model determines the number of hidden layer neurons by using the least squares fitting slope to ensure that the network adapts to the fluctuation patterns of port energy consumption. The specific implementation steps include: Step C1: Collect energy consumption data from the port yard for 24 consecutive hours and calculate the hourly energy consumption residual value. ;in, For the first Actual energy consumption per hour (unit: kW·h) For the first Hourly energy consumption fitted value (unit: kW·h); Step C2: Calculate the fitted slope using the least squares fitted line slope formula, which is: ; in, To fit the slope of the straight line, For time step , For the first Hourly energy consumption residual (unit: kW·h), where n is the number of fitted samples and ; Step C3: If the calculation yields... =0.0032, indicating that the energy consumption residual has moderate fluctuations, and the LSTM hidden layer is determined to be a 2-layer structure; the first layer contains 128 neurons to capture long-term energy consumption trends, and the second layer contains 64 neurons to refine short-term energy consumption fluctuations; both layers use the ReLU activation function, and the dropout coefficient is set to 0.2 to avoid overfitting.

[0070] In one embodiment, the No. 3 container yard of a port is used as an application scenario. The total area of ​​the yard is 60,000 square meters, which is divided into a ship berthing operation area and a cargo storage area. The ship berthing operation area has an area of ​​25,000 square meters and is equipped with 20 200W LED high mast lights, while the cargo storage area has an area of ​​35,000 square meters and is equipped with 28 180W LED high mast lights. 24-hour continuous operation lighting is required.

[0071] The energy consumption data acquisition uses smart energy meters with an accuracy class of 0.5S and a measurement range of 0-100A. These meters are installed in each of the lighting distribution boxes in the yard, totaling 8 distribution boxes (3 in the ship operation area and 5 in the storage area). The energy consumption data of a single box is collected every 5 minutes, and RS485 communication is supported.

[0072] Illuminance data acquisition uses a CMOS image sensor with a resolution of 1920×1080 and an illuminance measurement range of 0-2000lx. It is installed on the top of each high-mast lamp, 15m above the ground, with the lens facing downwards to cover a working area with a radius of 30m. It is powered by PoE.

[0073] For collecting operational parameters, the number of arriving ships is measured using millimeter-wave radar with a detection range of 0-500m and an angle of ±30°, which is installed on a 20m high monitoring tower at the entrance of the yard. The density of cargo locations is measured using infrared ranging sensors with a measurement range of 10-80cm, with one sensor per 100㎡ of cargo location, for a total of 600 sensors. These sensors are embedded in pre-drilled holes in the ground of the cargo location, with a depth of 5cm and the top flush with the ground.

[0074] Execution control module: Each LED high mast light is equipped with a dimming controller, which is connected to the PLC via a shielded cable. The controller is installed in the junction box at the bottom of the light pole and has an IP65 protection rating. After receiving dimming commands, it adjusts the output power of the light fixture.

[0075] Each module constructs a bidirectional communication network via industrial Ethernet. The smart energy meter, CMOS sensor, millimeter-wave radar, and infrared sensor are connected to the local area network via RS485 to Ethernet module and communicate with the data acquisition equipment through a switch.

[0076] Data acquisition runs continuously every 5 minutes. The edge gateway sends data request commands to each acquisition device according to a preset time sequence. It sends a command to the smart meter to read the current active power consumption, using standard industrial protocol function code 0x04, to obtain the cumulative energy consumption over the past 5 minutes in kW·h. It sends a command to the CMOS sensor to obtain the average illuminance; the sensor automatically captures three images of the work area, calculates the image grayscale values, converts them to illuminance, and returns the average value. It sends a command to the millimeter-wave radar to count the number of vessels; the radar scans continuously for 30 seconds, identifies the outlines of vessels moored in the yard, excludes small vessels smaller than 5000 tons, and returns the number. It sends a command to the infrared sensor to read the occupancy rate of cargo spaces; each sensor returns the occupancy or vacancy status of the corresponding cargo space, and the edge gateway calculates the total occupancy rate. The cargo space density equals the number of occupied cargo spaces divided by the total number of cargo spaces.

[0077] The specific steps include: Step D1: The edge gateway calls the preprocessing script, first performing outlier detection: taking the energy consumption data from the four most recent consecutive collection cycles as the sequence to be detected, denoted as x. i ,x i+1 ,x i+2 ,x i+3 , where i is the starting index. For example, at a certain detection time, the corresponding historical energy consumption values ​​are: 145 kWh (x i ), 150 kWh (x i+1 ), 320 kWh (x i+2 ), 155 kWh (x i+3 According to formula δ i =x i -3x i+1 +3x i+2 -x i+3 Calculate the third difference and substitute it into the example data to get: δ i =145−3×150+3×320−155=500.

[0078] Calculate the residual standard deviation of nearly 300 energy consumption data points. Take the actual measured value of the target parameter. With smooth value Calculated using a 5-point moving average. Substitute into the formula ,have to ,Pick .

[0079] because ,determination Time data For outliers, linear interpolation is used for completion. Outliers at the beginning and end, such as No time Data, take 3 consecutive normal data points. Mean completion, for example Abnormal illumination at any given time, padding value is: .

[0080] Step D2: Perform Z-score standardization on the completed data, according to the formula. ,in The mean and standard deviation of this parameter over the past 30 days represent energy consumption. , After standardization, the data range is controlled within [-3, 3].

[0081] Step D3: The edge gateway sends the standardized 9-dimensional input features to the server's LSTM energy consumption prediction model. The 9-dimensional features include... to Historical energy consumption per hour Real-time ambient light levels, number of ships arriving at port, and cargo density.

[0082] The LSTM module loads feature data and performs computation through two hidden layers: the first layer has 128 ReLU neurons, and the second layer has 64 ReLU neurons. Output the predicted energy consumption for the next 1-4 hours, such as the predicted energy consumption from 10:00 to 11:00. .

[0083] Synchronously retrieve data to obtain the measured values ​​of the five indicators required for evaluating the lighting effect at time t: The average illuminance of the work area is x=120 lx, the illuminance compliance rate is z=92%, the illuminance uniformity is y=0.82, the glare value is g=0.4, and the energy consumption per unit area is w=2.7 W / ㎡.

[0084] Step 1: Calculate the membership degree of each indicator Calculate the following based on the defined linear membership function: 1. Membership degree of average illuminance: Since 100 ≤ x < 200, therefore .

[0085] 2. Membership of Illuminance Compliance Rate: The preset compliance rate threshold range is [80%, 95%]. Since z=92% falls within this range, therefore... .

[0086] 3. Membership degree of illuminance uniformity: Since 0.7 ≤ y < 1.0, therefore .

[0087] 4. Glare Value Membership: The preset glare value threshold range is [0.5, 0.3] (the smaller the value, the better). Since g=0.4 falls within this range, .

[0088] 5. Energy consumption membership degree per unit area: The preset energy consumption threshold range is [4,2] W / m² (the smaller the value, the better). Since w=2.7 is within this range, .

[0089] Step 2: Calculate the comprehensive evaluation score based on the weights. The comprehensive weight vector of the indicators, W=(0.22, 0.18, 0.18, 0.12, 0.30), which is determined in advance using the analytic hierarchy process, corresponds to the above five indicators.

[0090] The lighting evaluation score f is based on the formula calculate: f=0.22×0.2+0.18×0.8+0.18×0.4+0.12×0.5+0.30×0.65=0.515.

[0091] By adjusting the membership function parameters (such as increasing the pass rate standard), the score in this example can be increased to 0.81.

[0092] Step D4: Energy-saving strategy generation and execution is completed within 5 minutes after the hourly prediction. The server calculates the energy consumption prediction difference. The average energy consumption over the same period of the past 30 days, such as Mondays from 10:00 to 11:00, is used as the benchmark: because Directly calculate the dimming coefficient .

[0093] By region Minimum value, ship berthing operation area Take 0.902, cargo storage area Take 0.902 and generate a dimming command in the format of area code + lamp number + dimming coefficient, such as 01-05-0.902.

[0094] The server sends the instruction to the PLC. After parsing, the PLC sends a PWM signal to the corresponding dimming controller via the bus. The controller adjusts the output power of the lamp. For example, the output power of a 200W lamp is approximately 180.4W (200 × 0.902 ≈ 180.4W).

[0095] Step D5: Closed-loop feedback control continuously adjusts every 5 minutes. A CMOS sensor collects the actual illuminance of the work area after dimming, such as the actual illuminance in the ship area. Instruction target illuminance ; Edge gateway calculates illuminance deviation percentage If the deviation is greater than 5%, as in the actual storage area, no adjustment is required; ,instruction , Then calculate the dimming coefficient adjustment amount. ,new The PLC resends the command; If the dimming controller fails to respond to the command 3 times consecutively or The PLC automatically switches to standby mode to maintain the illuminance in the ship area. Storage area maintained At the same time, it sends an alarm signal to the server containing the faulty device number.

[0096] In another embodiment, the algorithm is based on nearly 6 months of lighting energy consumption data (4320 hours) from a port yard to ensure its feasibility. Specifically: B-spline curve method for outlier detection. Data preparation: 744 data points of hourly energy consumption data from January 1st to January 31st, 2024 were collected and denoted as follows. to Partial data is as follows .

[0097] Third-order difference calculation: Calculate over 4 consecutive points ,like hour, ; hour, .

[0098] Calculation of residual standard deviation sigma: Perform 3rd B-spline smoothing on the original data, node vector.

[0099] Number of fits To obtain smoothed values ,like .

[0100] Calculate residuals , For the original data, such as .

[0101] Substitute into the formula , , The calculation yields: .

[0102] Outlier detection: Take ,because ,determination This is an outlier.

[0103] Outlier completion: Linear interpolation is used. The completed data sequence is .

[0104] Z-score standardization: Calculates the average energy consumption for the month. Standard deviation After standardization In another embodiment, the detailed construction of the LSTM energy consumption prediction model includes: (1) Determining the input feature dimension Correlation integral function calculation: Taking the energy consumption time series from January to March 2024, n=2160 hours, the correlation integral function is defined. ,in For phase point and Euclidean distance, when , when ,Pick , This represents the standard deviation of energy consumption data.

[0105] Iterate through the parameter combinations m=2-8 and t=1-6, and calculate the results for different combinations. For example, when hour, Calculated ;when Calculated Close to 0.

[0106] Analysis of the statistic S(m,r,t): Construction ,when hour, When the value approaches 0 for the first time, the optimal time interval t=4 is determined.

[0107] Embedding dimension calculation: Embedding window width ,Pick Substituting into .

[0108] Input feature determination: Based on the characteristics of port operations, a total of 9-dimensional input features were ultimately determined, including... to Six characteristics of historical energy consumption per hour One characteristic of ambient illuminance at any given time The number of ships arriving at port at time t is one characteristic, and the cargo density at time t is another characteristic.

[0109] (2) LSTM network architecture design Energy consumption residual fitting slope calculation: 24-hour energy consumption data collected on April 1, 2024. to ,unit ,like , .

[0110] Obtained by linear fitting Calculate the residual ,like .

[0111] Substituting into the least squares fitting slope formula ,in The calculation yields: The value is close to 0.0032, indicating that the energy consumption residual shows moderate fluctuation.

[0112] Network layer design: The input layer has 9 neurons, corresponding to 9-dimensional features. The shape of the input data is (batch_size, time_steps, input_dim) = (32, 4, 8), where time_steps = 4 corresponds to the embedding dimension. m =4.

[0113] Hidden layer 1 has 128 ReLU neurons. Return_sequences=True is set to pass sequence data to hidden layer 2. Dropout=0.2 randomly discards 20% of neurons to avoid overfitting.

[0114] Hidden layer 2 has 64 ReLU neurons, and return_sequences=False is set to pass single values ​​to the output layer, with dropout=0.2.

[0115] The output layer has one linear neuron, which outputs the predicted energy consumption value for the next hour.

[0116] Model training parameters: The loss function uses the root mean square error, and the formula is as follows: ; The optimizer uses an adaptive optimizer with an initial learning rate of 0.001, which decays by 10% every 10 epochs. The number of training rounds is set to 100. ; Training terminates when the validation set comprises 20% of the total data. 5 consecutive and Training has stopped.

[0117] (3) Validation of prediction accuracy Evaluation sample selection: Energy consumption data from May 1st to May 7th, 2024, totaling 168 hours, was used as the test set, representing actual energy consumption. Compared with the predicted value Some data is as follows: RMSE calculation: Substitute into the formula The calculation yields: MAPE calculation: Substitute into the formula: Calculated , The model's accuracy was deemed to meet the standard.

[0118] The detailed implementation of the lighting effect evaluation algorithm includes: 1. Detailed steps for determining weights using the Analytic Hierarchy Process (AHP) (1) Constructing a hierarchical structure model Target layer (O): Comprehensive evaluation of lighting effects.

[0119] Criterion Layer (C): Includes three criteria: lighting safety (C1), lighting comfort (C2), and energy efficiency (C3).

[0120] The indicator layer (I) includes five specific evaluation indicators: average illuminance of the work area (I1), illuminance compliance rate (I2), illuminance uniformity (I3), glare value (I4), and energy consumption per unit area (I5).

[0121] The attribution relationships between indicators and criteria are as follows: I1 and I2 belong to C1; I3 and I4 belong to C2; I5 belongs to C3.

[0122] (2) Construct the judgment matrix Five experts with an average of over eight years of experience in the port lighting field were invited to conduct pairwise importance comparisons and scores of each element using a 1-9 scale. The geometric mean of all expert opinions was then used to obtain the following judgment matrix: a) Criterion Layer Judgment Matrix A C(The relative importance of each criterion relative to objective O) b) Indicator-level judgment matrix (relative to the applicable criteria) The comparison matrix A between indicators I1 and I2 relative to criterion C1 (safety) C1 : Comparison matrix A between indicators I3 and I4 relative to criterion C2 (comfort). C2 : Compared to criterion C3 (energy efficiency), there is only one indicator I5, so its weight is 1.

[0123] (3) Calculate the weight vector and perform a consistency check. For each judgment matrix, calculate its largest eigenvalue λ. max The corresponding normalized eigenvectors (i.e., weight vectors) are then analyzed, and a consistency check is performed. If the consistency ratio CR = CI / RI < 0.10, the consistency of the judgment matrix is ​​considered acceptable.

[0124] a) Calculation of criterion layer weights: For matrix A C : Its largest eigenvalue was calculated. .

[0125] Consistency Indicators , where n=3 is the matrix order.

[0126] Querying the average random consistency index (RI) table, when n=3, RI=0.58.

[0127] Consistency ratio It passed the test.

[0128] The corresponding normalized feature vector (i.e., the criterion layer weights) is: W (C) =(0.4,0.3,0.3) T .Right now: Lighting safety (C1) weight: 0.4, lighting comfort (C2) weight: 0.3, energy efficiency (C3) weight: 0.3.

[0129] b) Calculation of indicator layer weights (under a single criterion): For matrix A C1 (Under the safety criteria), the calculated index weight is: W (C1) =(0.55,0.45) TThat is, I1 (average illuminance) has a weight of 0.55, and I2 (compliance rate) has a weight of 0.45.

[0130] For matrix AC2 (under the comfort criterion), the index weights are calculated as: W (C2) =(0.6,0.4) T That is, I3 (uniformity) has a weight of 0.6, and I4 (glare value) has a weight of 0.4.

[0131] For criterion C3 (energy efficiency), there is only index I5, so its weight is 1.

[0132] (4) Calculate the comprehensive weight of the indicator layer (overall ranking of the hierarchy) The weight of each indicator relative to its criterion is multiplied by the weight of the criterion relative to the overall goal to obtain the comprehensive weight of each indicator relative to the overall goal.

[0133] Therefore, the comprehensive weight vector of the five indicators used for fuzzy comprehensive evaluation is: W=(w1,w2,w3,w4,w5)=(0.22,0.18,0.18,0.12,0.30) The weight satisfies .

[0134] 2. Calculation of Fuzzy Membership Function (1) Average illuminance of the work area Weight 0.22: Measured Illuminance Substituting into the membership function, because ; (2) Illuminance compliance rate Weight 0.18: Define the compliance rate Actual measurement Design the membership function. hour , hour Calculated ; (3) Illuminance uniformity Weight 0.18: Measured uniformity Substituting into the membership function, because ; (4) Glare value Weight 0.12: Define glare value Actual measurement Design the membership function. hour hour hour Calculated ; (5) Energy consumption per unit area Weight 0.3: Actual energy consumption per unit area Design the membership function. hour , hour hour Calculated ; (6) Overall evaluation score: .

[0135] IV. Detailed Implementation of the Energy-Saving Strategy Generation Algorithm (1) The benchmark energy consumption is determined by the average lighting energy consumption during the same period from May 1 to May 30, 2024, such as from 14:00 to 15:00 every day. The calculation yields: (2) Calculation of energy consumption prediction difference and dimming coefficient: Energy consumption prediction difference LSTM module predicted energy consumption from 14:00 to 15:00 on May 31st ; Lighting effect evaluation score judgment: because First, adjust the average illuminance of the work area to... current The requirement of illuminance uniformity being ≥0.7 (currently 0.85) has been met and no further adjustment is needed. Dimming coefficient calculate: ; Zone dimming factor limit: Ship berthing operation area Take 0.956, cargo storage area Set the value to 0.956 and generate a dimming command. The dimming coefficient for the 20 lamps in the ship area is 0.956, and the dimming coefficient for the 28 lamps in the storage area is 0.956.

[0136] The above details the process of optimizing port yard resources. This method can also be implemented using a corresponding device, the structure and function of which will be described in detail below.

[0137] For example, such as Figure 2The diagram illustrates a modular flow chart of a port yard resource optimization method in one embodiment. A multi-dimensional data acquisition module collects port yard energy consumption, illuminance, and specific operational parameters, providing foundational data for subsequent modules. A data preprocessing module processes the collected data to ensure input data quality. A core calculation module performs data calculations and processing. An LSTM energy consumption prediction module outputs predicted future energy consumption values ​​to support precise energy consumption control. A lighting effect evaluation module constructs a three-dimensional evaluation system of "safety-comfort-energy saving" to constrain the safety of energy-saving strategies. An energy-saving strategy generation module generates zoned dimming commands to balance energy consumption and safety. An execution control module implements closed-loop feedback control to ensure strategy implementation.

[0138] Based on the same inventive concept, this invention also provides a port yard resource optimization device 300, see [link to relevant documentation]. Figure 3 The port yard resource optimization device 300 includes: The parameter acquisition module 302 is used to acquire the operating parameters of the port yard; the operating parameters include energy consumption value, illuminance value, number of ships arriving at the port and cargo density; The energy consumption prediction module 304 is used to input the operation parameters into the trained energy consumption prediction model and output the energy consumption prediction value for a future preset time period through the energy consumption prediction model. The lighting evaluation module 306 is used to input the energy consumption prediction value into the lighting effect evaluation model, and to score the lighting effect through the lighting effect evaluation model to obtain a lighting evaluation score; The dimming coefficient determination module 308 is used to determine whether the lighting evaluation score is greater than the preset score threshold. Based on the judgment result, the dimming coefficient of the port yard is calculated through the dimming coefficient calculation model so that the first dimming coefficient of the ship area is greater than the first preset value and the second dimming coefficient of the storage area is greater than the second preset value. The dimming coefficient is used to provide the control module for dimming control.

[0139] This invention also provides a computer storage medium storing computer-executable instructions, including a program for executing the above-described port yard resource optimization method. The computer-executable instructions can execute the methods in any of the above-described method embodiments.

[0140] The computer storage medium can be any available medium or data storage device that a computer can access, including but not limited to magnetic storage (such as floppy disks, hard disks, magnetic tapes, magneto-optical disks (MO), etc.), optical storage (such as CDs, DVDs, BDs, HVDs, etc.), and semiconductor storage (such as ROMs, EPROMs, EEPROMs, non-volatile memory (NAND flash), solid-state drives (SSDs)).

[0141] Figure 4A structural block diagram of an electronic device according to another embodiment of the present invention is shown. The electronic device 1100 may be a host server with computing capabilities, a personal computer (PC), or a portable computer or terminal, etc. The specific embodiments of the present invention do not limit the specific implementation of the electronic device.

[0142] The electronic device 1100 includes at least one processor 1110, a communications interface 1120, a memory array 1130, and a bus 1140. The processor 1110, the communications interface 1120, and the memory 1130 communicate with each other via the bus 1140.

[0143] The communication interface 1120 is used to communicate with network elements, including, for example, virtual machine management centers and shared storage.

[0144] Processor 1110 is used to execute programs. Processor 1110 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention.

[0145] Memory 1130 is used for executable instructions. Memory 1130 may include high-speed RAM and may also include non-volatile memory, such as at least one disk storage device. Memory 1130 may also be a memory array. Memory 1130 may also be divided into blocks, and the blocks may be combined into virtual volumes according to certain rules. The instructions stored in memory 1130 can be executed by processor 1110 to enable processor 1110 to execute the port yard resource optimization method in any of the above method embodiments.

[0146] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for optimizing port yard resources, characterized in that, include: Obtain the operational parameters of the port yard; the operational parameters include energy consumption, illuminance, number of ships arriving at the port, and cargo density; The operation parameters are input into the trained energy consumption prediction model, and the energy consumption prediction model outputs the predicted energy consumption value for a future preset time period. The predicted energy consumption value is input into the lighting effect evaluation model, and the lighting effect is scored by the lighting effect evaluation model to obtain a lighting evaluation score; The system determines whether the lighting evaluation score is greater than a preset score threshold. Based on the determination result, it calculates the dimming coefficient of the port yard using a dimming coefficient calculation model, so that the first dimming coefficient of the ship area is greater than the first preset value, and the second dimming coefficient of the storage area is greater than the second preset value. The dimming coefficient is used to provide the control module for dimming control.

2. The port yard resource optimization method according to claim 1, characterized in that, Before inputting the operation parameters into the trained energy consumption prediction model, the method further includes: Calculate the third-order difference of multiple consecutive operation parameter measurement points ;in, For the first The third-order difference of a series of consecutive points starting from a measurement point, where the measurement point refers to the energy consumption value or illuminance value; Determine the outlier detection threshold Calculate the standard deviation of the residuals ;make ,like The measurement point is then determined to be an outlier. Linear interpolation is used to complete outliers, and the completed data is then standardized using Z-score.

3. The port yard resource optimization method according to claim 1, characterized in that, The training process of the energy consumption prediction model includes: The root mean square error (RMSE) is defined as the evaluation index for prediction accuracy. ； The formula for mean absolute percentage error is: ； in, It is the root mean square error value. It is the mean absolute percentage error value. Energy consumption value The energy consumption prediction value is the output of the energy consumption prediction model, and n is the number of samples. The pre-collected data was divided into training, validation, and test sets in a 7:2:1 ratio. The energy consumption prediction model was trained using the pre-collected data. Train the loss function for the model; Set a termination condition for model training: when the validation set has been validated for 5 consecutive rounds. ≤0.35 and Training is complete when the value is ≤0.

02.

4. The port yard resource optimization method according to claim 1, characterized in that, The lighting effect evaluation model calculates the lighting evaluation score using the fuzzy comprehensive evaluation method, including: Construct an evaluation index set U={u1,u2,u3,u4,u5}; where u1 is the average illuminance of the work area, u2 is the illuminance compliance rate, u3 is the illuminance uniformity, u4 is the glare value, and u5 is the energy consumption per unit area. The weight vector W = (w1, w2, w3, w4, w5) for each indicator is determined using the analytic hierarchy process (AHP), satisfying the following conditions: ; The membership function of the index weight vector is quantified using a linear membership function. The membership function of the average illuminance x in the work area is as follows: ； The membership function of the illuminance uniformity y in the work area is: ； The membership functions of indicators u2, u4, and u5 are obtained by linearly mapping the actual measured values ​​to the preset threshold intervals. According to the formula Calculate the lighting evaluation score f; where w i Let u be the weight of the i-th indicator. i It is its corresponding membership value.

5. The port yard resource optimization method according to claim 4, characterized in that, The step of determining whether the lighting evaluation score is greater than a preset score threshold, and calculating the dimming coefficient of the port yard based on the determination result using a dimming coefficient calculation model, includes: The average lighting energy consumption over a preset period of time is used as the baseline energy consumption. ; according to Calculate the difference in energy consumption prediction ,in, Output energy consumption prediction values ​​for the energy consumption prediction model; If the lighting evaluation score f is greater than or equal to the preset score threshold, then according to Calculate the dimming factor k; If the lighting evaluation score f is less than the preset score threshold, the average illuminance of the port yard will be adjusted to no less than [a certain threshold]. Adjust the illuminance uniformity to not less than 0.7, and then proceed according to... Calculate the dimming factor k.

6. The port yard resource optimization method according to claim 1, characterized in that, The method further includes: Regularly collect the actual illuminance values ​​of the port yard operation area and calculate the deviation between the actual illuminance values ​​and the commanded illuminance values ​​determined by the dimming coefficient; If the deviation is greater than the preset deviation threshold, the dimming coefficient adjustment amount is calculated based on the dimming coefficient, and the dimming coefficient parameter is adjusted based on the dimming coefficient adjustment amount.

7. The method according to claim 6, characterized in that, The deviation between the actual illuminance value and the commanded illuminance value determined by the dimming coefficient is calculated, including: pass Calculate the percentage deviation; where, It's a percentage of deviation. This represents the actual illuminance value in the work area. The commanded illuminance value determined by the dimming factor; The calculation of the dimming coefficient adjustment amount based on the dimming coefficient includes: pass Calculate the dimming factor adjustment amount, where, It is the dimming factor adjustment amount. This is the dimming factor.

8. A port yard resource optimization device, characterized in that, include: The parameter acquisition module is used to acquire the operational parameters of the port yard; the operational parameters include energy consumption value, illuminance value, number of ships arriving at the port, and cargo density. The energy consumption prediction module is used to input the operation parameters into the trained energy consumption prediction model and output the energy consumption prediction value for a future preset time period through the energy consumption prediction model. The lighting evaluation module is used to input the energy consumption prediction value into the lighting effect evaluation model, and to score the lighting effect through the lighting effect evaluation model to obtain a lighting evaluation score; The dimming coefficient determination module is used to determine whether the lighting evaluation score is greater than a preset score threshold. Based on the determination result, the dimming coefficient of the port yard is calculated through the dimming coefficient calculation model so that the first dimming coefficient of the ship area is greater than the first preset value and the second dimming coefficient of the storage area is greater than the second preset value. The dimming coefficient is used to provide the control module for dimming control.

9. A computer storage medium, characterized in that, The computer storage medium stores computer-executable instructions for executing the port yard resource optimization method according to any one of claims 1 to 7.

10. An electronic device, characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the port yard resource optimization method according to any one of claims 1 to 7.

Images