Intelligent integration method for electric power material supply

Abstract

The invention discloses an intelligent integration method for electric power material supply, which belongs to the field of electric power, and remarkably improves the timeliness, economy and environmental protection property of electric power emergency material dispatching through intelligent storage data fusion and dynamic optimization decision. Wherein the innovative design of the material merging strategy and the dynamic multi-objective optimization breaks through the limitation of the traditional method. In a material merging link, a multi-dimensional feature fusion mechanism is provided, statistical distribution and power business attributes are combined, a mixed feature vector is constructed, a dynamic weighted similarity algorithm is introduced, key business features are endowed with higher weights through expert scoring, the merging precision is improved, and the false merging rate is reduced. According to the method, a dynamic weight function is innovatively designed, and the time efficiency and the carbon emission weight are adjusted in real time through parameters alpha and beta. And in combination with a multi-objective optimization algorithm, the timeliness, the cost and the carbon emission are globally balanced, and efficient and low-carbon operation of the electric power material supply chain is realized.

Description

Technical Field

[0001] The present invention relates to the electric power field, and in particular to an electric power material supply integration method based on intelligent warehousing. Background Art

[0002] In the current rapidly developing power industry, efficient management of power material supply is particularly important. However, under the traditional management model, there are multiple challenges in the power material supply process. First, due to the wide variety of materials and complex properties, inaccurate classification often occurs, which not only affects the efficiency of material management, but also brings inconvenience to subsequent warehousing and distribution. Secondly, low merging efficiency is also a significant problem. Existing methods are difficult to effectively identify material groups that can be combined for supply, resulting in resource waste and increased costs. In addition, the decision-making process is complex and relies on the experience and intuition of managers. It lacks scientific basis and system support, making it difficult to achieve optimal resource allocation. Faced with these challenges, the industry urgently needs a more intelligent and systematic solution to optimize the supply integration process of power materials. Summary of the invention

[0003] In view of the above-mentioned problems existing in the prior art, the purpose of the present invention is to provide an intelligent integration method for power material supply in the power field, which solves the problems of inaccurate classification, low merging efficiency and complex decision-making process in the power material supply process, realizes efficient and accurate management of power materials, and improves the utilization rate of storage space and the response speed of material supply.

[0004] The intelligent integration method for power material supply of the present invention realizes efficient and accurate management of power materials, improves storage space utilization and material supply response speed, through intelligent storage multi-source data collection, improved weighted K-means++ algorithm dynamic clustering, extracts characteristic distribution rules and identifies material groups that can be combined for supply, and establishes a multi-objective optimization model for supply combination decision-making. This innovation provides strong technical support for material management in the power industry. The specific method is as follows:

[0005] Step S1: Intelligent warehousing multi-source data collection;

[0006] Through the IoT terminal, real-time data on material inventory, transportation time, demand fluctuation rate, and storage environment are collected;

[0007] Construct a multidimensional data matrix X, which covers material characteristics, supply chain characteristics and environmental characteristics. Material characteristics include type, weight and expiration date; supply chain characteristics include supplier response time and transportation route complexity; environmental characteristics include warehouse temperature, humidity and geographical location;

[0008]

[0009] Wherein, X is a multi-dimensional data matrix, i is the number of material types, j is the feature dimension, including inventory, demand forecast error, and transportation cost; x ij is the original data of the j-th dimension feature of the i-th type of material.

[0010] Step S2: Dynamically cluster the data using an improved weighted K-means++ algorithm;

[0011] First, standardize the original data x ij of the j-th dimension feature of the i-th type of material in Step S1 to eliminate the dimension difference;

[0012]

[0013] Wherein, x' ij is the standardized data, μ j is the mean of the j-th dimension feature, and σ j is the standard deviation of the j-th dimension feature;

[0014] Then, construct a dynamic clustering function J using the improved weighted K-means++ algorithm to dynamically cluster the data;

[0015]

[0016] Wherein, J is a dynamic clustering function constructed using the improved weighted K-means++ algorithm; m is the total number of samples, k is the number of cluster centers; t is the current iteration number; w ic ∈[0,1] is the membership degree weight of the i-th sample to the c-th class; is the cluster center of the previous iteration; λ is a dynamic adjustment factor;

[0017]

[0018] μ c is the coordinate of the new cluster center of the c-th class.

[0019] Step S3: Based on the clustering result, extract the feature distribution rules of different material categories, calculate the inter-class similarity, and identify the material groups that can be combined for supply;

[0020] ① Feature distribution vector construction: Introduce multi-dimensional statistics and business features, and for each cluster c, calculate its material mixed feature distribution vector V c :

[0021] V c =[μ c ,σ c ,Skewnecc c ,Kurtosis c ,vj , v z

[0022] Statistical dimension: σ c is the standard deviation, Skewnecc c is the skewness, Kurtosis c is the kurtosis, used to capture the distribution shape;

[0023] Business dimension: Add the unique attributes of electric power materials, v j is the emergency level coefficient, v z is the inventory turnover rate;

[0024] ② Calculate the dynamic weighted similarity Sim(V a , V b )

[0025]

[0026] where V a and V b are the feature distribution vectors of clusters a and b; Sim ∈ [-1, 1] is the similarity value; μ aj and μ bj represent the cluster center coordinates of clusters a and b on the jth feature respectively;

[0027] The weight τ j is determined by expert scoring. Key electric power business features are given priority. When calculating the similarity, key electric power business features will be assigned higher weights;

[0028] ③ Based on the dynamic weighted similarity, conduct the merger of electric power materials to generate a set of candidate merger groups G = {g1, g2,..., g n}, g n represents the nth merged material group.

[0029] Step S4: Establish a multi-objective optimization model for supply merger decision-making;

[0030] Introduce a dynamic weight adjustment function to dynamically optimize the weights in combination with real-time data of traffic conditions and the proportion of emergency orders:

[0031]

[0032] a(t) is the weight coefficient function of transportation timeliness; α is the amplification or suppression intensity parameter that adjusts the influence of the proportion of emergency orders on the timeliness weight a(t); η is the proportion of emergency orders; ζ is the traffic congestion index; c(t) is the weight coefficient function of carbon emissions; β is the direct influence ratio parameter that adjusts the intensity of the carbon tax policy on the carbon emissions weight coefficient function c(t); θ is the intensity of the carbon tax policy; b(t) is the weight coefficient of the distribution distance. ​

[0033] Under the condition of meeting the constraints, find the optimal merging scheme that minimizes the total cost C merge ;

[0034]

[0035] In the formula:

[0036] C merge : The total cost of the merging scheme;

[0037] a(t), b(t), and c(t): The weight coefficients of transportation timeliness, distribution distance, and carbon emissions;

[0038] a(t) + b(t) + c(t) = 1;

[0039] T g : The average transportation timeliness of the g-th merging group, in hours;

[0040] D g : The total distribution distance of the g-th merging group, in kilometers;

[0041] E g : The carbon emissions directly generated by the energy consumption of transportation tools during the distribution of the g-th merging group;

[0042] Q min : The minimum carrying capacity per single distribution;

[0043] Q max : The maximum carrying capacity per single distribution;

[0044] The quantity of the p-th material in the g-th group;

[0045] Δt max : The maximum allowable timeliness difference threshold;

[0046] t p and t q : Two arbitrary time points during transportation.

[0047] Compared with the prior art, the intelligent integration method for power material supply provided by the present invention has the following advantages:

[0048] The present invention significantly improves the timeliness, economy and environmental protection of power emergency material dispatch through intelligent storage data fusion and dynamic optimization decision-making. The innovative design of material merging strategy and dynamic multi-objective optimization breaks through the limitations of traditional methods. In the material merging link, the traditional method relies on a single statistical feature or manual experience, and the merging error rate is high. The present invention proposes a multi-dimensional feature fusion mechanism, combines statistical distribution with power business attributes, constructs a hybrid feature vector, and introduces a dynamic weighted similarity algorithm. Through expert scoring, key business features are given higher weights, which improves the merging accuracy and reduces the mismerging rate. In the dynamic multi-objective optimization link, the traditional fixed weight model is difficult to adapt to emergency scenarios. The present invention innovatively designs a dynamic weight function to adjust the timeliness and carbon emission weights in real time through parameters α and β. Combined with the multi-objective optimization algorithm, the timeliness, cost and carbon emissions are globally balanced to achieve efficient and low-carbon operation of the power material supply chain. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The accompanying drawings herein are used to provide a further understanding of the present invention and constitute a part of the present application. The exemplary embodiments of the present invention and their descriptions are used to understand the present invention and do not constitute improper limitations of the present invention. In the accompanying drawings:

[0050] Figure 1 It is a flow chart of the intelligent integration method for power material supply proposed by the present invention. DETAILED DESCRIPTION

[0051] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the technical scheme in the present invention is clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the present invention is not limited to the following embodiments, and the specific implementation method can be determined according to the technical scheme of the present invention and the actual situation. In order to avoid confusing the essence of the present invention, the known methods, processes and procedures are not described in detail.

[0052] The present invention provides an intelligent integration method for power material supply, referring to Figure 1 As shown, the power material supply integration method includes:

[0053] Step S1: Intelligent warehousing multi-source data collection;

[0054] Through RFID tags, temperature and humidity sensors, GPS logistics tracking equipment and other IoT terminals, real-time data collection of material inventory, transportation time, demand volatility, storage environment parameters and other data is collected. A multidimensional data matrix X is constructed, covering material characteristics (such as type, weight, expiration date), supply chain characteristics (such as supplier response time, transportation route complexity) and environmental characteristics (such as warehouse temperature and humidity, geographical location).

[0055]

[0056] In the formula, X is a multi-dimensional data matrix, i is the number of material types, and j is the feature dimension (such as inventory, demand forecast error, transportation cost, etc.); x ij is the original data of the j-th feature of the i-th type of material.

[0057] Step S2: Use the improved weighted K-means++ algorithm to perform dynamic clustering on the data;

[0058] Perform standardization processing on the original data in Step S1 to eliminate the dimension difference.

[0059]

[0060] x' ij is the standardized data, μ j is the mean of the j-th dimension feature, and σ j is the standard deviation of the j-th dimension feature.

[0061] Then use the improved weighted K-means++ algorithm to construct a dynamic clustering function J and perform dynamic clustering on the data.

[0062]

[0063] J is a dynamic clustering function constructed by the improved weighted K-means++ algorithm; m is the total number of samples, k is the number of cluster centers; t is the current iteration number; w ic ∈[0,1] is the membership degree weight of the i-th sample to the c-th class; is the cluster center of the previous iteration; λ is a dynamic adjustment factor (to prevent sudden changes in the cluster center); μ c is the coordinate of the new cluster center of the c-th class (the updated value).

[0064] In the dynamic clustering process, the cluster center update rule is the core operation of the improved K-means++ algorithm. It solves the problems of sensitivity to noise and easy to fall into local optimum of the traditional algorithm by dynamically adjusting the position of the cluster center.

[0065]

[0066] Step S3: Based on the clustering results, extract the characteristic distribution rules of different material categories, calculate the inter-class similarity, and identify the material groups that can be combined for supply;

[0067] ① Construction of the characteristic distribution vector: Introduce multi-dimensional statistics and business characteristics, and for each cluster c, calculate its material mixed characteristic distribution vector V c

[0068] V c =[μ c ,σc , Skewnecc c , Kurtosis c , v j , v z

[0069] Statistical dimension: σ c is the standard deviation, Skewnec c is the skewness, Kurtosis c is the kurtosis, used to capture the distribution shape;

[0070] Business dimension: Incorporate the unique attributes of electric power materials, v j is the urgency coefficient, v z is the inventory turnover rate.

[0071] By fusing the statistical distribution and business attributes, the comprehensiveness of feature expression is improved, making it more suitable for the electric power material scenario.

[0072] ② Calculate the dynamic weighted similarity Sim(V a , V b ), to improve the merging accuracy

[0073] Traditional methods directly use cosine similarity without distinguishing the importance of features. The present invention designs a weighted cosine similarity, giving higher weights to electric power business features.

[0074]

[0075] V a , V b are the feature distribution vectors of clusters a and b; Sim ∈ [-1, 1] is the similarity value. μ aj , μ bj represent the cluster center coordinates of cluster a and cluster b on the j-th feature respectively.

[0076] The weight τ j is determined by expert scoring. Key electric power business features are prioritized. When calculating the similarity, key electric power business features will be given higher weights to reflect their importance in the electric power business.

[0077] ③ Based on the dynamic weighted similarity, perform electric power material merging to generate a set of candidate merging groups G = {g1, g2,..., g n}(each g n represents a possible material group to be merged)

[0078] Step S4: Establish a multi-objective optimization model for supply merging decision-making.

[0079] ​Traditional methods using fixed weight coefficients cannot adapt to real-time scenario changes. The present invention introduces a dynamic weight adjustment function to dynamically optimize weights in combination with real-time data (such as traffic conditions, proportion of emergency orders):

[0080]

[0081] a(t) is the weight coefficient function of transportation timeliness; α is the parameter for adjusting the amplification or suppression intensity of the "proportion of emergency orders" on the timeliness weight a(t); η is the proportion of emergency orders; ζ is the traffic congestion index; c(t) is the weight coefficient function of carbon emissions; β is the parameter for adjusting the direct influence proportion of the "carbon tax policy intensity" on the carbon emissions weight coefficient function c(t); θ is the carbon tax policy intensity; b(t) is the weight coefficient of the delivery distance.

[0082] Under the satisfaction of the constraint conditions, find the optimal merging scheme that minimizes the total cost C merge the smallest.

[0083]

[0084] In the formula:

[0085] C merge : the total cost of the merging scheme;

[0086] a(t), b(t), c(t): the weight coefficients of transportation timeliness, delivery distance, and carbon emissions;

[0087] a(t) + b(t) + c(t) = 1;

[0088] T g : the average transportation timeliness (hours) of the g-th merging group;

[0089] D g : the total delivery distance (kilometers) of the g-th merging group;

[0090] E g : the carbon emissions directly generated by the energy consumption (fuel or electricity) of the transportation tools (such as trucks, drones) during the delivery process of the g-th merging group;

[0091] Q min / Q max : the minimum / maximum carrying capacity per single delivery;

[0092] the quantity of the p-th material in the g-th merging group;

[0093] Δt max : the maximum allowable timeliness difference threshold;

[0094] t p 、t q: Two arbitrary time points in transportation.

[0095] In summary, the present invention significantly improves the timeliness, economy and environmental protection of power emergency material dispatch through intelligent storage data fusion and dynamic optimization decision-making. The innovative design of step S3 (material merging strategy) and step S4 (dynamic multi-objective optimization) breaks through the limitations of traditional methods. In the material merging link (step S3), the traditional method relies on a single statistical feature or manual experience, and the merging error rate is high. The present invention proposes a multi-dimensional feature fusion mechanism, combines statistical distribution with power business attributes, constructs a hybrid feature vector, and introduces a dynamic weighted similarity algorithm. Through expert scoring, key business features are given higher weights, which improves the merging accuracy and reduces the mismerging rate. In the dynamic multi-objective optimization link (step S4), the traditional fixed weight model is difficult to adapt to emergency scenarios. The present invention innovatively designs a dynamic weight function to adjust the timeliness and carbon emission weights in real time through parameters α and β. Combined with the multi-objective optimization algorithm, the timeliness, cost and carbon emissions are globally balanced, realizing efficient and low-carbon operation of the power material supply chain.

[0096] Obviously, the above embodiments of the present invention are only examples for clearly explaining the present invention, and are not intended to limit the implementation methods of the present invention, and the present invention is not limited to the above examples. For ordinary technicians in the relevant field, other different forms of changes or modifications can be made based on the above description. It is impossible to list all the implementation methods here. All obvious changes or modifications derived from the technical solution of the present invention are still within the scope of protection of the present invention.

Claims

1. An intelligent integration method for power material supply, characterized in that, Including: Step S1: Intelligent warehousing multi-source data collection; Step S2: Using the improved weighted K-means++ algorithm to dynamically cluster the data; Step S3: Based on the clustering results, extract the characteristic distribution laws of different material categories, calculate the similarity between classes, and identify the material groups that can be combined for supply; Step S4: Establish a multi-objective optimization model for supply merger decision-making.

2. The intelligent integration method of power material supply according to claim 1, wherein, Step 1 includes: Through the Internet of Things terminal, real-time collect the inventory of materials, transportation timeliness, demand volatility, and warehousing environment data; Construct a multi-dimensional data matrix X, which covers material characteristics, supply chain characteristics, and environmental characteristics. Among them, material characteristics include type, weight, and expiration date; supply chain characteristics include supplier response time and transportation path complexity, and environmental characteristics include warehouse temperature and humidity and geographical location; Wherein, X is a multi-dimensional data matrix, i is the number of material types, j is the feature dimension, including inventory, demand forecast error, and transportation cost; x ij is the original data of the j-th dimension feature of the i-th type of material.

3. The intelligent integration method of power material supply according to claim 2, wherein Step S2 includes: First, standardize the original data x of the j -th dimension feature of the i -th type of material in step S1 to eliminate the dimensional difference; ij ​ where \(x'\) ij is the standardized data, and \(\mu\) j is the mean of the \(j\)-th dimensional feature, and \(\sigma\) j is the standard deviation of the \(j\)-th dimensional feature; Then, use the improved weighted K-means++ algorithm to construct a dynamic clustering function J to dynamically cluster the data; where J is the dynamic clustering function constructed by using the improved weighted K-means++ algorithm; m is the total number of samples, k is the number of clustering centers; t is the current iteration number; w ic ∈[0,1] is the membership degree weight of the i-th sample to the c-th class; is the clustering center of the previous iteration; λ is the dynamic adjustment factor; μ c is the coordinate of the new cluster center for the c-th class.

4. The intelligent integration method of power material supply according to claim 3, characterized in that Step S3 includes the following process: ① Feature distribution vector construction: Introduce multi-dimensional statistics and business features, and for each cluster c, calculate its material mixture feature distribution vector V c : V c = [μ c , σ c , Skewnecc c , Kurtosis c , v j , v z ​ Statistical dimension: σ c is the standard deviation, Skewnecc c is the skewness, Kurtosis c is the kurtosis, used to capture the distribution shape; Business dimension: Incorporate the unique attributes of power materials, v j is the emergency level coefficient, v z is the inventory turnover rate; ② Calculate the dynamic weighted similarity Sim(V a ,V b ) where V a and V b are the feature distribution vectors of clusters a and b; Sim ∈ [-1, 1] is the similarity value; μ aj and μ bj represent the cluster center coordinates of cluster a and cluster b on the j-th feature, respectively; Weight τ j Determined by expert scoring, giving priority to ensuring key power business characteristics. When calculating similarity, key power business characteristics will be assigned higher weights; ③Based on the dynamic weighted similarity, power materials are merged to generate a set of candidate merged groups \(G = \{g_1, g_2,..., g_n\}\), where \(g_n\) represents the \(n\)th merged material group. n},where \(g_n\) n represents the \(n\)th merged material group.

5. The intelligent integration method of power material supply according to claim 4, characterized in that, Step S4 includes: Introduce a dynamic weight adjustment function, and dynamically optimize the weights in combination with real-time data such as traffic conditions and the proportion of emergency orders: a(t) is the weight coefficient function of transportation timeliness; α is the amplification or suppression intensity parameter that adjusts the influence of the proportion of emergency orders on the timeliness weight a(t); η is the proportion of emergency orders; ζ is the traffic congestion index; c(t) is the weight coefficient function of carbon emissions; β is the direct influence proportion parameter that adjusts the intensity of the carbon tax policy on the weight coefficient function c(t) of carbon emissions; θ is the intensity of the carbon tax policy; b(t) is the weight coefficient of the distribution distance; Under the condition of satisfying the constraints, find the optimal merging scheme that minimizes the total cost C merge ; In the formula: C merge : Total cost of the merging plan; a(t), b(t), and c(t): The weight coefficient of transportation timeliness, the weight coefficient of distribution distance, and the weight coefficient of carbon emissions; a(t) + b(t) + c(t) = 1; T g : The average transportation time efficiency of the g-th merging group, in hours; D g : The total distribution distance of the g-th merging group, in kilometers; E g : The carbon emissions directly generated by the energy consumption of the transportation vehicle during the distribution process of the g-th merging group; Q min : The minimum carrying capacity for a single distribution Q max : Maximum load capacity for single delivery; The quantity of the p-th material in the g-th merging group; Δt max : The maximum allowable aging difference threshold; t p and t q : Two arbitrary points in time during transportation.

Images