Cold chain warehouse inventory management system based on big data analysis

Abstract

The invention relates to the technical field of cold chain warehouse inventory management, in particular to a cold chain warehouse inventory management system based on big data analysis, which comprises a correlation factor determination module used for classifying a cold storage energy consumption related parameter set into a non-correlation parameter set and a correlation parameter set through a correlation model, and the correlation model is constructed based on an Apriori algorithm; the data filling module is used for generating filling parameters from the uncorrelated parameter set through a Lagrange interpolation method; the energy consumption prediction module is used for generating and predicting the energy consumption of the refrigeration house through an energy consumption prediction model by using the filling parameters and the associated parameter set, and the energy consumption prediction model is based on architecture construction of a time convolutional network fused gating circulation unit and an attention mechanism; and the cold chain warehouse inventory management module is used for managing the cold chain warehouse inventory according to the predicted energy consumption of the refrigeration house. According to the invention, intelligent cold-control inventory management of cold chain storage according to prediction of the energy consumption of the refrigeration house is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of cold chain warehousing inventory management, and particularly to a cold chain warehousing inventory management system based on big data analysis. Background Art

[0002] Cold chain warehouses, as the core link in the processing, storage, and transportation of cold chain foods, are becoming increasingly prominent. According to the data released by the Cold Chain Alliance, the cold storage capacity has shown a steady growth trend, and this growth momentum is expected to continue in the coming period.

[0003] However, the rapid development of the cold chain industry is also accompanied by challenges such as high energy consumption and high costs. Therefore, in-depth research on the energy consumption of cold chain warehouses and strengthening energy efficiency inventory management are of crucial significance for improving economic benefits.

[0004] Currently, through the investigation and analysis of the energy consumption of a 30,000-ton cold chain warehouse, it is found that the energy consumption of the refrigeration system accounts for 83.3%, of which the energy consumption of the fans accounts for 13.3%, the energy consumption for cooling accounts for 70.0%, the energy consumption for lighting accounts for 2.3%, and the energy consumption for others accounts for 14.4%. Therefore, it can be seen that reasonably utilizing the energy consumption of the refrigeration system in cold chain warehousing inventory management is crucial for improving energy use efficiency.

[0005] Therefore, how to accurately predict the energy consumption of cold chain warehouses to achieve the transformation of cold chain warehousing from passive refrigeration inventory management to intelligent cooling control inventory management based on the predicted energy consumption is a technical problem to be solved currently. Summary of the Invention

[0006] For this purpose, the present invention provides a cold chain warehousing inventory management system based on big data analysis. By using the Apriori algorithm to determine the real-time cold storage energy consumption correlation parameter set, and then by complementing the parameters that have no correlation with the cold storage energy consumption and inputting them into the energy consumption prediction model based on the time series convolution fusion bidirectional memory network, a more accurate predicted cold storage energy consumption is generated, and then intelligent cooling control inventory management based on energy consumption prediction is realized.

[0007] To achieve the above object, the present invention proposes a cold chain warehousing inventory management system based on big data analysis, including:

[0008] A correlation factor determination module for classifying the cold storage energy consumption related parameter set into an uncorrelated parameter set and a correlated parameter set through a correlation model, wherein the correlation model is constructed based on an improved Apriori algorithm with a set dynamic threshold, and the dynamic threshold is dynamically adjusted according to the cold storage state change amount and / or the correlation rule set eigenvalue;

[0009] A data filling module, which is connected to the associated factor determination module, and is used to generate filling parameters for the unassociated parameter set through Lagrange interpolation method;

[0010] An energy consumption prediction module, which is connected to the associated factor determination module and the data filling module, and is used to generate predicted cold storage energy consumption by the energy consumption prediction model with the filling parameters and the associated parameter set. The energy consumption prediction model is constructed based on the architecture of a time convolutional network integrating a gated recurrent unit and an attention mechanism. The time convolutional network is used to generate a time series parameter sequence according to the filling parameters and the associated parameter set. The gated recurrent unit is used to generate the predicted cold storage energy consumption as a hidden state according to the time series parameter sequence. The attention mechanism is used to control the update of the hidden state by the gated recurrent unit;

[0011] A cold chain warehousing inventory management module, which is connected to the energy consumption prediction module, and is used to manage the inventory optimization, the storage location of goods, and the inbound and outbound strategies of cold chain warehousing inventory according to the predicted cold storage energy consumption.

[0012] Further, the time convolutional network has at least two residual blocks, and each residual block includes a dilated causal convolutional layer, a weight normalization layer, a LeakyReLU activation function layer, and a regularization layer;

[0013] The gated recurrent unit is used to generate the predicted cold storage energy consumption as a hidden state according to the time series parameter sequence;

[0014] The attention mechanism is used to control the update of the hidden state by the gated recurrent unit.

[0015] Further, the hidden state includes a previous hidden state, a current hidden state, and a transfer hidden state, and the attention mechanism includes a weight generation subunit and a transfer subunit;

[0016] The weight generation subunit is used to generate attention weights through the attention mechanism according to the current hidden state and a target vector;

[0017] The transfer subunit is used to generate the transfer hidden state according to the attention weights and the previous hidden state, and transfer the transfer hidden state to the gated recurrent unit.

[0018] Further, the energy consumption prediction module further includes a target function unit;

[0019] The target function unit is used to control the optimized prediction of the gated recurrent unit through a target function, where the target function is constructed based on cross entropy and an activation function.

[0020] In the above solution, the similarity between the user's current hidden state and the target item is calculated through the attention mechanism to update the hidden state of the gated recurrent unit, which combines the potential characteristics of the time series parameter sequence and can effectively improve the performance and accuracy of the cold storage energy consumption prediction task.

[0021] Further, the association factor determination module includes a preprocessing unit, an association rule generation unit, and a judgment unit;

[0022] The preprocessing unit is used to process the cold storage energy consumption related parameter set and generate an interval representation;

[0023] The association rule generation unit is used to iterate the interval representation to generate an association rule set associated with the comprehensive cold storage energy consumption;

[0024] The judgment unit is used to classify the cold storage energy consumption related parameter set according to the association rule set to generate the non-associated parameter set and the associated parameter set.

[0025] Further, the dynamic threshold includes a support threshold and a confidence threshold, the cold storage state change amount includes the cold storage external temperature change amount and the goods storage change amount, the association rule set eigenvalue is the average lift of the association rule set, and the judgment unit includes a support dynamic adjustment subunit and a confidence dynamic adjustment subunit;

[0026] The support dynamic adjustment subunit is used to set the support threshold according to the cold storage external temperature change amount and the goods storage change amount;

[0027] The confidence dynamic adjustment subunit is used to set the confidence threshold according to the average lift of the association rule set.

[0028] Further, the preprocessing unit includes a data cleaning subunit, a standardization subunit, and a clustering and grouping subunit;

[0029] The data cleaning subunit is used to clean the cold storage energy consumption related parameter set to generate a cleaned cold storage energy consumption related data set;

[0030] The standardization subunit is used to standardize the cleaned cold storage energy consumption related data set to generate a standard cold storage energy consumption related data set;

[0031] The clustering and grouping subunit is used to discretize the standard cold storage energy consumption related data set through a clustering algorithm to generate the interval representation;

[0032] Among them, the cold storage energy consumption related parameter set includes cold chain storage internal and external environment parameters, cold chain storage switch parameters, cold chain storage inventory parameters, and cold chain storage refrigeration equipment parameters.

[0033] Furthermore, the data filling module includes a historical data set updating unit and a filling unit;

[0034] The historical data set updating unit is used to update the cold storage historical data set according to the cold storage operation stability;

[0035] The filling unit is used to determine the corresponding parameters in the cold storage historical data set according to the uncorrelated parameter set, and generate filling parameters for the corresponding parameters by Lagrange interpolation method.

[0036] In the above solution, the problem that the correlation relationships of various parameters in the cold storage energy consumption related parameter set are not clear enough and are prone to change over time is overcome. Furthermore, the current correlation relationship is accurately mined, and the dynamic and accurate correlation degree judgment of the cold storage energy consumption related parameter set is realized. As a result, the data input into the energy consumption prediction model of the time series convolution fusion bidirectional memory network has a relatively high correlation, and the cold storage energy consumption can be accurately predicted.

[0037] Furthermore, the cold storage energy consumption related parameters include the inventory goods category, and the cold chain storage inventory management module further includes an inventory optimization unit and a goods storage location planning unit;

[0038] The inventory optimization unit is used to adjust the upper limit of the regional inventory quantity of the cold chain storage according to the predicted cold storage energy consumption;

[0039] The goods storage location planning unit is used to determine the recommended storage area for inventory goods according to the predicted cold storage energy consumption and the inventory goods category.

[0040] Furthermore, the cold chain storage inventory management module includes an inbound and outbound strategy unit;

[0041] The inbound and outbound strategy unit is used to take the time period when the predicted cold storage energy consumption is in the low power consumption interval as the recommended inbound and outbound time period.

[0042] In the above solution, the dynamic management of the upper limit of the inventory quantity, the goods storage location and the inbound and outbound strategy of the cold chain storage inventory is realized through the predicted cold storage energy consumption.

[0043] Compared with the prior art, the beneficial effects of the present invention are as follows.

[0044] The real-time cold storage energy consumption correlation parameter set is determined by the Apriori algorithm. Furthermore, the parameters that have no correlation with the cold storage energy consumption are complemented and input into the energy consumption prediction model based on the time series convolution fusion bidirectional memory network to generate a more accurate predicted cold storage energy consumption, and then the intelligent cold storage inventory management based on energy consumption prediction is realized.

[0045] The hidden state of the gated recurrent unit is updated by calculating the similarity between the current hidden state of the user and the target item through the attention mechanism, which combines the potential characteristics of the time series parameter sequence and can effectively improve the performance and accuracy of the cold storage energy consumption prediction task.

[0046] It overcomes the problem that the correlation relationships of various parameters in the cold storage energy consumption related parameter set are not clear enough and are prone to change over time, thereby realizing the accurate mining of the current correlation relationships, achieving a dynamic and accurate correlation degree judgment for the cold storage energy consumption related parameter set, and then enabling the data of the energy consumption prediction model of the input time series convolution fusion bidirectional memory network to have a high correlation and accurately generating the predicted cold storage energy consumption.

[0047] It realizes the dynamic management of the upper limit of the inventory quantity, the storage location of goods, and the inbound and outbound strategies of the cold chain storage inventory by predicting the cold storage energy consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a schematic structural diagram of the cold chain storage inventory management system based on big data analysis according to an embodiment of the present invention;

[0049] Figure 2 It is a schematic flow diagram of the cold chain storage inventory management system based on big data analysis according to an embodiment of the present invention;

[0050] Figure 3 It is a schematic structural diagram of the residual block of the energy consumption prediction model of the cold chain storage inventory management system based on big data analysis according to an embodiment of the present invention;

[0051] Figure 4 It is a detailed schematic structural diagram of the bidirectional long short-term memory network unit and the attention mechanism of the energy consumption prediction model of the cold chain storage inventory management system based on big data analysis according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0052] In order to make the purpose and advantages of the present invention clearer and more understandable, the present invention will be further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0053] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principle of the present invention and do not limit the protection scope of the present invention.

[0054] It should be noted that in the description of the present invention, the terms indicating the direction or positional relationship such as "upper", "lower", "left", "right", "inner", "outer", etc. are based on the direction or positional relationship shown in the drawings. This is only for convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation to the present invention.

[0055] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installation", "connection", and "connection" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0056] As Figures 1 to 4 shown, the present invention provides a cold chain warehousing inventory management system based on big data analysis. By using the Apriori algorithm to determine the real-time cold storage energy consumption correlation parameter set, and then by complementing the parameters that have no correlation with the cold storage energy consumption and inputting them into the energy consumption prediction model based on the time series convolution fusion bidirectional memory network, a more accurate prediction of the cold storage energy consumption is generated, and then the intelligent cold storage inventory management based on energy consumption prediction is realized.

[0057] As Figures 1 to 4 shown, this embodiment proposes a cold chain warehousing inventory management system based on big data analysis, including:

[0058] An association factor determination module for classifying the cold storage energy consumption related parameter set into an uncorrelated parameter set and a correlated parameter set through an association degree model. Among them, the association degree model is constructed based on an improved Apriori algorithm with a set dynamic threshold, and the dynamic threshold is dynamically adjusted according to the cold storage state change amount and / or the eigenvalue of the association rule set;

[0059] A data filling module connected to the association factor determination module for generating filling parameters from the uncorrelated parameter set by using the Lagrange interpolation method;

[0060] An energy consumption prediction module, which is connected to the associated factor determination module and the data filling module, is used to generate a predicted cold storage energy consumption through an energy consumption prediction model using the filling parameters and the associated parameter set. The energy consumption prediction model is constructed based on an architecture that combines a temporal convolutional network, a gated recurrent unit, and an attention mechanism. The temporal convolutional network is used to generate a temporal parameter sequence according to the filling parameters and the associated parameter set. The gated recurrent unit is used to generate the predicted cold storage energy consumption as a hidden state according to the temporal parameter sequence. The attention mechanism is used to control the update of the hidden state by the gated recurrent unit;

[0061] A cold chain storage inventory management module, which is connected to the energy consumption prediction module, is used to manage the inventory optimization, the storage location of goods, and the inbound and outbound strategies of the cold chain storage inventory according to the predicted cold storage energy consumption.

[0062] It can be understood that the cold storage energy consumption prediction based on TCN-GRU (temporal convolutional fusion bidirectional memory network) uses the cold storage energy consumption and multiple other cold storage energy consumption-related parameters in the previous period to predict the cold storage energy consumption at the current time. There needs to be a strong correlation / causal relationship between the input quantity and the predicted quantity. However, this correlation may change over time. For example, when the outdoor temperature is low, the average opening duration of the cold storage door and the inventory have a weak correlation with the cold storage energy consumption, and thus should be screened out and filled with data through the Lagrange interpolation method to ensure a clear causal relationship between the input variables and output variables of the model and an accurate control strategy for cold chain storage inventory management.

[0063] Specifically, the Lagrange interpolation method fills and replaces the unassociated parameter set according to the closest normal historical data.

[0064] Furthermore, as Figure 3 shown, the temporal convolutional network has at least two residual blocks, and each residual block includes a dilated causal convolutional layer, a weight normalization layer, a LeakyReLU activation function layer, and a regularization layer.

[0065] Specifically, as Figure 3 shown, the temporal convolutional network (TCN) unit includes: a causal convolution, which is used to ensure the causality of the temporal parameter sequence, and uses a one-dimensional convolutional layer with a convolution kernel size of 9, and each output point is composed of the inputs of the previous 9 time instants; a dilated convolution, which is used to expand the receptive field through skip interval sampling to capture long-period features; and residual block connections (ResidualConnection), preferably set to 4, which are used to alleviate the problem of gradient disappearance and accelerate model convergence. Therefore, the temporal convolutional network (TCN) unit is suitable for periodic prediction of different cold storage energy consumption.

[0066] Furthermore, referring to Figure 4, the hidden state includes a previous hidden state, a current hidden state, and a transmitted hidden state, and the attention mechanism includes a weight generation subunit and a transmission subunit;

[0067] The weight generation subunit is used to generate attention weights through the attention mechanism according to the current hidden state and the target vector;

[0068] The transmission subunit is used to generate the transmitted hidden state according to the attention weights and the previous hidden state, and transmit the transmitted hidden state to the gated recurrent unit.

[0069] See Figure 4 , specifically, the weight generation subunit is:

[0070]

[0071] In the formula, represents the output value of the attention mechanism, represents that the softmax activation function controls whether the target vector at the t-th moment is updated, and respectively represent the current hidden state, and W represents the attention weight.

[0072] See Figure 4 , specifically, the transmission subunit is:

[0073]

[0074] In the formula, respectively represent the previous hidden state and the transmitted hidden state, represents the input embedding at the t-th moment, ReLU is the ReLU activation function, W represents the attention weight, is a learnable weight vector, and

[0075] are both learnable bias terms. Figure 4 See

[0076]

[0077] In the formula, respectively represent the next hidden state, the previous hidden state, and the transmitted hidden state, and

[0078] Specifically, see Figure 4, the gated recurrent unit (GRU) in the energy consumption prediction model is an improved recurrent neural network structure. Its core design realizes the modeling and prediction of the dynamic characteristics of time series through the update gate and reset gate mechanisms. The expression of the gated recurrent unit is as follows:

[0079]

[0080] In the formula, respectively represent the current hidden state, the previous hidden state, and the transmitted hidden state, respectively represent the update gate at the t-th moment, the input embedding at the t-th moment, and the reset gate at the t-th moment. ELU is the ELU (Exponential Linear Unit) activation function, are all learnable weight vectors, are all learnable bias terms. Among them, the previous hidden state and the transmitted hidden state are updated according to the update gate, and their sum is used to obtain the current hidden state , the value of the transmitted hidden state is updated through the transmission sub-unit, and the reset gate is used to judge whether the input embedding and the previous hidden state need to be updated to generate the transmitted hidden state. Specifically, take Figure 3 The output of the time convolutional network (TCN) shown as Figure 4 The input embedding of the gated recurrent unit (GRU) in

[0081] Furthermore, the energy consumption prediction module further includes an objective function unit;

[0082] The objective function unit is used to control the optimized prediction of the gated recurrent unit through the objective function, where the objective function is constructed based on cross-entropy, specifically:

[0083]

[0084] In the formula, L is the objective function, y represents the label value, p represents the predicted value, F is the output of the weight generation sub-unit, W is the learnable weight vector, and b is the learnable bias term.

[0085] In the above solution, the similarity between the user's current hidden state and the target item is calculated through the attention mechanism to update the hidden state of the gated recurrent unit, combining the potential time series parameter sequence features, which can effectively improve the performance and accuracy of the cold storage energy consumption prediction task.

[0086] Furthermore, as Figure 2 shown, the correlation factor determination module includes a preprocessing unit, a correlation rule generation unit, and a judgment unit;

[0087] The preprocessing unit is used to process the data of the cold storage energy consumption related parameter set and generate an interval representation;

[0088] The association rule generation unit is used to iterate the interval representation to generate an association rule set associated with the comprehensive cold storage energy consumption;

[0089] The judgment unit is used to classify the cold storage energy consumption related parameter set according to the association rule set to generate the non-associated parameter set and the associated parameter set.

[0090] Specifically, the Apriori algorithm has scalability and good performance, and is especially suitable for single-dimensional attributes. The process of mining association rules is as follows: iterate to find all item sets in the transaction database that are greater than or equal to the user-specified minimum support threshold, that is, frequent item sets; use the frequent item sets to mine strong association rules that meet the user's requirements, that is, an association rule set with support and confidence greater than or equal to the corresponding thresholds. Let the interval representation D be a set of data transaction databases, and both A and B are item sets, where the item set A is the interval representation and the item set B is the comprehensive cold storage energy consumption. The association rule set contains multiple association rules, and the association rule can be expressed in the form that if the item set A appears, then the item set B may also appear, and has a support s and a confidence c in the transaction set D. Among them, the support s refers to the percentage of the item sets A and B in the transaction set D, and the confidence c is the percentage of the item set B in the transaction set D while also containing the item set A.

[0091] Among them, the item set A of the interval representation includes the grade values and combinations of grade values of multiple parameters in the cold storage energy consumption related parameter set. For example, the elements in the item set A include the temperature interval representation in the cold chain warehouse from one to three, the humidity interval representation in the cold chain warehouse from one to three, the temperature interval representation outside the cold chain warehouse from one to three, and combinations between the above various elements, such as the temperature interval representation in the cold chain warehouse one + the humidity interval representation in the cold chain warehouse two + the temperature interval representation outside the cold chain warehouse one. The interval representation range is, for example, the temperature interval representation in the cold chain warehouse one is the low temperature interval [-19.5, -19.2] degrees Celsius, the temperature interval representation in the cold chain warehouse two is the medium temperature interval (-19.2, -18.8] degrees Celsius, and the temperature interval representation in the cold chain warehouse three is the high temperature interval (-18.8, -18.4] degrees Celsius.

[0092] It should be noted that the comprehensive cold storage energy consumption is the sum of the refrigeration electricity consumption (REC) and the direct electricity consumption (DEC). The refrigeration electricity consumption includes the power consumption of core equipment such as compressors and condensers, and the direct electricity consumption includes the power consumption of auxiliary equipment such as lighting, cooling fans, defrosting, and air curtains.

[0093] Furthermore, as Figure 2As shown, the dynamic thresholds include a support threshold and a confidence threshold. The cold storage state change amount includes the cold storage external temperature change amount and the goods storage change amount. The associated rule set eigenvalue is the average lift of the associated rule set. The judgment unit includes a support dynamic adjustment subunit and a confidence dynamic adjustment subunit;

[0094] The support dynamic adjustment subunit is used to set the support threshold according to the cold storage external temperature change amount and the goods storage change amount;

[0095] The confidence dynamic adjustment subunit is used to set the confidence threshold according to the average lift of the associated rule set.

[0096] Specifically, the support dynamic adjustment subunit is expressed as:

[0097]

[0098] In the formula, represents the support threshold, respectively represent the temperature change value and the temperature change threshold, and S is the goods storage volume.

[0099] Specifically, the average lift is:

[0100]

[0101] In the formula, is the average lift of the associated rule set, is the lift calculation process, where X is the interval representation, Y represents the comprehensive energy consumption of the cold storage, Support represents the support, and N is the number of rules in the rule set.

[0102] It can be understood that the support (Support(A→B)) reflects the co-occurrence frequency of the item set A and the item set B of the comprehensive energy consumption of the cold storage. Item sets with low support may be occasional but critical abnormal patterns, such as low-temperature intervals and high humidity leading to a sharp increase in defrost energy consumption. Therefore, when the average lift is greater than the high lift threshold, the confidence threshold is increased by the first set ratio to avoid misjudging weak associations due to the common occurrence of high-frequency combinations; when the average lift is greater than the high lift threshold, the confidence threshold is decreased by the second set ratio to retain low-frequency but high-value abnormal association rules. Preferably, the high support threshold is 20%, the low support threshold is 1.8%, the first set ratio is 5%, and the second set ratio is 3%.

[0103] Further, the determination unit is configured to set the parameters that exist at least in two of the association rule sets as the associated parameter set, otherwise as the irrelevant parameters. For example, if the internal and external environment parameters of the cold chain storage simultaneously satisfy the rules: the interval representation of the internal and external environment parameters of the cold chain storage is one to the comprehensive energy consumption of the cold storage, the interval representation of the internal and external environment parameters of the cold chain storage is two, and the interval representation of the internal humidity of the cold chain storage is one to the comprehensive energy consumption of the cold storage, then the internal and external environment parameters of the cold chain storage are used as the associated parameter set.

[0104] Further, the preprocessing unit includes a data cleaning subunit, a standardization subunit, and a clustering and grouping subunit;

[0105] The data cleaning subunit is configured to perform data cleaning on the cold storage energy consumption related parameter set to generate a cleaned cold storage energy consumption related data set;

[0106] The standardization subunit is configured to perform standardization processing on the cleaned cold storage energy consumption related data set to generate a standard cold storage energy consumption related data set;

[0107] The clustering and grouping subunit is configured to perform discretization processing on the standard cold storage energy consumption related data set through a clustering algorithm to generate the interval representation;

[0108] Wherein, the cold storage energy consumption related parameter set includes the internal and external environment parameters of the cold chain storage, the cold chain storage switch parameters, the cold chain storage inventory parameters, and the cold chain storage refrigeration equipment parameters, and the interval representation is the level of each parameter in the cold storage energy consumption related parameter set.

[0109] Specifically, the internal and external environment parameters include the internal temperature of the cold chain storage, the internal humidity of the cold chain storage, the external temperature of the cold chain storage, and the external humidity of the cold chain storage. The cold chain storage switch parameter is the average value of the opening and closing door duration at present on the current day. The cold chain storage inventory parameters include the frequency of goods in and out, the number of times the cold storage door is opened and closed, and the inventory quantity. The cold chain storage refrigeration equipment parameters include the running duration of the compressor, the refrigerant pressure, and the equipment aging coefficient.

[0110] Specifically, data cleaning is to delete the obviously distorted data in the cold storage energy consumption related parameter set, such as the pump pressure and dry pressure that are accidentally detected by the compressor and are close to 0. Since the original data volume is large and the proportion of such data is small and has little impact, it is deleted.

[0111] Specifically, the standardization processing is the maximum-minimum standardization to eliminate the dimension difference in the data.

[0112] Specifically, the clustering algorithm is the K-means clustering algorithm, and the number of clusters is set to 3 levels.

[0113] Further, the data filling module includes a historical data set update unit and a filling unit;

[0114] The historical dataset updating unit is used to update the cold storage historical dataset according to the operation stability of the cold storage;

[0115] The filling unit is used to determine the corresponding parameters in the cold storage historical dataset according to the uncorrelated parameter set, and generate filling parameters for the corresponding parameters by Lagrange interpolation method.

[0116] Specifically, the historical dataset updating unit is used to calculate the statistical correlation coefficient between historical data and the comprehensive energy consumption of the cold storage, and store the historical data with the correlation coefficient greater than the lower limit value and use it to update the cold storage historical dataset regularly. It can be understood that if only the statistical correlation coefficient is used to judge the relevance of data, the data input into the energy consumption prediction model will be too small to accurately predict the energy consumption. Therefore, only the statistical correlation coefficient is used to ensure the reliability of the filled data.

[0117] Specifically, the filling unit generates filling parameters through interpolation calculation with multiple corresponding parameters.

[0118] In the above solution, the problem that the correlation relationships of various parameters in the cold storage energy consumption related parameter set are not clear enough and are prone to change with time is overcome. Furthermore, the current correlation relationship is accurately mined, and the dynamic and accurate correlation degree judgment of the cold storage energy consumption related parameter set is realized. As a result, the data input into the energy consumption prediction model constructed by integrating TCN and GRU has a high correlation, and the cold storage energy consumption can be accurately predicted.

[0119] Furthermore, the cold storage energy consumption related parameters include the types of stored goods, and the cold chain storage inventory management module further includes an inventory optimization unit and a goods storage location planning unit;

[0120] The inventory optimization unit is used to adjust the upper limit of the regional inventory quantity of the cold chain storage according to the predicted cold storage energy consumption;

[0121] The goods storage location planning unit is used to determine the recommended storage area for the stored goods according to the predicted cold storage energy consumption and the types of stored goods.

[0122] Furthermore, the cold chain storage inventory management module includes an inbound and outbound strategy unit;

[0123] The inbound and outbound strategy unit is used to take the time period when the predicted cold storage energy consumption is in the low power consumption interval as the recommended inbound and outbound time period.

[0124] Specifically, the inventory optimization unit calculates the inventory capacity mapping through the energy consumption prediction value for the next 24 hours, and when the capacity mapping exceeds the safety threshold, it triggers a downward adjustment of the upper limit of the regional inventory quantity.

[0125] Specifically, the goods storage location planning unit calculates the matching degree with the candidate area according to the category of goods (frozen / refrigerated), temperature sensitivity (such as storage temperature ±2°C), and shelf life:

[0126]

[0127] In the formula, Score represents the matching degree, respectively represent the ideal storage temperature and the current actual temperature of goods i, represents the upper limit value of the storage temperature, respectively represent the predicted energy consumption and the reference energy consumption.

[0128] The goods storage location planning unit performs priority allocation. High-value goods (such as medicines, seafood) are preferentially allocated to the core area with stable temperature control; ordinary goods are dynamically adjusted to the edge area, and an elastic inventory upper limit is set; to avoid odor cross-contamination, such as isolating seafood from dairy products, and dynamically avoiding high-energy consumption periods: if the predicted energy consumption in a certain area exceeds the limit in the next 2 hours, new goods allocation is suspended.

[0129] The inbound and outbound strategy unit is used to determine the recommended low-energy consumption periods according to the hourly energy consumption prediction values in the next 24 hours.

[0130] In this embodiment, real-time collection is carried out through IoT sensors (temperature and humidity, access control switch times), and it is integrated and set in the cold chain warehousing business system, such as WMS (Warehouse Management System), ERP (Enterprise Resource Planning).

[0131] Specifically, after converting the data format, the Apriori function is called, the initial support threshold is set to 7.5%, and the initial confidence threshold is set to 40%. The typical opening frequency of the cold storage is analyzed, and the relationship between the working day type and the cold storage temperature is obtained, laying a foundation for linking the cold storage opening behavior with the energy consumption analysis. The cold storage opening behavior determines the operating mode of the cooling fan, which in turn affects the energy consumption level of the cold storage to a certain extent. To establish a reasonable energy consumption prediction model, this article combines the opening behavior of the cold storage, the operating state of the cooling fan, and the temperature inside the freezer for a period of time. When the cooling fan is in the shutdown state and the door is opened continuously for 140s, the cold storage temperature rises by 0.64°C; a single door opening behavior of 10 to 20s causes the temperature fluctuation of the cold storage not to exceed 0.1°C. In fact, if the cold storage has frequent short door openings within a period of time, the temperature fluctuation of the cold storage will also increase accordingly.

[0132] In the above solution, dynamic management of the upper limit of the inventory quantity, the goods storage location, and the inbound and outbound strategy of the cold chain warehousing inventory is realized through predicting the cold storage energy consumption.

[0133] In this embodiment, the real-time cold storage energy consumption correlation parameter set is determined by the Apriori algorithm. Then, the parameters that have no correlation with the cold storage energy consumption are complemented and input into the energy consumption prediction model constructed by fusing TCN and GRU to generate a more accurate predicted cold storage energy consumption. Furthermore, the intelligent cold storage management of the cold chain storage is realized according to the predicted cold storage energy consumption. By calculating the similarity between the current hidden state of the user and the target item through the attention mechanism, the hidden state of the gated recurrent unit is updated, combining the potential time series parameter sequence features, which can effectively improve the performance and accuracy of the cold storage energy consumption prediction task. It overcomes the problem that the correlation relationships of various parameters in the cold storage energy consumption-related parameter set are not clear enough and are prone to change over time. Furthermore, it realizes the accurate mining of the current correlation relationships, realizes the dynamic and accurate correlation degree judgment of the cold storage energy consumption-related parameter set. As a result, the data input into the energy consumption prediction model of the temporal convolutional fusion bidirectional memory network has a high correlation, and the predicted cold storage energy consumption can be accurately generated. It realizes the dynamic management of the upper limit of the inventory quantity, the storage location of goods, and the inbound and outbound strategies of the cold chain storage inventory according to the predicted cold storage energy consumption.

[0134] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the protection scope of the present invention.

[0135] The above are only the preferred embodiments of the present invention and are not used to limit the present invention; for those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A cold chain warehouse inventory management system based on big data analysis, characterized in that, Including: An associated factor determination module, configured to classify a cold storage energy consumption related parameter set into an unassociated parameter set and an associated parameter set through an association degree model, wherein the association degree model is constructed based on an improved Apriori algorithm with a set dynamic threshold, and the dynamic threshold is dynamically adjusted according to a cold storage state change amount and / or an association rule set eigenvalue; A data filling module, connected to the associated factor determination module, configured to generate filling parameters for the unassociated parameter set through Lagrange interpolation; An energy consumption prediction module, connected to the associated factor determination module and the data filling module, configured to generate a predicted cold storage energy consumption for the filling parameters and the associated parameter set through an energy consumption prediction model, wherein the energy consumption prediction model is constructed based on an architecture that fuses a time convolutional network, a gated recurrent unit, and an attention mechanism. The time convolutional network is configured to generate a time series parameter sequence according to the filling parameters and the associated parameter set, the gated recurrent unit is configured to generate the predicted cold storage energy consumption as a hidden state according to the time series parameter sequence, and the attention mechanism is configured to control the update of the hidden state by the gated recurrent unit; A cold chain warehousing inventory management module, connected to the energy consumption prediction module, configured to manage the inventory optimization, the goods storage location, and the inbound and outbound strategies of the cold chain warehousing inventory according to the predicted cold storage energy consumption.

2. The cold chain warehousing inventory management system based on big data analysis according to claim 1, characterized in that The time convolutional network has at least two residual blocks, and each residual block includes an extended causal convolutional layer, a weight normalization layer, a LeakyReLU activation function layer, and a regularization layer.

3. The cold chain storage inventory management system based on big data analysis according to claim 1, characterized in that The hidden state includes a previous hidden state, a current hidden state, and a transfer hidden state, and the attention mechanism includes a weight generation subunit and a transfer subunit; The weight generation subunit is configured to generate attention weights through the attention mechanism according to the current hidden state and a target vector; The transfer subunit is configured to generate the transfer hidden state according to the attention weights and the previous hidden state, and transfer the transfer hidden state to the gated recurrent unit.

4. The cold chain storage inventory management system based on big data analysis according to claim 1, characterized in that, The energy consumption prediction module further includes an objective function unit; The objective function unit is configured to control the optimized prediction of the gated recurrent unit through an objective function, wherein the objective function is constructed based on cross entropy and an activation function.

5. The cold chain warehousing inventory management system based on big data analysis according to claim 1, characterized in that, The associated factor determination module includes a preprocessing unit, an association rule generation unit, and a judgment unit; The preprocessing unit is configured to perform data processing on the cold storage energy consumption related parameter set and generate an interval representation; The association rule generation unit is configured to iterate on the interval representation to generate an association rule set associated with the comprehensive cold storage energy consumption; The judgment unit is configured to classify the cold storage energy consumption related parameter set according to the association rule set to generate the unassociated parameter set and the associated parameter set.

6. The cold chain warehousing inventory management system based on big data analysis according to claim 5, characterized in that The dynamic threshold includes a support threshold and a confidence threshold. The cold storage state change amount includes a cold storage external temperature change amount and a goods storage change amount. The association rule set eigenvalue is the average lift of the association rule set. The judgment unit includes a support dynamic adjustment subunit and a confidence dynamic adjustment subunit; The support degree dynamic adjustment subunit is used to set the support degree threshold according to the change amount of the external temperature of the cold storage and the change amount of the goods storage; The confidence degree dynamic adjustment subunit is used to set the confidence degree threshold according to the average lift of the association rule set.

7. The cold chain storage inventory management system based on big data analysis according to claim 5, characterized in that The preprocessing unit includes a data cleaning subunit, a standardization subunit, and a clustering and grouping subunit; The data cleaning subunit is used to clean the cold storage energy consumption related parameter set to generate a cleaned cold storage energy consumption related data set; The standardization subunit is used to perform standardization processing on the cleaned cold storage energy consumption related data set to generate a standard cold storage energy consumption related data set; The clustering and grouping subunit is used to perform discretization processing on the standard cold storage energy consumption related data set through a clustering algorithm to generate the interval representation; Among them, the cold storage energy consumption related parameter set includes cold chain storage internal and external environment parameters, cold chain storage switch parameters, cold chain storage inventory parameters, and cold chain storage refrigeration equipment parameters.

8. The cold chain warehousing inventory management system based on big data analysis according to claim 1, characterized in that, The data filling module includes a historical data set update unit and a filling unit; The historical data set update unit is used to update the cold storage historical data set according to the cold storage operation stability; The filling unit is used to determine the corresponding parameters in the cold storage historical data set according to the uncorrelated parameter set, and generate filling parameters for the corresponding parameters by Lagrange interpolation method.

9. The cold chain warehousing inventory management system based on big data analysis according to any one of claims 1 to 8, characterized in that, The cold storage energy consumption related parameters include the categories of stored goods, and the cold chain storage inventory management module further includes an inventory optimization unit and a goods storage location planning unit; The inventory optimization unit is used to adjust the upper limit of the regional inventory volume of the cold chain storage according to the predicted cold storage energy consumption; The goods storage location planning unit is used to determine the recommended storage area for the inventory goods according to the predicted cold storage energy consumption and the categories of the inventory goods.

10. The cold chain warehousing inventory management system based on big data analysis according to any one of claims 1 to 8, characterized in that, The cold chain storage inventory management module includes an inbound and outbound strategy unit; The inbound and outbound strategy unit is used to use the time period when the predicted cold storage energy consumption is in the low power consumption interval as the recommended inbound and outbound time period.

Images