Warehouse environment adjusting method and system based on industrial Internet of Things

Abstract

The invention relates to the technical field of environment control, in particular to a warehouse environment adjusting method and system based on the industrial Internet of Things. The method comprises the following steps: respectively acquiring indoor temperature data and outdoor temperature data of each independent storage area in a warehouse in a set time sequence and a state sequence formed by door opening or closing states at each moment; based on the indoor temperature data, the outdoor temperature data and the state sequence of the set time sequence, the predicted indoor temperature of the next moment is predicted; and outputting a control signal at the next moment by using a fuzzy control algorithm based on the predicted indoor temperature. According to the scheme, the warehouse environment can be timely and accurately adjusted.

Description

Technical Field

[0001] The present invention relates to the field of environmental control technology. More specifically, the present invention relates to a warehouse environment adjustment method and system based on the Industrial Internet of Things. Background Art

[0002] Currently, maintaining a stable warehouse environment is crucial for ensuring the quality and safety of stored goods. This is especially true for warehouses storing hazardous chemicals, pharmaceuticals, and high-value commodities, which have stringent environmental requirements. For example, most vaccines require refrigeration at temperatures between 2°C and 8°C. Therefore, in warehouses storing these specialized goods, environmental control not only impacts the storage of these goods but also the overall safety of the warehouse.

[0003] Traditional warehouse environmental control methods rely primarily on manual inspections and simple sensors collecting data periodically. While this approach can monitor and regulate the warehouse environment to a certain extent, it suffers from significant regulatory limitations. Therefore, environmental control for goods stored in warehouses is particularly important.

[0004] In the related technology, such as the Chinese patent application document with publication number CN119472896A and invention name "Warehouse environment adjustment method, system, equipment and medium based on industrial Internet of Things", it is disclosed that based on the collected image data in the warehouse, the stacked materials are divided to obtain multiple stacking units in the warehouse, and by analyzing the environmental compatibility of each stacking unit, it is determined whether the environment in the warehouse needs to be adjusted.

[0005] The above solution only considers the environmental compatibility of the goods themselves to achieve warehouse environmental control. However, it fails to consider that the environmental control of independent storage areas within the warehouse may be affected by external events, such as the opening or closing of doors. This makes it impossible to accurately control the temperature of independent storage areas in a timely and effective manner, resulting in safety risks for goods stored in independent storage areas. Therefore, how to timely and effectively monitor and adjust the warehouse environment is particularly important. Summary of the Invention

[0006] The purpose of the present invention is to propose a warehouse environment adjustment method and system based on the Industrial Internet of Things to solve the problem in the existing technology that it is impossible to accurately control the environment of independent storage areas in the warehouse in a timely and effective manner; to this end, the present invention provides solutions in the following two aspects.

[0007] In a first aspect, the present invention provides a warehouse environment adjustment method based on the Industrial Internet of Things, comprising: Obtain the indoor temperature data, outdoor temperature data and door opening or closing status of each independent storage area in the warehouse within the set time sequence; Based on the set time series of indoor temperature data, outdoor temperature data, and state sequence, predict the predicted indoor temperature at the next moment; Based on the predicted indoor temperature, the fuzzy control algorithm is used to output the control signal for the next moment; Among them, the rule weight of the fuzzy control algorithm is: Obtaining a historical data set for each independent storage area, where each data point in the historical data set includes a historical impact, a difference between the indoor temperature and the target temperature, and a rule weight; the historical impact represents the change in indoor temperature at any historical moment; Clustering all data points in the historical data set based on the historical influence and difference to obtain multiple clusters; calculating the rule weight of each data point in each cluster under the corresponding control rule; the rule weight is negatively correlated with the control time under the corresponding control rule; Determine the cluster to which the predicted temperature and influence degree at the next moment belong, and select the rule weights under the corresponding control rules in the cluster as the rule weights at the next moment.

[0008] In the above scheme, by analyzing the historical impact and temperature difference of each data point in the historical data set of each independent storage area, moments with similar historical impact and temperature difference can be grouped into clusters. The rule weights of multiple control rules under the same similar environment can then be obtained. This allows the rule weights of the fuzzy control algorithm to be determined for the next moment in order to perform temperature control at that moment. In other words, the scheme of the present invention can address the dynamic impact of the door opening or closing state on the temperature within the independent storage area, and accurately control the temperature of the independent storage area in a timely and effective manner.

[0009] Optionally, the indoor temperature data is collected by sensors deployed near the door inside each independent storage area; the outdoor temperature data is collected by sensors deployed near the door outside each independent storage area.

[0010] Optionally, the influence for: ; in, The end of the window corresponds to The predicted indoor temperature at the moment, The first end of the window corresponds to The indoor temperature at the moment, is the window length, is the door opening state ratio, which is the ratio of the number of door opening states within the window length; The door opening or closing state at a certain moment is the same as the door opening or closing state at the previous moment.

[0011] The above solution provides a method for accurately calculating the influence.

[0012] Optionally, the multiple clusters are obtained by a K-mean clustering algorithm.

[0013] Through clustering methods, data with similar environments are grouped into one category, providing a basis for obtaining subsequent rule weights.

[0014] Optionally, the rule weight is a normalized value of the inverse of the control time under the corresponding control rule; the control time is the time interval from the start of control to reaching the target temperature.

[0015] By adjusting the speed of the control time, we can evaluate the control effects under different control rules in similar environments, and then obtain the rule weights of each control rule in the corresponding cluster.

[0016] Optionally, outputting a control signal at a next moment includes: The difference between the predicted indoor temperature and the target temperature is input into the fuzzy control algorithm, and the control signal at the next moment is output.

[0017] Optionally, the predicted indoor temperature is obtained using a recurrent neural network model.

[0018] The recurrent neural network model can be used to predict the indoor temperature of the independent storage area at the next moment.

[0019] Optionally, the training process of the recurrent neural network model is: Obtain a sample data set; wherein the sample data set includes historical indoor temperature data, historical outdoor temperature data, and historical state sequence within a historical set time series; The sample data set is input into the recurrent neural network model for training, and the cross entropy loss is used to calculate the difference. The parameters of the recurrent neural network model are adjusted using the gradient descent algorithm to adjust the parameters of the network model until the cross entropy loss value between the output prediction result and the true result is less than the difference threshold, thereby obtaining a trained temperature prediction model.

[0020] Optionally, it also includes: performing time alignment processing on the collected indoor temperature data and outdoor temperature data to obtain indoor temperature data after a delay time corresponding to the indoor temperature data; the delay time used in the time alignment processing is the time difference between the two parameters corresponding to the maximum mutual correlation coefficient.

[0021] Time alignment of indoor and outdoor temperature data ensures the consistency and accuracy of the collected indoor and outdoor temperature data, provides a more reliable data basis for subsequent temperature prediction and control, and improves the stability and reliability of environmental regulation in independent storage areas within the warehouse.

[0022] In the second aspect, a warehouse environment adjustment system based on the industrial Internet of Things includes: processor; The memory stores computer instructions for adjusting the warehouse environment based on the industrial Internet of Things. When the computer instructions are executed by the processor, the system executes the above-mentioned warehouse environment adjustment method based on the industrial Internet of Things.

[0023] The beneficial effects of the present invention are: The solution of the present invention can comprehensively collect and analyze the indoor and outdoor temperature data and door opening or closing status information of each independent storage area in the warehouse, fully considering the impact of various factors on the temperature of the independent storage area. Compared with the traditional temperature adjustment method that only considers a single factor or a simple threshold, it can more accurately reflect the actual temperature changes in the warehouse and achieve more reasonable temperature control. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 The following is a schematic diagram showing the steps of the warehouse environment adjustment method based on the industrial Internet of Things in this embodiment; Figure 2 The structural block diagram of the warehouse environment adjustment system based on the industrial Internet of Things in this embodiment is schematically shown. DETAILED DESCRIPTION

[0025] The technical solutions in the embodiments of the present invention will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present invention.

[0026] Since different independent storage areas are easily affected by outdoor temperature, door opening or closing operations (for example, frequent door opening will allow outdoor hot air to enter the warehouse and affect the indoor temperature), the present invention is aimed at temperature control of different independent storage areas in the warehouse.

[0027] like Figure 1 As shown, the warehouse environment adjustment method based on the industrial Internet of Things in this embodiment includes the following steps: Step S1, respectively obtaining indoor temperature data, outdoor temperature data and state sequence of each independent storage area in the warehouse within a set time sequence.

[0028] In this embodiment, indoor temperature data is collected by sensors deployed near the door inside each independent storage area; outdoor temperature data is collected by sensors deployed near the door outside each independent storage area.

[0029] Among them, the set time sequence includes the current moment and the moment before the current moment; at the same time, labels are set for the data points consisting of the indoor temperature and outdoor temperature at each moment, label 1 is the door open state, and label 0 is the door closed state. In this way, the state sequence of the door open or closed state at all times can be obtained.

[0030] Furthermore, the collected indoor and outdoor temperature data are time-aligned. Specifically, the cross-correlation coefficient between the indoor and outdoor temperature data is calculated, and the time difference between the two parameters corresponding to the maximum cross-correlation coefficient is found as the delay time. For example, a sliding window approach is used to calculate the cross-correlation coefficient within a certain delay range (e.g., from time t1 to time t1+T, where T is the set interval). The time difference between the two data points corresponding to the maximum cross-correlation coefficient is determined as the delay time.

[0031] Based on the determined delay time, the indoor temperature data and the outdoor temperature data are time-aligned to ensure that the indoor temperature data and the outdoor temperature data are accurately aligned in time, providing more accurate data for subsequent temperature prediction and control in independent storage areas.

[0032] Step S2: predicting the indoor temperature at the next moment based on the indoor temperature data, outdoor temperature data, and state sequence of the set time series.

[0033] The indoor temperature is predicted using a recurrent neural network model, specifically an LSTM model.

[0034] The training process of the recurrent neural network model is: First, a sample data set is obtained, wherein the sample data set includes historical indoor temperature data, historical outdoor temperature data, and historical state sequences within a plurality of historical set time series, wherein the label is the indoor temperature at the next moment after the last moment in each historical set time series.

[0035] The time span of the historical time series data in the sample dataset should be long enough to cover warehouse temperature changes across different seasons and time periods. For example, indoor and outdoor temperature data and corresponding historical state sequences for different time periods every day over the past year should be collected to ensure representative sample data.

[0036] Secondly, the sample data set is input into the recurrent neural network model for training, and the cross entropy loss is used to calculate the difference. The parameters of the recurrent neural network model are adjusted using the gradient descent algorithm to adjust the parameters of the network model until the cross entropy loss value between the output prediction result and the true result is less than the difference threshold, and a trained temperature prediction model is obtained.

[0037] The difference threshold mentioned above can be 0 or a value close to 0, such as 0.01. Through multiple experiments and adjustment of model parameters, the model can achieve better prediction results.

[0038] In this embodiment, after obtaining the trained recurrent neural network model, the indoor temperature data, outdoor temperature data, and state sequence with set time series are input into the recurrent neural network model to obtain the indoor temperature at the next moment and the door open or closed state of the current moment, thereby realizing the prediction of the temperature of independent storage areas in the warehouse.

[0039] By comprehensively considering multiple factors such as indoor and outdoor temperature, state sequence (door opening and closing frequency and duration of opening or closing), and utilizing a recurrent neural network model, the above method can accurately predict the indoor temperature of each independent storage area at the next moment. Compared with traditional methods, it can predict temperature change trends in advance and reduce the risk of cargo damage due to abnormal temperature.

[0040] Step S3: Based on the predicted indoor temperature, a fuzzy control algorithm is used to output a control signal for the next moment.

[0041] Among them, the fuzzy control algorithm mainly includes several steps: (1) defining input and output variables; (2) fuzzification; (3) establishing a fuzzy rule base; (4) fuzzy reasoning; (5) defuzzification; and (6) outputting control signals.

[0042] The fuzzification step mainly involves calculating membership, usually using a triangular membership function. Fuzzy reasoning, on the other hand, combines rules and rule weights, aggregating the activation strength of each rule based on membership and fuzzy rules to form a fuzzy output set.

[0043] The rule weight is the weight coefficient of the fuzzy rule, ranging from 0 to 1.

[0044] The process of obtaining the rule weights of the fuzzy control algorithm in this embodiment is as follows: First, a historical data set of each independent storage area is obtained. Each data point in the historical data set includes a historical impact, a difference between the indoor temperature and the target temperature, a door opening or closing state sequence, and a corresponding control rule.

[0045] Among them, historical influence for: ; in, is the indoor temperature at the end of the window at time t1, The first end of the window corresponds to The indoor temperature at the moment, is the window length, The ratio of door open state.

[0046] The continuous ratio of door-open states is the ratio of the number of door-open states within the window length.

[0047] By analyzing the temperature changes at time t1 and the proportion of door-open states at all times within the window, we can determine the degree of influence at time t1. Specifically, a greater rate of temperature change within the window indicates that the isolated storage area is likely affected by the outdoor temperature, and thus, the degree of influence is greater. A greater proportion of door-open states indicates that the number of door-open states within the window is high, possibly indicating that the door was continuously open for a period of time. Therefore, over time, the outdoor temperature has a greater impact on the indoor temperature.

[0048] The control rule is a control strategy executed by the devices in the independent storage area under the conditions of the historical impact degree and the difference between the indoor temperature and the target temperature.

[0049] For example, for environmental regulation when any independent storage area in a warehouse is a cold storage, when performing fuzzy control, the fuzzy rules are as shown in Table 1.

[0050] Table 1

[0051] For example, if the target temperature is 2°C, NL means the temperature is much lower than the target temperature, such as the current temperature is -5°C; NS means the temperature is slightly lower than the target temperature, such as the current temperature is -1°C; ZE means the temperature is close to the target temperature, such as the current temperature is 1°C; PS means the temperature is slightly higher than the target temperature, such as the current temperature is 3°C; PL means the temperature is much higher than the target temperature, such as the current temperature is 5°C.

[0052] It should be noted that different independent storage areas can be configured with fuzzy rule bases according to actual conditions and are not limited to the above fuzzy rule bases. The fuzzy rule bases of the same independent storage area are the same when controlled at different times.

[0053] Secondly, all data points in the historical data set are clustered based on the historical influence and the difference between the indoor temperature and the target temperature to obtain multiple clusters.

[0054] In this embodiment, based on the historical influence and the difference between the indoor temperature and the target temperature, a K-means algorithm is used to cluster all data points in the historical data set to obtain multiple clusters.

[0055] Each cluster represents a group of similar environmental states.

[0056] Then, the cluster to which the predicted indoor temperature and influence degree at the next moment belong is determined, and the rule weights under the corresponding control rules in the cluster are selected as the rule weights at the next moment.

[0057] In this embodiment, the cluster to which the predicted indoor temperature and influence degree at the next moment belong is determined by calculating the Euclidean distance between the influence degree at the next moment and the difference between the predicted indoor temperature and the target temperature and the center of each cluster.

[0058] The smaller the Euclidean distance, the more likely the environment at the next moment belongs to the corresponding cluster. The influence degree represents the change in indoor temperature, which covers the impact of outdoor temperature on indoor temperature and the change in the indoor temperature environment itself.

[0059] The influence degree of the predicted indoor temperature in this embodiment is positively correlated with the rate of change of the indoor temperature and the door opening frequency.

[0060] In one embodiment, the influence for: ; Among them, The end of the window corresponds to The predicted indoor temperature at the moment, The first end of the window corresponds to The indoor temperature at the moment, is the window length, The ratio of door open state.

[0061] The door open state ratio is the ratio of the number of door open states within the window length. The window length can be 5 or 10 time points.

[0062] The door-open or door-closed state at the moment of the predicted indoor temperature can be considered to be the same as the door-open or door-closed state at the previous moment. For example, if the door-open state was at the previous moment, then the door-open state is at the moment of the predicted indoor temperature. This is because the two adjacent moments are close, so whether the predicted indoor temperature is at the door-open or door-closed state, the instantaneous change between the two moments is theoretically slow. Therefore, in this embodiment, the state at the previous moment is used as the state at the next moment.

[0063] It should be noted that since the same cluster contains multiple data points, the control rules corresponding to each data point may be the same or different. If they are different, the cluster includes multiple different control rules; if they are the same, the cluster includes a single control rule. Therefore, after determining the cluster to which the predicted temperature and influence at the next moment belong, if there are multiple different control rules in the cluster, the control rule for the predicted indoor temperature at the moment includes all control rules in the cluster, and the rule weight for each control rule can be the normalized reciprocal of the control time under the corresponding control rule.

[0064] The control time is the time interval from the start of control to the target temperature. The smaller the time interval, the better the control effect of the corresponding control rule, and the greater the rule weight of the control rule.

[0065] The above normalization can adopt the maximum and minimum value normalization method; or the sum of the reciprocals of the control times under all different control rules can be set to 1 to obtain the corresponding rule weight.

[0066] When there is only one control rule in the cluster, there is no need to consider the control effect and the corresponding rule weight is directly set to 1.

[0067] After determining the rule weight corresponding to the predicted indoor temperature, in this embodiment, the difference between the predicted indoor temperature and the target temperature is input into the fuzzy control algorithm to activate the fuzzy rules, and the selected rule weight is applied to perform weighted output on the membership under each control rule to achieve temperature control at the next moment.

[0068] Since the fuzzy control algorithm is an existing technology, it will not be described in detail here.

[0069] The solution of the present invention can be dynamically adjusted according to actual environmental changes to adapt to complex and changing warehouse environments, thereby improving the precise control of independent storage areas where sensitive goods are located.

[0070] The present invention also provides a warehouse environment adjustment system based on the industrial Internet of Things. Figure 2 As shown, the system includes a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, the warehouse environment adjustment method based on the industrial Internet of Things according to the present invention is implemented.

[0071] In the present invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, the computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store the required information and can be accessed by an application, module, or both. Any such computer storage medium can be part of, accessible to, or connectable to a device. Any application or module described in the present invention can be implemented by computer-readable / executable instructions stored or otherwise retained by such a computer-readable medium.

[0072] In the description of this specification, “a plurality of” means at least two, for example, two, three or more, etc., unless otherwise clearly defined.

[0073] Although this specification has shown and described several embodiments of the present invention, it is obvious to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and substitutions without departing from the idea and spirit of the present invention.

Claims

1. A warehouse environment adjustment method based on industrial Internet of Things, characterized in that: include: Obtain the indoor temperature data, outdoor temperature data and door opening or closing status of each independent storage area in the warehouse within the set time sequence; Based on the set time series of indoor temperature data, outdoor temperature data, and state sequence, predict the predicted indoor temperature at the next moment; Based on the predicted indoor temperature, the fuzzy control algorithm is used to output the control signal for the next moment; Among them, the rule weight of the fuzzy control algorithm is: Obtaining a historical data set for each independent storage area, where each data point in the historical data set includes a historical impact and a difference between the indoor temperature and the target temperature; the historical impact represents a change in the indoor temperature at any historical moment; Clustering all data points in the historical data set based on the historical influence and difference to obtain multiple clusters; calculating the rule weight of each data point in each cluster under the corresponding control rule; the rule weight is negatively correlated with the control time under the corresponding control rule; Determine the cluster to which the predicted temperature and influence degree at the next moment belong, and select the rule weights under the corresponding control rules in the cluster as the rule weights at the next moment.

2. The warehouse environment adjustment method based on industrial Internet of Things according to claim 1 is characterized in that: The indoor temperature data is collected by sensors deployed near the door of each independent storage area; the outdoor temperature data is collected by sensors deployed near the door outside each independent storage area.

3. The warehouse environment adjustment method based on industrial Internet of Things according to claim 1 is characterized in that: The impact for: ； in, The end of the window corresponds to The predicted indoor temperature at the moment, The first end of the window corresponds to The indoor temperature at the moment, is the window length, is the door opening state ratio, which is the ratio of the number of door opening states within the window length; The door opening or closing state at a certain moment is the same as the door opening or closing state at the previous moment.

4. The warehouse environment adjustment method based on industrial Internet of Things according to claim 1 is characterized in that: The multiple clusters are obtained by using a K-mean clustering algorithm.

5. The warehouse environment adjustment method based on industrial Internet of Things according to claim 1 is characterized in that: The rule weight is the normalized value of the inverse of the control time under the corresponding control rule; the control time is the time interval from the start of control to reaching the target temperature.

6. The warehouse environment adjustment method based on industrial Internet of Things according to claim 1 is characterized in that: The outputting of the control signal at the next moment includes: The difference between the predicted indoor temperature and the target temperature is input into the fuzzy control algorithm, and the control signal at the next moment is output.

7. The warehouse environment adjustment method based on industrial Internet of Things according to claim 1 is characterized in that: The predicted indoor temperature is obtained using a recurrent neural network model.

8. The warehouse environment adjustment method based on industrial Internet of Things according to claim 7 is characterized in that: The training process of the recurrent neural network model is: Obtain a sample data set; wherein the sample data set includes historical indoor temperature data, historical outdoor temperature data, and historical state sequences within a plurality of historical set time series; The sample data set is input into the recurrent neural network model for training, and the cross entropy loss is used to calculate the difference. The parameters of the recurrent neural network model are adjusted using the gradient descent algorithm to adjust the parameters of the network model until the cross entropy loss value between the output prediction result and the true result is less than the difference threshold, thereby obtaining a trained temperature prediction model.

9. The warehouse environment adjustment method based on industrial Internet of Things according to claim 1, characterized in that: The method further includes: performing time alignment processing on the collected indoor temperature data and outdoor temperature data to obtain indoor temperature data after a delay time corresponding to the indoor temperature data; the delay time used in the time alignment processing is the time difference between the two parameters corresponding to the maximum mutual correlation coefficient.

10. The warehouse environment adjustment system based on industrial Internet of Things is characterized by: include: processor; A memory storing computer instructions for adjusting the warehouse environment based on the industrial Internet of Things. When the computer instructions are executed by the processor, the system executes the warehouse environment adjustment method based on the industrial Internet of Things according to any one of claims 1 to 9.

Images