An energy-saving power supply method and system for energy consumption analysis in a workshop

Abstract

The present application relates to an energy-saving power supply method and system for workshop energy consumption analysis, wherein the method comprises: collecting a material image sequence and a sensor parameter sequence within a preset time window in the past in the workshop, and predicting the amount of material to be transmitted at a preset time in the future; then performing a prediction error analysis based on the sensor parameter sequence to obtain an adjusted prediction error coefficient; using the adjusted prediction error coefficient, fusing and error-compensating the predicted material quantity to obtain a compensated material quantity; optimizing the equipment operating power based on the compensated material quantity to obtain the optimal operating power and implement energy-saving power supply. The present application solves the technical problem in the prior art that plastic crushing equipment still operates at high power in the case of discontinuous material supply, resulting in power waste, and achieves the technical effect of dynamically optimizing the equipment operating power by predicting the material quantity and performing error compensation, reducing power loss while ensuring the quality of crushing processing.

Description

Technical Field

[0001] The present invention relates to the field of energy-saving power supply, and in particular to an energy-saving power supply method and system for workshop energy consumption analysis. Background Art

[0002] Plastic crushing equipment is widely used in industrial production and resource recycling to crush waste plastic products into small particles for reuse. In the prior art, plastic crushing equipment usually adopts a constant power operation mode, that is, regardless of the amount of material on the conveyor belt, the equipment continues to operate at a preset high power. However, the material supply in the workshop is usually discontinuous, with the amount of material varying from time to time, and there may even be no material at all for a short period of time. However, the existing crushing equipment control lacks the ability to perceive and respond to changes in material quantity, resulting in the equipment maintaining a high power operation state when the amount of material is small or there is no material, resulting in a large amount of waste of electricity resources. Summary of the Invention

[0003] The present invention aims to solve the technical problem in the prior art that plastic crushing equipment still runs at high power when the material supply is discontinuous, resulting in power waste, and provides an energy-saving power supply method and system for workshop energy consumption analysis to solve the problem.

[0004] The technical solution of the present invention to solve the above technical problems is as follows:

[0005] In a first aspect, the present invention provides an energy-saving power supply method for workshop energy consumption analysis, comprising: in the workshop, collecting a material image sequence and a sensor parameter sequence within a preset time window in the past, predicting the transmission material quantity at a preset moment in the future, and obtaining a first predicted material quantity and a second predicted material quantity; performing a prediction error analysis based on the sensor parameter sequence to obtain a prediction error coefficient, and adjusting the prediction error coefficient based on the first predicted material quantity and the second predicted material quantity to obtain an adjusted prediction error coefficient; using the adjusted prediction error coefficient, performing fusion processing and error compensation on the first predicted material quantity and the second predicted material quantity to obtain a compensated material quantity; optimizing the equipment operation energy consumption power based on the compensated material quantity to obtain the optimal operating power and perform energy-saving power supply.

[0006] Optionally, in the workshop, by collecting material image sequences and sensor parameter sequences within a past preset time window, including: in the workshop, by using image monitoring equipment and weight sensors installed on the equipment, collecting material images and sensor parameters within a past preset time window, wherein the equipment is a plastic crushing equipment; according to the timestamps of the collected material images and sensor parameters, integrating them in chronological order to obtain material image sequences and sensor parameter sequences.

[0007] Optionally, a prediction of the amount of material to be transmitted at a preset moment in the future is performed to obtain a first predicted material amount and a second predicted material amount, including: training a first material amount prediction branch and a second material amount prediction branch to obtain a material amount predictor; inputting the material image sequence and the sensor parameter sequence into the first material amount prediction branch and the second material amount prediction branch in the material amount predictor respectively, predicting the amount of material to be transmitted at a preset moment in the future, and outputting the first predicted material amount and the second predicted material amount.

[0008] Optionally, training the first material quantity prediction branch and the second material quantity prediction branch to obtain a material quantity predictor includes: collecting a sample material image sequence set and a sample sensor parameter sequence set based on data records of material transmission processing performed by the equipment within a historical period, and collecting the transmission material quantity at a preset moment after each sample material image sequence and sample sensor parameter sequence, and marking them as a sample material quantity set; using a convolutional neural network and a feedforward neural network to construct a first material quantity prediction branch and a second material quantity prediction branch; using the sample material image sequence set and the sample sensor parameter sequence set as input features, and using the sample material quantity set as output, respectively, to train and test the first material quantity prediction branch and the second material quantity prediction branch, respectively, until they are qualified; combining the first material quantity prediction branch and the second material quantity prediction branch to obtain a material quantity predictor.

[0009] Optionally, a prediction error analysis is performed based on the sensing parameter sequence to obtain a prediction error coefficient, including: obtaining multiple sensing parameters in the sensing parameter sequence, converting and processing to obtain multiple historical material quantities; using the multiple historical material quantities to search within a set of historical material quantities recorded during equipment operation to obtain multiple material quantity occurrence rates; obtaining an average material quantity occurrence rate of different material quantities in the historical material quantity set, calculating the average of the ratios of the average material quantity occurrence rate to the multiple material quantity occurrence rates, and obtaining an occurrence rate coefficient; obtaining a basic prediction error coefficient for transmission material quantity prediction, and using the occurrence rate coefficient to perform correction calculation on the basic prediction error coefficient to obtain a prediction error coefficient.

[0010] Optionally, the prediction error coefficient is adjusted according to the first predicted material quantity and the second predicted material quantity to obtain an adjusted prediction error coefficient, including: calculating the average according to the first predicted material quantity and the second predicted material quantity to obtain the predicted material quantity; obtaining the historical average material quantity during the operation of the equipment; calculating the ratio of the predicted material quantity to the historical average material quantity, and adjusting the prediction error coefficient to obtain the adjusted prediction error coefficient.

[0011] Optionally, the adjusted prediction error coefficient is used to fuse the first predicted material quantity and the second predicted material quantity and perform error compensation to obtain the compensated material quantity, including: calculating the mean fusion processing based on the first predicted material quantity and the second predicted material quantity to obtain the predicted material quantity; using the sum of the adjusted prediction error coefficient and 1 as the error compensation coefficient, performing error compensation calculation on the predicted material quantity, and obtaining the compensated material quantity.

[0012] Optionally, according to the amount of compensation material, the equipment operation energy consumption power is optimized to obtain the optimal operating power and perform energy-saving power supply, including: randomly configuring the operating power of the equipment; obtaining the current real-time power of the equipment, combining the operating power processing to obtain the switching energy consumption parameter, and analyzing and obtaining the crushing particle size qualification according to the operating power and the amount of compensation material; and calculating the energy-saving adaptability according to the switching energy consumption parameter, the operating power and the crushing particle size qualification, as shown in the following formula: ; Among them, SAf is energy-saving adaptability, 、 and is the weight, 、 and The sum of is 1, To preset the switching energy consumption parameters, is the switching energy consumption parameter, P is the operating power, is the preset operating power of the equipment, K is the qualified degree of crushed particle size, The qualified degree of crushed particle size is preset; the operating power of the equipment is continuously randomly configured, and iterative optimization is performed until convergence, and the operating power with the largest energy-saving adaptability is retained as the optimal operating power for energy-saving power supply.

[0013] Optionally, the current real-time power of the equipment is obtained, and the switching energy consumption parameters are obtained in combination with the operating power processing, and the crushing particle size qualification is obtained by analysis based on the operating power and the compensation material quantity, including: according to the power adjustment data record of the equipment in the historical time, according to the real-time power and operating power, the switching energy consumption parameters are retrieved; according to the material crushing data record of the equipment, a sample material quantity set and a sample operating power set are collected, and the proportion of the particle size that meets the requirements after material crushing under different sample material quantities and sample operating powers is collected, marked as the sample crushing particle size qualification, and the sample crushing particle size qualification set is obtained; the sample material quantity set, sample operating power set and sample crushing particle size qualification set are used as training data and test data, and a crushing predictor is trained based on a feedforward neural network; the compensation material quantity and operating power are input into the crushing predictor, and the crushing particle size qualification is obtained as output.

[0014] In the second aspect, the present invention provides an energy-saving power supply system for workshop energy consumption analysis, including: a material quantity prediction module, which is used to collect material image sequences and sensor parameter sequences within a preset time window in the past in the workshop, and predict the transmission material quantity at a preset moment in the future to obtain a first predicted material quantity and a second predicted material quantity; an error coefficient acquisition module, which is used to perform prediction error analysis based on the sensor parameter sequence to obtain a prediction error coefficient, and adjust the prediction error coefficient based on the first predicted material quantity and the second predicted material quantity to obtain an adjusted prediction error coefficient; a compensation material quantity module, which is used to use the adjusted prediction error coefficient to perform fusion processing and error compensation on the first predicted material quantity and the second predicted material quantity to obtain a compensated material quantity; a power optimization module, which is used to optimize the equipment operation energy consumption power based on the compensated material quantity, obtain the optimal operating power, and perform energy-saving power supply.

[0015] The beneficial effects of the present invention are:

[0016] By collecting sequences of material images and sensor parameter sequences within a preset time window in the past within the workshop, the material quantity to be transferred at a preset future moment is predicted, obtaining a first predicted material quantity and a second predicted material quantity, providing data support for subsequent power optimization. Prediction error analysis is performed based on the sensor parameter sequence to obtain a prediction error coefficient. This prediction error coefficient is then adjusted based on the first and second predicted material quantities to obtain an adjusted prediction error coefficient, ensuring the processing quality of critical materials. Using the adjusted prediction error coefficient, the first and second predicted material quantities are fused and error-compensated to obtain a compensated material quantity, ensuring sufficient processing capacity even when the material quantity forecast may be low. Equipment operating power consumption is optimized based on the compensated material quantity to obtain optimal operating power, enabling energy-saving power supply and dynamic energy conservation.

[0017] Through the above technical solution, the present application realizes accurate prediction of the material quantity of the crushing equipment, error compensation and optimized control of the operating power, effectively solving the problem of power waste of the crushing equipment under the condition of discontinuous material supply, while ensuring the quality of crushing processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic flow chart of an energy-saving power supply method for workshop energy consumption analysis provided by the present invention;

[0019] Figure 2 This is a structural schematic diagram of an energy-saving power supply system for workshop energy consumption analysis provided by the present invention.

[0020] In the accompanying drawings, the components represented by the reference numerals are as follows:

[0021] Material quantity prediction module 11, error coefficient acquisition module 12, material quantity compensation module 13, power optimization module 14. DETAILED DESCRIPTION

[0022] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0023] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of the specified features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0024] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or illustration". Any embodiment of the present invention described as "for example" is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any person skilled in the art to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed herein.

[0025] Example 1, as Figure 1 As shown, an embodiment of the present invention provides an energy-saving power supply method for workshop energy consumption analysis, including:

[0026] S100: In a workshop, a material image sequence and a sensor parameter sequence within a preset time window in the past are collected, and a transmission material quantity at a preset time in the future is predicted to obtain a first predicted material quantity and a second predicted material quantity.

[0027] Specifically, within the operating environment of a plastic crushing plant, the crushing equipment's material conveyor belt is monitored in real time using image monitoring devices and weight sensors installed on the crushing equipment. For example, image data of the material on the conveyor belt and the corresponding weight sensor parameter data are continuously collected within a preset time window (e.g., 10 seconds). These data are then integrated into a material image sequence and sensor parameter sequence in timestamp order.

[0028] After acquisition, the material image sequence and sensor parameter sequence are input into a pre-trained material quantity prediction engine. This engine comprises a first and second material quantity prediction branches. The first branch analyzes the material image sequence, while the second branch analyzes the sensor parameter sequence. The engine processes these two inputs to predict the material quantity on the conveyor belt at a predetermined future time (e.g., two seconds later), outputting a first and second predicted material quantity, respectively.

[0029] Through the dual-path prediction mechanism, visual information and sensor information can be comprehensively utilized to provide basic data support for subsequent error analysis and operating power optimization, thereby achieving precise control of crushing equipment and energy consumption optimization.

[0030] S200: performing prediction error analysis according to the sensing parameter sequence to obtain a prediction error coefficient, and adjusting the prediction error coefficient according to the first predicted material quantity and the second predicted material quantity to obtain an adjusted prediction error coefficient.

[0031] Specifically, the sensor parameters in the sensor parameter sequence are first converted into corresponding historical material quantities. A search and match is then performed within a pre-stored set of historical material quantities from equipment operation, and the occurrence rate index for each material quantity is calculated. The occurrence rate coefficient is calculated by comparing the ratio of the current material quantity occurrence rate to the historical average material quantity occurrence rate. This coefficient reflects the representativeness of the current material state in the historical data—a lower occurrence rate indicates a rarer state and a greater potential for prediction error. The occurrence rate coefficient is then corrected and calculated with the basic prediction error coefficient to obtain a preliminary prediction error coefficient.

[0032] Next, the average of the first and second predicted material quantities is calculated to obtain the predicted material quantity, which is then compared with the historical average material quantity. A larger material quantity indicates a more critical crushing task and requires greater processing redundancy to ensure crushing quality. Therefore, based on the ratio of the predicted material quantity to the historical average material quantity, the prediction error coefficient is further adjusted to obtain the adjusted prediction error coefficient.

[0033] Through a dual adjustment mechanism based on historical data analysis and current material properties, prediction errors can be evaluated more accurately, providing reliable parameter support for subsequent error compensation and power optimization.

[0034] S300: Using the adjusted prediction error coefficient, the first predicted material quantity and the second predicted material quantity are fused and error compensated to obtain a compensated material quantity.

[0035] Specifically, after obtaining the adjusted prediction error coefficient, the first predicted material quantity and the second predicted material quantity are fused and error compensated to obtain an accurate compensated material quantity.

[0036] First, the arithmetic mean of the first and second predicted material quantities is calculated to obtain a fused predicted material quantity, thereby comprehensively leveraging the advantages of image analysis and sensor data to reduce the deviations that may be caused by a single data source. Subsequently, an error compensation coefficient is constructed based on the obtained adjusted prediction error coefficient. Specifically, the adjusted prediction error coefficient is added to 1 to form the final error compensation coefficient. This error compensation coefficient reflects the processing redundancy that needs to be additionally considered under the current material conditions to ensure that the crushing quality is not affected by the material prediction error. After that, the error compensation coefficient is multiplied by the fused predicted material quantity to calculate the compensated material quantity. The compensated material quantity fully considers the prediction error and the importance of material processing, providing a more reliable decision-making basis for subsequent equipment power optimization.

[0037] Error compensation can effectively address uncertainties in material quantity prediction and improve stability and reliability.

[0038] S400: Optimizing the energy consumption of equipment operation according to the amount of compensation material, obtaining optimal operating power, and performing energy-saving power supply.

[0039] Specifically, after obtaining the amount of compensation material, the optimal equipment operating power is determined based on the amount of compensation material, achieving energy-saving power supply while ensuring the quality of crushing processing.

[0040] First, a series of possible operating power values are randomly configured as the initial operating power scheme. For each configured operating power scheme, the current real-time power data of the equipment is obtained, and the energy consumption parameters required for power switching are calculated. At the same time, the pre-trained crushing predictor is used to predict the crushing particle size qualification when processing the compensation material amount at this operating power, reflecting the quality level of the crushing process. Subsequently, an energy-saving adaptability evaluation index is constructed, which comprehensively considers the influence of three factors: switching energy consumption parameters, operating power, and crushing particle size qualification, and comprehensively evaluates the performance of each power configuration scheme. This index balances the relationship between energy consumption and crushing quality through weighted calculation. Afterwards, through iterative optimization, new operating power schemes are continuously tried and their energy-saving adaptability is calculated until the algorithm converges. Finally, the operating power scheme with the highest energy-saving adaptability is selected as the optimal operating power, and the energy-saving power supply control of the crushing equipment is carried out accordingly.

[0041] Through the optimization mechanism, the energy consumption of the equipment can be minimized while ensuring the quality of crushing processing, the goal of intelligent energy saving can be achieved, and the waste of electricity caused by the equipment still running at high power when the amount of material is small or there is no material can be avoided.

[0042] Furthermore, in the workshop, by collecting the material image sequence and sensor parameter sequence within the past preset time window, including:

[0043] S110: In a workshop, an image monitoring device and a weight sensor installed on equipment are used to collect images and sensor parameters of materials within a preset time window, wherein the equipment is a plastic crushing equipment;

[0044] S120: Integrate the collected material images and sensor parameters according to the timestamps in chronological order to obtain a material image sequence and a sensor parameter sequence.

[0045] In one alternative embodiment, in a plastic crushing plant, image monitoring equipment and weight sensors deployed on the crushing equipment collect real-time images and sensor parameters of the material within a preset time window (e.g., 10 seconds). The image monitoring equipment can be a high-speed industrial camera that can clearly capture the shape, density, and distribution of the material on the conveyor belt. The weight sensor measures the weight changes of the material on the conveyor belt in real time, providing a direct numerical indicator of the material quantity.

[0046] The acquired discrete material images and sensor parameters are then subjected to time-series integration. Specifically, the timestamp information for each frame of the material image and each set of sensor parameters is read. These data are then arranged and integrated in a strict chronological order based on the timestamps, forming a material image sequence and a sensor parameter sequence. This timestamp-based sequence integration method ensures consistency and synchronization between the two heterogeneous data sets in the temporal dimension, providing high-quality input data for subsequent material quantity prediction.

[0047] Through the above steps, key monitoring data during the operation of the plastic crushing equipment can be efficiently obtained, and a comprehensive perception capability of the material transmission status can be established, laying a data foundation for accurate material quantity prediction and energy-saving control.

[0048] Furthermore, the transmission material quantity at a preset time in the future is predicted to obtain a first predicted material quantity and a second predicted material quantity, including:

[0049] S130: training the first material quantity prediction branch and the second material quantity prediction branch to obtain a material quantity predictor;

[0050] S140: Input the material image sequence and the sensor parameter sequence into the first material quantity prediction branch and the second material quantity prediction branch of the material quantity predictor respectively, predict the transmission material quantity at a preset time in the future, and output a first predicted material quantity and a second predicted material quantity.

[0051] In a preferred embodiment, a material quantity predictor is first constructed through offline training. This predictor consists of two parallel prediction branches: the first branch specifically processes image information, extracting implicit information related to the material's transport capacity from its visual features; the second branch specifically processes sensor parameter information, inferring transport capacity trends from measured data of the material's physical properties. This dual-branch design leverages the complementary advantages of multimodal data to improve prediction robustness and accuracy. Then, the real-time collected and integrated material image sequence and sensor parameter sequence are input into the corresponding prediction branches of the pre-trained predictor. Specifically, the material image sequence is input into the first branch, where it undergoes feature extraction and time series modeling to output a first predicted material quantity. Simultaneously, the sensor parameter sequence is input into the second branch, where it undergoes parameter analysis and pattern recognition to output a second predicted material quantity. These two prediction branches operate independently, predicting the transported material quantity at a preset future time (e.g., two seconds later) based on different data modalities.

[0052] Through the dual-channel parallel prediction mechanism, it is possible to maintain a high level of prediction accuracy in actual production environments where material transmission conditions are complex and changeable, providing a reliable decision-making basis for subsequent error analysis and power optimization.

[0053] Furthermore, the first material quantity prediction branch and the second material quantity prediction branch are trained to obtain a material quantity predictor, including:

[0054] S131: Based on the data records of material transfer processing performed by the equipment in the historical time, a sample material image sequence set and a sample sensor parameter sequence set are collected, and the amount of transferred material at a preset time after each sample material image sequence and sample sensor parameter sequence is collected and marked as a sample material amount set;

[0055] S132: Using a convolutional neural network and a feedforward neural network, construct a first material quantity prediction branch and a second material quantity prediction branch;

[0056] S133: Using the sample material image sequence set and the sample sensor parameter sequence set as input features and the sample material quantity set as output, respectively, the first material quantity prediction branch and the second material quantity prediction branch are trained and tested until they pass the test;

[0057] S134: Combine the first material quantity prediction branch and the second material quantity prediction branch to obtain a material quantity predictor.

[0058] In a preferred embodiment, to train a material quantity predictor, training data is first collected from the historical operation records of the plastic crushing equipment. Specifically, a set of sample material image sequences and a set of sample sensor parameter sequences are extracted from a database of material transfer processes performed by the equipment over a historical period of time, serving as input features for the material quantity predictor. Furthermore, for each set of input data, the actual material quantity transferred after a preset time (e.g., 2 seconds) is collected and labeled as a sample material quantity set, which serves as the output label for the material quantity predictor. This historical data-based labeling method ensures the authenticity and representativeness of the training data. Then, prediction branches are constructed using corresponding deep learning architectures based on the characteristics of different data modalities. Specifically, the first material quantity prediction branch, which processes the material image sequence, uses a convolutional neural network as its underlying architecture, which effectively extracts spatial features and material morphological information from the image. The second material quantity prediction branch, which processes the sensor parameter sequence, uses a feedforward neural network as its underlying architecture, which efficiently analyzes complex relationships between numerical parameters. Through targeted network design, the information value of different data modalities can be maximized.

[0059] Next, the two prediction branches are independently trained and performance tested. Specifically, the first material quantity prediction branch uses a sample material image sequence set as input and a sample material quantity set as output labels for training; the second material quantity prediction branch uses a sample sensor parameter sequence set as input and is also trained with a sample material quantity set as output labels. During the training process, the backpropagation algorithm is used to optimize the network parameters, and the performance of the material quantity predictor is evaluated through cross-validation. Iterations are repeated until the performance of the material quantity predictor meets the preset qualification standards. Afterwards, the trained first and second material quantity prediction branches are combined and integrated to form a complete material quantity predictor. This material quantity predictor can process two data modalities in parallel, outputting prediction results based on different information sources, respectively, providing a multi-dimensional decision-making basis for subsequent fusion processing and error compensation.

[0060] Through the above steps, a deep learning-based dual-channel material quantity predictor was constructed, which can fully tap the information value of multimodal data, significantly improve the accuracy and reliability of predictions, and lay the foundation for achieving precise energy-saving control.

[0061] Furthermore, a prediction error analysis is performed based on the sensing parameter sequence to obtain a prediction error coefficient, including:

[0062] S210: Acquire multiple sensing parameters in the sensing parameter sequence, and convert and process them to obtain multiple historical material quantities;

[0063] S220: Using the multiple historical material quantities, searching within a set of historical material quantities recorded during equipment operation to obtain multiple material quantity appearance rates;

[0064] S230: Obtaining an average material quantity appearance rate of different material quantities in the historical material quantity set, calculating an average of the ratios of the average material quantity appearance rate to the multiple material quantity appearance rates, and obtaining an appearance rate coefficient;

[0065] S240: Obtain a basic prediction error coefficient for the transmission material quantity prediction, and use the occurrence rate coefficient to perform correction calculation on the basic prediction error coefficient to obtain a prediction error coefficient.

[0066] In a preferred embodiment, multiple sensor parameter data points are first acquired from a real-time sensor parameter sequence and converted into corresponding historical material quantities. For example, based on the sensor calibration curve, the material weight on the conveyor belt at different times is accurately reflected, providing a precise numerical representation of the material transport status. The converted multiple historical material quantity values are then searched and matched against a pre-stored set of historical material quantities for the device. This search operation determines the frequency of occurrence of these material quantity values in the historical operating data, and calculates an occurrence rate index for each material quantity value, resulting in multiple material quantity occurrence rates. These material quantity occurrence rates directly reflect the typicality and representativeness of the current material state in historical operation.

[0067] Next, the average material quantity occurrence rate for all different material quantity values in the historical material quantity set is obtained as a baseline reference value. Next, the ratio of this average material quantity occurrence rate to the multiple obtained material quantity occurrence rates is calculated, and the arithmetic mean of these ratios is taken to obtain the occurrence rate coefficient. This coefficient measures the specificity of the current material state in the statistical distribution. Next, a pre-set basic prediction error coefficient for the transmitted material quantity forecast is obtained, which represents the inherent error level of the forecast under standard conditions. The obtained occurrence rate coefficient is then used to correct the basic prediction error coefficient to generate a prediction error coefficient that is tailored to the characteristics of the current material state. Specifically, when the occurrence rate coefficient is small, the prediction error coefficient is increased accordingly to compensate for the prediction uncertainty caused by rare material states; otherwise, the prediction error coefficient is decreased.

[0068] Through the above steps, an adaptive error assessment mechanism based on historical data analysis was established, which can dynamically adjust the prediction error assessment standard according to the statistical characteristics of the current material state, improve the pertinence and effectiveness of error compensation, and provide a more reliable decision-making basis for subsequent energy-saving control.

[0069] Furthermore, adjusting the prediction error coefficient according to the first predicted material quantity and the second predicted material quantity to obtain an adjusted prediction error coefficient includes:

[0070] S250: Calculating an average based on the first predicted material quantity and the second predicted material quantity to obtain a predicted material quantity;

[0071] S260: Obtain the historical average material quantity during equipment operation;

[0072] S270: Calculate the ratio of the predicted material quantity to the historical average material quantity, and perform adjustment calculation on the prediction error coefficient to obtain an adjusted prediction error coefficient.

[0073] In one feasible implementation, the first and second predicted material quantities, respectively output by the first and second material quantity prediction branches, are first merged. Specifically, the arithmetic mean of these two prediction results is calculated to obtain a comprehensive predicted material quantity, thereby balancing the strengths and weaknesses of the two prediction branches and improving the overall reliability of the prediction results. Next, historical material quantity data recorded during the long-term operation of the equipment is obtained from a historical database, and the average of this historical data is calculated to obtain a historical average material quantity. This historical average material quantity reflects the material loading level of the plastic crushing equipment under normal operating conditions and serves as a benchmark for assessing the relative importance of the current material status.

[0074] The ratio of the predicted material volume to the historical average volume is then calculated, and the prediction error coefficient is adjusted based on this ratio to obtain the adjusted prediction error coefficient. For example, when the predicted material volume is greater than the historical average volume, it indicates that the current crushing task is heavy and the processing is of high importance. Therefore, the prediction error coefficient is increased accordingly to provide greater processing redundancy and ensure crushing quality. Conversely, when the predicted material volume is less than the historical average volume, the prediction error coefficient is decreased accordingly to optimize energy efficiency.

[0075] Through the adaptive error adjustment mechanism based on the material load level, the prediction error evaluation standard can be dynamically adjusted according to the relative importance of the current material processing task, the relationship between processing quality and energy efficiency can be balanced, and more scientific parameter support can be provided for achieving precise energy-saving control.

[0076] Furthermore, the first predicted material quantity and the second predicted material quantity are fused and error compensated by using the adjusted prediction error coefficient to obtain the compensated material quantity, including:

[0077] S310: Calculating the mean fusion process based on the first predicted material quantity and the second predicted material quantity to obtain the predicted material quantity;

[0078] S320: Using the sum of the adjusted prediction error coefficient and 1 as the error compensation coefficient, performing error compensation calculation on the predicted material quantity to obtain the compensated material quantity.

[0079] In an optional embodiment, first, the first predicted material quantity and the second predicted material quantity output by two different prediction branches are fused. Specifically, the arithmetic mean method is used to calculate the mean of the two predicted values to obtain the fused predicted material quantity. This mean fusion strategy can comprehensively utilize the two prediction results based on image analysis and sensor parameter analysis, reduce the random errors and systematic errors that may be caused by a single prediction method, and improve the stability and reliability of the prediction results. Then, an error compensation mechanism is constructed based on the obtained adjusted prediction error coefficient. Specifically, the adjusted prediction error coefficient is added to 1 to form the final error compensation coefficient, thereby ensuring that the error compensation coefficient is always greater than 1, and positive error compensation can be achieved, providing the necessary redundancy for material processing. Subsequently, the error compensation coefficient is multiplied by the obtained predicted material quantity, and the error compensation calculation is performed to obtain the compensated material quantity.

[0080] Through forward error compensation, processing redundancy can be appropriately increased based on the predicted material quantity, effectively addressing the uncertainty caused by forecast errors and material fluctuations. Especially in the case of large material quantities, the corresponding increase in the adjusted forecast error coefficient provides greater processing redundancy, ensuring the processing quality of important crushing tasks. In contrast, in the case of smaller material quantities, the processing redundancy provided is relatively small, which helps improve energy efficiency. Through adaptive error compensation strategies, compensation strategies can be dynamically adjusted based on material state characteristics and processing importance, achieving an optimal balance between ensuring processing quality and energy efficiency, providing a decision-making basis for subsequent equipment power optimization.

[0081] Furthermore, according to the amount of compensation material, the energy consumption of the equipment is optimized to obtain the optimal operating power and perform energy-saving power supply, including:

[0082] S410: Randomly configure the operating power of the device;

[0083] S420: Obtain the current real-time power of the equipment, combine it with the operating power to obtain the switching energy consumption parameter, and analyze and obtain the crushing particle size qualification according to the operating power and the amount of compensation material;

[0084] S430: Calculate the energy-saving adaptability according to the switching energy consumption parameters, operating power and crushed particle size qualification, as shown in the following formula:

[0085] ;

[0086] Among them, SAf is the energy-saving adaptability, 、 and is the weight, 、 and The sum of is 1, To preset the switching energy consumption parameters, is the switching energy consumption parameter, P is the operating power, is the preset operating power of the equipment, K is the qualified degree of crushed particle size, The qualified degree of crushed particle size is preset;

[0087] S440: Continue to randomly configure the operating power of the device and perform iterative optimization until convergence. The operating power with the highest energy-saving adaptability is retained as the optimal operating power for energy-saving power supply.

[0088] In a preferred embodiment, first, a random search strategy is used to initialize the operating power configuration of the device. Specifically, a set of operating powers is randomly generated within the power range allowed by the device as the initial solution space of the optimization algorithm. This random initialization method helps to avoid the optimization process from falling into a local optimal solution and improves the global search capability. Then, the current real-time power of the device is obtained, and compared and analyzed with the randomly configured operating power to calculate the energy consumption parameters required for power switching. This parameter reflects the additional energy required to switch the device from the current power state to the candidate operating power state. At the same time, the randomly configured operating power and the obtained compensation material amount are used as input parameters, and analyzed by a pre-trained crushing predictor to predict the crushing particle size qualification when processing the compensation material amount at the operating power. The crushing particle size qualification reflects the crushing processing quality of the equipment under the current parameter configuration.

[0089] Then, based on the switching energy consumption parameters, operating power and crushing particle size qualification, the energy-saving adaptability is calculated to comprehensively evaluate the performance of the randomly configured operating power. The specific formula for calculating the energy-saving adaptability is as follows:

[0090] ;

[0091] Among them, energy-saving adaptability It is obtained by weighted calculation of three key factors, namely, the switching energy consumption parameter ratio , operating power ratio and crushed particle size qualification ratio .in, 、 and are the corresponding weight coefficients and their sum is 1, To preset the switching energy consumption parameters, is the switching energy consumption parameter, P is the operating power, is the preset operating power of the equipment, K is the qualified degree of crushed particle size, This multi-objective weighted evaluation method can simultaneously consider energy efficiency and processing quality, achieving a balanced optimization between the two.

[0092] Afterwards, an iterative optimization strategy is employed to continuously refine the power configuration. Specifically, a new operating power configuration is randomly generated, and the evaluation process in steps S420 and S430 is repeated. The energy-saving fitness of the newly configured operating power is calculated and compared with the currently optimal operating power configuration. If the energy-saving fitness of the newly configured operating power is higher, the optimal operating power is updated. This iterative process continues until the algorithm converges and reaches the stopping condition. Ultimately, the operating power configuration with the highest energy-saving fitness is selected as the optimal operating power, and energy-saving power supply control for the crushing equipment is implemented accordingly.

[0093] Through a random search optimization mechanism, equipment energy consumption can be minimized while ensuring crushing quality, achieving intelligent energy conservation. This approach is particularly effective in plastic crushing applications where material quantities fluctuate significantly. This method dynamically adjusts equipment power based on real-time material conditions, avoiding the waste of electricity caused by high-power operation when material quantities are low or absent, significantly improving energy efficiency.

[0094] Furthermore, the current real-time power of the equipment is obtained, and the switching energy consumption parameters are obtained by combining the operating power processing. According to the operating power and the amount of compensation material, the crushing particle size qualification is analyzed and obtained, including:

[0095] S421: Retrieving switching energy consumption parameters based on the power adjustment data records of the device in the historical period and the real-time power and the operating power;

[0096] S422: Based on the material crushing data records of the equipment, a sample material amount set and a sample operating power set are collected, and the proportion of the material particle size meeting the requirements under different sample material amounts and sample operating powers is collected and marked as the sample crushing particle size qualification, thereby obtaining a sample crushing particle size qualification set;

[0097] S423: using the sample material amount set, the sample operating power set, and the sample crushing particle size qualification set as training data and test data, and training a crushing predictor based on a feedforward neural network;

[0098] S424: Input the compensation material amount and operating power into the crushing predictor, and output the obtained crushing particle size qualification.

[0099] In a preferred embodiment, first, the energy consumption characteristics of power switching are analyzed based on the historical operating data of the equipment. Specifically, from the power adjustment data records within the historical time, based on the current real-time power and the candidate operating power as retrieval conditions, the energy consumption data under similar power switching scenarios are queried to obtain the corresponding switching energy consumption parameters. The retrieval method based on historical records can accurately reflect the actual energy consumption level of the equipment in different power switching states, and provide a reliable parameter basis for energy-saving adaptability calculation. Then, a training data set for the crushing predictor is constructed. Specifically, from the material crushing data records of the equipment, a sample material quantity set and a sample operating power set are collected as input features, and the proportion of the particle size of the material after crushing that meets the specification requirements under different combinations of sample material quantity and sample operating power is collected, marked as the sample crushing particle size qualification, and a sample crushing particle size qualification set is formed as the output label. This multidimensional data collection method comprehensively covers the joint impact of material quantity and operating power on crushing quality, and provides rich learning samples for the training of the crushing predictor.

[0100] Afterwards, a crushing predictor is constructed and trained based on the collected training data. Specifically, the sample material quantity set and the sample operating power set are used as input features, and the sample crushing particle size qualification set is used as the output label. The crushing predictor architecture is constructed based on the feedforward neural network architecture, and the parameter optimization training is performed through supervised learning methods. During the training process, the data set is divided into a training set and a test set to ensure that the model has good generalization ability. The neural network-based prediction method can effectively capture the complex nonlinear relationship between material quantity, operating power and crushing quality, and achieve high-precision crushing effect prediction. Afterwards, the trained crushing predictor is used for real-time prediction. Specifically, the obtained compensation material quantity and the currently evaluated operating power are used as input parameters and input into the crushing predictor. After calculation, the crushing particle size qualification is output. This prediction result directly reflects the expected quality level of plastic crushing processing under the current material state and power configuration.

[0101] Through the crushing quality prediction mechanism based on historical data analysis and deep learning, it is possible to accurately evaluate the crushing processing effect under different material load and power configuration conditions, provide a scientific basis for power optimization decisions, and achieve the best balance between processing quality and energy efficiency.

[0102] Example 2, as Figure 2 As shown, based on the same inventive concept as the energy-saving power supply method for workshop energy consumption analysis provided in Example 1, an embodiment of the present invention further provides an energy-saving power supply system for workshop energy consumption analysis, including:

[0103] The material quantity prediction module 11 is used to collect a sequence of material images and a sequence of sensor parameters within a preset time window in the past in the workshop, and predict the amount of material to be transported at a preset time in the future to obtain a first predicted material quantity and a second predicted material quantity;

[0104] an error coefficient acquisition module 12, configured to perform a prediction error analysis based on the sensing parameter sequence to obtain a prediction error coefficient, and adjust the prediction error coefficient based on the first predicted material quantity and the second predicted material quantity to obtain an adjusted prediction error coefficient;

[0105] A compensation material quantity module 13 is configured to use the adjusted prediction error coefficient to perform fusion processing and error compensation on the first predicted material quantity and the second predicted material quantity to obtain a compensation material quantity;

[0106] The power optimization module 14 is used to optimize the energy consumption of the equipment according to the amount of compensation material, obtain the optimal operating power, and perform energy-saving power supply.

[0107] Furthermore, the execution steps of the material quantity prediction module 11 include:

[0108] In the workshop, the image monitoring device and weight sensor installed on the equipment are used to collect material images and sensor parameters within a preset time window, wherein the equipment is a plastic crushing equipment;

[0109] According to the timestamps of the collected material images and sensor parameters, they are integrated in chronological order to obtain the material image sequence and sensor parameter sequence.

[0110] Furthermore, the execution steps of the material quantity prediction module 11 also include:

[0111] Train the first material quantity prediction branch and the second material quantity prediction branch to obtain a material quantity predictor;

[0112] The material image sequence and the sensor parameter sequence are respectively input into the first material quantity prediction branch and the second material quantity prediction branch in the material quantity predictor to predict the transmission material quantity at a preset time in the future, and the first predicted material quantity and the second predicted material quantity are output.

[0113] Furthermore, the execution steps of the material quantity prediction module 11 also include:

[0114] According to the data records of the material transmission and processing of the equipment in the historical time, a sample material image sequence set and a sample sensor parameter sequence set are collected, and the transmission material quantity at a preset time after each sample material image sequence and sample sensor parameter sequence is collected and marked as a sample material quantity set;

[0115] A convolutional neural network and a feedforward neural network are used to construct a first material quantity prediction branch and a second material quantity prediction branch;

[0116] Using the sample material image sequence set and the sample sensor parameter sequence set as input features, and using the sample material quantity set as output, respectively, the first material quantity prediction branch and the second material quantity prediction branch are trained and tested until they pass;

[0117] The first material quantity prediction branch and the second material quantity prediction branch are combined to obtain a material quantity predictor.

[0118] Furthermore, the error coefficient acquisition module 12 executes the following steps:

[0119] Acquire multiple sensing parameters in the sensing parameter sequence, and convert and process them to obtain multiple historical material quantities;

[0120] Using the multiple historical material quantities, searching within a set of historical material quantities recorded during equipment operation to obtain multiple material quantity occurrence rates;

[0121] Obtaining an average material quantity appearance rate of different material quantities in the historical material quantity set, calculating an average of the ratios of the average material quantity appearance rate to the multiple material quantity appearance rates, and obtaining an appearance rate coefficient;

[0122] A basic prediction error coefficient for the transmission material quantity prediction is obtained, and the occurrence rate coefficient is used to perform correction calculation on the basic prediction error coefficient to obtain a prediction error coefficient.

[0123] Furthermore, the error coefficient acquisition module 12 may further execute the following steps:

[0124] Calculating an average based on the first predicted material quantity and the second predicted material quantity to obtain a predicted material quantity;

[0125] Obtain the historical average material quantity during equipment operation;

[0126] The ratio of the predicted material quantity to the historical average material quantity is calculated, and the prediction error coefficient is adjusted and calculated to obtain an adjusted prediction error coefficient.

[0127] Furthermore, the execution steps of the material quantity compensation module 13 include:

[0128] Calculating the mean fusion process based on the first predicted material quantity and the second predicted material quantity to obtain the predicted material quantity;

[0129] The sum of the adjusted prediction error coefficient and 1 is used as the error compensation coefficient to perform error compensation calculation on the predicted material quantity to obtain the compensated material quantity.

[0130] Furthermore, the power optimization module 14 executes the following steps:

[0131] The operating power of the randomly configured device;

[0132] Obtain the current real-time power of the equipment, combine it with the operating power processing to obtain the switching energy consumption parameters, and analyze and obtain the crushing particle size qualification based on the operating power and the amount of compensation material;

[0133] According to the switching energy consumption parameters, operating power and crushing particle size qualification, the energy saving adaptability is calculated as follows:

[0134] ;

[0135] Among them, SAf is the energy-saving adaptability, 、 and is the weight, 、 and The sum of is 1, To preset the switching energy consumption parameters, is the switching energy consumption parameter, P is the operating power, is the preset operating power of the equipment, K is the qualified degree of crushed particle size, The qualified degree of crushed particle size is preset;

[0136] Continue to randomly configure the operating power of the device and perform iterative optimization until convergence. Keep the operating power with the highest energy-saving adaptability as the optimal operating power for energy-saving power supply.

[0137] Furthermore, the execution steps of the power optimization module 14 also include:

[0138] According to the power adjustment data record of the device in the historical time, according to the real-time power and the operating power, the switching energy consumption parameter is retrieved;

[0139] According to the material crushing data records of the equipment, a sample material amount set and a sample operating power set are collected, and the proportion of the particle size that meets the requirements after material crushing under different sample material amounts and sample operating powers is collected and marked as the sample crushing particle size qualification, thereby obtaining the sample crushing particle size qualification set;

[0140] Using the sample material amount set, the sample operating power set and the sample crushing particle size qualification set as training data and test data, a crushing predictor is trained based on a feedforward neural network;

[0141] The compensation material amount and operating power are input into the crushing predictor, and the crushing particle size qualification is output.

[0142] It should be noted that, in the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0143] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0144] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0145] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0146] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0147] Although preferred embodiments of the present invention have been described, additional changes and modifications to these embodiments may occur to those skilled in the art once the basic inventive concepts become known.

[0148] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the present invention and its equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. An energy-saving power supply method for workshop energy consumption analysis, characterized in that: The method comprises: In the workshop, a material image sequence and a sensor parameter sequence within a preset time window in the past are collected to predict the amount of material to be transported at a preset time in the future, thereby obtaining a first predicted material amount and a second predicted material amount; Performing a prediction error analysis according to the sensing parameter sequence to obtain a prediction error coefficient, and adjusting the prediction error coefficient according to the first predicted material quantity and the second predicted material quantity to obtain an adjusted prediction error coefficient, including: Acquire multiple sensing parameters in the sensing parameter sequence, and convert and process them to obtain multiple historical material quantities; Using the multiple historical material quantities, searching within a set of historical material quantities recorded during equipment operation to obtain multiple material quantity occurrence rates; Obtaining an average material quantity appearance rate of different material quantities in the historical material quantity set, calculating an average of the ratios of the average material quantity appearance rate to the multiple material quantity appearance rates, and obtaining an appearance rate coefficient; Obtaining a basic prediction error coefficient for the transmission material quantity prediction, and using the occurrence rate coefficient to perform correction calculation on the basic prediction error coefficient to obtain a prediction error coefficient; Using the adjusted prediction error coefficient, the first predicted material quantity and the second predicted material quantity are fused and error compensated to obtain a compensated material quantity; According to the amount of compensation material, the energy consumption power of the equipment operation is optimized to obtain the optimal operating power and perform energy-saving power supply.

2. The energy-saving power supply method for workshop energy consumption analysis according to claim 1 is characterized in that: In the workshop, by collecting material image sequences and sensor parameter sequences within the past preset time window, including: In a workshop, an image monitoring device and a weight sensor installed on a device are used to collect images and sensor parameters of materials within a preset time window, wherein the device is a plastic crushing device; According to the timestamps of the collected material images and sensor parameters, they are integrated in chronological order to obtain the material image sequence and sensor parameter sequence.

3. The energy-saving power supply method for workshop energy consumption analysis according to claim 1 is characterized in that: Predicting the amount of material to be transported at a preset future moment to obtain a first predicted amount of material and a second predicted amount of material includes: Train the first material quantity prediction branch and the second material quantity prediction branch to obtain a material quantity predictor; The material image sequence and the sensor parameter sequence are respectively input into the first material quantity prediction branch and the second material quantity prediction branch in the material quantity predictor to predict the transmission material quantity at a preset time in the future, and the first predicted material quantity and the second predicted material quantity are output.

4. The energy-saving power supply method for workshop energy consumption analysis according to claim 3 is characterized in that: Training the first material quantity prediction branch and the second material quantity prediction branch to obtain a material quantity predictor includes: According to the data records of the material transmission and processing of the equipment in the historical time, a sample material image sequence set and a sample sensor parameter sequence set are collected, and the transmission material quantity at a preset time after each sample material image sequence and sample sensor parameter sequence is collected and marked as a sample material quantity set; A convolutional neural network and a feedforward neural network are used to construct a first material quantity prediction branch and a second material quantity prediction branch; Using the sample material image sequence set and the sample sensor parameter sequence set as input features, and using the sample material quantity set as output, respectively, the first material quantity prediction branch and the second material quantity prediction branch are trained and tested until they pass; The first material quantity prediction branch and the second material quantity prediction branch are combined to obtain a material quantity predictor.

5. The energy-saving power supply method for workshop energy consumption analysis according to claim 1, characterized in that: Adjusting the prediction error coefficient according to the first predicted material quantity and the second predicted material quantity to obtain the adjusted prediction error coefficient includes: Calculating an average based on the first predicted material quantity and the second predicted material quantity to obtain a predicted material quantity; Obtain the historical average material quantity during equipment operation; The ratio of the predicted material quantity to the historical average material quantity is calculated, and the prediction error coefficient is adjusted and calculated to obtain an adjusted prediction error coefficient.

6. The energy-saving power supply method for workshop energy consumption analysis according to claim 1, characterized in that: The adjusted prediction error coefficient is used to fuse the first predicted material quantity and the second predicted material quantity and perform error compensation to obtain the compensated material quantity, including: Calculating the mean fusion process based on the first predicted material quantity and the second predicted material quantity to obtain the predicted material quantity; The sum of the adjusted prediction error coefficient and 1 is used as the error compensation coefficient to perform error compensation calculation on the predicted material quantity to obtain the compensated material quantity.

7. The energy-saving power supply method for workshop energy consumption analysis according to claim 1, characterized in that: According to the amount of compensation material, the energy consumption of the equipment is optimized to obtain the optimal operating power and perform energy-saving power supply, including: The operating power of the randomly configured device; Obtain the current real-time power of the equipment, combine it with the operating power processing to obtain the switching energy consumption parameters, and analyze and obtain the crushing particle size qualification based on the operating power and the amount of compensation material; According to the switching energy consumption parameters, operating power and crushing particle size qualification, the energy saving adaptability is calculated as follows: ； Among them, SAf is the energy-saving adaptability, 、 and is the weight, 、 and The sum of is 1, To preset the switching energy consumption parameters, To switch the energy consumption parameters, is the operating power, The preset operating power of the device. The crushing particle size qualification is The qualified degree of crushed particle size is preset; Continue to randomly configure the operating power of the device and perform iterative optimization until convergence. Keep the operating power with the highest energy-saving adaptability as the optimal operating power for energy-saving power supply.

8. The energy-saving power supply method for workshop energy consumption analysis according to claim 7, characterized in that: Obtain the current real-time power of the equipment, combine it with the operating power processing to obtain the switching energy consumption parameters, and analyze and obtain the crushing particle size qualification based on the operating power and the amount of compensation material, including: According to the power adjustment data record of the device in the historical time, according to the real-time power and the operating power, the switching energy consumption parameter is retrieved; According to the material crushing data records of the equipment, a sample material amount set and a sample operating power set are collected, and the proportion of the particle size that meets the requirements after material crushing under different sample material amounts and sample operating powers is collected and marked as the sample crushing particle size qualification, thereby obtaining the sample crushing particle size qualification set; Using the sample material amount set, the sample operating power set and the sample crushing particle size qualification set as training data and test data, a crushing predictor is trained based on a feedforward neural network; The compensation material amount and operating power are input into the crushing predictor, and the crushing particle size qualification is output.

9. An energy-saving power supply system for workshop energy consumption analysis, characterized in that: The energy-saving power supply method for implementing the workshop energy consumption analysis according to any one of claims 1 to 8, the system comprising: The material quantity prediction module is used to collect a sequence of material images and a sequence of sensor parameters within a preset time window in the past in the workshop, and predict the amount of material to be transported at a preset time in the future to obtain a first predicted material quantity and a second predicted material quantity; an error coefficient acquisition module, configured to perform a prediction error analysis based on the sensing parameter sequence to obtain a prediction error coefficient, and adjust the prediction error coefficient based on the first predicted material quantity and the second predicted material quantity to obtain an adjusted prediction error coefficient; a compensation material quantity module, configured to use the adjusted prediction error coefficient to perform fusion processing and error compensation on the first predicted material quantity and the second predicted material quantity to obtain a compensation material quantity; The power optimization module is used to optimize the energy consumption of equipment operation according to the amount of compensation material, obtain the optimal operating power, and perform energy-saving power supply.

Images