Peak reaching path planning method suitable for ship building and repairing enterprises and related equipment

Abstract

The invention provides a peak reaching path planning method suitable for a ship building and repairing enterprise and related equipment, and is applied to the technical field of data processing. According to the method, to-be-assessed enterprise information is processed, and peak-reaching driving parameter information of a to-be-assessed enterprise is generated; processing the regional information of the enterprise to be evaluated, and generating peak reaching measure parameter information of a target region and energy conservation and emission reduction parameter information of the target region; processing the carbon emission information of the to-be-evaluated enterprise in a preset time period, and generating a peak-reaching prediction feature vector; processing a preset peak accounting planning model based on the training sample set with the target feature information to generate a target peak accounting planning model; and processing the peak reaching prediction feature vector, the peak reaching driving parameter information of the to-be-evaluated enterprise, the peak reaching measure parameter information of the target area and the energy conservation and emission reduction parameter information of the target area based on a target peak reaching accounting planning model to generate target peak reaching path planning information.

Description

[0001] The present application claims priority to Chinese patent application CN202510004583.3, filed on January 02, 2025, the contents of the specification, drawings and claims of which are incorporated herein by reference in their entirety for all purposes. TECHNICAL FIELD

[0002] The present application relates to the technical field of data processing, in particular to a peak-reaching path planning method suitable for shipbuilding and repairing enterprises and related equipment. BACKGROUND

[0003] Under the background of global sustainable development today, energy and environmental problems have become increasingly prominent and have become the focus of international attention. Achieving the carbon peak and carbon neutral (double carbon) goals has become a key measure to address climate change and promote green transformation of the economy and society. Among them, the energy-saving and low-carbon strategy, as one of the core paths to achieve the double carbon goal, is self-evident. However, in actual application scenarios, the energy-saving and low-carbon strategy presents a diversified characteristic, covering energy efficiency improvement, clean energy replacement, production process optimization, carbon capture and storage, and many other fields, and each field contains numerous specific technologies and methods.

[0004] There is a problem of information asymmetry of energy-saving technologies and equipment in the shipbuilding and repairing industry. First, enterprises do not know what energy-saving technologies and equipment are available in the industry. Second, even if the technology is installed, it does not match the existing situation of the enterprise and cannot be used. Third, although the enterprise has applied a single energy-saving technology, it lacks a design for systematic application and a long-term carbon reduction route. Various energy-saving technologies such as high-efficiency energy conversion technology, advanced energy storage technology, smart grid technology, and waste heat recovery technology in industrial production are emerging; low-carbon technologies include various clean energy power generation technologies (such as solar, wind, water, and nuclear energy), low-carbon building material technologies, low-carbon transportation technologies, and carbon capture, utilization, and storage (CCUS) technologies. These technologies differ significantly in principle, application scenarios, technology maturity, cost-effectiveness, and other aspects. Against this background, with the determination of the double carbon goal, it is crucial for enterprises that urgently need to plan a scientific carbon reduction path to accurately predict their long-term carbon emission trends and carbon neutralization process, scientifically and effectively develop carbon emission execution routes, peak-reaching time execution routes, and other routes based on the actual situation of the enterprise.

[0005] It should be noted that the information disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, and therefore can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0006] The purpose of the present application is to provide a peak path planning method and related equipment and system suitable for shipbuilding and repairing enterprises, at least to some extent, to overcome the problems existing in the prior art. By obtaining enterprise information to be evaluated, regional information, carbon emission information, a preset model, and a training sample set. Then, the enterprise information is processed to obtain peak driving parameter information, the regional information is processed to obtain peak measures and energy saving and emission reduction parameter information, and the carbon emission information is processed to obtain a prediction subtask. Then, a feature vector is extracted from the prediction subtask, and a sample set with target feature information is obtained by processing the training sample set. Finally, based on the target peak accounting planning model, the peak prediction feature vector, the enterprise and regional related parameters are integrated to generate an energy and industrial structure adjustment path, and then the preset peak path, the peak time, the peak value, and the target peak path planning information are obtained.

[0007] Other characteristics and advantages of the present application will become apparent from the following detailed description, or will be learned by practice of the application.

[0008] According to one aspect of the present application, a peak path planning method suitable for shipbuilding and repairing enterprises is provided, comprising: obtaining enterprise information to be evaluated, regional information of the enterprise to be evaluated, carbon emission information of the enterprise to be evaluated in a preset time period, a preset peak accounting planning model, and a training sample set; processing the enterprise information to be evaluated to generate peak driving parameter information of the enterprise to be evaluated; processing the regional information of the enterprise to be evaluated to generate peak measure parameter information of a target region and energy saving and emission reduction parameter information of the target region; processing the carbon emission information of the enterprise to be evaluated in the preset time period to generate an interval peak prediction subtask and a sequential access peak prediction subtask; processing the interval peak prediction subtask and the sequential access peak prediction subtask to generate a peak prediction feature vector; processing the training sample set to generate a training sample set with target feature information, wherein the target feature information is used to represent risk factors of the emission reduction effect in an abnormal state; processing the preset peak accounting planning model based on the training sample set with target feature information to generate a target peak accounting planning model; processing the peak prediction feature vector, the peak driving parameter information of the enterprise to be evaluated, the peak measure parameter information of the target region, and the energy saving and emission reduction parameter information of the target region based on the target peak accounting planning model to generate target peak path planning information.

[0009] In another aspect of the present application, a peak path planning device for a shipbuilding and repairing enterprise comprises an acquisition module configured to acquire enterprise information to be evaluated, regional information of the enterprise to be evaluated, carbon emission information of the enterprise to be evaluated within a preset time period, a preset peak accounting planning model, and a training sample set; and a processing module configured to process the enterprise information to be evaluated to generate peak driving parameter information of the enterprise to be evaluated, process the regional information of the enterprise to be evaluated to generate peak measure parameter information of a target region and energy saving and emission reduction parameter information of the target region, process the carbon emission information of the enterprise to be evaluated within the preset time period to generate an interval peak prediction subtask and a sequential access peak prediction subtask, process the interval peak prediction subtask and the sequential access peak prediction subtask to generate a peak prediction feature vector, process the training sample set to generate a training sample set with target feature information, wherein the target feature information is used to represent risk factors of an abnormal state of emission reduction effect, process the preset peak accounting planning model based on the training sample set with the target feature information to generate a target peak accounting planning model, and process the peak prediction feature vector, the peak driving parameter information of the enterprise to be evaluated, the peak measure parameter information of the target region, and the energy saving and emission reduction parameter information of the target region based on the target peak accounting planning model to generate target peak path planning information.

[0010] According to still another aspect of the present application, an electronic device comprises a first processor and a memory configured to store executable instructions of the first processor, wherein the first processor is configured to execute the executable instructions to implement the peak path planning method for a shipbuilding and repairing enterprise.

[0011] According to still another aspect of the present application, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a second processor to implement the peak path planning method for a shipbuilding and repairing enterprise.

[0012] The peak path planning method for a shipbuilding and repairing enterprise and related devices provided by the present application comprises the following steps: a server acquires enterprise information to be evaluated, regional information, carbon emission information, a preset model, and a training sample set. Then, the enterprise information is processed to obtain peak driving parameter information, the regional information is processed to obtain peak measure and energy saving and emission reduction parameter information, and the carbon emission information is processed to obtain a prediction subtask. Then, a feature vector is extracted from the prediction subtask, and a sample set with target feature information is obtained by processing the training sample set. Finally, based on a target peak accounting planning model, an energy and industrial structure adjustment path is generated by comprehensively processing the peak prediction feature vector, the enterprise and regional related parameters, and a preset peak path, a peak time, a peak value, and target peak path planning information are obtained.

[0013] It should be understood that the general description above and the detailed description below are only exemplary and explanatory, and are not limiting to the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0014] Figure 1 A flowchart of a peak-reached path planning method suitable for a shipbuilding and repairing enterprise according to an embodiment of the present application is shown; Figure 2 A structural schematic diagram of a peak-reached path planning device suitable for a shipbuilding and repairing enterprise according to an embodiment of the present application is shown; Figure 3 A peak-reached curve effect diagram according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0015] The preferred embodiments of the present application are described below in conjunction with the accompanying drawings, and it should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application, and are not used to limit the present application.

[0016] The peak-reached path planning method suitable for a shipbuilding and repairing enterprise according to the exemplary embodiments of the present application is described below in conjunction with Figure 1 It should be noted that the following application scenarios are only shown for the convenience of understanding the spirit and principles of the present application, and the embodiments of the present application are not limited in this respect. On the contrary, the embodiments of the present application can be applied to any applicable scenario.

[0017] In one embodiment, the present application further provides a peak-reached path planning method suitable for a shipbuilding and repairing enterprise and related equipment. Figure 1 A flowchart of a peak-reached path planning method suitable for a shipbuilding and repairing enterprise according to an embodiment of the present application is shown. As shown in Figure 1 The method is applied to a server and includes: S101, obtaining to-be-evaluated enterprise information, regional information of the to-be-evaluated enterprise, carbon emission information of the to-be-evaluated enterprise in a preset time period, a preset peak-reached accounting planning model, and a training sample set.

[0018] In one implementation, the enterprise basic profile information is as follows: enterprise name: XX shipbuilding and repair factory, type: manufacturing industry (shipbuilding and repair). Enterprise size: 2000 employees, 500,000 square meters of land, annual turnover of 1 billion yuan, located in coastal areas, near the port, convenient transportation, but facing the pressure of marine ecological protection. Production capacity: 50 ships can be built and 200 ships can be repaired per year, products are diverse, and the complexity of different shipbuilding and repair processes affects energy consumption and carbon emissions. Equipment situation: 5 large docks, 20 cranes, 100 welding equipment, 30 coating equipment, etc., each equipment has different energy consumption characteristics, and the energy consumption and carbon emissions of each process of shipbuilding and repair are different, such as high energy consumption of steel pretreatment and VOCs emission in coating process.

[0019] The regional information is as follows: climate: subtropical monsoon climate, high temperature and rain in summer, mild and little rain in winter, affecting enterprise energy demand, such as high energy consumption for refrigeration in summer and precipitation affecting wastewater treatment. Energy: abundant wind energy resources, lack of traditional energy, prompting enterprises to consider renewable energy utilization. Geographic location: located in developed coastal areas, near ports and transportation hubs, but subject to marine environmental regulations, surrounded by marine, wetland and a small amount of forest ecosystems, and enterprises need to avoid damaging the environment during production. The local government has strict marine ecological protection policies, and the enterprise needs to take environmental protection measures. Flat terrain is conducive to construction and transportation, but there is a risk of sea level rise, which needs to be considered for flood and tide prevention. Developed transportation leads to increased transportation energy consumption and carbon emissions, and the enterprise needs to optimize logistics. Carbon emission information is as follows: total amount trend: total carbon emissions have fluctuated and increased over the past five years, reaching 500,000 tons of carbon dioxide equivalent in 2018 and 600,000 tons in 2022. Carbon emission composition: direct carbon emissions mainly come from energy combustion during shipbuilding and repair, diesel combustion accounts for 70%, welding gas emissions account for 15%, and coating solvent volatilization accounts for 10%, etc.; indirect carbon emissions mainly come from electricity consumption, accounting for about 30% of total carbon emissions, and showing an upward trend. Unit product value carbon emission intensity is also increasing, and the enterprise needs to strengthen energy saving and emission reduction.

[0020] The preset peak-reaching accounting planning model is taken as an example, which is a mixed integer linear programming model. The model is constructed by defining the objective function and setting the constraint conditions, and the decision variables (such as energy consumption, industrial development scale, etc.) are quantitatively optimized under the premise of meeting the actual operation restrictions of the enterprise, and finally the specific target (such as the shortest peak-reaching time, the lowest peak value, etc.) is realized. The objective function adopts a weighted combination form, which comprehensively reflects the priority demand of the enterprise for the peak-reaching time and the peak value, and the formula is as follows: , wherein Z is the objective function value (comprehensive optimization target); is the peak-reaching time (unit: year); is the peak value (unit: ten thousand tons of carbon dioxide equivalent); , Weighting coefficients ( =1), set by the company according to its strategic needs. For example, if the company prioritizes "reaching the peak as early as possible", then... =0.6、 =0.4; If priority is given to controlling the "peak size", then set =0.3、 =0.7.

[0021] The constraints cover key limitations throughout the entire enterprise operation process, specifically as follows: Regarding energy balance constraints: Ensure that energy supply meets production needs at all times, as shown in the formula... ( ),in, This represents the consumption of energy type k in time period t (unit: 10,000 tons of standard coal). Let m be the total energy demand of the enterprise in time period t (unit: 10,000 tons of standard coal), and m be the quantity of energy type (such as coal, electricity, natural gas, etc.).

[0022] Regarding constraints on industrial development: The scale of the industry is limited by the upper limit of enterprise production capacity and market demand, as shown in the formula: ,in, The output of product type s in time period t (unit: 10,000 ships / 10,000 tons). , These are the minimum guaranteed output (to meet basic order requirements) and the maximum production capacity (limited by equipment and site conditions) for the product.

[0023] Regarding emission reduction potential constraints: The maximum feasible effect of emission reduction measures is determined based on the enterprise's technological level, using the following formula: ( ,i), where, This represents the actual emission reduction of the i-th emission reduction measure in time period t (unit: 10,000 tons of carbon dioxide). The maximum technical emission reduction potential of this measure is determined by equipment energy efficiency and process level. For example, the maximum emission reduction of "upgrading welding equipment" can be calculated as 20,000 tons per year through equipment parameters.

[0024] Regarding environmental capacity constraints: Compliance with regional environmental emission standards, the formula is as follows: ( ),in, This represents the actual carbon emissions of the enterprise in time period t (unit: 10,000 tons of CO2 equivalent). The carbon emission allowance for the region during this period (as determined by the local environmental protection department).

[0025] The input variables of the model are as follows: For energy consumption data: such as annual coal consumption (unit: ten thousand tons), annual electricity consumption (unit: million kilowatt-hours), natural gas price (unit: yuan / cubic meter). For production process parameters: such as energy utilization efficiency of welding process (unit: %), VOCs emission coefficient of painting process (unit: kg / m2), shipyard turnover rate (unit: times / year). For regional energy supply: such as regional wind energy developable installed capacity (unit: ten thousand kilowatts), total annual solar radiation (unit: megajoules / square meter). For emission reduction measures data: such as investment cost of "photovoltaic roof construction" (unit: ten thousand yuan), expected annual emission reduction (unit: ten thousand tons of carbon dioxide), investment payback period (unit: years).

[0026] The output variables of the model are as follows: For core results: peak time (accurate to quarter, such as "2028Q3"), peak value (accurate to 1 decimal place, such as "58.3 million tons of carbon dioxide equivalent"). For energy structure: the proportion of coal, electricity, and clean energy in each year (such as "in 2025, coal accounts for 35%, electricity accounts for 40%, and clean energy accounts for 25%"). For industry scale: shipbuilding / repair output in each year (such as "in 2026, 18 bulk carriers are built and 190 ships are repaired"). For emission reduction plan: annual list of emission reduction measures (such as "in 2024, welding equipment upgrade is implemented, and in 2025, 50,000 kilowatt photovoltaic power station is built").

[0027] Examples of training sample sets are as follows: Industry data: relevant data of 100 shipbuilding and repair enterprises are collected, which are distributed in different regions and have different scales, covering various types from small repair shipyards to large shipbuilding enterprises, with strong representativeness. Historical data accumulation: including detailed production and operation data, energy consumption data, carbon emission data and records of energy saving and emission reduction measures of the enterprise in the past 10 years, which can reflect the characteristics and trends of the enterprise in different development stages, providing rich internal information for model training. Enterprise scale characteristics: such as number of employees, total assets, annual turnover, etc., which can reflect the production capacity and economic strength of the enterprise, and have important influence on the carbon emission and peak reaching path of the enterprise. For example, large enterprises usually have more production equipment and higher energy consumption, but they also have advantages in technology research and development and capital investment, and may be more capable of implementing large-scale energy saving and emission reduction projects.

[0028] Technical level features: including the advancement of production process, energy utilization efficiency, degree of automation of equipment, etc. Advanced production process and efficient equipment can often reduce the energy consumption and carbon emissions per unit of product, and the investment and achievements of enterprises in technological innovation can be reflected through these features, which are also key factors affecting the peak path. Regional features: such as the energy resource endowment of the region where the enterprise is located (the richness of resources such as coal, natural gas, and renewable energy), the strictness of environmental policy (such as carbon emission standards, environmental tax collection standards, etc.), the level of economic development (regional GDP, per capita income, etc.). Regional features will affect the energy procurement cost, environmental pressure and market demand of the enterprise, thus affecting the peak strategy of the enterprise. Carbon emission related features: including total carbon emissions, carbon emission intensity, carbon emission source structure, etc. These are the core target features of the model training. Through analyzing the relationship between these features and other sample features, the model can learn the key factors and rules that affect carbon emissions, so as to predict the peak path of the enterprise under different circumstances.

[0029] S102, processing the to-be-evaluated enterprise information to generate the peak driving parameter information of the to-be-evaluated enterprise.

[0030] In an implementation manner, the to-be-evaluated enterprise information is processed to generate attribute information of the to-be-evaluated enterprise, wherein the attribute information of the to-be-evaluated enterprise includes equipment list information, capacity data information, energy consumption parameters, and production process parameters. For example, the equipment list information is as follows. Large dock: 5 in number, with a size of 300 meters in length, 50 meters in width, and 10 meters in depth, built between 2010 and 2015, using an advanced hydraulic system to control the lifting of the dock, mainly consuming electric power for pumping and equipment operation, and the maximum carrying capacity of a single dock is 100,000 tons. Crane: including portal cranes and tower cranes, a total of 20, among which the maximum lifting capacity of the portal crane is 500 tons, the maximum lifting height of the tower crane is 100 meters, the service life is 5-15 years, the energy consumption is mainly electric power, and some old cranes have low energy utilization efficiency. Welding equipment: including 100 sets of manual arc welding machines, gas shielded welding machines, and submerged arc welding machines, with different welding current and voltage ranges for different welding materials and processes, and the energy consumption is electric power. New welding machines have certain improvement in energy saving and welding quality. Coating equipment: including 30 sets of high-pressure airless spraying machines and electrostatic spraying equipment, with a maximum spraying area of 500 square meters per hour, a service life of 3-10 years, and a main energy consumption of electric power. Some devices need to be improved in paint utilization rate, and volatile organic compounds (VOCs) are emitted during operation.

[0031] For example, the shipbuilding capacity is 50 ships per year, including 20 bulk carriers, 15 container ships, 10 tankers and 5 special-purpose ships. The construction period varies for different types of ships, with an average of 12 months for bulk carriers, 10 months for container ships, 15 months for tankers and 18-24 months for special-purpose ships depending on the specific design and process requirements. The ship repair capacity is 200 ships per year, mainly including hull structure repair, power system maintenance and ship painting and refurbishment. The average repair time for each ship is 1-3 months depending on the extent of damage and the complexity of repair.

[0032] For example, the energy consumption parameters are as follows: electricity consumption accounts for 60% of the total energy consumption, mainly used for production equipment operation (such as dock pumping, crane operation, welding equipment power supply, coating equipment driving, etc.), workshop lighting and office electricity. Diesel consumption accounts for 40% of the total energy consumption, mainly used for ship trial, transportation vehicles and part of emergency power generation equipment. Energy consumption intensity: in the past year, the enterprise consumed 100,000 kilowatt-hours of electricity and 50 tons of diesel for every 100,000 yuan of output value created. The energy consumption intensity varies greatly in different production links, for example, the electricity consumption intensity is higher in the welding link of shipbuilding process, while the diesel consumption intensity is higher in the ship trial stage.

[0033] The production process parameters are as follows: steel pretreatment process: using shot blasting and paint spraying pretreatment method, the shot blasting speed of the shot blasting equipment is 10-15 meters per minute, the paint spraying thickness is controlled at 100-200 microns, the energy consumption of this process is mainly concentrated in the motor drive of the shot blasting equipment and the air compression link of the paint spraying equipment, and a certain amount of waste steel shot and paint slag and other waste will be generated. Subsection manufacturing process: using assembly line operation mode, ship sections are assembled and welded in different stations, the welding process mainly uses gas shielded welding and submerged arc welding combined, the welding speed is between 30-60 cm per minute according to the plate thickness and welding requirement, the key of quality control of this process is the accurate adjustment of welding parameters such as welding current, voltage, welding speed, and the energy consumption mainly comes from the welding equipment. Shipbuilding berth closing process: the sections are hoisted to the berth for closing by large cranes, high-precision measuring equipment such as total station is used for positioning and accuracy control, the energy consumption in the closing process is mainly the running energy consumption of the crane, and the construction precision and process requirement are relatively high to ensure the overall structural strength and watertightness of the ship. Outfitting process: including ship internal equipment installation, pipeline laying, cable laying and other work, using modular outfitting technology to improve outfitting efficiency, but the installation and debugging of some complex equipment need experienced technicians to operate, the energy consumption is relatively low, mainly concentrated in lighting and small electric tool power consumption. Painting process: using high-pressure airless spraying and electrostatic spraying combined method, the coating thickness is between 200-500 microns according to the ship use environment and corrosion prevention requirement, the environmental impact of painting process is larger, VOCs emission is the main problem, and the utilization rate of paint and drying process have certain influence on energy consumption, such as using hot air drying method will consume a lot of electricity.

[0034] The device list information, capacity data information, energy consumption parameters and production process parameters are processed to generate target parameter information of the enterprise to be evaluated and weight information matched with the target parameter information. For example, the device energy efficiency parameter: according to the device type and service life in the device list information, combined with the energy consumption parameters, the energy utilization efficiency of each device is calculated. For example, the energy utilization efficiency of a new crane is 80% (i.e. 80% of the input energy is converted into effective work energy), while the energy utilization efficiency of an old crane may only be 60%. For welding equipment, the energy utilization efficiency of different models varies between 50%-70%. Through the statistics and analysis of these device energy efficiency parameters, the energy utilization level of the overall equipment of the enterprise can be determined, providing a basis for evaluating the energy saving and emission reduction potential of the enterprise. The capacity and energy consumption correlation parameter: the capacity data information is combined with the energy consumption parameters to calculate the energy consumption per unit product. For example, 500,000 kilowatt-hours of electricity and 200 tons of diesel are consumed to build one bulk carrier, and 100,000 kilowatt-hours of electricity and 50 tons of diesel are consumed to repair one ship. These parameters can reflect the energy consumption intensity of the enterprise in different production activities, and help to analyze the relationship between the production efficiency and energy consumption of the enterprise, providing a reference for formulating targeted energy saving and emission reduction measures. Process carbon emission parameter: based on the production process parameters and energy consumption parameters, the carbon dioxide emissions of each production process link are calculated. For example, in the steel pretreatment process, the carbon dioxide emissions generated by the electric power consumption in the shot blasting process are 100 tons per year, and the carbon dioxide emissions generated by the solvent evaporation and energy consumption in the paint spraying process are 200 tons per year; in the shipbuilding process, the carbon dioxide emissions generated by the diesel combustion in the crane lifting process are 300 tons per year. Through the analysis of the process carbon emission parameter, the main source of carbon emissions of the enterprise can be determined, providing a direction for optimizing the production process and reducing carbon emissions.

[0035] The weight information example is a Japanese system (determined by the analytic hierarchy process), and a judgment matrix is constructed: through data analysis, the relative importance of each target parameter information to the enterprise peak driving is compared, and a judgment matrix is constructed. Calculate the weight vector: solve the maximum eigenvalue of the judgment matrix and the corresponding eigenvector, and normalize to get the weight vector. After calculation, the weight of the device energy efficiency parameter is 0.25, the weight of the capacity and energy consumption correlation parameter is 0.25, and the weight of the process carbon emission parameter is 0.5. These weights reflect the relative importance of each target parameter information in influencing the enterprise peak driving, which will play an important role in the subsequent calculation of the peak driving parameter information.

[0036] The target parameter information of the enterprise to be evaluated and the weight information matched with the target parameter information are processed to generate peak-reaching driving parameter information of the enterprise to be evaluated. For example, the peak-reaching driving parameter calculation is as follows: comprehensive parameter calculation: according to the target parameter information and the weight information, the peak-reaching driving comprehensive parameter of the enterprise is calculated. For example, if the peak-reaching driving comprehensive parameter is, then the equipment energy efficiency parameter is the capacity and energy consumption associated parameter is the process carbon emission parameter. If the equipment energy efficiency parameter of an enterprise is 0.7 (indicating that the overall equipment energy utilization efficiency is relatively high), the capacity and energy consumption associated parameter is 0.6 (indicating that the unit product energy consumption is at a medium level), and the process carbon emission parameter is 0.8 (indicating that the process carbon emission is relatively high), then the peak-reaching driving comprehensive parameter is. Driving factor contribution analysis: by calculating the contribution of each target parameter information to the peak-reaching driving comprehensive parameter, the key driving factors of the enterprise to reach the peak are analyzed. In the above example, the contribution of the equipment energy efficiency parameter to is, the contribution of the capacity and energy consumption associated parameter is, and the contribution of the process carbon emission parameter is. It can be seen that the process carbon emission parameter has the greatest impact on the driving of the enterprise to reach the peak, and the enterprise should focus on the optimization of the production process and the reduction of carbon emission when formulating the peak-reaching strategy, such as adopting more environmentally friendly coating process, improving the energy utilization efficiency of the welding process, etc.; at the same time, the improvement of the equipment energy efficiency and the reasonable matching of the capacity and energy consumption are also of great significance, which can be realized through equipment updating and modification and production process optimization, etc.

[0037] S103, processing the regional information of the enterprise to be evaluated to generate peak-reaching measure parameter information of the target region and energy-saving and emission-reducing parameter information of the target region.

[0038] In one implementation, the regional information of the enterprise to be evaluated is processed to generate natural environment factor information of the target region, ecosystem factor information of the target region, and geographical factor information of the target region. Examples of natural environment factor information (taking a coastal area as an example), climate conditions: the region belongs to a subtropical monsoon climate, with an average annual temperature of about 22°C, an average summer temperature of 30°C, and an average winter temperature of 10°C. The annual precipitation is abundant, with an average precipitation of 1800 millimeters, and the precipitation is mainly concentrated in summer (May-September), accounting for about 70% of the annual precipitation. The prevailing wind direction changes significantly with the seasons, with southeast winds in summer and northwest winds in winter, with an average wind speed of 3-5 meters per second. Such climate conditions have a significant impact on the energy demand of enterprises, with high summer temperatures prompting enterprises to increase cooling energy consumption, while abundant wind energy resources provide potential for renewable energy development. Natural resources: the region is rich in natural resources, with abundant water resources, with multiple rivers running through the region, with an annual runoff of about 5 billion cubic meters, providing a certain guarantee for enterprise production water, but also requiring enterprises to manage wastewater discharge and water resources. In terms of energy resources, in addition to traditional energy sources such as coal and oil that need to be imported externally, wind and solar energy resources have great potential. According to estimates, the wind energy development capacity of the coastal area is about 10 million kilowatts, with an annual effective utilization of more than 2000 hours; the total solar radiation is about 5000 megajoules per square meter, providing good conditions for large-scale development and utilization of solar energy.

[0039] Examples of ecosystem factor information include the following: surrounding ecosystem types: the surrounding area of the enterprise is mainly composed of marine ecosystems, wetland ecosystems, and terrestrial forest ecosystems. The marine ecosystem is an important ecological asset in the region, with a coastline of 200 kilometers and abundant marine biodiversity, with a variety of fish, shellfish, and marine mammals. The wetland ecosystem is distributed in the river estuary and coastal mudflats, with an area of about 500 square kilometers, and has important ecological functions such as climate regulation, water purification, and habitat provision. The terrestrial forest ecosystem is mainly distributed in the mountainous area, with a forest coverage rate of about 30%, playing a key role in soil and water conservation, water conservation, and maintenance of regional ecological balance. Ecological protection requirements: based on the rich ecosystem resources, the local government has introduced strict ecological protection policies. In terms of marine ecological protection, a marine ecological protection zone has been designated, prohibiting any activities that may damage the marine ecosystem within the protection zone, such as prohibiting wastewater discharge, limiting fishing intensity, controlling ship speed to reduce disturbance to marine life, etc. For the wetland ecosystem, a wetland restoration and protection project is implemented, requiring enterprises to strictly control wastewater discharge to ensure that the water quality meets the requirements of the wetland ecosystem, and to avoid damaging the wetland ecosystem. In terms of terrestrial forest ecological protection, a forest cutting quota system is implemented, and enterprises must undergo strict approval and take appropriate ecological compensation measures if they need to occupy forest land for infrastructure construction or production activities.

[0040] Geographical factors information example as follows, landform: the region terrain is mainly plains and hills, the terrain is relatively flat, the altitude is between 0-500 meters. Coastal areas have some tidal flats and shallow sea areas, suitable for the construction of ports and the development of marine industries. The plain area is convenient for transportation, which is conducive to the construction of factory buildings and logistics transportation for enterprises, but there is a certain risk of flooding, especially in the rainy season, enterprises need to strengthen the construction of flood control facilities. The hilly area is rich in forest resources, but the development and construction are relatively difficult, which has a certain impact on the infrastructure construction and production and operation layout of enterprises. Transportation network: the transportation network in the region is developed, with highways, railways and water transportation connecting each other. The total length of highways reaches 5000 kilometers, with expressways connecting major cities around the region. The main railway line connects the main railway hubs in the country, with an annual freight volume of 50 million tons. There are many coastal ports with multiple 10,000-ton berths, with an annual cargo throughput of 200 million tons, which is an important channel for the import of raw materials and the export of products for enterprises. The convenient transportation network provides good logistics conditions for enterprise development, but also leads to traffic congestion and increased energy consumption and carbon emissions during transportation, so enterprises need to optimize logistics management and improve transportation efficiency.

[0041] The natural environmental factors information of the target area is processed to generate the evaluation information of the renewable energy resources in the region. The wind energy resource assessment example is as follows: wind energy resource distribution: according to meteorological data and field measurement, the wind energy resource distribution map in the region is drawn. It is determined that the coastal area is a wind energy resource rich area with high average wind speed and wind power density of 300-500 watts per square meter; the inland hilly area has relatively low wind speed and wind power density of 100-300 watts per square meter. Through the analysis of the wind energy resource distribution, the basis for the site selection of wind power generation field is provided. Wind energy development potential estimation: combined with the landform, land use planning and environmental impact assessment, the developable wind power installed capacity in the region is estimated. Considering that the coastal tidal flats and part of the hilly areas are suitable for the construction of large wind turbine generators, it is estimated that the total developable installed capacity is 8 million kilowatts. At the same time, the utilization efficiency of wind energy resources in different wind speed sections is analyzed to determine the optimal wind turbine selection and layout scheme to improve the economic and environmental benefits of wind energy development.

[0042] For example, the solar energy resource assessment is as follows: solar radiation analysis: using meteorological station observation data and satellite remote sensing data, the spatial and temporal distribution characteristics of solar radiation in the region are analyzed. The regions with high annual total solar radiation are mainly concentrated in the open areas of plains and hills, with annual total radiation of 4500-5500 megajoules per square meter. According to the solar radiation in different regions, the solar energy resource grade is divided, which provides reference for the planning of solar photovoltaic power generation project. Solar energy utilization potential calculation: combined with land resources and roof resources investigation, the area available for solar photovoltaic power generation in the region is calculated. For example, the roof area of industrial factory is about 1 million square meters, the roof area of residential housing is about 2 million square meters, and the area of wasteland and idle land is about 50 square kilometers. Considering the conversion efficiency of solar panels and the installation angle and other factors, the potential annual power generation of solar photovoltaic power generation in the region is estimated to be 2 billion kilowatt-hours, which provides data support for enterprises to develop renewable energy utilization plan.

[0043] The ecosystem factor information of the target region is processed to generate regional restriction parameter information, wherein the regional restriction parameter information includes industrial structure planning parameter information and enterprise energy saving and emission reduction potential parameter information. For example, the industrial structure planning parameter information is as follows: industrial development direction: according to the regional resource endowment and ecological protection requirements, the industrial structure adjustment plan is made. It is clear that the low-carbon and green industries such as marine high-end equipment manufacturing, new energy industry and energy-saving and environmental protection industry should be developed, and traditional industries with high energy consumption and high pollution such as small chemical industry and printing and dyeing industry should be gradually eliminated. It is planned that the output value proportion of marine high-end equipment manufacturing industry will increase from the current 20% to 30% in the next 5 years, the output value proportion of new energy industry will increase from 10% to 20%, and the output value proportion of high energy consumption industry will decrease from 30% to less than 20%. Industrial access threshold: in order to control the increase of carbon emissions, strict industrial access threshold is set. For newly introduced projects, it is required that the unit energy consumption is less than 50% of the regional average level and the unit carbon dioxide emission is less than 60% of the regional average level. At the same time, enterprises are encouraged to adopt advanced production processes and technologies to improve resource utilization efficiency, for example, the energy utilization efficiency of newly introduced projects is required to reach more than 80% of the industry advanced level, and the water resource reuse rate is required to reach more than 90%.

[0044] For example, the potential parameters of enterprise energy saving and emission reduction are as follows: technical improvement potential: analyze the existing production process and equipment of the enterprise, and evaluate the potential of energy saving and emission reduction through technical improvement. For example, if a new welding technology is used in shipbuilding and repair enterprises, the welding efficiency can be improved by 20% and the energy consumption can be reduced by 15%. If the coating equipment is upgraded and transformed, low VOC content paint and high efficiency spraying process can be used to reduce VOC emissions by more than 30%. According to the technical level and equipment condition of different enterprises, the direction and potential of technical improvement are determined to provide guidance for enterprises to develop energy saving and emission reduction measures. Management optimization potential: evaluate the optimization space of enterprise energy management and production management. For example, the establishment of a perfect energy management system can realize real-time monitoring and fine management of energy consumption, which is expected to reduce energy consumption by 5%-10%. By optimizing the production process, reasonably arranging the production plan, reducing the idle time and standby time of the equipment, the production efficiency can be improved by 10%-15%, and the energy consumption and carbon emissions can be reduced. According to the management status of the enterprise, the corresponding management optimization scheme is developed to tap the energy saving and emission reduction potential of the enterprise.

[0045] The evaluation information of regional renewable energy resources and the parameter information of industrial structure planning are processed to generate the peak reaching measure parameter information of the target region. For example, the energy structure adjustment strategy is as follows: renewable energy development target: based on the evaluation information of regional renewable energy resources, the development target of renewable energy is formulated. It is planned to increase the proportion of renewable energy in energy consumption structure from the current 10% to 30% in the next 10 years. Specific measures include building large offshore wind farms in coastal areas with a total installed capacity of 5 million kilowatts; promoting distributed solar photovoltaic power generation projects in plain and hilly areas with an installed capacity of 2 million kilowatts; at the same time, exploring the development and utilization of other renewable energy sources such as biomass energy and geothermal energy, such as building biomass power generation projects and geothermal energy heating demonstration projects. Traditional energy replacement plan: gradually reduce the dependence on traditional high-carbon energy such as coal and oil, and develop a traditional energy replacement plan. Strengthen cooperation with energy suppliers and increase the proportion of clean energy such as natural gas in energy consumption. It is planned to increase the proportion of natural gas in industrial energy consumption from the current 20% to 30% in the next 5 years. Promote clean and efficient utilization technologies of coal, such as coal washing, efficient combustion technology, etc., to reduce carbon emissions during the use of coal, and encourage enterprises to transform energy and use clean energy such as electricity and natural gas to replace coal as production energy.

[0046] For example, the government can introduce a series of policies to support industrial transformation, including financial subsidies, tax incentives, and financial support. For enterprises investing in new energy industries and energy-saving and environmental protection industries, a financial subsidy of 10%-20% of the total investment is given. Enterprises that meet the conditions are subject to the "three exemptions and half reduction" policy of income tax. A green industry development fund is established to provide low-interest loans and financing guarantees to encourage enterprises to increase investment in low-carbon industries. Technical innovation support: Establish a platform for industry-university-research cooperation to strengthen cooperation between enterprises, universities, and research institutions to jointly develop low-carbon technology research and innovation. For example, support enterprises and research institutions to jointly develop marine new energy equipment manufacturing technology research and development to improve the enterprise's independent innovation capability. A science and technology innovation award fund is established to reward enterprises and individuals who have made outstanding achievements in energy-saving and emission-reduction technology innovation, and to stimulate enterprise innovation enthusiasm.

[0047] The geographical factor information and enterprise energy-saving and emission-reduction potential parameter information of the target area are processed to generate energy-saving and emission-reduction parameter information of the target area. For example, the impact of geographical factors on energy-saving and emission-reduction potential is as follows: Traffic energy-saving and emission-reduction: According to the characteristics of regional traffic network, develop traffic energy-saving and emission-reduction measures. Optimize traffic organization, promote intelligent traffic system, improve traffic operation efficiency, reduce vehicle congestion and idling time, and reduce traffic energy consumption by 5%-10%. Encourage the development of public transportation, increase public transportation lines and vehicles, and improve public transportation share. It is planned to increase the public transportation share from the current 30% to 40% within the next 3 years. At the same time, promote the application of new energy vehicles in logistics transportation and official vehicles, build supporting facilities such as charging piles, and gradually increase the proportion of new energy vehicles in transportation. Industrial layout optimization: Optimize industrial layout based on topography and traffic conditions to reduce enterprise logistics costs and energy consumption. For example, high-energy-consuming enterprises are concentrated in areas with convenient energy supply, such as near ports or railway freight stations, to reduce the transportation distance of raw materials and products; low-energy-consuming, high-value-added enterprises are located in the suburbs of cities or areas with good ecological environment to form an industrial cluster effect and improve resource utilization efficiency. Through industrial layout optimization, it is expected to reduce the overall energy consumption of enterprises in the region by 5%-8%.

[0048] The enterprise energy saving and emission reduction potential comprehensive assessment example is as follows, enterprise individual energy saving and emission reduction target setting: according to the enterprise energy saving and emission reduction potential parameter information, the individual energy saving and emission reduction target of each enterprise is set. For example, for the shipbuilding and repairing enterprise with large energy consumption, it is required to reduce the energy consumption per unit output value by 15% and the carbon dioxide emission by 20% in the next 3 years; for small and medium-sized manufacturing enterprises, according to their production scale and technical level, corresponding energy saving and emission reduction indicators are formulated, such as reducing the energy consumption per unit output value by 10% and the carbon dioxide emission by 15%. Regional energy saving and emission reduction overall planning: considering the energy saving and emission reduction potential of all enterprises in the region, the regional energy saving and emission reduction overall planning is made. It is expected that through the implementation of the above energy saving and emission reduction measures, in the next 5 years, the energy consumption per unit GDP in the region will be reduced by 20%, and the carbon dioxide emission will be reduced by 25%, so as to realize the regional energy saving and emission reduction target and promote the green development of regional economy, and lay a solid foundation for realizing the carbon emission peak.

[0049] In one embodiment, the carbon emission information of the enterprise to be evaluated in a preset time period is processed to generate an interval peak prediction subtask and a sequential access peak prediction subtask.

[0050] In one embodiment, the carbon emission information of the enterprise to be evaluated in a preset time period is processed to generate a carbon emission trend factor, an interval peak evaluation factor, carbon emission target node information, peak sequence information of the carbon emission target node, and peak time information of the carbon emission target node. An example of carbon emission trend factor (taking the past 5 years as an example), through linear regression analysis of the carbon emission data of the enterprise in the past 5 years, assuming that the year is the independent variable and the total carbon emission is the dependent variable, the regression equation is obtained. If the calculated (unit: million tons of carbon dioxide / year) is 5, it indicates that the carbon emission of the enterprise presents a linear growth trend of 50,000 tons of carbon dioxide per year, and this coefficient is one of the carbon emission trend factors, reflecting the basic growth trend of the total carbon emission of the enterprise with time. The carbon emission data of each year is decomposed by quarter or month to observe whether there is obvious seasonal fluctuation. For example, it is found that the carbon emission of the enterprise in the second quarter (April-June) of each year is usually 10%-15% higher than that in other quarters due to the increase of production task, and this seasonal fluctuation characteristic is also part of the carbon emission trend factor and can be used for adjustment of carbon emission in different time periods in subsequent prediction.

[0051] The interval peak assessment factor example is as follows, considering the influence of the double carbon policy introduced by the country and the local government on the enterprise. For example, the local government implemented a more stringent carbon emission quota system in a certain year, requiring enterprises to reduce carbon emission quota by 5% per year. This policy factor will be an interval peak assessment factor to prompt enterprises to accelerate the pace of emission reduction and affect the enterprise's peak time prediction. The energy-saving and emission-reducing technical improvement measures taken by the enterprise itself will also affect the peak assessment. For example, the enterprise introduced a new energy management system in a certain year, which is expected to reduce energy consumption by 10% and thus reduce carbon emissions. The effect of this technical improvement measure (10% emission reduction) can be used as an interval peak assessment factor to reflect the change in the enterprise's emission reduction capacity under the promotion of technology, and has a positive impact on peak prediction.

[0052] The carbon emission target node information example (taking a shipbuilding and repair enterprise as an example) is as follows, taking the main production links such as steel pretreatment, segmented manufacturing, shipbuilding, outfitting, and painting in the shipbuilding process as the carbon emission target nodes. For example, the painting link will produce volatile organic compound (VOCs) emissions due to the use of a large amount of paint, which is one of the important sources of carbon emissions of the enterprise, accounting for about 20% of the total emissions of the enterprise. Therefore, the painting link is taken as a key carbon emission target node, and its emission reduction potential and peak are focused on. According to the energy consumption structure of the enterprise, power consumption and diesel consumption are taken as the main carbon emission target nodes. For example, the power consumption of the enterprise accounts for 60% of the total energy consumption, mainly used for production equipment operation and lighting, and diesel consumption accounts for 40%, mainly used for ship trial and transportation vehicles. Analysis of the carbon emission of these energy consumption nodes helps to develop targeted energy-saving and emission-reducing measures to achieve the peak target.

[0053] The peak sequence information of the carbon emission target node is as follows, the peak sequence is determined by analyzing the technical feasibility, cost-effectiveness, and other factors of each carbon emission target node. For example, for a shipbuilding and repair enterprise, it may start with the power consumption node which is relatively easy to achieve emission reduction, and through measures such as optimizing equipment operation and adopting energy-saving equipment, it is expected that the carbon emission of the power consumption node will peak and start to decline within 2-3 years; while for the painting link, due to the complex problems such as process improvement and paint replacement, it may take 3-5 years to achieve carbon emission peak, so the power consumption node is ranked before the painting link to achieve peak. Combined with the sequence of the enterprise's production process, the peak sequence of the carbon emission target node is determined. For example, in the shipbuilding process, the steel pretreatment link is at the front end of the production process, and its carbon emission peak will affect the carbon emission of the subsequent production links, so it is given priority to achieve peak as early as possible, followed by the segmented manufacturing, shipbuilding, and other links to ensure that the carbon emission of the entire production process gradually peaks and declines in an orderly manner.

[0054] The peak time information of the carbon emission target node is as follows: assuming that the enterprise plans to comprehensively upgrade and reform the main production equipment within the next 3 years, and expects that the energy utilization efficiency of the equipment after the reform can be improved by 20%, which will make the peak time of some carbon emission target nodes (such as the power consumption node) advance. According to the equipment reform plan and the expected emission reduction effect, it is calculated that the power consumption node may reach the carbon emission peak after 2 years, which is 1 year earlier than without the implementation of the reform plan. If the enterprise predicts that the market demand for environmentally friendly ships will increase in the future, it will promote the enterprise to speed up the production process improvement and emission reduction measure implementation to meet the market demand. For example, in order to enter the environmentally friendly ship market in advance, the enterprise decides to replace all high-VOCs coatings in the coating process with low-VOCs environmentally friendly coatings within the next 2 years, which will make the carbon emission peak time of the coating link advance to 1.5 years, about 1 year earlier than the original plan.

[0055] The carbon emission trend factor and the interval peak evaluation factor are processed to generate an interval peak prediction subtask. The prediction based on the trend factor and the evaluation factor is as follows: according to the carbon emission trend factor (such as a linear growth trend), it is assumed that without other major influencing factors, the future carbon emission is predicted according to the current growth rate. For example, extrapolating the trend of increasing 50,000 tons of carbon dioxide per year in the past 5 years, it is predicted that the total carbon emission in the next 3 years will be the current emission plus 50,000, 150,000 and 250,000 tons of carbon dioxide respectively. However, the influence of the interval peak evaluation factor (such as a policy requiring a 5% annual carbon emission quota reduction) needs to be considered, and the prediction result needs to be adjusted. In this case, the total carbon emission in the first year after the current emission plus 50,000 tons of carbon dioxide is multiplied by to obtain the adjusted prediction value, and the prediction emission in the next 2-3 years is calculated in the same way, so as to generate an interval peak prediction subtask, that is, to focus on how the carbon emission gradually approaches the peak value in each time period and achieves the peak target under the policy and the enterprise's own efforts, and the possible peak time range.

[0056] The carbon emission trend and evaluation factor changes under different scenarios are considered. For example, a baseline scenario (developing according to the existing trend), a policy strengthening scenario (the government introduces stricter policies) and a technology breakthrough scenario (the enterprise makes a major energy-saving and emission-reducing technological breakthrough) are set. In the policy strengthening scenario, it is assumed that the carbon emission quota reduction rate is increased from 5% to 8% per year, and the future carbon emission prediction value is recalculated; in the technology breakthrough scenario, it is assumed that the enterprise adopts a new clean energy replacement technology, which can reduce carbon emission by 20%, and the prediction result is adjusted again. Through the analysis of different scenarios, multiple interval peak prediction subtasks are generated to provide reference for the enterprise to develop peak strategies to deal with different situations, for example, in the policy strengthening scenario, the enterprise needs to speed up the implementation of emission reduction measures to adapt to stricter policy requirements; in the technology breakthrough scenario, the enterprise can optimize the peak path according to the time node and effect of the technology application.

[0057] The carbon emission target node information, the peak reaching order information of the carbon emission target node, and the peak reaching time information of the carbon emission target node are processed to generate a sequentially accessed peak reaching prediction subtask. For example, based on the target node information, the peak reaching time and peak value of each node are predicted in sequence according to the peak reaching order information of the carbon emission target node. Taking the steel pretreatment node as an example, the carbon emission change curve of the node in each time period in the future is predicted by combining the current carbon emission level, the emission reduction measure plan (such as using more efficient rust removal equipment), and the expected peak reaching time (such as reaching the peak after 2 years), and the carbon emission peak value of the node when reaching the peak is determined. Then, according to the production process order, similar prediction is performed on the segmented manufacturing, ship berth closing, outfitting, painting, and other nodes in sequence to generate a sequentially accessed peak reaching prediction subtask, that is, to simulate the process of sequentially reaching the peak of each key node in the production process of the enterprise, and to analyze the mutual influence between the nodes and the feasibility of the overall peak reaching path.

[0058] The mutual correlation between the carbon emission target nodes is considered. For example, the emission reduction measures (such as using environmentally friendly paint) in the painting link may affect the production efficiency and energy consumption of the outfitting link, and then affect the carbon emission. When predicting the peak reaching of the painting link, the correlation influence needs to be considered comprehensively, and when predicting the peak reaching of the outfitting link, the influence of the emission reduction measures in the painting link also needs to be considered in the analysis. Through this node correlation analysis, the peak reaching time and peak value of each node are more accurately predicted, and a sequentially accessed peak reaching prediction subtask that is more consistent with the actual production situation is generated, which provides a basis for the enterprise to develop a comprehensive and effective peak reaching scheme, and ensures that the overall carbon emission can reach the peak value in an orderly and efficient manner and realize subsequent decline in the process of reaching the peak of each node.

[0059] S105, processing the interval peak reaching prediction subtask and the sequentially accessed peak reaching prediction subtask to generate a peak reaching prediction feature vector.

[0060] In one implementation, feature extraction processing is performed on the interval-to-peak prediction subtask to generate access time interval sequence features, access interval number features, and access interval information features. Generating an example of access time interval sequence features (assuming a quarterly time interval), assuming that the carbon emission data of the enterprise in the past 5 years (2018-2022) is counted by quarter, and taking the first quarter of 2018 as the starting point, the time interval sequence is 1-20 (a total of 20 quarters). For each quarter, calculate its time interval from the starting point (such as 2 for the second quarter of 2018, 5 for the first quarter of 2019, etc.), and record the total carbon emissions of that quarter. For example, the total carbon emissions are 100,000 tons when the time interval is 2 (the second quarter of 2018), and the total carbon emissions are 120,000 tons when the time interval is 5 (the first quarter of 2019), etc. These time interval and corresponding carbon emission data constitute the access time interval sequence feature, which reflects the changes in carbon emissions over time at different intervals, and helps to analyze the long-term trends and seasonal fluctuations of carbon emissions.

[0061] Generating an example of access interval number features, for example, continuing with the above quarterly data, assigning an access interval number to each quarter, such as 1-4 representing the four quarters of each year (1 for the first quarter, 2 for the second quarter, etc.). For the carbon emission data of each quarter, in addition to recording its time interval sequence feature, the corresponding quarter number is also recorded. For example, the access interval number for the first quarter of 2018 is 1, and the total carbon emissions are 80,000 tons; the access interval number for the second quarter of 2018 is 2, and the total carbon emissions are 100,000 tons, etc. The access interval number feature can help identify the regularity of carbon emission data in different seasons or specific time periods, such as whether there are seasonal differences (such as some enterprises have a production peak in the second quarter, and the interval data corresponding to the number 2 may show higher values).

[0062] Generating an example of access interval information features, for example, the access interval information feature can contain more detailed information about carbon emissions in each time interval, in addition to the total carbon emissions, the growth rate of carbon emissions, the change amount compared to the previous interval, etc. For example, the growth rate of carbon emissions in the second quarter of 2018 relative to the first quarter of 2018 is 25%, and the change amount is 10-8=20,000 tons; the growth rate of carbon emissions in the first quarter of 2019 relative to the fourth quarter of 2018 is 33.3%, and the change amount is 12-9=30,000 tons, etc. These growth rate and change amount data, combined with the time interval sequence and number features, can more comprehensively describe the dynamic changes of enterprise carbon emissions, providing more rich information for subsequent analysis.

[0063] The access interval information features are processed to generate adjacent interval difference sequence information, difference sequence mean, and difference sequence variance. An example of generating adjacent interval difference sequence information is as follows: based on the carbon emission total amount change data in the access interval information features described above, the difference sequence of adjacent intervals is calculated. For example, the carbon emission total amount change between the second quarter of 2018 and the first quarter of 2018 is 20,000 tons, and the carbon emission total amount change between the third quarter of 2018 and the second quarter of 2018 is 15,000 tons (assuming), so the adjacent interval difference sequence is 2, 1.5, etc. This difference sequence reflects the increasing or decreasing trend of the carbon emission total amount in adjacent time intervals, which helps to analyze the stability and change rate of carbon emission changes. For example, if the difference sequence fluctuates less, it means that the carbon emission growth is more stable, and vice versa.

[0064] Examples of calculating the difference sequence mean and variance are as follows: the mean of the adjacent interval difference sequence is calculated. Assuming that the above difference sequence 2, 1.5, etc. has 19 data (because there are 20 time intervals, and the adjacent interval difference has 19), the mean is obtained by adding these data and dividing by 19. For example, the mean is 18,000 tons (assuming), which represents the average change of the carbon emission total amount between adjacent quarters in the past 5 years, and can be used as an important indicator to measure the trend of carbon emission changes. When calculating the variance, the average of the square of the difference between each difference value and the mean is calculated according to the variance formula. The variance reflects the dispersion degree of the adjacent interval difference, and a smaller variance indicates that the change of the carbon emission total amount between adjacent quarters is relatively stable, and a larger variance indicates that the change is more dramatic, which may be affected by some special factors (such as production process adjustment, market demand fluctuation, etc.).

[0065] The sequential access peak prediction sub-task is processed to generate a one-dimensional prediction feature vector, and the one-dimensional prediction feature vector is processed to generate target access list information, mean of adjacent access information similarity sequence, and variance of adjacent access information similarity sequence. An example of generating a one-dimensional prediction feature vector (using the principal component analysis (PCA) method) is as follows: assuming that in the sequential access peak prediction sub-task, multiple feature variables of the enterprise in different production links (such as raw material procurement, production processing, product transportation, etc.) are considered, such as energy consumption, raw material usage, pollutant emission, production efficiency, etc. Each link has 5 feature variables, and there are 3 links, so the original data is a 3x5 matrix. Through the PCA method, the multi-dimensional data is projected into a one-dimensional space to obtain a one-dimensional prediction feature vector. For example, after PCA calculation, the obtained one-dimensional prediction feature vector is [0.5, 0.3, 0.2] (assuming), which integrates the information of the original multiple feature variables and represents the overall characteristics of the enterprise in the production process in a new way, facilitating subsequent integration and analysis with other features, while reducing the complexity of the data and highlighting the main information.

[0066] An example of generating target access list information (taking the production link as an example) is to determine the key target access nodes or links according to the enterprise production process and carbon emission situation, and form the target access list information. For example, for shipbuilding and repairing enterprises, the steel pretreatment, welding, painting and other links are determined as key target access nodes, because these links have greater impact on carbon emissions. In the target access list, record the relevant information of these nodes, such as node name, expected carbon emission peak (estimated according to historical data and prediction model, such as steel pretreatment link expected carbon emission peak of 50,000 tons, welding link of 30,000 tons, painting link of 40,000 tons, etc.), peak time range (such as steel pretreatment link is expected to peak in the next 2-3 years, welding link in 1-2 years, etc.). The target access list information provides the focus object for subsequent analysis, which helps to develop more targeted emission reduction measures and peak reaching strategies.

[0067] Calculate the similarity between adjacent access nodes, for example, use the cosine similarity method to calculate the similarity between steel pretreatment link and welding link in energy consumption, carbon emission and other characteristics. Assuming that the similarity of the two is 0.6, then the similarity between welding link and painting link is 0.4, etc., forming the similarity sequence of adjacent access information [0.6, 0.4] (assuming). The mean of the similarity sequence is 0.5, and the variance is 0.01 (assuming). The mean reflects the average level of similarity between adjacent access nodes, and the variance represents the dispersion degree of similarity. Smaller variance indicates that the similarity between adjacent nodes is relatively stable, while larger variance indicates that there is a large difference between adjacent nodes, which helps to analyze the correlation and change rule between different links in the production process of the enterprise, and provides basis for optimizing the production process and emission reduction measures.

[0068] The access time interval sequence feature, access interval number feature, access interval information feature, target access list information, mean of adjacent access information similarity sequence, and variance of adjacent access information similarity sequence are processed to generate a peak reaching prediction feature vector. The generated access time interval sequence feature (e.g., 100,000 tons of carbon emissions when the time interval is 2), access interval number feature (e.g., quarterly number and corresponding carbon emission data), access interval information feature (e.g., carbon emission growth rate, change amount, etc.), target access list information (e.g., key nodes and related parameters), mean of adjacent access information similarity sequence (e.g., 0.5), and variance of adjacent access information similarity sequence (e.g., 0.01) are integrated. For example, a vector [10, 2, 0.25, 5, 0.5, 0.01] is constructed (where 10 is the total carbon emissions when the time interval is 2, 2 is the corresponding quarterly number, 0.25 is the carbon emission growth rate of the quarter relative to the previous quarter, 5 is the predicted carbon emission peak of the steel pretreatment link, 0.5 is the mean of the adjacent access information similarity sequence, and 0.01 is the variance), and this vector is the peak reaching prediction feature vector. It integrates various key information extracted from the interval peak reaching prediction subtask and the sequential access peak reaching prediction subtask, and can comprehensively reflect the characteristics and trends of enterprise carbon emissions, providing an important data basis for peak reaching prediction and development of emission reduction strategies based on the model.

[0069] In S106, the training sample set is processed to generate a training sample set with target feature information, wherein the target feature information is used to represent risk factors of the emission reduction effect being in an abnormal state.

[0070] In an implementation, the training sample set is grouped and processed to generate a grouped training sample set, wherein the grouped training sample set includes peak reaching feature information of different regions and enterprises. The collected training samples are grouped according to different regions, for example, into eastern coastal areas, central regions, and western regions. The eastern coastal areas may have a developed economy, advanced technology level, and relatively strict environmental protection policies. Enterprises in this area face fierce market competition and have a high acceptance of energy-saving and emission reduction technologies, but the land and energy costs are also relatively high. The central region has a medium level of economic development, and the industrial structure is being optimized. Some enterprises face certain challenges in technology upgrading and energy-saving and emission reduction. The western region is rich in resources but relatively underdeveloped economically. Some enterprises may need to improve energy utilization efficiency and environmental awareness. Through such regional division, the differences in enterprise peak reaching characteristics under different regional environments can be analyzed. For example, enterprises in the eastern coastal areas may prefer to use clean energy and advanced environmental protection technologies to achieve carbon emission peak reaching, and their peak reaching time may be relatively early, while enterprises in the western region may be limited by economic and technical conditions, and their peak reaching time may be relatively late.

[0071] According to the industry and scale of the enterprise, it is divided into large manufacturing enterprises (such as shipbuilding and repair, steel production, etc.), small manufacturing enterprises (such as mechanical processing, plastic products, etc.) and service industry enterprises (such as logistics transportation, software development, etc.). The production and operation mode and carbon emission characteristics of different types of enterprises are very different. Large manufacturing enterprises usually have large energy consumption and high total carbon emissions, but they also have strong financial and technological strength to implement energy-saving and emission-reducing measures; small manufacturing enterprises may have relatively low total carbon emissions, but due to limited technology and management level, the unit carbon emission of production value is high; the carbon emissions of service industry enterprises are mainly concentrated in energy consumption (such as office electricity, transportation, etc.), which is different from the carbon emission sources and emission reduction ways of manufacturing enterprises. This grouping helps to develop individualized peak strategies for different types of enterprises, for example, large manufacturing enterprises focus on production process improvement and energy structure adjustment, small manufacturing enterprises focus on improving management efficiency and adopting small energy-saving equipment, and service industry enterprises mainly start from optimizing operation management and promoting green office methods.

[0072] Taking large manufacturing enterprises in the eastern coastal area as an example, this group of samples contains relevant data of multiple enterprises, such as enterprise A is a shipbuilding and repair enterprise, its peak characteristics information includes that the total carbon emissions have shown a trend of rising first and then tending to be stable in the past 10 years, reaching a relatively high plateau around 2015, and is currently working to achieve carbon emission peak through technological transformation and energy transformation; enterprise B is a steel production enterprise, its total carbon emissions have been at a high level for a long time, in recent years, with the strengthening of environmental protection policies, it has begun to take a series of emission reduction measures such as waste heat recovery and utilization, and the use of high-efficiency desulfurization and denitrification equipment, and is expected to achieve carbon emission peak in the next few years. The peak characteristics information of these enterprises (carbon emission trend, emission reduction measure implementation, peak expectation, etc.) constitutes the training sample set of large manufacturing enterprises in the eastern coastal area, providing a rich data basis for subsequent analysis, which can reflect the commonness and characteristics of the region and enterprise type in the process of carbon emission peak.

[0073] Feature extraction is performed on the grouped training sample set to generate an original feature library. For example, the enterprise size features include total assets, number of employees, annual turnover, etc. Total assets reflect the economic strength and production scale of the enterprise. Enterprises with larger total assets usually have greater ability to invest in equipment upgrades, technology research and development, and energy-saving and emission-reduction projects. The number of employees affects the energy consumption and carbon emission level of the enterprise. For example, labor-intensive enterprises may generate more carbon emissions in personnel commuting and office electricity use. The annual turnover is closely related to the production and operation activities of the enterprise. Enterprises with high turnover may mean higher production intensity and energy demand, and also have stronger financial recycling capacity to support energy-saving and emission-reduction work. Energy consumption features: involving total energy consumption, energy consumption structure (such as the proportion of different energies such as coal, oil, natural gas, and electricity), and energy consumption per unit of output value. Total energy consumption is directly related to the total carbon emissions and is an important indicator for measuring the carbon emission level of an enterprise. Energy consumption structure reflects the dependence of the enterprise on different energies. For example, an energy structure dominated by coal usually leads to higher carbon emissions, while increasing the proportion of clean energy helps to reduce carbon emissions. Energy consumption per unit of output value reflects the energy utilization efficiency of the enterprise. Enterprises with low energy consumption per unit of output value have better performance in energy-saving and emission-reduction and are more likely to achieve carbon emission peak earlier.

[0074] Production process features: covering the advancement of production process, the complexity of production process, the degree of automation of equipment, etc. Advanced production processes can often improve raw material utilization, reduce energy consumption and reduce pollutant emissions. For example, the use of advanced intelligent manufacturing technology can achieve precise production, reduce energy waste and waste generation. The complexity of the production process affects the distribution of energy consumption and carbon emissions in the enterprise. A complex production process may lead to more intermediate energy consumption and carbon emissions. Enterprises with high degree of equipment automation can more accurately control energy consumption in the production process, improve production efficiency, and reduce energy waste and carbon emissions caused by human factors. Regional environmental features: including regional energy supply stability, environmental policy strictness, regional economic development level, etc. Regional energy supply stability affects the energy procurement cost and energy structure adjustment strategy of the enterprise. Energy supply stability and diversification in the region are conducive to the selection of cleaner and more efficient energy by the enterprise. The strictness of environmental policy is an important driving force for enterprises to implement energy-saving and emission-reduction measures. Strict policies will encourage enterprises to increase environmental protection investment and improve production processes. The regional economic development level is closely related to market demand, technological innovation capability and environmental awareness of the enterprise. Enterprises in economically developed areas usually pay more attention to sustainable development and have stronger initiative and ability in energy-saving and emission-reduction.

[0075] The generation of the original feature library is shown in the following example. The above-mentioned various types of features extracted from different grouping samples are integrated to form the original feature library. For example, for the large manufacturing enterprise group in the eastern coastal area, the enterprise size features (such as total assets of 10 billion yuan for enterprise A, 5,000 employees, annual sales of 8 billion yuan, total assets of 20 billion yuan for enterprise B, 8,000 employees, annual sales of 120 billion yuan, etc.), energy consumption features (total energy consumption of 500,000 tons of standard coal for enterprise A, coal proportion of 30%, electricity proportion of 40%, unit energy consumption of 0.6 tons of standard coal per 10,000 yuan, total energy consumption of 800,000 tons of standard coal for enterprise B, coal proportion of 40%, electricity proportion of 30%, unit energy consumption of 0.8 tons of standard coal per 10,000 yuan, etc.), production process features (production process advancedness score of 7 for enterprise A, higher production process complexity, equipment automation degree of 80%, production process advancedness score of 8 for enterprise B, high production process complexity, equipment automation degree of 90%, etc.), and regional environmental features (stable energy supply in the eastern coastal area, strict environmental policy, high economic development level) are integrated together to form the original feature data of this group. The same feature extraction and integration is performed for other groups (such as the small manufacturing enterprise group in the central region, the service enterprise group in the western region, etc.), and finally the original feature library containing the features of all grouping samples is formed, providing comprehensive data support for subsequent model training and analysis.

[0076] The original feature library is processed to generate a training set and a validation set. Usually 70%-80% of the data is used as the training set and 20%-30% of the data is used as the validation set. For example, in the original feature library with a total of 1000 samples, 700 samples are randomly selected as the training set to train the prediction model, so that the model learns the relationship between the features and the emission reduction effect; the remaining 300 samples are used as the validation set to evaluate and adjust the model during training, to prevent overfitting of the model and ensure that the model has good generalization ability to accurately predict the emission reduction effect of new data. When dividing the training set and the validation set, the distribution of the data should be representative, i.e. the proportion of each type of sample (different regions, enterprise types, etc.) in the training set and the validation set should be approximately the same as that in the original feature library. For example, the proportion of large manufacturing enterprise samples in the eastern coastal area in the original feature library is 30%, the proportion of small manufacturing enterprise samples in the central region is 40%, and the proportion of service enterprise samples in the western region is 30%, so the proportions of the three types of samples in the divided training set and validation set should also be close to 30%, 40% and 30% to ensure that the model can fully learn the features and rules of different types of samples and improve the accuracy and reliability of the model.

[0077] The classifier is used to predict the validation set and generate a prediction result. The training set is trained based on a preset algorithm to generate the prediction result of the validation set. The decision tree classifier is selected to predict the validation set. The decision tree classifier classifies data by constructing a tree structure, judges the sample feature values (such as enterprise size features, energy consumption features, etc.) on the nodes of the tree, and gradually classifies the samples into different categories (such as good, general, and poor emission reduction effects). For example, for an enterprise sample in the validation set, the decision tree classifier first judges the total energy consumption, if the total energy consumption is high (greater than a certain threshold), it further judges the coal proportion in the energy consumption structure, if the coal proportion is also high, it further judges the strictness of the environmental policy in the region where the enterprise is located, and finally classifies the enterprise sample into the corresponding emission reduction effect category to generate the prediction result. The prediction result can be represented as the probability of each sample belonging to different emission reduction effect categories, for example, the probability of a certain enterprise sample belonging to the good emission reduction effect category is 0.3, the probability of belonging to the general category is 0.5, and the probability of belonging to the poor category is 0.2.

[0078] The support vector machine algorithm is used to train the training set. The support vector machine algorithm classifies data by finding an optimal hyperplane, maps samples in the training set to a high-dimensional space, and maximizes the separation between different categories. During training, the parameters of the hyperplane are adjusted according to the characteristics of the samples (such as enterprise size, energy consumption, production process, and regional environment) and the corresponding emission reduction effect labels (such as achieving emission reduction targets and not achieving emission reduction targets), so that the hyperplane can accurately separate different categories of samples. For example, for enterprise samples in the training set, their enterprise size features (total assets, number of employees, etc.), energy consumption features (total amount, structure, unit consumption, etc.), production process features (advancedness, complexity, automation, etc.), and regional environment features (energy supply, policy, economic level, etc.) are used as input vectors, and the corresponding emission reduction effect labels are used as output. The parameters of the hyperplane are calculated by the support vector machine algorithm to achieve high classification accuracy on the training set. After training, the trained model is used to predict the validation set to generate the prediction result of the validation set, which is compared and analyzed with the prediction result based on the classifier.

[0079] The prediction results and the validation set class prediction results are processed to generate target feature information, which is used to represent risk factors of the emission reduction effect being in an abnormal state. The prediction results based on the classifier (such as the decision tree classifier) are compared with the validation set prediction results based on the preset algorithm (such as the support vector machine algorithm). For example, it is found that for some enterprise samples, the decision tree classifier predicts that the emission reduction effect is good, but the support vector machine algorithm predicts that the emission reduction effect is general. Further analysis of the characteristics of these samples finds that it may be because these enterprises have some special circumstances in the production process characteristics, such as the use of a new but not yet widely used energy-saving production process. The decision tree classifier may cause prediction deviation when processing this new feature due to imperfect rules, while the support vector machine algorithm can better capture the complex relationship between this feature and the emission reduction effect in high-dimensional space.

[0080] According to the prediction result difference analysis, the risk factors representing the emission reduction effect being in an abnormal state are determined as target feature information. For example, for the enterprise samples described above that use new energy-saving production processes but have different prediction results, “uncertainty of the application and effect of new energy-saving production processes” is determined as a target feature information. In addition, if it is found that there are large fluctuations in the prediction of the emission reduction effect during the adjustment of the energy consumption structure of the enterprise (such as during the transition from a high coal proportion to a high clean energy proportion), “instability in the energy structure transition process” can also be a target feature information. These target feature information can help enterprises and decision-makers pay more attention to and solve key problems that may cause the emission reduction effect to be abnormal, such as strengthening the monitoring and evaluation of new energy-saving production processes and optimizing energy structure transition strategies, thereby improving the effectiveness and reliability of energy saving and emission reduction work, and providing strong support for achieving the scientific and reasonable carbon emission peak and carbon neutralization targets, which is consistent with the overall content of the document about using target feature information to improve the peak accounting planning model.

[0081] S107, based on the training sample set with target feature information, processing the preset peak accounting planning model to generate a target peak accounting planning model.

[0082] In an embodiment, the preset peak accounting planning model is initially constructed based on linear regression, and the core prediction formula is as follows: the peak time prediction is ; the peak value prediction is ; wherein, is a linear feature affecting the peak time (such as energy-saving technology investment intensity), is a linear feature affecting the peak value (such as the proportion of coal in the energy structure), , , , For linear regression coefficients. But in practical applications, it is found that when the enterprise is in the energy structure transformation period (such as the proportion of coal from 50% to 20%), the linear model prediction deviation is significant (error rate is more than 20%), the core reason is that "energy structure transformation" has a nonlinear impact - in the early stage (coal proportion 40%-50%) emission reduction effect is slow, in the middle stage (30%-40%) emission reduction is accelerated, and in the later stage (20%-30%) effect tends to be flat, and the linear model cannot capture this dynamic change.

[0083] To solve the above problems, the following adjustments are made to the model: introduce the square term of energy structure change rate, modify the peak time and peak value prediction formula: optimized peak time prediction ; optimized peak peak prediction ; Wherein, is the annual energy structure change rate (for example, "clean energy proportion in 2025 is 5% higher than in 2024, then = 5%, c, d are nonlinear term coefficients (fitted by historical transformation enterprise data, for example, the fitted value of shipbuilding and repair industry c = -0.02, d = -0.3, negative sign indicates the inhibitory effect of structure optimization on peak time and peak value). Redistribute the weight: based on the target feature information ("instability in the energy structure transformation process"), increase the weight of related feature variables. For example, increase the weight of "clean energy proportion change" from the original 0.15 to 0.25, and increase the weight of "energy transformation investment growth rate" from 0.08 to 0.18, to ensure that the model pays attention to key factors in the transformation period.

[0084] Using the adjusted model, the training sample set with target feature information is used again. For training sample set processing, the training sample set with target feature information needs to include special data of "energy structure transformation enterprises", for example, the sample coverage: 30 shipbuilding and repair enterprises in the transformation period (coal proportion is more than 40% before transformation, and decreases to less than 25% after transformation). Label "transformation fluctuation coefficient" (value 0-1, 0 means stable transformation, 1 means severe transformation, for example, a certain enterprise's coal proportion decreases by 15% within a year due to policy adjustment, and is labeled as 0.8).

[0085] The gradient descent method is used to minimize the prediction error (mean square error MSE), and the model parameters are iteratively updated. , , c, d, etc.), and the iteration termination condition is that the MSE decreases by less than 0.001 for 5 consecutive iterations. The optimized model (peak-reach accounting planning model) can accurately capture the nonlinear impact of energy structure transformation, significantly improve the accuracy of peak-reach prediction for enterprises in the transformation period, and provide reliable support for subsequent generation of targeted peak-reach path planning.

[0086] In S108, the peak-reach prediction feature vector, the peak-reach driving parameter information of the enterprise to be evaluated, the peak-reach measure parameter information of the target region, and the energy saving and emission reduction parameter information of the target region are processed based on the target peak-reach accounting planning model to generate target peak-reach path planning information.

[0087] In an embodiment, based on the target peak-reach accounting planning model, the peak-reach prediction feature vector, the peak-reach driving parameter information of the enterprise to be evaluated (current total energy consumption of 500,000 tons of standard coal, production scale of 1,200,000 deadweight tons / year), the peak-reach measure parameter information of the target region (photovoltaic power generation cost of 0.35 yuan / kWh, wind power on-grid price of 0.42 yuan / kWh), and the energy saving and emission reduction parameter information (regional renewable energy subsidy policy, carbon trading price of 80 yuan / ton CO2) are input to generate an energy structure adjustment path for shipbuilding and repair enterprises. As shown in Figure 3 The initial energy structure of the enterprise to be evaluated is as follows: coal accounts for 50% (used for welding preheating, painting drying, etc.), diesel accounts for 30% (ship power test, transportation vehicle energy consumption), and purchased electricity accounts for 20%. The total carbon emissions in 2020 reached 486,000 tons of CO2, with an annual growth rate of 8.2%. Combined with the resource conditions of the target region, such as the annual effective utilization hours of wind energy of 2,200 hours and the annual solar radiation of 52,000 megajoules / square meter, a clean energy replacement scheme with "photovoltaic + wind power + energy storage" as the core is determined, and shore power and intelligent energy consumption systems are matched to improve efficiency.

[0088] In the first 2 years, 8000 kW of roof photovoltaic power station (covering the roof of the hull workshop and painting workshop) will be built, 12 diesel welding machines will be transformed into electric welding machines, 2 MW / 4 MWh energy storage power station will be put into use to smooth photovoltaic fluctuations, the target is to make solar energy account for 10%, coal 44%, diesel 27%, and electricity 19%, it is expected to reduce 32,000 tons of CO2 emissions per year, reduce the carbon emission growth rate to 5.1%, invest 120 million yuan, apply for 30 million yuan of regional new energy subsidies, and connect with State Grid construction enterprise micro-grid access point. In the third and fourth years, 15 MW of distributed wind power projects will be introduced (co-constructed with regional energy companies), diesel transport vehicles will be completely replaced by electric vehicles, 5000 kW of shore power systems will be put into use (covering 3 ship repair wharfs), the target is to achieve 15% wind power, 15% solar energy, 30% coal, 20% diesel, and 20% electricity, a total of 128,000 tons of CO2 emissions will be reduced, the carbon emission growth rate will be reduced to 1.3%, an investment of 250 million yuan will be made, 40 million yuan will be offset by carbon trading income, and energy-saving reconstruction such as intelligent air compression station will be completed. By the fifth year, the photovoltaic power station will be expanded to 12,000 kW, the wind power will be expanded to 20 MW, the production energy will be 100% electrified (emergency diesel engine group is reserved), and the water reuse system coverage will be 80% to reduce indirect energy consumption, the target is 25% wind power, 20% solar energy, 20% coal, 10% diesel, and 25% electricity, the total carbon emissions will be reduced to 423,000 tons of CO2, a decrease of 6.2% from the peak, an investment of 180 million yuan will be made, green credit will be used to solve the funding gap, and the volatile organic compound emission compliance rate will be 100%. Through the implementation of this path, the peak curve presents a trend of "slow growth - peak stability - gradual decline": the carbon emission growth rate tends to be 0 in the third year, reaches a peak of 491,000 tons of CO2 in the fourth year, and starts to decline in the fifth year, a decrease of 4.3% from the initial trend, and the peak time is advanced by 2 years.

[0089] In combination with the target area industrial planning (encouraging new energy equipment manufacturing, green ship repair), the industrial structure transformation path is generated based on the target peak accounting planning model. The initial evaluation enterprise is mainly engaged in traditional ship segmentation manufacturing (accounting for 70% of production value), supplemented by steel component processing (accounting for 30% of production value), with unit production value carbon emission of 1.8 tons of CO2 / 10,000 yuan, which is much higher than the advanced level of the industry (1.1 tons of CO2 / 10,000 yuan). If the current situation is maintained, the peak time will be delayed to 2035, and the peak value will break through 600,000 tons of CO2. In the first to second year, two low-efficiency steel pretreatment production lines are eliminated (10% of production capacity is reduced), welding robots are introduced, and the utilization rate of steel is improved to 95%; at the same time, a green ship technology laboratory is co-built with scientific research institutions (an investment of 50 million yuan), and 80 technical personnel for LNG fuel tank installation are reserved. The goal is that traditional industry accounts for 90%, and emerging industry accounts for 10%, so that the unit production value carbon emission is reduced to 1.6 tons of CO2 / 10,000 yuan. In the third to fourth year, the steel processing capacity is reduced by 20%, focusing on high value-added ship segmentation, and the coating process is upgraded, so that VOCs emission is reduced by 30%; 210,000 tons of LNG dual-fuel bulk carrier key components are produced in small batches (tank installation technology is broken through), and wind power component capacity reaches 5,000 tons / year (supporting regional wind power projects), the goal is that traditional industry accounts for 60%, and emerging industry accounts for 40% (new energy equipment accounts for 30%), so that the unit production value carbon emission is reduced to 1.3 tons of CO2 / 10,000 yuan. By the fifth year, the general steel processing is completely withdrawn, and high-end ship segmentation manufacturing is retained (accounting for 30% of production value), and green transformation accounts for 60% in ship repair business; LNG dual-fuel ship component capacity is increased to 20,000 tons / year, ammonia fuel ship orders are accepted, and wind power component supporting rate covers 80% of the regional market, the goal is that traditional industry accounts for 30%, and emerging industry accounts for 70% (new energy equipment accounts for 50%), so that the unit production value carbon emission is reduced to 1.0 tons of CO2 / 10,000 yuan.

[0090] The industrial structure transformation and energy structure adjustment form a synergistic effect: the emerging industry value accounts for 40% in the fourth year, which drives the unit energy consumption to decrease by 18.3%, and the clean energy replacement effect makes the peak value decrease by 8.5% compared with the single energy adjustment path; in the fifth year, the high value-added new energy equipment manufacturing (gross profit rate of 25%, traditional business is 12%) provides financial support for subsequent carbon capture technology investment, and further consolidates the emission reduction results.

[0091] Based on the energy-industry collaborative transformation logic, combined with different implementation intensity and risk preference, three types of preset paths are derived through the target peak accounting planning model. Path A (aggressive) concentrates an investment of 420 million yuan in the first three years, making the photovoltaic + wind power installed capacity reach 23 MW (2 years ahead of schedule), starting the production line of LNG components in the first year, and the emerging industry accounting for 40% in the third year (1 year ahead of schedule), with an average annual investment intensity of 210 million yuan and a medium technology risk level. Path B (moderate) implements according to the benchmark pace, with the supporting energy storage ratio increased to 50% (enhancing energy stability), focusing on technology research and development in the first two years, starting to release capacity in the fourth year, and the emerging industry accounting for 40% in the fifth year, with an average annual investment intensity of 110 million yuan and a low technology risk level. Path C (emphasis) prioritizes the development of photovoltaic (75% of clean energy) in the first three years, and introduces wind power (reduces grid connection pressure) in the last two years, balancedly promotes ship green transformation and wind power component manufacturing, with an average annual growth rate of emerging industries of 15%, an average annual investment intensity of 150 million yuan, and a medium technology risk level.

[0092] The carbon emission curve of each path is calculated by the target peak accounting planning model. Path A peaks in the fourth year, with a peak value of 472,000 tons of CO2, a peak value of 19.0% lower than the benchmark, a cumulative investment of 105 million yuan in five years, an investment recovery period of 8.2 years, and a technology feasibility score of 7.2 (10-point system). Path B peaks in the fifth year, with a peak value of 498,000 tons of CO2, a peak value of 14.6% lower than the benchmark, a cumulative investment of 58 million yuan in five years, an investment recovery period of 11.5 years, and a technology feasibility score of 8.5. Path C peaks in the fourth and a half year, with a peak value of 485,000 tons of CO2, a peak value of 16.8% lower than the benchmark, a cumulative investment of 76 million yuan in five years, an investment recovery period of 9.8 years, and a technology feasibility score of 8.0. The benchmark scenario (without adjustment) peaks in the sixth year, with a peak value of 583,000 tons of CO2.

[0093] Path A achieves "peak as soon as possible, peak value is the lowest" through high-intensity investment, but the investment pressure is larger, which is suitable for enterprises with strong financial strength and sufficient technology reserves; Path B achieves the emission reduction target at a lower cost, which is suitable for enterprises with conservative risk preference; Path C balances the emission reduction effect and implementation difficulty, providing the best choice for most shipbuilding and repair enterprises. The quantitative data output by the model can directly support enterprises to choose the appropriate path according to their own resource endowments.

[0094] The peak time and peak value of different preset peak paths are generated by processing several preset peak paths respectively. For path A, according to the enterprise energy consumption data, production plan, and energy structure and industrial structure adjustment plan, combined with the algorithm in the target peak calculation planning model, the carbon emission curve over time is calculated. By analyzing the curve, it is found that the total carbon emission reaches a peak in the fourth year, i.e. the peak time is 4 years. This is because in path A, the energy structure adjustment and industrial structure transformation measures that are rapidly promoted in the early stage begin to play a significant role in reducing emissions in the fourth year, effectively curbing the growth trend of carbon emissions and making it start to decline. For path B, after a similar calculation process, the peak time is 5 years. Because its energy structure adjustment and industrial structure transformation are relatively stable, the emission reduction effect is not enough to make carbon emissions reach a peak until the fifth year. For path C, the peak time is calculated to be 4.5 years. The combination of the early-stage focus on solar power development and the late-stage introduction of wind power generation, as well as the steady promotion of industrial structure transformation measures, causes the carbon emission peak to appear in the 4.5th year.

[0095] When calculating the peak value of path A, according to the total carbon emission in the fourth year, it is assumed that the total carbon emission of the enterprise in that year is 800,000 tons of carbon dioxide equivalent (taking into account the comprehensive impact of energy structure adjustment and industrial structure transformation, such as the reduction of carbon emissions due to the replacement of part of fossil energy by clean energy, but there may be certain transition costs in the early stage of industrial transformation, resulting in incomplete carbon emission reduction according to the ideal state). For path B, the total carbon emission is calculated to be 750,000 tons of carbon dioxide equivalent when it peaks in the fifth year. Because of its stable strategy, the downward trend of carbon emissions is relatively gentle, and the peak is relatively low. Path C peaks in the 4.5th year, with a total carbon emission of 780,000 tons of carbon dioxide equivalent, which is between path A and path B, reflecting the influence of its unique energy structure and industrial structure adjustment strategy on the peak value.

[0096] The peak time and peak value of different preset peak paths are processed to generate target peak path planning information. Considering factors such as the economic bearing capacity of the enterprise, technical feasibility, market demand, and regional environmental requirements, the peak time and peak value of different preset peak paths are evaluated. Assuming that the enterprise wants to achieve peak in a short time to enhance its image and competitiveness in the current market competition environment, but also needs to consider that the economic cost cannot be too high. Comparing the three preset peak paths, path A has the shortest peak time (4 years), but the early-stage investment is large, which may put a lot of pressure on the enterprise's cash flow; path B has a longer peak time (5 years), although the economic cost is relatively low, but it may not meet the enterprise's demand for rapid peak; path C has a peak time of 4.5 years and a peak value of 780,000 tons of carbon dioxide equivalent, its early-stage investment is relatively small compared to path A, and it can also better meet the enterprise's demand for peak time. After comprehensive evaluation, path C is chosen as the target peak path.

[0097] The target peak path planning information includes detailed implementation steps and schedules. For example, in terms of energy structure adjustment, in the first to second year, 50 million yuan is invested in the construction of solar photovoltaic power generation systems, with an installed solar panel area of 20,000 square meters, which is expected to reduce coal consumption by 10,000 tons and carbon dioxide emissions by 25,000 tons per year; in the third to fourth year, wind power generation is introduced, and a cooperation agreement is signed with energy suppliers to invest in supporting power transmission and transformation facilities with an investment of 80 million yuan, which is expected to reduce coal consumption by 20,000 tons and diesel consumption by 10,000 tons per year, and carbon dioxide emissions by 60,000 tons per year, etc. In terms of industrial structure transformation, 20 million yuan is invested in technology research and development and talent recruitment in the first to second year, and a research and development center is established in cooperation with scientific research institutions; in the third to fourth year, wind turbine parts are started to be produced, and the production scale is gradually expanded, which is expected to increase annual output value by 50 million yuan, while reducing carbon emissions by 30,000 tons caused by steel production; in the fifth year, the scale of new energy equipment manufacturing industry is further expanded, with an output value of 200 million yuan, accounting for more than 50% of the total output value of the enterprise, and the total carbon emissions continue to decrease. In addition, the target peak path planning information also includes risk assessment and countermeasures in the implementation process, such as policy risks (such as changes in subsidy policies) and technical risks (such as unanticipated power generation efficiency) that new energy power generation projects may face, as well as corresponding countermeasures (such as closely following policy developments, strengthening technology research and development and monitoring, etc.), providing a comprehensive and feasible guidance scheme for the enterprise to achieve carbon emissions peak.

[0098] In another embodiment, to accurately calculate the peak time of shipbuilding and repair enterprises, the method constructs a dynamic calculation formula that integrates technical, policy, and economic multi-dimensional influencing factors. The formula is derived based on the dynamic principle of Kaya model, fully considering the periodicity of shipbuilding and repair enterprise production and the timeliness of emission reduction measures. The specific expression is as follows, and the calculation formula of peak time is: ; wherein, represents the carbon emission peak time (year) of the enterprise to be evaluated, represents the baseline time (year), combined with the carbon emission data statistical period of shipbuilding and repair enterprises; represents the technical influence coefficient, which reflects the emission reduction efficiency of key energy-saving technologies in shipbuilding and repair, with a value range of 0.1-0.9 (the larger the efficiency, the larger the value), calibrated based on industry measured data of welding energy-saving equipment, painting waste gas recovery, etc. represents the implementation progress coefficient, which represents the proportion of technology landing completion, with a value range of 0-1 (the higher the completion degree, the larger the value), quantified according to the stages of equipment installation, debugging, personnel training, and stable operation. represents the annual carbon emission reduction of a single technology (tons CO2 / year), obtained through equipment energy consumption comparison test and production process simulation calculation. The policy intensity coefficient represents the binding and incentive effects of regional carbon emission reduction policies, with a value range of 0.2-0.8 (the more comprehensive the policy, the larger the value). It is used to assess the overall strength of policies such as carbon quota allocation, new energy subsidies, and environmental penalties. The policy implementation coefficient represents the effectiveness of policy implementation at the enterprise level. Its value ranges from 0 to 1 (the more effective the implementation, the higher the value). It is calculated based on the policy application approval rate and the emission reduction task completion rate. This represents the annual carbon emission reduction (tons of CO2 / year) driven by policy, converted by combining carbon trading revenue and emission reduction investment leveraged by subsidies. The value represents the sensitivity of economic growth to carbon emissions, ranging from 0.3 to 0.7 (the higher the sensitivity, the larger the value), and is based on the historical correlation analysis results between the growth rate of shipbuilding and repair industry output and the growth rate of carbon emissions; G represents the average annual output growth rate of enterprises (%), which is determined based on the development plan of the shipbuilding industry in the target region and the order forecast of enterprises. , These represent the rate of technological iteration and the frequency of policy updates (% / year), respectively, with a default value of 3% / year set based on the industry's technological upgrade cycle and policy adjustment cycle.

[0099] Taking a medium-sized shipbuilding and repair enterprise (repairing 50 ships and building 3 ships annually) as an example, we selected the frequently used "welding robot + waste heat recovery" combined energy-saving technology (n=1), and calculated it in conjunction with the regional "dual carbon" policy package: [Baseline time] In 2020, the average annual output growth rate of enterprises was G=6.5%, and the sensitivity to economic growth was... =0.52 (corresponding to the average level in coastal shipbuilding-intensive areas); Technology Influence Coefficient =0.75 (actual measurements show it can reduce welding energy consumption by 35%), current technology implementation progress =0.5 (50 robots installed, 30 remaining to be deployed), annual emission reduction for a single technology =1200 tons of CO2 / year (calculated based on equipment operation data); Policy intensity coefficient =0.6 (The region implements a dual policy of carbon trading and new energy subsidies), policy implementation coefficient =0.8 (100% of enterprises enjoy subsidies and 100% carbon quota compliance rate), policy-driven annual emission reduction =800 tons of CO2 / year (calculated based on energy-saving renovation investment using carbon trading revenue).

[0100] Substitute the above parameters into the formula to calculate. =2020+ The results show that under the selected technology and policy combination, the enterprise is expected to peak carbon emissions in 2040, 15 years earlier than the baseline scenario without measures (peak in 2055), in line with the "Fourteenth Five-Year" emission reduction plan requirements of the shipbuilding and repair industry.

[0101] Based on the characteristics of shipbuilding and repair enterprises, such as "energy consumption is concentrated, and industrial links are closely related", a peak calculation formula is constructed based on the logic of industrial-energy coordinated emission reduction, which comprehensively reflects the superposition effect of structural adjustment and technological upgrading. The specific expression is as follows: Peak calculation formula , where, represents the carbon emission peak of the enterprise to be evaluated (tons of CO2); represents the initial carbon emission level in the base year (tons of CO2), which is calculated based on the fossil energy consumption and production process emission data in the enterprise energy audit report; represents the industrial structure coefficient, which reflects the emission reduction potential of industrial transformation, with a value range of 0.2-0.8 (the higher the proportion of high-value-added industries, the higher the value), and reference is made to the correlation analysis of the industrial composition and unit value carbon emission of shipbuilding and repair enterprises; represents the industrial transformation progress coefficient, which represents the completion proportion of industrial structure adjustment, with a value range of 0-1 (the more in-depth the transformation, the higher the value), which is quantified according to the proportion of emerging industry output value and the elimination rate of backward production capacity.

[0102] represents the annual carbon emission reduction of industrial structure adjustment (tons of CO2 / year), which is calculated by the difference between the unit value carbon emissions of new and old industries and the output value scale; represents the energy structure coefficient, which reflects the emission reduction effect of energy substitution, with a value range of 0.3-0.9 (the higher the proportion of clean energy, the higher the value), which is calculated based on the carbon emission factor of different energy categories; represents the energy structure adjustment progress coefficient, which represents the completion proportion of clean energy substitution, with a value range of 0-1 (the more complete the substitution, the higher the value), which is calculated according to the proportion of clean energy consumption; represents the annual carbon emission reduction of energy structure adjustment (tons of CO2 / year), which is calculated by the difference between the amount of fossil energy substitution and the carbon emission factor; represents the lag time of emission reduction measures (years), which reflects the effectiveness period of technology landing and structural adjustment, with a default value of 3 years for the shipbuilding and repair industry; represents the transformation implementation period (years), which is set in combination with the actual period of enterprise capacity renewal and equipment modification, with a default value of 5 years.

[0103] Continuing the above case of a medium-sized shipbuilding and repair enterprise, the industrial structure transformation and energy structure adjustment plan (m=1, i.e. single transformation combination scheme) is calculated. For the basic parameter value, the initial carbon emission level in the base year =32000 tons of CO2 (2020 measured value), lag time of emission reduction measures =3 years, transformation implementation period =5 years. For the relevant parameters of the industrial structure, the industrial structure coefficient =0.65 (planning to develop new energy ship component manufacturing, reducing carbon emissions by 40% per unit of output value), current transformation progress =0.4 (new energy component output value accounts for 25%), annual emission reduction amount of industrial adjustment =1800 tons of CO2 / year (based on the replacement scale of old and new capacity). For the relevant parameters of the energy structure, the energy structure coefficient =0.7 (planning photovoltaic + wind power accounts for 45%), current adjustment progress =0.3 (clean energy accounts for 15%), annual emission reduction amount of energy adjustment =2200 tons of CO2 / year (based on the amount of coal and diesel replaced).

[0104] Substituting the above parameters into the formula gives, 32000 x (1+0.52 x 6.5% x 3) - (0.65 x 0.4 x 1800 x 5 + 0.7 x 0.3 x 2200 x 5) = 32000 x 1.101 - (2340 + 2310) = 35232 - 4650 = 30582 tons of CO2. This result shows that under the industrial-energy collaborative transformation scheme, the enterprise carbon emission peak is about 30582 tons of CO2, which is 20.8% lower than the baseline scenario peak (38600 tons of CO2), verifying the emission reduction effectiveness of the "new energy equipment manufacturing + clean energy replacement" combination strategy. The error is less than 0.1% compared with the peak prediction result (30600 tons of CO2) in the path planning in the foregoing.

[0105] The application obtains enterprise information to be evaluated, regional information, carbon emission information, a preset peak reaching calculation planning model and a training sample set. Then, key parameters and tasks are generated by processing these data: enterprise information is processed to obtain peak reaching driving parameter information, which includes obtaining attribute information (such as equipment, capacity, energy consumption and production process parameters) from enterprise information, and then generating target parameter information and weight information, and finally obtaining peak reaching driving parameter information; regional information is processed to obtain peak reaching measure parameter information and energy saving and emission reduction parameter information of the target region, which involves extracting natural, ecological and geographical factor information from regional information, and further generating renewable energy evaluation information, regional restriction parameter information and the like; carbon emission information is processed to obtain interval peak reaching prediction subtasks and sequential access peak reaching prediction subtasks.

[0106] Then, the feature vectors are extracted from the prediction subtasks, including feature extraction on the interval peak prediction subtask, dimension reduction processing on the sequential access peak prediction subtask, and the like to obtain the peak prediction feature vectors. The training sample set is processed to generate a training sample set with target feature information, including grouping, feature extraction, generation of a training set and a validation set, classifier prediction, and algorithm training, and the like. Finally, based on the target peak accounting planning model, the peak prediction feature vectors, enterprise peak driving parameters, regional peak measures, and energy saving and emission reduction parameters are combined to generate an energy and industrial structure adjustment path, and further obtain a preset peak path, a peak time, and a peak value, and finally determine the target peak path planning information. Based on the embedded industry emission reduction technology database, the optimization target and constraint conditions are established to generate energy-saving technologies referenced by shipbuilding and repairing enterprises, and energy-saving technology reference conditions and judgment logic are generated. The optimal execution solution set and optimization scheme are proposed to realize self-optimization and intelligent optimization with the "adaptive" feature, and the optimal carbon emission execution route, the best peak cost execution route, the optimal economic benefit execution route, and the optimal peak time execution route are proposed. The key energy consumption links of enterprise production are digitally integrated, and scientific reference conditions and judgment logic are proposed for the innovative energy-saving technologies and energy-saving equipment suitable for the industry, and the information of energy-saving technologies and energy-saving equipment is successfully integrated, shared, and bidirectionally matched. Goodbye to the "single and one-sided" situation, the enterprise carbon reduction route has a systematic analysis and scientific diagnosis, which greatly improves the scientificity and systematicness of the enterprise carbon reduction path, and also sets an example for the digital transformation of the industry.

[0107] In an embodiment, as shown in Figure 2 The application also provides a peak path planning device suitable for shipbuilding and repairing enterprises, comprising: The acquisition module 201 is configured to acquire the to-be-evaluated enterprise information, the regional information of the to-be-evaluated enterprise, the carbon emission information of the to-be-evaluated enterprise in a preset time period, a preset peak accounting planning model, and a training sample set. The processing module 202 is configured to process the to-be-evaluated enterprise information to generate peak-reaching driving parameter information of the to-be-evaluated enterprise, process regional information of the to-be-evaluated enterprise to generate peak-reaching measure parameter information of a target region and energy-saving and emission-reducing parameter information of the target region, process carbon emission information of the to-be-evaluated enterprise in a preset time period to generate an interval peak-reaching prediction subtask and a sequential access peak-reaching prediction subtask, process the interval peak-reaching prediction subtask and the sequential access peak-reaching prediction subtask to generate a peak-reaching prediction feature vector, process the training sample set to generate a training sample set with target feature information, wherein the target feature information is used to represent risk factors of an abnormal state of emission reduction effect, process the preset peak-reaching accounting planning model based on the training sample set with the target feature information to generate a target peak-reaching accounting planning model, and process the peak-reaching prediction feature vector, the peak-reaching driving parameter information of the to-be-evaluated enterprise, the peak-reaching measure parameter information of the target region, and the energy-saving and emission-reducing parameter information of the target region based on the target peak-reaching accounting planning model to generate target peak-reaching path planning information.

[0108] Each of the embodiments in the present application is described in a related manner, and the same or similar parts between the embodiments can be referred to each other. Each of the embodiments focuses on the difference from other embodiments. In particular, the peak-reaching path planning method, the electronic device, the electronic equipment, and the readable storage medium for evaluating the shipbuilding and repairing enterprise are basically similar to the peak-reaching path planning method for the shipbuilding and repairing enterprise described above, so the description is relatively simple, and the relevant parts can be referred to the description of the peak-reaching path planning method for the shipbuilding and repairing enterprise.

Claims

1. A peak-reaching path planning method applicable to shipbuilding and repair enterprises, characterized in that, include: Obtain information on the companies to be assessed, their regional information, their carbon emission information within a preset time period, a preset peak emission accounting planning model, and a training sample set; The information of the enterprise to be evaluated is processed to generate the peak driving parameter information of the enterprise to be evaluated; The regional information of the enterprise to be evaluated is processed to generate peak emission control parameter information and energy conservation and emission reduction parameter information for the target region. The carbon emission information of the enterprise to be evaluated within a preset time period is processed to generate an interval peak prediction subtask and a sequential access peak prediction subtask. The interval peak prediction subtask and the sequential access peak prediction subtask are processed to generate a peak prediction feature vector. The training sample set is processed to generate a training sample set with target feature information, wherein the target feature information is used to characterize the risk factors that cause the emission reduction effect to be in an abnormal state. The preset peak calculation planning model is processed based on the training sample set with target feature information to generate the target peak calculation planning model. Based on the target peak achievement accounting planning model, the peak achievement prediction feature vector, the peak achievement driving parameter information of the enterprise to be evaluated, the peak achievement measure parameter information of the target area, and the energy conservation and emission reduction parameter information of the target area are processed to generate target peak achievement path planning information.

2. The method as described in claim 1, characterized in that, The information of the enterprise to be evaluated is processed to generate peak-driving parameter information for the enterprise to be evaluated, including: The information of the enterprise to be evaluated is processed to generate the attribute information of the enterprise to be evaluated, wherein the attribute information of the enterprise to be evaluated includes equipment list information, production capacity data information, energy consumption parameters, and production process parameters; The equipment list information, the production capacity data information, the energy consumption parameters, and the production process parameters are processed to generate target parameter information and weight information that matches the target parameter information for the enterprise to be evaluated. The target parameter information and the weight information that matches the target parameter information of the enterprise to be evaluated are processed to generate the peak driving parameter information of the enterprise to be evaluated.

3. The method as described in claim 1, characterized in that, The regional information of the enterprises to be evaluated is processed to generate peak emission reduction parameter information and energy conservation and emission reduction parameter information for the target region, including: The regional information of the enterprise to be evaluated is processed to generate natural environmental factor information, ecosystem factor information, and geographical factor information of the target area. The natural environmental factors information of the target area is processed to generate assessment information on regional renewable energy resources; The ecosystem factor information of the target area is processed to generate regional limiting parameter information, which includes industrial structure planning parameter information and enterprise energy conservation and emission reduction potential parameter information. The assessment information of renewable energy resources in the region and the industrial structure planning parameters are processed to generate peak-reaching measure parameters for the target region. The geographical factors of the target area and the energy conservation and emission reduction potential parameters of the enterprise are processed to generate energy conservation and emission reduction parameter information for the target area.

4. The method as described in claim 1, characterized in that, The carbon emission information of the enterprise to be evaluated within a preset time period is processed to generate an interval peak prediction subtask and a sequential access peak prediction subtask, including: The carbon emission information of the enterprise to be evaluated within a preset time period is processed to generate carbon emission trend factor, interval peak assessment factor, carbon emission target node information, peaking sequence information of carbon emission target nodes, and peaking time information of carbon emission target nodes. The carbon emission trend factor and the interval peak assessment factor are processed to generate an interval peak prediction subtask. The carbon emission target node information, the peaking sequence information of the carbon emission target nodes, and the peaking time information of the carbon emission target nodes are processed to generate a sequential access peaking prediction subtask.

5. The method as described in claim 4, characterized in that, The interval peak prediction subtask and the sequential access peak prediction subtask are processed to generate a peak prediction feature vector, including: The interval peak prediction subtask is subjected to feature extraction processing to generate access time interval sequence features, access interval number features, and access interval information features; The access interval information features are processed to generate adjacent interval difference sequence information, difference sequence mean, and difference sequence variance; The sequential access peak prediction subtask is dimensionality reduced to generate a one-dimensional prediction feature vector. The one-dimensional predicted feature vector is subjected to feature extraction processing to generate target access list information, mean of adjacent access information similarity sequence, and variance of adjacent access information similarity sequence; The mean and variance of the access time interval sequence features, access interval number features, access interval information features, target access list information, and adjacent access information similarity sequences are processed to generate a peak prediction feature vector.

6. The method as described in claim 1, characterized in that, The training sample set is processed to generate a training sample set with target feature information, including: The training sample set is grouped to generate a grouped training sample set, wherein the grouped training sample set includes peak characteristic information of different regions and enterprises; The grouped training sample set is processed to extract features and generate the original feature library. The original feature library is processed to generate a training set and a validation set; The validation set is then used to perform prediction processing based on the classifier to generate prediction results. The training set is trained based on a preset algorithm to generate prediction results for the validation set. The prediction results and the validation set prediction results are processed to generate target feature information, which is used to characterize the risk factors that cause the emission reduction effect to be in an abnormal state.

7. The method as described in claim 1, characterized in that, Based on the target peak achievement accounting planning model, the peak achievement prediction feature vector, the peak achievement driving parameter information of the enterprise to be evaluated, the peak achievement measure parameter information of the target area, and the energy conservation and emission reduction parameter information of the target area are processed to generate target peak achievement path planning information, including: Based on the target peak accounting planning model, the peak prediction feature vector, the peak driving parameter information of the enterprise to be evaluated, the peak measure parameter information of the target area, and the energy conservation and emission reduction parameter information of the target area are processed to generate energy structure adjustment path and industrial structure transformation path; The energy structure adjustment path and the industrial structure transformation path are processed to generate several preset peak-reaching paths; Several preset peak-reaching paths are processed to generate peak-reaching times and peak values ​​for different preset peak-reaching paths; The peak time and peak value of different preset peak-reaching paths are processed to generate target peak-reaching path planning information; The method also includes a calculation formula for obtaining the peak time, the calculation formula being: ； in, Represents the base time. Represents the technological impact coefficient. Represents the implementation progress coefficient. Represents carbon emission reduction. This represents the sensitivity of economic growth to carbon emissions. Represents the policy intensity coefficient; The method also includes a calculation formula for obtaining the peak value, the calculation formula being: ； in, Represents the initial carbon emission level. Represents the industrial structure coefficient. Represents the progress coefficient of industrial transformation. This represents the emission reductions resulting from industrial restructuring. Represents the energy structure coefficient. This represents the progress coefficient of energy structure adjustment.

8. A peak-reaching path planning device suitable for shipbuilding and repair enterprises, characterized in that, The device includes: The acquisition module is used to acquire information about the enterprise to be assessed, the regional information of the enterprise to be assessed, the carbon emission information of the enterprise to be assessed within a preset time period, the preset peak accounting planning model, and the training sample set. The processing module is used to process the information of the enterprise to be evaluated to generate peak emission driving parameter information for the enterprise; process the regional information of the enterprise to be evaluated to generate peak emission measure parameter information and energy conservation and emission reduction parameter information for the target region; process the carbon emission information of the enterprise to be evaluated within a preset time period to generate interval peak emission prediction sub-tasks and sequential access peak emission prediction sub-tasks; process the interval peak emission prediction sub-tasks and the sequential access peak emission prediction sub-tasks to generate peak emission prediction feature vectors; process the training sample set to generate a training sample set with target feature information, wherein the target feature information is used to characterize the risk factors of abnormal emission reduction effects; process the preset peak emission accounting planning model based on the training sample set with target feature information to generate a target peak emission accounting planning model; and process the peak emission prediction feature vector, the peak emission driving parameter information of the enterprise to be evaluated, the peak emission measure parameter information of the target region, and the energy conservation and emission reduction parameter information of the target region based on the target peak emission accounting planning model to generate target peak emission path planning information.

9. An electronic device, characterized in that, include: First processor; and memory for storing executable instructions of the first processor; The first processor is configured to execute the peak path planning method for shipbuilding and repair enterprises according to any one of claims 1 to 7 by executing the executable instructions.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the second processor, it implements the peak-reaching path planning method for shipbuilding and repair enterprises as described in any one of claims 1 to 7.

Images