CN-121094238-B - Multi-source comprehensive energy system energy-saving optimization platform based on data mining

CN121094238BCN 121094238 BCN121094238 BCN 121094238BCN-121094238-B

Abstract

The invention relates to the technical field of data mining, and discloses a multi-source comprehensive energy system energy-saving optimization platform based on data mining, which comprises the following steps of acquiring upstream energy components, node flow, settlement reference values, system routing weights and other characteristic parameters through a data acquisition module; the data processing module identifies the propagation characteristics of each node component, generates an energy characteristic time sequence and constructs a system level deviation and energy reconciliation residual sequence, the association analysis module extracts hysteresis characteristic quantity under characteristic frequency, establishes association relation between the hysteresis characteristic quantity and system energy consumption, generates an energy consumption association value, the optimization decision module takes system parameters as variables, minimizes the energy cost and the energy consumption association value under operation constraint, and the execution module applies an optimal decision result to the system to realize energy conservation and operation optimization.

Inventors

WANG KEWEN
LI BO
GUO HONGJIE
Hou Jiake
LI YU

Assignees

北京一控系统技术有限公司

Dates

Publication Date: 20260508
Application Date: 20251028

Claims (5)

1. The multi-source comprehensive energy system energy-saving optimization platform based on data mining is characterized by comprising: the data acquisition module acquires characteristic parameters of a target system, wherein the characteristic parameters comprise an upstream energy component time sequence, each node flow time sequence, a settlement reference value time sequence and a system routing weight time sequence; The data processing module is used for identifying component transmission characteristic parameters of each node by combining the characteristic parameters, generating a node energy characteristic time sequence, and constructing a system level deviation time sequence and an energy reconciliation residual time sequence, and comprises the following steps: Constructing a system level deviation time series P xt (t): ; Wherein P xt (t) is the value of the system level deviation time sequence at the time t, mu jd,i (t) is the value of the system routing weight time sequence at the time t of the ith node, Q jd,i (t) is the value of the volume heat value time sequence of the ith node at the time t, and R (t) is the value of the settlement reference value time sequence at the time t; Constructing an energy reconciliation residual time sequence L xt (t): ; Wherein, L xt (t) is the value of the energy reconciliation residual time sequence at the time t, and F jd,i (t) is the value of the traffic time sequence of the ith node at the time t; The association analysis module extracts characteristic parameters of the system level deviation time sequence and the energy reconciliation residual time sequence under characteristic frequency, and determines corresponding hysteresis characteristic quantity, and the association analysis module comprises the following steps: The projection quantity is calculated in a preset time window T ck for the system level deviation time sequence and the energy reconciliation residual time sequence, and the projection quantity is specifically as follows: ; ; ; ; Wherein S zp is the sine projection amount of P xt (t) at the preset characteristic frequency f tz , S yp is the cosine projection amount of P xt (t) at the preset characteristic frequency f tz , S zc is the sine projection amount of L xt (t) at the preset characteristic frequency f tz , and S yc is the cosine projection amount of L xt (t) at the preset characteristic frequency f tz ; based on the projection amount, the main frequency amplitude, phase and phase difference under the characteristic frequency are calculated: ; ; ; ; ; Wherein A pc is the main frequency amplitude of P xt (t) at the preset characteristic frequency f tz , A cc is the main frequency amplitude of L xt (t) at the preset characteristic frequency f tz , For the dominant frequency phase of P xt (t) at the preset characteristic frequency f tz , For the dominant frequency phase of L xt (t) at the preset characteristic frequency f tz , For the phase difference, N ck represents the number of sampling instants within the preset time window T ck ; based on the main frequency amplitude and the phase difference, calculating hysteresis characteristic quantity of a preset time window: ; ; Wherein S zh is hysteresis area, and K zh is main axis ratio; Combining the hysteresis area and the spindle ratio to obtain hysteresis characteristic quantity; establishing an association relation between hysteresis characteristic quantity and system energy consumption based on historical data, and generating an energy consumption association value; The optimization decision module takes the system routing weight time sequence and the settlement reference value adjustment amount time sequence as decision variables, and minimizes the system energy cost and the energy consumption correlation value under the constraint of system operation to obtain target decision variables; and the execution module is used for executing the target decision variable to the target system.
2. The data mining-based multi-source integrated energy system energy conservation optimization platform of claim 1, wherein the generating the node energy characteristic time sequence by identifying component propagation characteristic parameters of each node in combination with the characteristic parameters comprises: calculating normalized cross-correlation coefficient ρ jd,i (τ): ; Wherein C (t) represents an upstream energy component time series, F jd,i (t+τ) represents a flow time series of the i-th node, Represents the average value of the upstream energy component time series in the preset time window T ck , The average value of the flow time sequence of the ith node in a preset time window Tck is represented, t represents the sampling moment, and tau is the set time offset; Taking τ which enables ρ jd,i (τ) to reach the maximum value as a delay parameter τ jd,i of the ith node; Obtaining a first-order hysteresis time constant T jd,i of an ith node based on fitting of a discrete first-order hysteresis equation, wherein the discrete first-order hysteresis equation is as follows: ; wherein y jd,i (t) is a component time sequence of the ith node, and Δt represents a sampling time interval; Determining a time constant objective function based on the discrete first order lag equation as follows: ; Obtaining a first-order lag time constant T jd,i by minimizing a time constant objective function, and further generating a component time sequence y jd,i (T) of the ith node; Generating a node energy characteristic time sequence, wherein the node energy characteristic time sequence comprises a node component time sequence y jd,i (t), a node volume heat value time sequence Q jd,i (t) and a node Wobbe index time sequence W jd,i (t); Node volume heating value time series ; Wherein Q base represents a volume heating value reference constant, and k r represents a volume heating value sensitivity coefficient; Node Wash ratio index time series ; Wherein W base represents the wobbe index reference constant, k w represents the wobbe index sensitivity coefficient; the node component time sequence reflects the component state of the energy source at the ith node, the node volume heat value time sequence reflects the heat characteristic of the energy source unit volume at the ith node, and the node Wobbe index time sequence reflects the combustion stability related characteristic of the energy source at the ith node.
3. The data mining-based multi-source integrated energy system energy-saving optimization platform of claim 2, wherein establishing a relationship between hysteresis characteristic quantity and system energy consumption based on historical data, generating an energy consumption relationship value, comprises: determining the energy consumption of a target system at each moment to form a system energy time sequence H xt (t); Determining the total system energy consumption E all of the system energy time sequence H xt (T) within a preset time window T ck ; obtaining M history calculation windows in a history time period, and extracting hysteresis characteristic quantity corresponding to the mth history calculation window, namely hysteresis area And main axis ratio Wherein the size of the history calculation window is equal to a preset time window; The design matrix X ls and the observation vector y ls are constructed based on hysteresis characteristic quantity of each historical calculation window: , ; calculating a parameter vector using least squares ; Wherein, the Representing a transpose operation; extracting a first element in a parameter vector theta gl to obtain a Chang Liangxiang parameter theta cl ; extracting a second element in the parameter vector theta gl to obtain a hysteresis area coefficient theta mj ; extracting a third element in the parameter vector theta gl to obtain a spindle ratio coefficient theta zz ; Based on the hysteresis area and the spindle ratio, calculating an energy consumption correlation value by combining the parameter vector theta gl : ; wherein V nhgl represents an energy consumption correlation value of a preset time window.
4. The data mining-based multi-source integrated energy system energy-saving optimization platform of claim 3, wherein the system energy cost and energy consumption correlation value are minimized under the constraint of system operation by taking a system routing weight time sequence and a settlement reference value adjustment amount time sequence as decision variables, and the target decision variables are obtained, comprising: Setting decision variables, wherein the decision variables comprise a value mu jd,i (t) of an ith node at a time t and a value delta R (t) of a settlement reference value adjustment quantity time sequence at the time t in a system routing weight time sequence; defining post-adjustment time series for settlement references ; Constructing a system level deviation adjustment version time sequence and an energy reconciliation residual error adjustment version time sequence for optimization calculation: ; ; Wherein P xttz (t) is the value of the system level deviation adjustment version time sequence at the time t, L xttz (t) is the value of the energy reconciliation residual error adjustment version time sequence at the time t, Under a preset time window and a preset characteristic frequency, respectively calculating a corresponding hysteresis area adjusting value S tzzh and a corresponding spindle ratio adjusting value K tzzh based on the system level deviation adjusting version time sequence and the energy reconciliation residual error adjusting version time sequence; Calculating a corresponding adjustment plate energy consumption related value based on the hysteresis area adjustment value S tzzh and the spindle ratio adjustment value K tzzh ; Defining the energy cost of the system within a preset time window: ; Wherein C nl represents system energy cost; Constructing a decision variable objective function J: ; Wherein lambda nl represents the weight coefficient of the energy cost of the system, lambda gl represents the weight coefficient of the energy consumption correlation value of the adjusting plate; The system operation constraints include: ; ; ; wherein B min (t) represents the minimum value of the settlement reference value adjustment amount, and B max (t) represents the maximum value of the settlement reference value adjustment amount; Solving a time series of routing weights that minimizes the decision variable objective function J And settlement reference value adjustment amount time series As a target decision variable.
5. The data mining-based multi-source integrated energy system energy conservation optimization platform of claim 4, wherein executing the target decision variable on the target system comprises: and issuing the target decision variable to a target system as a control instruction.

Description

Multi-source comprehensive energy system energy-saving optimization platform based on data mining Technical Field The invention relates to the technical field of data mining, in particular to a multi-source comprehensive energy system energy-saving optimization platform based on data mining. Background Currently, the operation management of multi-source energy systems relies on energy management systems or comprehensive energy optimization platforms. Such platforms are typically based on static models, with energy balance calculations being performed on the basis of the energy conversion coefficients of each energy carrier being unchanged. For example, in an electricity-gas-heat cogeneration system, the gas heating value is considered to be constant, and in actual operation, the above-described basis is often distorted. The energy content per unit volume varies with time due to differences in gas sources (e.g., different natural gas wells, hydrogen loading ratios, or fluctuations in gas quality). In addition, the gas storage and pressure change in the pipe network also introduce transmission delay, so that the effective energy actually obtained by the downstream user is inconsistent with the upstream settlement caliber. The above phenomenon leads to systematic deviations in energy system scheduling and energy efficiency assessment: 1. The same energy supply strategy shows differentiated energy efficiency under different component conditions; 2. the energy-saving reconstruction measures are difficult to verify accurately; 3. the cost and energy efficiency accounting across energy types lacks a unified, dynamic energy mapping basis. As the specific gravity of renewable energy and hydrogen energy in integrated energy systems increases, the volatility and time-varying nature of the gas source increases significantly. When the composition of the upstream gas source changes, the heating value of the gas and its contribution to the overall energy balance also changes. However, the scheduling and settlement algorithms of current systems do not reflect the changes in real time, but rather follow a constant basis for energy conversion coefficients. Such deviations can accumulate and amplify within the system due to the energy transfer relationship between the different energy carriers. The concrete steps are as follows: 1. The power side load prediction is based on outdated energy conversion coefficients; 2. the distribution of the thermal side energy does not consider the real-time change of the heat value of the gas; 3. the energy cost term in the economic dispatch objective function is disjointed from the actual energy. When these deviations are superimposed, the system may experience a significant energy mismatch, i.e. the energy recorded by the settlement layer does not correspond to the effective energy actually consumed by the terminal. This not only results in an unrealistic reflection of the energy saving potential, but also may lead to misleading of the scheduling strategy, resulting in energy waste and economic loss. Disclosure of Invention The invention provides a multi-source comprehensive energy system energy-saving optimization platform based on data mining, which solves the technical problems of how to identify energy mismatching rules caused by energy component change from multi-source energy operation data by a data mining means and optimize energy distribution and settlement parameters at a system level so as to reduce energy-saving loss caused by energy deviation. The invention provides a multi-source comprehensive energy system energy-saving optimization platform based on data mining, which comprises the following steps: the data acquisition module acquires characteristic parameters of a target system, wherein the characteristic parameters comprise an upstream energy component time sequence, each node flow time sequence, a settlement reference value time sequence and a system routing weight time sequence; The data processing module is used for identifying component transmission characteristic parameters of each node by combining the characteristic parameters, generating a node energy characteristic time sequence, and constructing a system level deviation time sequence and an energy reconciliation residual time sequence; The correlation analysis module is used for extracting characteristic parameters of the system level deviation time sequence and the energy reconciliation residual error time sequence under characteristic frequency and determining corresponding hysteresis characteristic quantity; The optimization decision module takes the system routing weight time sequence and the settlement reference value adjustment amount time sequence as decision variables, and minimizes the system energy cost and the energy consumption correlation value under the constraint of system operation to obtain target decision variables; and the execution module is used for executing the target decision variable to