Spectral matching analysis method for correlated response of building multiple loads

Abstract

This invention discloses a spectral matching analysis method for the correlation response of multiple building loads, comprising the following steps: constructing a pairwise spectral analysis model of the predicted load curves of various internal equipment within the building energy system, including electric vehicle charging load and air conditioning load, and the photovoltaic load curve; establishing an internal spectral transfer mechanism within the pairwise spectral analysis model to form normalized response similarity weights, and extracting the single building equipment load corresponding to the maximum value from the response weight array to form an optimal synchronous match with the photovoltaic load; based on the mismatch difference between the single load with the maximum response similarity and the photovoltaic load, formulating a dynamic positive difference spectral iterative combination method to achieve coordinated operation of internal building equipment and maximize the photovoltaic absorption rate.

Description

TECHNICAL FIELD

[0001] The application belongs to the research field of building demand response, and particularly relates to a spectrum matching analysis method for building multi-element load correlation response. BACKGROUND

[0002] In recent years, the global energy resource technology revolution is being promoted to solve the most common energy and environmental problems at the present stage, so as to further achieve the goal of reducing energy consumption and carbon emission. About 40% of the total energy consumption in the world is related to the building industry, and reducing building energy consumption has become an important part of the process of reducing energy consumption and carbon emission. With the increasingly broad application prospect of renewable energy, the connection between the supply side and the demand side of building energy is becoming closer, and the application of photovoltaic systems in buildings is considered an important method for reducing energy consumption and carbon emission. More and more researchers are paying attention to the combination of photovoltaic resources and building operation, and are researching and developing economic and efficient and controllable building energy demand response plans.

[0003] With the popularity of charging facilities in buildings and large-scale grid-connected photovoltaic power generation, in order to improve the local consumption of photovoltaic power, electric vehicles, air conditioners and other building sub-equipment workloads are often coordinated with photovoltaic energy operation, but at this time the relationship between the building power loads becomes more complex. Quantitative evaluation of the flexibility potential and response level of building energy systems is very important in energy management. In order to evaluate the consumption of building photovoltaic power, the time domain analysis method is used to qualitatively evaluate the response matching degree of each device sub-load and photovoltaic power generation, which has been difficult to meet the demand, and the method also has some limitations in the characterization of the characteristics of each load and their mutual matching relationship, which to some extent affects the selection and execution of comprehensive optimization scheduling strategy. Therefore, considering the balance between supply and demand, the complementary effect of multiple types of loads, it is of great significance to establish a demand matching model for building multi-element load correlation response to realize the optimal operation of power load scheduling. SUMMARY

[0004] Therefore, the application provides a spectrum matching analysis method for building multi-element load correlation response, which cooperates with the correlation response operation of building multi-element load, maximizes the improvement of photovoltaic consumption effect, and realizes the optimal operation of power load scheduling.

[0005] To achieve the above purpose, the application provides a spectrum matching analysis method for building multi-element load correlation response, which includes the following steps:

[0006] Step 1: Based on the predicted curve of different equipment sub-power load in building energy system, the paired spectral analysis model between building socket sub-power load including air conditioning load and electric vehicle charging load and photovoltaic power curve is constructed. The paired spectral analysis model contains the spectral density distribution, coherence and phase relationship of paired load sequence. The spectral density distribution is used to reflect the change degree of paired load amplitude with frequency distribution, as shown in equation (1):

[0007]

[0008] In the formula, D Lxy (f) XY (s) is the cross-covariance function; s is the time delay; f is the frequency.

[0009] The expression form of the paired spectral density distribution is transformed by using polar coordinate representation method, as shown in equations (2)-(4):

[0010] D Lxy (f) Lxy (f)exp{2πW Lxy (f)} (2)

[0011]

[0012] W Lxy (f) Lxy (f) / m Lxy (f)] (4)

[0013] Where, since D Lxy (f) is a complex number, the real part is m Lxy (f), and the imaginary part is n Lxy (f); A Lxy (f) is the amplitude relationship between paired load sequences; W Lxy (f) is the phase relationship between paired load sequences.

[0014] Then, the frequency domain correlation of paired load curve is identified based on the coherence spectrum, as shown in equation (5):

[0015]

[0016] In the formula, D Lx (f) and D Ly (f) are single load spectral density distributions; G Lxy (f) is the coherence spectrum of paired load curve, reflecting the linear correlation degree of two sequences at different frequencies. The closer the coherence spectrum value is to 1, the greater the correlation of paired load sequence at the frequency is.

[0017] Meanwhile, the phase spectrum is used to analyze the frequency domain sequence of the paired load, as shown in equation (4). The spectrum parameter reflects the phase difference of the paired load sequence at each frequency, i.e., the sequence of the paired load. The closer the phase spectrum value is to 0, the weaker the leading or lagging degree of the paired load curve at the frequency.

[0018] Finally, the three spectrum parameters are comprehensively analyzed to study the correlation characteristics between the power sub-load of various devices in the building and the photovoltaic power generation curve.

[0019] Step 2: After obtaining the spectrum density distribution, coherence, and phase relationship three parameters, the internal frequency spectrum transfer mechanism of the paired spectrum analysis model is established. First, the error of the paired spectrum density value of the building different sub-device operation load sequence and the photovoltaic power generation curve at the frequency is calculated, and the common frequency band interval under the set error σ is taken. Then, the correlation and phase relationship of the paired load sequence are further compared, the coherence spectrum and phase spectrum values corresponding to the common frequency band interval are intercepted, and the extracted data mean value is processed and then normalized to form the normalized response similarity weight Δ i , as shown in equation (6).

[0020]

[0021] , where Δ i is the coherence and phase normalized response similarity weight of the i-th paired load sequence, Δ i ∈ [0, 1]; G i is the coherence mean value of the i-th paired load sequence spectrum analysis, G i ∈ [0, 1]; W i is the phase mean value of the i-th paired load sequence spectrum analysis, W i ∈ [0, 1]; and n is the number of paired load sequence combinations.

[0022] Finally, the power load of multiple building devices is paired with the photovoltaic power generation curve for spectrum combination calculation to obtain the normalized response similarity weight array, and the maximum value corresponding to the single building device load in the array is extracted to form the optimal synchronous matching with the photovoltaic load, as shown in equation (7).

[0023] L x = max{Δ1, Δ2,..., Δ n} (7)

[0024] , where L x is the single building device load sequence corresponding to the maximum value extracted in the array.

[0025] Step 3: Based on the spectrum transfer mechanism, if the normalized response similarity weight Δ ≥ 0.5000, the maximum response similarity single load sequence and the mismatch difference of the photovoltaic power load are calculated to obtain the time-domain mismatch load difference curve, as shown in formula (8).

[0026]

[0027] Where, P(-L x ) is the time-domain mismatch load difference curve, P(t) is the photovoltaic output prediction curve, L x (t) is the maximum response similarity single load prediction curve.

[0028] Then, the remaining sub-power load curve and the mismatch load difference curve are subjected to paired spectrum analysis calculation, steps 1) and 2) are executed, until the remaining paired load sequence maximum normalized response similarity weight Δ < 0.5000, the dynamic positive difference spectrum iteration combination is completed, and the optimal combination scheme of building multi-load and photovoltaic power load is obtained to realize the maximum photovoltaic consumption effect.

[0029] Advantages

[0030] (1) The present application can characterize the characteristics of each load in the building energy system and the matching relationship therebetween from the perspective of spectrum analysis, and can quantitatively evaluate the response matching degree of each device sub-load and photovoltaic power generation, thereby providing guidance for building energy demand response planning.

[0031] (2) The present application has high efficiency, can consider supply-demand balance, complementary action of various types of loads, and the associated response between building multi-device load and photovoltaic output load, and can realize optimal operation of power load dispatching and greatly improve the photovoltaic consumption rate. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 It is a technical roadmap of the spectrum matching analysis method of building multi-load associated response of the present application.

[0033] Figure 2 It is three types of load curves matched with the photovoltaic output curve in an embodiment of the present application.

[0034] Figure 3 It is a spectrum diagram of three paired load sequences matched with the photovoltaic output curve in an embodiment of the present application.

[0035] (1) Building air conditioning load and photovoltaic output curve spectrum matching;

[0036] (2) Electric vehicle load and photovoltaic output curve spectrum matching;

[0037] (3) Water heater and photovoltaic output curve spectrum matching.

[0038] Figure 4 This is a curve showing the mismatch difference between building air conditioning load and photovoltaic power load in one embodiment of the present invention. Detailed Implementation

[0039] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0040] Figure 1 This is a technical roadmap for a spectral matching analysis method for the multi-element load correlation response of a building according to the present invention.

[0041] This paper takes a building as the research object, with a total area of ​​5000m². 2 It has four floors above ground, with a total height of 23m, and solar photovoltaic panels are installed on the roof.

[0042] The building's parking lot is equipped with 10 slow-charging stations, each with a rated charging power of 7kW.

[0043] The operating scenario is based on winter conditions. The hourly cooling load of the building, the power consumption of individual equipment items, and the power consumption of electric vehicles in the parking lot are predicted based on historical load data obtained from the Internet of Things energy consumption information monitoring platform, and the predicted curves of the load of various equipment in the building and the photovoltaic output are obtained.

[0044] This article sets up three paired load curves for the building in this case study, including the building air conditioning load, electric vehicle charging load, and water heater power load, which are paired with the photovoltaic output curve respectively.

[0045] like Figure 2 The three types of load curves shown are matched with the photovoltaic output curves.

[0046] Based on the principle of pairwise spectral matching analysis, the three sets of paired load time-domain diagrams matched with the photovoltaic output curve are transformed into spectral diagrams, such as... Figure 3 As shown.

[0047] By analyzing the spectral density distribution of the paired photovoltaic power output and building power load curves of the three groups, the error percentage of the two spectral density values ​​at frequency f is calculated, and a frequency band with an error percentage σ ≤ 50% is set.

[0048] As shown in Table 1, within the selected frequency band, the spectral density difference of the paired load sequence curves should be kept as small as possible.

[0049] Table 1. Percentage of Spectral Density Error (%)

[0050]

[0051] In order to further compare the correlation and phase relationship, the corresponding coherence spectrum and phase spectrum values under the common frequency band interval are intercepted, and the mean values of the extracted data are processed and then normalized. The normalized response similarity weight analysis results of different pairs of sequences are shown in Table 2.

[0052] Table 2 Normalized response similarity weight analysis data table of paired load sequences

[0053]

[0054] From the spectrum analysis data in Table 2, it can be seen that the normalized response similarity weight of the building air conditioner load and the photovoltaic output curve is the largest, indicating that the correlation degree of the photovoltaic output and the building air conditioner load in the main period is the largest, and the period co-variation of the two load curves is also the strongest. The building air conditioner load can more effectively and timely consume the photovoltaic load.

[0055] Then, the mismatch difference between the building air conditioner load and the photovoltaic power load is calculated to obtain the time-domain mismatch load difference curve, as shown in Figure 4 .

[0056] Steps 1) and 2) are performed, the remaining electric vehicle charging load and the water heater electricity load curve are respectively paired with the mismatch load difference curve for spectrum analysis calculation, and the normalized response similarity weight of the remaining load sequence is shown in Table 3.

[0057] Table 3 Normalized response similarity weight table of remaining load sequences

[0058]

[0059] From the remaining load pair spectrum analysis data in Table 3, it can be seen that the matching degree of the water heater electricity load and the photovoltaic mismatch load is higher than that of the electric vehicle charging load and the photovoltaic mismatch load curve. The water heater electricity load can effectively use the photovoltaic power that the air conditioner load cannot use, further consume the photovoltaic load, and improve the photovoltaic consumption rate. Therefore, for this embodiment, the combination of the building air conditioner load and the water heater electricity load can further improve the photovoltaic utilization degree within one day of the demand response period.

[0060] It should be understood that the embodiments and cases discussed herein are only for illustration, and those skilled in the art can make improvements or changes, and all these improvements and changes shall fall within the protection scope of the appended claims of the present application.

Claims

1. A spectral matching analysis method for building poly-loads correlation response, characterized in that, The method comprises the following steps: Step 1): based on the predicted curve of different equipment sub-power load in the building energy system, a paired spectral analysis model between the building socket sub-load including air conditioning power load and electric vehicle charging load and the photovoltaic power curve is constructed; The paired spectral analysis model contains the spectral density distribution, coherence and phase relationship of the paired load sequence, the spectral density distribution is used to reflect the change degree of the paired load amplitude with the frequency distribution, the frequency domain correlation of the paired load curve is identified based on the coherence spectrum, and the frequency domain sequence of the paired load is discussed by means of the phase spectrum, the correlation characteristics between the building sub-power load and the photovoltaic power generation curve are studied based on the three spectral parameters; Step 2): based on the three parameters of spectral density distribution, coherence and phase relationship obtained in step 1), the internal frequency spectrum transfer mechanism of the paired spectral analysis model is established, the normalized response similarity weight is formed, the multiple building equipment power load is respectively paired with the photovoltaic power generation curve for spectral combination calculation, the normalized response similarity weight array is obtained, and the maximum normalized response similarity weight corresponding to the single building equipment load is extracted in the array, and the optimal synchronous matching is formed between the single building equipment load and the photovoltaic load; Step 3): according to the maximum normalized response similarity weight extracted in step 2), the mismatch difference value of the corresponding single load and the photovoltaic load is calculated, the time domain localization curve of the mismatch load difference is obtained, and a dynamic positive difference spectrum iterative combination method is developed, which is as follows: Based on the frequency spectrum transfer mechanism, if the normalized response similarity weight Δ is greater than or equal to 0.5000, the mismatch difference value of the single load sequence with the maximum response similarity and the photovoltaic power load is calculated, and the time domain mismatch load difference curve is obtained, as shown in formula (8); where P(-L x ) is the time-domain mismatch penalty difference curve, P(t) is the photovoltaic power output prediction curve, L x (t) is the maximum response similar single load prediction curve; Then, the remaining sub-power load curve and the mismatch load difference curve are calculated by paired spectral analysis, steps 1) and 2) are executed, until the maximum normalized response similarity weight Δ of the remaining paired load sequence is less than 0.5000, the dynamic positive difference spectrum iterative combination is completed, and the optimal combination scheme of the building multiple load and the photovoltaic power generation load is obtained.

2. The spectral matching analysis method of building multi-load correlation response according to claim 1, wherein, The step 1) is specifically: The spectral density distribution is used to reflect the change degree of the paired load amplitude with the frequency distribution, as shown in formula (1): where D Lxy (f) is the pairwise spectral density distribution; R XY (s) is the cross-covariance function; s is the time delay; f is the frequency; The polar coordinate expression form of the paired spectral density distribution is transformed, as shown in formulas (2)-(4): D Lxy (f) = A Lxy (f)exp{2πW Lxy (f)} (2) W Lxy (f) = arctan[-n Lxy (f) / m Lxy (f)](4) wherein, due to D Lxy (f) is a complex number, taking the real part as m Lxy (f) and the imaginary part as n Lxy (f); A Lxy (f) is the amplitude relationship between pairs of load sequences; W Lxy (f) is the phase relationship between pairs of load sequences Then, the frequency domain correlation of the paired load curve is identified based on the coherence spectrum, as shown in formula (5): Among them, D Lx (f) and D Ly (f) represent the spectral density distribution of individual loads; G Lxy (f) is the coherence spectrum of the paired load curves, which reflects the degree of linear correlation between the two sequences at different frequencies. The closer the coherence spectrum value is to 1, the greater the correlation between the paired load sequences at that frequency. Meanwhile, the frequency domain sequence of the paired load is analyzed by means of the phase spectrum, as shown in formula (4); The spectral parameters reflect the phase difference size of the paired load sequence at each frequency, that is, the sequence relationship, the closer the phase spectrum value is to 0, the weaker the leading or lagging degree of the paired load curve at the frequency is; Finally, the correlation characteristics between the building equipment power sub-load and the photovoltaic power generation curve are studied by comprehensively analyzing the three spectral parameters.

3. The spectral matching analysis method of building multi-load correlation response according to claim 1, wherein, The step 2) is specifically: Firstly, the error of the paired spectral density value of the building different sub-equipment operation load sequence and the photovoltaic power generation curve at the frequency is calculated, and the common frequency band interval under the set error σ is taken; Then, the correlation and phase relationship of the pair of load sequences are further compared, the corresponding coherence spectrum and phase spectrum values under the common frequency band interval are intercepted, and the extracted data mean values are processed and then normalized to calculate the normalized response similarity weight Δ i As shown in equation (6): where Δ i is the similarity weight of the coherence and phase normalized response of the ith pair of load sequences, Δ i ∈ [0, 1]; G i is the coherence mean of the spectral analysis of the ith pair of load sequences, G i ∈ [0, 1]; W i is the phase mean of the spectral analysis of the ith pair of load sequences, W i ∈ [0, 1]; and n is the number of combinations of pairs of load sequences. Finally, the normalized response similarity weight array is obtained by the pair spectrum combination calculation of the building equipment power load and the photovoltaic power generation curve, and the maximum value corresponding to the single building equipment load is extracted in the array to form the optimal synchronous matching with the photovoltaic load, as shown in equation (7): L x = max {Δ1, Δ2,..., Δ n} (7) where L x is the single building equipment load sequence corresponding to the maximum value extracted from the array.

Images