CN103021632A - Dedicated transformer for distributed generation and vibration measurement method of dedicated transformer - Google Patents

Abstract

The invention discloses a special transformer for distributed power generation and a vibration measurement method thereof, belonging to the technical field of transformer design and manufacture and signal monitoring. The transformer uses amorphous alloy as the core to make an amorphous alloy distribution transformer, which can effectively reduce the no-load loss of the transformer and improve the efficiency of solar power generation. The core structure adopts a three-phase three-column structure, which can not only be used for connection group Dyn11, but also meet the wiring requirements of Yyn0. The internal winding adopts a three-phase split winding structure, including a high-voltage winding and two split low-voltage windings, which can effectively limit the short-circuit current of the network and save construction investment costs and floor space. The invention adopts soft measurement technology, and constructs an RBF neural network model by capturing real-time voltage, current, magnetic flux, magnetic flux leakage and other relevant information of the transformer to realize online monitoring of vibration information. The achievements of the invention are widely used in distributed power generation, especially in solar power generation applications.

Description

A special transformer for distributed power generation and its vibration measurement method

technical field

The invention relates to a special transformer for distributed power generation and a vibration measurement method thereof, and belongs to the technical field of transformer design and manufacture and signal monitoring.

Background technique

In today's world energy structure, the energy used by human beings is mainly fossil energy such as coal, natural gas and oil. The large-scale exploitation and use of fossil energy has caused the energy situation to become increasingly tense, causing the world energy crisis, and at the same time causing serious problems. environmental pollution. Therefore, distributed generation has been paid more and more attention in recent years. The most important point of distributed power generation is to improve its power generation efficiency and power conversion rate, which requires that the no-load loss of the transformer should be as low as possible. At present, the transformers used in solar power generation are mainly 10KV distribution transformers, which still have many shortcomings in terms of reducing no-load loss and improving power generation efficiency.

In the field of power transformers, as a new type of high-tech transformer core material, amorphous alloy has many unique performance characteristics, such as excellent magnetic properties, corrosion resistance, wear resistance, high hardness, high strength, high resistivity, etc. The iron core made of crystal alloy strip has high saturation magnetic induction, low coercive force, low loss, low excitation current, and good temperature stability. It has attracted more and more attention and is one of the energy-saving products promoted by the country.

At present, the imported amorphous alloy strip has only three widths of 142mm/170mm/213mm. The core structure is a three-phase five-column or three-phase three-column structure. The cross-sectional area of the core is rectangular, and the transformer winding is rectangular with rounded corners. The short-circuit resistance of rectangular windings is poor, and the winding process is complicated, which limits the development of large-capacity amorphous alloy transformers. Circular windings have strong short-circuit resistance, and the manufacturing process of winding windings is mature (large transformers with silicon steel sheet cores are all circular windings), especially suitable for large-capacity amorphous alloy transformers. The maximum rectangular cross-sectional area is only 63.66% of the circular cross-sectional area, and the filling rate is low, which increases the cost of large-capacity amorphous alloy transformers. At the same time, due to the vibration of the iron core caused by the magnetostriction of the amorphous alloy strip, and the vibration of the winding caused by the dynamic electromagnetic force between the strips, the mechanical performance will be affected and noise will be generated. At present, there is no effective transformer vibration measurement method.

Contents of the invention

In view of the defects and deficiencies in the field of distributed power transformers in the prior art, the present invention introduces a special transformer for distributed power generation and a vibration measurement method thereof. The transformer structure breaks through the disadvantages of difficult manufacturing process and high manufacturing cost of the amorphous alloy transformer; the supporting use of the monitoring method can quickly, real-time and accurately monitor the vibration information of the transformer dedicated to distributed power generation.

The technical scheme of the special transformer for distributed power generation and its vibration measurement method of the present invention is:

A special transformer for distributed power generation, including an iron core and an internal winding. The iron core is an amorphous alloy iron core and adopts a three-phase three-column structure; the internal winding adopts a three-phase split winding structure, which has a high-voltage winding and two For low-voltage windings, the sum of their capacities is equal to the total capacity of the transformer.

A method for measuring vibration of a special transformer for distributed power generation, the method adopts soft sensor technology to realize online measurement of transformer vibration information.

As a further improvement of the present invention, the RBF neural network is used to construct the soft sensor model, and the vibration information measured online includes vibration amplitude and vibration frequency.

As a further improvement of the present invention, the basic variables selected by the RBF neural network include rated voltage, current, rated capacity, frequency, magnetic flux, magnetic flux leakage, load current, current density, core diameter, resistivity, ambient temperature, load factor .

As a further improvement of the present invention, the output layer of the RBF neural network has 2 neurons.

As a further improvement of the present invention, the base function of the RBF network adopts Gaussian function; the learning center of the base function is obtained by K-means clustering algorithm; the learning of weight value adopts LMS algorithm.

The beneficial effects of the present invention are: adopting the amorphous alloy transformer as the special transformer for distributed power generation, since the amorphous alloy transformer itself has the characteristics of ultra-low no-load loss, it can effectively reduce the loss of solar power generation and improve the power generation efficiency. Energy-saving products. At the same time, it adopts a three-phase three-column structure, which can not only be used to connect group Dyn11, but also meet the wiring requirements of Yyn0, which is more flexible. The use of three-phase split winding structure can effectively limit the short-circuit current of the network, and can save construction investment costs and floor space. The real-time and accurate online vibration measurement system guarantees the operation and maintenance of the transformer, which has practical significance and practical value.

Description of drawings

Fig. 1 is a structural schematic diagram of a transformer dedicated to distributed power generation of the present invention;

Figure 2 is a wiring diagram of split transformer coils;

Figure 3 is a flow chart of RBF neural network modeling.

In the figure: 1. Low-voltage end; 2. First low-voltage winding; 3. High-voltage terminal; 4. Low-voltage split air gap; 5. High-voltage winding; 6. Second low-voltage winding.

Detailed ways

The present invention will be described in further detail below in conjunction with the accompanying drawings and embodiments.

The structure of the dedicated transformer for distributed power generation of the present invention is shown in Figure 1, including: low-voltage terminal 1, first low-voltage winding 2, high-voltage terminal 3, low-voltage split air gap 4, high-voltage winding 5, and second low-voltage winding 6. The transformer uses amorphous alloy as the iron core to make an amorphous alloy iron core distribution transformer, which can reduce the no-load loss of the transformer to about 25% of the SCB11 type, and its efficiency can reach more than 98%. The iron core adopts a three-phase three-column structure, which can not only be used for connection group Dyn11, but also meet the wiring requirements of Yyn0, and has a wide range of applications. The internal winding adopts a three-phase split winding structure, with one high-voltage winding and two low-voltage windings, and the sum of their capacities is equal to the total capacity of the transformer. After the low-voltage winding is split, the distance between the high-voltage coil and each low-voltage coil will be greatly increased, as well as the short-circuit impedance value between the parts after the low-voltage coil is split. This will limit the short-circuit current of the network, save construction investment costs and occupy an area. certain practical economic significance.

The wiring method of the split transformer coil is shown in Figure 2. The Y/Δ-Δ-11-11 wiring method is adopted, that is, the high-voltage coil of the transformer is connected in a star shape, and the two split low-voltage coils are connected in a delta shape. In addition, there are other connections Way. In Figure 2, A, B, and C are high-voltage coil terminals, a1, b1, and c1 are low-voltage coil terminals, and O is the neutral point of the coil.

The vibration measuring method of the special transformer for distributed power generation of the present invention adopts the soft measurement technology to realize the online measurement of the vibration information of the special transformer for distributed power generation. A neural network is used to construct a soft sensor model, which can realize online measurement of vibration information including: vibration amplitude, vibration frequency, etc. The RBF neural network is used to construct the model. The basic variables selected by the neural network are: rated voltage, current, rated capacity, frequency, magnetic flux, magnetic flux leakage, load current, current density, core diameter, resistivity, ambient temperature, load factor, etc. The number of input layer neurons of this RBF neural network is determined according to the number of selected input variables, and the number of output layer neurons is determined according to the number of variables that need to be detected. In the present invention, it is necessary to measure the vibration amplitude, vibration frequency, so a total of 2 output neurons are required. The basis function of RBF network adopts Gaussian function; the learning center of basis function is obtained by K-means clustering algorithm; the learning of weight value adopts LMS algorithm.

This method can adapt to different transformers and different working environments by adjusting the number of input and output layer neurons, output weights and related parameters. The algorithm termination condition can be set according to the actual required precision, or the maximum number of iterations can be set.

For online measurement of vibration information, as shown in Figure 3, the K-means clustering algorithm is used to obtain the learning center of the basis function; the weight value is obtained by using the LMS algorithm. Among them, the specific steps of the K-means clustering algorithm are as follows:

(i=1,2,…,N),设置迭代步数n=0。 The first step is to initialize the clustering center, that is, randomly select N samples from the training sample set as the initial center based on experience

(i=1,2,...,N), set the number of iteration steps n=0.

。 The second step is to randomly input training samples

.

离哪个中心最近，即找到

使其满足

，i=1,2,…,N;式中

，是第n次迭代时基函数的第i个中心。 The third step is to find training samples

Which center is closest, i.e. find

make it satisfy

, i=1,2,…,N; where

, is the ith center of the base function at the nth iteration.

The fourth step is to adjust the center formula:

 即为RBF网络最终的基函数的中心。 The fifth step is to judge whether all the training samples have been learned and the central distribution is not changing, if so, end, otherwise n=n+1 go to the second step. finally got

It is the center of the final basis function of the RBF network.

Basis function variance determination:

为所选取中心之间的最大距离。 In the above formula, N is the number of hidden units,

is the maximum distance between the selected centers.

Claims (6)

1. A special transformer for distributed power generation, comprising an iron core and an internal winding, characterized in that: the iron core is an amorphous alloy iron core and adopts a three-phase three-column structure; the internal winding adopts a three-phase split winding structure, with One high-voltage winding and two low-voltage windings, the sum of their capacities is equal to the total capacity of the transformer. 2. A vibration measurement method for a transformer dedicated to distributed power generation according to claim 1, characterized in that: the online measurement of transformer vibration information is realized by using soft sensor technology. 3. The vibration measurement method of a dedicated transformer for distributed power generation according to claim 2, characterized in that: the RBF neural network is used to construct a soft sensor model, and the vibration information that can be measured online includes vibration amplitude and vibration frequency. 4. The vibration measurement method of the special transformer for distributed power generation according to claim 3, characterized in that: the basic variables selected by the RBF neural network include rated voltage, current, rated capacity, frequency, magnetic flux, magnetic flux leakage, load Current, current density, core diameter, resistivity, ambient temperature, load factor. 5. The method for measuring vibration of a dedicated transformer for distributed power generation according to claim 3 or 4, characterized in that: the output layer of the RBF neural network has 2 neurons. 6. the vibration measurement method of the special transformer for distributed generation according to claim 5, is characterized in that: the basis function of described RBF network adopts Gaussian function; The learning center of basis function obtains with K-means clustering algorithm; Weight The learning uses the LMS algorithm.

Images