SOLR result are not too accurate

Question

I'm trying to improve my SOLR scheme to search in chinese field. In my index I have document with name_t_zh value AB考試課程3 Scheme for field type looks like this:

<fieldType name="text_general_zh" class="solr.TextField">
    <analyzer>
      <tokenizer class="solr.ICUTokenizerFactory"/>
      <filter class="solr.CJKWidthFilterFactory"/>
      <filter class="solr.LowerCaseFilterFactory"/>
    </analyzer>
</fieldType>

For that field I'm doing query:

name_t_zh:(AB考試課程3)

Problem is with the query which SOLR makes:

 "parsedquery":"name_t_zh:ab name_t_zh:考試 name_t_zh:課程 name_t_zh:3"

And in result I'm getting more not accurate result which contains number 1 and the rest is totally different or part of Chinese character. I want get one document with exactly the same name or the same text with other character at the end or beginning.